zekeZZ/wmdp-bio-retain-corpus-datamodel-sorted

text stringlengths 2 1.78M
The authors confirm that all data underlying the findings are fully available without restriction. The sequence data generated in this study are available in the Gene Expression Omnibus database at NCBI under the series accession number GSE57326. Whole C. roseus transcriptome sequence along with functional annotation, expression profiling and identified SSRs are available at the Catharanthus Transcriptome web page (<http://nipgr.res.in/mjain.html?page=catharanthus>). Introduction {#s1} ============ Catharanthus roseus, popularly known as Madagascar periwinkle, is a medicinal plant which belongs to family Apocyanaceae. The plant is diploid (2n = 16) and native to islands of Madagascar, but now grown in many tropical countries as ornamental plant [@pone.0103583-Magnotta1]. C. roseus is well known for its pharmacological importance as it produces more than 130 terpenoid indole alkaloids (TIAs) including vinblastine and vincristine, which are widely used in anti-cancer chemotherapies [@pone.0103583-VanderHeijden1], [@pone.0103583-Guimaraes1]. Most tissues of C. roseus are known to produce alkaloids and no other single plant is known to produce such a wide spectrum of alkaloids [@pone.0103583-Blasko1]. The plant is known to treat diabetes also, due to hypoglycemic properties in its tissue extracts [@pone.0103583-Nammi1]. Moreover, roots of C. roseus are known to accumulate ajmalicine and serpentine which help controlling blood pressure and cardio-vascular disorders [@pone.0103583-Svoboda1]. Alkaloid biosynthetic pathways are highly branched and complex, with wide differences in alkaloid composition between underground and aerial tissues. TIAs have high commercial value because they are produced by plants in very low amounts and its infusion is very difficult. The common precursor of TIAs, strictosidine, is the central intermediate formed by the condensation of tryptamine (product of shikimate pathway) and secologanin (product of non-mevalonate pathway) involving strictosidine synthase (STR). Pharmacologically important alkaloids, vinblastine and vincristine (found only in aerial tissues) are synthesized in vivo by the condensation of vindoline and catharanthine, both of which are obtained from branch-point intermediate cathenamine. Biochemical pathway resulting in formation of vindoline is specifically present in well differentiated aerial tissues of the plant, but not in roots and cell cultures, thereby marking the presence of tissue-specific TIA pathway in C. roseus [@pone.0103583-Rischer1]--[@pone.0103583-Jaggi1]. In recent years, next generation sequencing has become the method of choice for fast and cost-effective transcriptome characterization for non-model plants [@pone.0103583-Morozova1]--[@pone.0103583-Garg1]. Earlier, a large-scale transcriptomic resource from three medicinal plants (Camptotheca acuminate, Catharanthus roseus and Rauvolfia serpentina) have been developed for elucidating monoterpene indole alkaloid (MIA) pathways [@pone.0103583-GongoraCastillo1]. Recently, Van Moerkercke et al. [@pone.0103583-VanMoerkercke1] used RNA-seq approach to construct metabolic pathway database, CathaCyc, for C. roseus. Such databases can facilitate identification of key regulator(s) for metabolic pathway engineering. To add on to existing resources, in this study, we generated C. roseus transcriptome by assembling RNA-seq data generated from C. roseus tissues (leaf, flower and root) and merged it with previously reported C. roseus transcripts. The updated comprehensive C. roseus transcriptome was screened for simple sequence repeats (SSRs) which might be helpful in development of functional molecular markers. We also identified transcription factor (TF) encoding transcripts in C. roseus transcriptome as only few TFs are known, which regulate TIA pathway genes. Expression analysis of genes involved in TIA pathways was also undertaken to reveal their tissue-specific expression. Gene ontology (GO) enrichment analysis highlighted the tissue-preferential/specific expression of transcripts in various biological processes. These data will provide a framework for further functional analysis of genes involved in biosynthesis of important alkaloids. Results {#s2} ======= Transcriptome sequencing and preprocessing of data {#s2a} -------------------------------------------------- To generate the transcriptome of C. roseus, three tissue samples (leaf, root and flower) were subjected to next generation sequencing using Illumina platform. Of the total ∼347 million reads generated from all the tissue samples, about 343 million reads were found to be of high quality after filtering with NGS QC Toolkit ([Table S1](#pone.0103583.s008){ref-type="supplementary-material"}), having average Phred quality score of at least 30 at each base position. As the short reads obtained may be redundant (due to PCR amplification at library preparation step) and their assembly needs a high-end server with high random access memory (RAM). Therefore, duplicate reads from each sample were removed and about 230 million non-redundant (NR) reads were obtained ([Table S1](#pone.0103583.s008){ref-type="supplementary-material"}). Optimization and validation of transcriptome assembly {#s2b} ----------------------------------------------------- To generate an optimal transcriptome assembly of C. roseus, we systematically compared the performance of various de novo short read assembly tools, including Velvet, Oases and ABySS. De novo assembly of total (343,384,084) and NR (230,715,698) high quality reads was performed employing a two-step approach. In the first step, primary assembly (best k-mer assembly) was generated using Velvet, Oases and ABySS at different k-mer lengths ranging from 31 to 95 ([Table S2](#pone.0103583.s009){ref-type="supplementary-material"} and [S3](#pone.0103583.s010){ref-type="supplementary-material"}). On the basis of several parameters described earlier [@pone.0103583-Garg2], assemblies obtained from respective assemblers at different k-mers were compared. Assemblies generated by Velvet showed a gradual increase in N50 and average read length with the k-mer length, best being at k-93 ([Table S2A and S3A](#pone.0103583.s009){ref-type="supplementary-material"}). Similarly, assembly at k-93 (from total reads; [Table S2C](#pone.0103583.s009){ref-type="supplementary-material"}) and k-87 (from NR reads; [Table S3C](#pone.0103583.s010){ref-type="supplementary-material"}) had higher N50 and average read lengths, and were considered to be the best assembly generated from ABySS. On the other hand, choosing the best assembly generated from Oases (using NR dataset) was a tricky task, as there was not much difference in N50 and average lengths at different k-mers ([Table S2B](#pone.0103583.s009){ref-type="supplementary-material"}). Finally, assembly at k-57 was selected, which had an optimal assembly size (number of contigs; [Table S3B](#pone.0103583.s010){ref-type="supplementary-material"}) and minimum redundant unigenes. Whereas, assembly of Oases at k-61 generated from total dataset had higher N50 and average lengths ([Table S2B](#pone.0103583.s009){ref-type="supplementary-material"}). By taking all the assembly parameters into consideration along with BLAST results, assembly generated by Oases at k-57 using NR short read dataset (NR-Oases k-57), was considered to be the best, which generated 42909 contigs (≥250 bp) of 1161 bp average length and 1990 bp N50 length ([Table 1](#pone-0103583-t001){ref-type="table"}). 10.1371/journal.pone.0103583.t001 ###### Assembly optimization/validation of Illumina data of C. roseus. ![](pone.0103583.t001){#pone-0103583-t001-1} Total high quality reads (best k-mer) Non-redundant high quality reads (best k-mer) MPGR assembly Merged assembly[2](#nt102){ref-type="table-fn"} -------------------------------------------------------------------------- --------------------------------------- ----------------------------------------------- --------------- ------------------------------------------------- ------- ------- ------- -------- -------- -------- Number of contigs 133650 39010 128716 106736 71190 39276 53017 42909 86726 59220 Total size (Mb) 72.78 25.38 165.01 160.92 57.62 25.62 51.86 49.81 107.79 76.03 Minimum length (bp) 100 185 100 250 100 185 107 250 251 250 Maximum length (bp) 15524 10893 17071 17071 15524 10913 17112 17112 12048 17141 Average length (bp) 544.6 650.7 1282 1507.8 809.5 652.3 978.2 1160.8 1243 1283.9 N50 length (bp) 1087 848 2161 2205 1400 841 1911 1990 1683 2115 Contigs with significant similarity [1](#nt101){ref-type="table-fn"} 63706 28142 76866 73292 41643 28477 22593 21275 66788 32666 Similarity search was done against TAIR10 proteome. Assembly generated by merging best k-mer with MPGR transcriptome using TGICL. Gongora-Castillo et al. [@pone.0103583-GongoraCastillo1] followed a robust approach to generate a comprehensive C. roseus transcriptome from different tissues and treatments. Our second step involved merging of primary assembly (best k-mer; NR-Oases k-57) with previously existing C. roseus assembly (MPGR de novo assembly; [@pone.0103583-GongoraCastillo1]). Our earlier studies have shown that TGICL software generates optimal merged assemblies [@pone.0103583-Garg1], [@pone.0103583-Agarwal1]. Unigenes from both the assemblies were size selected (≥250 bp) and transcript isoforms from MPGR assembly were removed (retaining the longest isoform) before subjecting to assembly using TGICL program. The merged assembly resulted in a total of 59220 contigs with improved average length (1284 bp) and increased N50 read length (2115 bp; [Table 1](#pone-0103583-t001){ref-type="table"}). Overall, the merged assembly was found to be much better than those reported previously [@pone.0103583-GongoraCastillo1], [@pone.0103583-Kumar2]. To assess the quality of C. roseus transcriptome thus obtained, we checked for the presence of publicly available C. roseus sequences in recently assembled transcriptome. Out of 287 full-length protein sequences (downloaded from NCBI), 220 (77%) were found to be present in the assembled transcriptome. Moreover, we also checked for the genes involved in various biochemical pathways and found that all the 108 genes previously reported by Van Moerkercke et al. [@pone.0103583-VanMoerkercke1] were represented in our transcriptome data. Similarly, all the 30 enzymes (genes) known to be involved in TIA bio-synthesis [@pone.0103583-VanMoerkercke1], were also present in the C. roseus transcriptome generated in this study. As compared with earlier reported transcriptome assemblies of C. roseus, we obtained nearly 19% (MPGR assembly) [@pone.0103583-GongoraCastillo1] and 42% (CathaCyc) [@pone.0103583-VanMoerkercke1] novel transcripts in our assembly. C. roseus belongs to the clade Asterids, and genome of three plant species of this clade, including Solanum tuberosum (potato), Solanum lycopersicum (tomato) and Sesamum indicum (sesame) have been sequenced so far. A BLAST analysis of C. roseus transcriptome against proteomes of tomato, potato, cucumber, grapevine and Arabidopsis, and transcriptomes of six known alkaloid producing plants (Atropa belladonna, C. acuminata, Cannabis sativa, R. serpentine, Rosmarinus officinalis and Valeriana officinalis) revealed higher similarity of C. roseus transcripts with R. officinalis (60.9%) followed by tomato (58.7%), potato (56.3%), cucumber (56.2%) and grapevine (56.1%) ([Fig. S1](#pone.0103583.s001){ref-type="supplementary-material"}). Further, reciprocal BLAST analysis with annotated protein sequences from closest reference genomes (tomato, potato, cucumber and grapevine) and Arabidopsis showed that, although Arabidopsis happens to be distantly related to C. roseus in phylogenetic tree but had the highest number of orthologs (15252) as compared to tomato (12118) and potato (11263), which belong to the same clade (Asterids) as that of C. roseus ([Fig. S2](#pone.0103583.s002){ref-type="supplementary-material"}). This may be due to availability of better genome annotation of the model plant Arabidopsis. Cucumber and grapevine had the least number of orthologs in C. roseus, because they belong to different clades. C. roseus transcripts generated above were designated as C. roseus tentative consensus (Cr_TC) and were assigned a unique identifier number from Cr_TC00001 to Cr_TC59220. The whole transcriptome sequence is available at Catharanthus Transcriptome Sequence web page (<http://nipgr.res.in/mjain.html?page=catharanthus>). The total size of transcriptome is ∼76 Mb with nearly 65% of the transcripts longer than 500 bp and more than 40% transcripts larger than 1000 bp ([Fig. S3](#pone.0103583.s003){ref-type="supplementary-material"}). Average GC content of C. roseus transcriptome was little lower (40.65%) than Arabidopsis (42.5%; [Fig. S4](#pone.0103583.s004){ref-type="supplementary-material"}), and comparable with that of soybean (40.9%) and chickpea (40.3%) [@pone.0103583-Garg2]. Whereas, average GC content of rice was much higher (55%) with respect to C. roseus and other dicot plant species analyzed. Functional annotation of C. roseus transcriptome {#s2c} -------------------------------------------------- For comprehensive annotation of C. roseus transcripts, similarity search was performed against several public databases sequentially. We were able to annotate 38380 (65%) unigenes with confidence (e-value≤1E-05), while others were considered to be C. roseus - specific which may be involved in various important biochemical pathways, whose intermediates and enzymes involved have not been catalogued in public repositories as of now. The putative function assigned to the transcripts is available at Catharanthus Transcriptome Sequence web page. Based on their similarity with Arabidopsis genes, C. roseus transcripts were assigned GOSlim terms under biological process, molecular function and cellular components categories ([Fig. S5A](#pone.0103583.s005){ref-type="supplementary-material"}). Among the biological process category, maximum number of C. roseus transcripts were assigned with the specific term, protein metabolism (13.17%) followed by response to stress (12.02%). GOSlim terms, nucleotide binding (9.81%), hydrolase activity (8.32%), transferase activity (8.03%) and protein binding (7.36%) were most represented under molecular function category. Among the cellular component category, nucleus (19%) followed by other cytoplasmic component (18.8%) were most represented ([Fig. S5A](#pone.0103583.s005){ref-type="supplementary-material"}). Based on COG (cluster of orthologous groups) classification, at least 33422 (56.43%) transcripts could be classified into 25 COG categories. Among the 25 COG categories, the cluster for general function prediction represented the largest group (7124; 21.31%), followed by post-translational modification, protein turnover, chaperones (3809; 11.40%) and signal transduction mechanisms (3380; 10.11%). In addition, 1941 (5.8%) of C. roseus transcripts were assigned into the cluster of unknown function ([Fig. S5B](#pone.0103583.s005){ref-type="supplementary-material"}). To elucidate various biochemical pathways represented in the transcriptome, C. roseus transcripts were searched against Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database, which aids in studying the role of gene(s) in complex metabolic pathways. In total, 4436 genes (4738 transcripts) were found to be involved in one of the 318 different KEGG metabolic pathways. Few of the metabolic pathways represented with higher number of genes were, ribosome (117 genes), spliceosome (98 genes), biosynthesis of amino acids (96 genes), RNA transport (87 genes), purine metabolism (82 genes), carbon metabolism (80 genes), oxidative phosphorylation (73 genes), pyrimidine metabolism (71 genes) and protein processing in endoplasmic reticulum (71 genes). Along with plant hormone signal transduction (38 genes) pathway, we found genes involved in various alkaloid biosynthesis pathways also, like terpenoid backbone biosynthesis (26 genes), phenylalanine tyrosine and tryptophan biosynthesis (23 genes), ubiquinone and other terpenoid-quinone biosynthesis (17 genes), tryptophan metabolism (10 genes), tropane piperidine and pyridine alkaloid biosynthesis (eight genes), diterpenoid biosynthesis (11 genes), isoquinoline alkaloid biosynthesis (seven genes), sesquiterpenoid and triterpenoid biosynthesis (five genes), monoterpenoid biosynthesis (two genes) and indole alkaloid biosynthesis (one gene). We also identified TF-encoding genes in C. roseus transcriptome and found 1820 transcripts representing 73 TF families. Among the 73 families represented, the MYB-domain (116) family TFs were the most abundant followed by AP2-EREBP- (106), WRKY- (78), bHLH- (75) and HB- (75) domain TFs ([Fig. 1](#pone-0103583-g001){ref-type="fig"}). TFs play an important role in secondary metabolite and TIA accumulation [@pone.0103583-Jaggi1], [@pone.0103583-VanMoerkercke1]. TFs known to be involved in TIA pathway like ORCA2, ORCA3, WRKY, MYC2 and zinc finger DNA-binding protein 1 and 2, were represented in the assembled transcriptome of C. roseus. ![Number of transcripts representing different transcription factor families in C. roseus transcriptome.](pone.0103583.g001){#pone-0103583-g001} Identification of simple sequence repeats {#s2d} ----------------------------------------- Transcriptome resources have been harnessed for mining of SSRs in several plant species. EST-SSRs provide an insight on the density of SSRs in the transcribed region of genome and have higher rates of transferability across species [@pone.0103583-Varshney1]. C. roseus transcriptome was screened for SSRs using MISA search tool and a total of 11620 SSRs were identified in 8644 (14.6%) C. roseus transcripts. Among the identified EST-SSRs, di-nucleotide repeats were most represented (55.55%), followed by tri-nucleotide repeats with 41.67% (4842; [Fig. 2A](#pone-0103583-g002){ref-type="fig"}). Among di-nucleotide repeats, AG/CT showed highest occurrence (60.84%), followed by AT/AT (31.73%), AC/GT (7.41%) and CG/CG (0.03%). In case of tri-nucleotide repeats, occurrence of various motifs was uniform except for AAG/CTT, which showed highest frequency (34%) and CCG/CGG being the least abundant (2.15%) ([Fig. 2B](#pone-0103583-g002){ref-type="fig"}). Further, we developed a comprehensive SSR marker resource for C. roseus by designing forward and reverse primers from their flanking sequences. In total, we could design primers for 7158 (61.6%) SSR repeat-motifs identified, which can be used for the generation of functionally relevant markers in C. roseus. The complete list of SSRs identified in C. roseus along with primer sequences are available at Catharanthus Transcriptome Sequence web page. ![Identification of simple sequence repeats (SSRs) in C. roseus transcriptome.\ (A) Distribution of SSRs in different classes (B) Frequency of most common SSRs motifs is shown by bar graph.](pone.0103583.g002){#pone-0103583-g002} Differential gene expression and gene ontology enrichment analysis {#s2e} ------------------------------------------------------------------ RNA-Seq has been considered to be the method of choice for differential gene expression studies at whole genome level [@pone.0103583-Ozsolak1], [@pone.0103583-Jain2]. In total, approximately 87--90% short reads mapped onto C. roseus transcriptome and nearly 84--87% mapped uniquely ([Table 2](#pone-0103583-t002){ref-type="table"}). DESeq package, was used to identify the genes differential expressed in different tissue samples [@pone.0103583-Anders1]. We identified differentially expressed genes among different tissues of C. roseus via pairwise comparisons ([Fig. 3A](#pone-0103583-g003){ref-type="fig"}). Leaf and root being the primary site for alkaloid production, differential gene expression in these tissues as compared to flower was analyzed in more detail. A total of 2443 and 2153 genes were differentially expressed in leaf and root, respectively, as compared to flower tissue ([Fig. 3A, B](#pone-0103583-g003){ref-type="fig"}). In leaves, higher number of genes (1635) were down-regulated than 808 up-regulated genes, whereas in roots, there were nearly equal number of up- (1125) and down-regulated (1028) genes ([Fig. 3A](#pone-0103583-g003){ref-type="fig"}). Out of the total 4596 differentially expressed genes in leaf and roots, 679 were common. Among 679 genes commonly differentially expressed in roots and leaves, 72 and 552 genes were up- and down-regulated, respectively ([Fig. 3B](#pone-0103583-g003){ref-type="fig"}). Fifty five genes were found to be up-regulated in one tissue and down-regulated in another tissue. The genes related to photosynthesis were up-regulated in leaf, whereas genes annotated as DNA binding proteins (TFs), disease-resistant proteins and wound response proteins were up-regulated in roots. A heat-map of 1861 up-regulated genes in at least one tissue is represented in [Fig. S6](#pone.0103583.s006){ref-type="supplementary-material"}. Among 1861 up-regulated genes, at least 148 were found to encode for TFs, which are key regulatory proteins. TFs are known to play an important role in accumulation of secondary metabolites in plants [@pone.0103583-Broun1]--[@pone.0103583-Patra1]. Among 148 TFs exhibiting significant differential expression in leaf and root tissues, majority of the transcription factors belonging to AP2-EREBP, HB, MYB, NAC, Tify and WRKY families were up-regulated in root tissues ([Fig. 3C](#pone-0103583-g003){ref-type="fig"}). ![Differential expression analysis of C. roseus transcriptome.\ (A) Number of differentially expressed genes in different tissues in pairwise comparisons. Number of up-regulated genes are in bold, while down-regulated are in normal font. (B) Venn diagram showing number of up- and down-regulated (in parentheses) genes in leaf and root tissues as compared to flower. Asterisk represents genes up-regulated in one tissue and down-regulated in another tissue. (C) Heat-map showing expression patterns of differentially up-regulated TF encoding genes in different tissues. The scale at the bottom represents log~2~ fold change. (D) Graphical view showing GO terms associated with biological process enriched in up-regulated genes of leaf. The GO enrichment was performed using BiNGO. Node size is proportional to the number of genes in each category and shades represent the scale denoting significance level (white- no significant difference).](pone.0103583.g003){#pone-0103583-g003} 10.1371/journal.pone.0103583.t002 ###### Mapping of non-redundant high-quality reads on C. roseus transcriptome. ![](pone.0103583.t002){#pone-0103583-t002-2} Tissue samples High quality reads Total mapped reads (%) Uniquely mapped reads (%) ---------------- -------------------- ------------------------ --------------------------- Leaf 79025564 70974531 (89.81) 68635709 (86.85) Flower 78728416 69955703 (88.86) 67774853 (86.09) Root 72961718 63571877 (87.13) 61747179 (84.63) We performed GO enrichment analysis to explore the major functional categories in up-regulated genes in leaves and roots. GO terms associated with various biological processes, such as metabolic process, nucleic acid metabolic process and cellular metabolic process were found to be enriched in up-regulated genes of leaf ([Fig. S7A](#pone.0103583.s007){ref-type="supplementary-material"}). Leaves being actively participating in photosynthesis, GO terms associated with photosynthetic process were also significantly enriched in leaves. Apart from these, biological process GO terms, like cellular response to jasmonic acid stimulus, ion homeostasis, carotenoid, isoprenoid and tertraterpenoid metabolic process were significantly enriched in genes up-regulated in leaf ([Fig. 3D](#pone-0103583-g003){ref-type="fig"}). Likewise, response to jasmonic acid stimulus was also significantly enriched in up-regulated genes of root. Thus, supporting the previous remarks that plant hormone jasmonic acid is one of the main drivers of TIA synthesis in C. roseus and plant secondary metabolism in general [@pone.0103583-Rischer1], [@pone.0103583-DeGeyter1]. Roots being an underground tissue is subjected to various biotic stresses present in the rhizosphere. GO terms like response to stress, response to biotic stimulus, defense response and response to fungus were significantly enriched in up-regulated root genes under biological process category. Regulation of transcription, cellular amino-acid derivative metabolic process, jasmonic acid biosynthesis, response to chemical stimulus, response to endogenous stimulus and response to salicylic acid were few other biological process GO terms, which were significantly enriched in up-regulated genes of roots ([Fig. S7B](#pone.0103583.s007){ref-type="supplementary-material"}). Expression profiling and validation of genes involved in TIA pathway {#s2f} -------------------------------------------------------------------- For expression analysis, we mapped the short reads of individual sample from our study and previous study [@pone.0103583-GongoraCastillo1] onto our C. roseus transcriptome and analyzed expression profile using DESeq software. Complete expression analysis of genes is available at Catharanthus Transcriptome Sequence web page. This resource can be utilized by researchers to look for expression profile of their gene(s) of interest. C. roseus (L.) var. Prabal is well known for its high alkaloid content [@pone.0103583-Dwivedi1] and we were interested in expression analysis of important alkaloids (TIA) biosynthesis pathway genes. We used CathaCyc database [@pone.0103583-VanMoerkercke1] for the analysis of these pathways. The common precursor of TIAs, strictosidine, is the central intermediate formed by the coupling of tryptamine (shikimate pathway) and monoterpene secologanin (methyl erythritol phosphate pathway) as shown in [Fig. 4A](#pone-0103583-g004){ref-type="fig"}. The alkaloid, vinblastine, is synthesized by coupling of vindoline and catharanthine, both of which are obtained from branch-point intermediate cathenamine ([Fig. 4A](#pone-0103583-g004){ref-type="fig"}). We identified C. roseus transcripts encoding for most of the enzymes catalyzing different reactions involved in these pathways ([Fig. 4A](#pone-0103583-g004){ref-type="fig"}). ![Expression patterns of transcripts involved in TIA biosynthesis.\ (A) Vindoline biosynthetic pathway showing important enzymes involved in different reactions. The IDs of C. roseus transcript encoding for the respective enzymes are also indicated. The important intermediates have been highlighted in bold font. (B) Heat-map showing expression patterns of TIA genes in different tissues and treatment. The scale at the bottom of each study represents log~2~ value of RPKM. Transcript IDs are given at left side and their putative annotation is on right side. (C) The correlation of gene expression results obtained from RNA-seq and real time RT-PCR analysis (D) Heat-map showing expression pattern of TF encoding genes in different tissues and treatment. The scale at the bottom of each study represents log~2~ value of RPKM. Transcript IDs are given at left side and their putative annotation is on right side.](pone.0103583.g004){#pone-0103583-g004} We identified 30 genes well-known to be involved in TIA biosynthetic pathways. The BLAST analysis showed that all of these genes were conserved in the sequenced genomes from Asterid clade (tomato and potato) and Arabidopsis at the protein level. However, only 17 (∼57%) and 22 (∼73%) of them exhibited significant similarity with annotated coding region sequences of Arabidopsis and Asterids (tomato and potato), respectively, at the nucleotide level. Further, we analyzed the expression of TIA biosynthetic pathway genes using RNA-Seq data in different tissue samples and treatments reported in our study and previous studies [@pone.0103583-GongoraCastillo1], [@pone.0103583-VanMoerkercke1]. As shown in the [Fig. 4B](#pone-0103583-g004){ref-type="fig"}, majority of the genes of TIA pathway were up-regulated in leaf and root tissues implying that these alkaloids are synthesized mainly in leaves and roots. Similar pattern of expression is also visible in developmental tissues used in the study by Gongora-Castillo et al. [@pone.0103583-GongoraCastillo1]. Both leaf and root tissues share more or less a common gene expression pattern for most TIA pathway genes, except for tabersonine 16-hydroxylase (Cr_TC35206) and deacetylvindoline 4-O-acetyltransferase (Cr_TC35622), which are highly down-regulated in roots and tabersonine 19-hydroxylase (Cr_TC04217), which is highly down-regulated in leaves ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). Apart from tabersonine 16-hydroxylase and deacetylvindoline 4-O-acetyltransferase, decreased expression of desacetoxyvindoline 4-hydroxylase was seen in root and hairy root cultures ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). Down-regulation of these enzymes in root tissues and hairy root cultures is in agreement with previous studies [@pone.0103583-Shukla1], [@pone.0103583-StPierre1], as they participate in terminal reactions for vindoline biosynthesis, which is restricted to aerial tissues. On the other hand, tabersonine 19-hydroxylase, which was found to be up-regulated in roots, further endorses previous finding that it helps to operate an alternate mechanism for tabersonine metabolism in roots by side-chain hydroxylation [@pone.0103583-Giddings1]. Gene expression analysis using RNA-seq data revealed that many of the genes involved in TIA pathway were differentially expressed in root and leaf tissues. To validate these findings, quantitative RT-PCR was performed for at least 10 genes of TIA pathway detected to be differentially expressed in root and leaf tissues. Real-time RT-PCR analysis revealed similar expression patterns of all the selected genes as observed in RNA-seq data. Moreover, the statistical analysis also showed a very good correspondence (correlation coefficient of 0.80) among the results of real time RT-PCR and RNA-seq data analysis as shown in [Fig. 4C](#pone-0103583-g004){ref-type="fig"}. Suspension culture supplemented with yeast extract does not seem to be an attractive approach for TIA production due to lower gene expression as seen in gene expression profile ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). On the other hand, seedlings treated with MeJa showed increased expression, after exposure for longer duration i.e. 5 and 12 days. Expression of TIA genes was relatively higher in hairy root cultures, however it escalated when subjected to MeJa treatment ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). Expression profiling of TFs known to be involved in TIA pathway across different tissues and treatments revealed that their expression is highly up-regulated in hairy root culture, stem and root ([Fig. 4D](#pone-0103583-g004){ref-type="fig"}). Similar to earlier observations, there was a growth related decrease in TIA transcripts in C. roseus and accumulation of bisindole alkaloid content depends on tissue maturity [@pone.0103583-Shukla1], [@pone.0103583-StPierre1], [@pone.0103583-Naaranlahti1], we also found that the expression of TIA genes diminished in the mature leaf tissue. Discussion {#s3} ========== C. roseus is widely known for its pharmaceutical potential and has become one of the extensively studied medicinal plants. It is considered to be single biological source of the anti-cancer compounds, vinblastine and vincristine [@pone.0103583-VanderHeijden1], [@pone.0103583-ElSayed1]. Although many good efforts have been made to elucidate the complete pathway of TIA biosynthesis, but few complex steps and intermediate compounds are still unknown. Transcriptome studies, with the advent of next generation sequencing technologies, can help addressing few of these problems via gene discovery. Here, we performed high-throughput sequencing of transcriptome from different tissues of C. roseus and used short read assembly tools (Velvet, Oases and ABySS) for de novo assembly optimization. A two-step strategy involving merging of best k-mer assembly (from out data) with earlier reported MPGR assembly [@pone.0103583-GongoraCastillo1] was employed to obtain a robust C. roseus transcriptome. Based upon various parameters [@pone.0103583-Garg2], assembly generated from Oases at k-mer length of 57 taking NR short reads was considered to be the best. This is in agreement with a previous study by Ghangal et al. [@pone.0103583-Ghangal1] who also reported that assembly generated from NR reads was better than total reads. Merging of best k-mer assembly (NR-Oases-k-57) with MPGR assembly using TGICL further improved the assembly assessment parameters, such as N50 length (2115 bp), average transcript length (1283 bp) and sequence similarity with closely related species. As sequence similarity also marks the completeness of the transcriptome, 77% of the known full-length C. roseus proteins were found to be present in our C. roseus transcriptome. We also found all the previously reported [@pone.0103583-VanMoerkercke1] genes involved in TIA bio-synthesis (30 genes) represented in our assembled transcriptome. The presence of already reported important alkaloid biosynthetic genes and full-length proteins marks the quality of C. roseus transcriptome. Moreover, BLAST analysis with earlier transcriptome sequences of C. roseus [@pone.0103583-GongoraCastillo1], [@pone.0103583-VanMoerkercke1] and other related plant proteome and transcriptome sequences revealed a better transcriptome assembly presented in our study. For comprehensive annotation, C. roseus transcriptome was subjected to similarity search against various known protein databases. We were able to annotate about 65% of C. roseus transcripts. Recently discovered genes encoding geraniol synthase (GES) [@pone.0103583-Simkin1] and iridoid synthase (IS) [@pone.0103583-GeuFlores1], known to be involved in biosynthesis of secologanin (a monoterpenoid alkaloid) from geranyl pyrophosphate were also present in our C. roseus transcriptome. Overall, more than 56% of transcripts were classified into 25 COG categories, which is quite higher than other studies [@pone.0103583-Lai1]--[@pone.0103583-Wei1]. We found the category "general function prediction" to be the most represented in COG classification accounting for its need for basic physiological and metabolic functions. The GC content of C. roseus transcriptome was found to be very much similar to other dicot plants. Our results concord with earlier findings that there is only marginal variation in average GC content between dicots like Arabidopsis, soybean, tomato, potato, pea and tobacco [@pone.0103583-Carels1]. Many studies have been undertaken to characterize and differentiate different C. roseus cultivars using various molecular markers, such as AFLP [@pone.0103583-Kim1], [@pone.0103583-ElDomyati1], RAPD [@pone.0103583-Kim1], [@pone.0103583-Shaw1], ISSR [@pone.0103583-ElDomyati1] and SSRs [@pone.0103583-Shokeen1]--[@pone.0103583-Mishra1]. Microsatellites are co-dominant molecular markers used for marker-assisted selection studies and their identification from high-throughput transcriptome studies have been reported in large number of plant species. A total of 11620 SSRs of 2--6 nucleotides were predicted in C. roseus transcripts with di-nucleotides repeats being most abundant followed by tri-nucleotide repeat. This is in accordance with previous studies on C. roseus, who also observed more di-nucleotide repeats than tri-nucleotide repeats in the EST datasets [@pone.0103583-Mishra1]. The availability of a large number of SSRs with primer sequences can help large-scale genotyping studies for various applications. Existence of genetic diversity in C. roseus have been demonstrated by developing STMS markers [@pone.0103583-Shokeen1], [@pone.0103583-Shokeen2]. Thus, availability of transcriptome for screening of SSRs hold an immense potential for high-throughput genotyping applications in C. roseus. TFs are key regulators that can alter the gene expression of several target genes, thereby can regulate metabolic flux. Members of some TF families, such as MYB, AP2-EREBP, WRKY, MYB-related and bHLH, are known to regulate secondary metabolism in plants [@pone.0103583-Kato1]--[@pone.0103583-Zhang1]. Members of the plant-specific AP2-EREBP TF family, namely octadecanoid-derivative responsive Catharanthus AP2-domain protein (ORCA2 and ORCA3), which are known to activate expression of several genes (enzymes) involved in TIA biosynthesis, were identified in our C. roseus transcriptome. Recently, Suttipanta et al. [@pone.0103583-Suttipanta1] characterized CrWRKY1 and reported its involvement in the transcriptional regulation of TIA pathway in C. roseus. Apart from AP2-domain and WRKY proteins, we also found MYC2, zinc-finger DNA binding protein 1 and 2 in the C. roseus transcriptome, which were reported by Van Moerkercke et al. [@pone.0103583-VanMoerkercke1] to be involved in regulation of TIA biosynthesis. Digital expression profiling, a powerful and efficient approach for in-silico analysis of gene expression, was employed to determine the expression of genes involved in TIA biosynthesis. When compared with flower, genes involved in TIA biosynthesis were highly active in leaf and root tissues. Except few genes, majority of the TIA pathway genes were up-regulated in leaves and roots, which are the prime source of anticancer and antihypertensive alkaloids, respectively. Our study further confirms previous findings that some of the enzymes involved in late reactions of vindoline biosynthesis are not expressed in cell cultures or in tissues unable to produce vindoline [@pone.0103583-DeLuca1], [@pone.0103583-De1]. As reported earlier by Goklany et al. [@pone.0103583-Goklany1], TIA pathway genes are up-regulated in hairy root cultures elicited with MeJa [@pone.0103583-Kumar1], [@pone.0103583-Raina1], we also observed an increase in gene expression of TIA pathway genes in MeJa induced hairy root cultures. Elicitor, like MeJa are compounds, which induce plant stress response and thereby increasing gene expression and alkaloid biosynthesis. Looking at overall digital expression profiling of TIA pathway genes, we conclude that expression of genes are dependent on plant maturity and are highly expressed in MeJa-elicited hairy root cultures. Nearly equal number of genes were differentially expressed in leaf and roots. Enrichment of GO terms, performed on differentially expressed genes showed that photosynthesis related genes were up-regulated in leaves. GO terms, like jasmonic acid biosynthesis, response to jasmonic acid stimulus, isoprenoid and tetraprenoid metabolic process were also enriched in differentially expressed genes. This further adds on to the findings that MeJa induces expression of TIA pathway genes. In conclusion, we assembled and annotated C. roseus transcriptome. Many transcripts harboring microsatellite repeats were identified, which can be used for marker-assisted breeding in C. roseus. Differential gene expression and GO enrichment analyses revealed the enrichment of genes involved in secondary metabolite production in leaf tissues, which is the prime source of bisindole alkaloids. Further, expression profiling of TIA genes determined that vindoline exclusively accumulates in aerial tissue of C. roseus and exposure to MeJa increases its production. However, deeper understanding of regulatory network governing TIA biosynthesis could help in successful metabolic engineering of alkaloid biosynthesis. The transcriptome resource generated in this study can facilitate understanding of regulatory and metabolic pathways underlying the biosynthesis of alkaloids. Materials and Methods {#s4} ===================== RNA isolation, sequencing and quality filtering {#s4a} ----------------------------------------------- Leaf, root and flower tissues of C. roseus L. var. Prabal were harvested from the adult plants grown in field. The tissues were harvested from the plants grown under natural environmental conditions in the experimentation field (28°31′55.3′′N 77°09′54.9′′E) of the National Institute of Plant Genome Research, New Delhi. The field experiments conducted in this study did not involve endangered or protected species and no specific permission was required for these location/activities. The tissues were snap frozen in liquid nitrogen and stored in −80°C until further use. RNA was isolated from tissue samples using TRI reagent (Sigma Life Science, USA). Quantity and quality of RNA samples were measured using Nanodrop (Thermo Fisher Scientific) and Agilent Bioanalyzer (Agilent technologies, Singapore). Sequencing was performed using HiSeq 2000 platform generating paired-end reads of 100 bp length. Stringent quality check was performed on short read datasets by using NGS QC Toolkit v2.3 [@pone.0103583-Patel1] to remove the low quality reads and those having primer/adaptor contamination. Duplicate reads from the dataset were removed using CLC Genomic Workbench (v4.7.2, <http://maq.sourceforge.net/index.shtml>) to obtain NR dataset. De novo short read assembly and validation {#s4b} -------------------------------------------- De novo transcriptome assembly was performed by using three commonly used short read assemblers, Velvet (v1.2.01) [@pone.0103583-Zerbino1], Oases (v0.2.04) [@pone.0103583-Schulz1] and ABySS (v1.2.6) [@pone.0103583-Simpson1]. All the three assemblers used in this study were run at different k-mer lengths, ranging from 31--95. We employed a two-step approach, the best k-mer assembly obtained from different assemblers was merged with MPGR C. roseus transcriptome [@pone.0103583-GongoraCastillo1] and subjected to second round of assembly using TGICL (v2.0) [@pone.0103583-Pertea1] with minimum and maximum overlap length of 40 and 90, respectively. Various assembly parameters kept in consideration for marking the best assembly has been described previously by Garg et al. [@pone.0103583-Garg2]. GC content analysis was done using in-house perl script. To validate the quality of assembled transcriptome, we performed simple and reciprocal BLAST searches at an E-value cut-off of ≤1e-05 for identification of best significant match. The proteome sequences of tomato, potato, cucumber, grapevine and Arabidopsis were downloaded from phytozome v9.1 ([www.phytozome.net](http://www.phytozome.net)) and transcriptome sequences of alkaloid producing plants (A. belladonna, C. acuminata, C. sativa, R. serpentine, R. officinalis and V. officinalis) were downloaded from Medicinal plant genomics resource ([www.medicinalplantgenomics.msu.edu](http://www.medicinalplantgenomics.msu.edu)). Functional annotation {#s4c} --------------------- One of the most common approach for annotating transcriptome assembly is similarity search via BLAST. C. roseus transcripts were searched against TAIR10 proteome, Uniref90, Uniref100 and non-redundant protein (NCBI-nr) data sets at an E-value cut-off of ≤1e-05 for identification of best significant match. GOSlim terms for molecular function, biological process and cellular component were assigned to each C. roseus transcripts on the basis of their best match Arabidopsis protein. Similarity search against COG database classified C. roseus transcripts among different categories of COG classification system. To look for the genes involved in various pathways, assignment of KEGG Orthology (KO) terms and KEGG pathway construction was performed using KAAS (KEGG Automatic Annotation Server) [@pone.0103583-Moriya1] at default parameters. Read mapping and gene expression analysis {#s4d} ----------------------------------------- For gene expression analysis, high-quality short reads were mapped on to C. roseus transcriptome assembly using RNA-seq analysis utility of CLC Genomics Workbench. A maximum of two mismatches were permitted for alignments. Unique read counts for each tissue sample were normalized by calculating the read per kilo-base per million (RPKM) for each transcript. DESeq (v1.10.1) [@pone.0103583-Anders1], a software of R package, was used for differential gene expression analysis. It measures gene expression based on the negative binomial distribution with variance and mean linked by local regression. We calculated the size factor for each sample for normalization of read count data using DESeq. A p-value cut-off of ≤0.05 and at least two-fold change in gene expression was used to identify differentially expressed genes. RPKM values were log~2~ transformed and heat-map showing expression profiles for genes involved in TIA pathway were generated using MultiExperiment Viewer (MeV, v4.8). Hierarchical clustering was performed using Pearson correlation metrics and average linkage rule using MeV. Real-time PCR analysis {#s4e} ---------------------- For real-time PCR analysis, gene-specific primers ([Table S4](#pone.0103583.s011){ref-type="supplementary-material"}) were designed using Primer Express (v3.0) software (Applied Biosystems, USA). Actin was used as an internal control. At least three independent biological replicates with three technical replicates of each biological replicate for each tissue sample were used for analysis. Real-time PCR reactions were carried out essentially following the protocol described previously [@pone.0103583-Garg3]. The correlation between expression profiles of selected genes obtained from real-time RT-PCR and RNA-seq data analysis was determined in R program. Identification of SSR and transcription factors {#s4f} ----------------------------------------------- C. roseus transcriptome was screened for the presence of microsatellites (SSRs) using MISA [@pone.0103583-Thiel1]. The number of repeating units considered in this study was, six for di-nucleotides, and five for tri-, tetra-, penta- and hexa-nucleotides. We did not consider mono-nucleotide repeats in this study. Primers for all the identified SSRs were designed using BatchPrimer3 v1.0 (probes.pw.usda.gov/batchprimer3). TFs encoding C. roseus transcripts were identified based on the Hidden Markov Model (HMM) profile search of conserved domain present in each TF family as described earlier [@pone.0103583-Garg2]. GO enrichment analysis {#s4g} ---------------------- For GO enrichment analysis, similarity search (BLASTX) was carried out against Arabidopsis proteome and the best hit corresponding to each C. roseus transcripts was identified. GO enrichment of different sets of genes was performed using BiNGO tool [@pone.0103583-Maere1] as described previously [@pone.0103583-Singh1]. Supporting Information {#s5} ====================== ###### Number of *C. roseus* transcripts showing significant similarity with proteome/transcriptome sequences of closely related/alkaloid producing plants. (PDF) ###### Click here for additional data file. ###### Reciprocal BLAST analysis of *C. roseus* transcripts showing number of orthologous genes in closely related plant species. (PDF) ###### Click here for additional data file. ###### Length distribution of transcripts in the *C. roseus* transcriptome. (PDF) ###### Click here for additional data file. ###### GC content distribution in the *C. roseus* and *A. thaliana* transcripts. (PDF) ###### Click here for additional data file. ###### Functional annotation of *C. roseus* transcripts. (A) GOSlim term assignment to the C. roseus transcripts in different categories of biological process, molecular function and cellular component. (B) COG function classification of C. roseus transcripts. (PDF) ###### Click here for additional data file. ###### Heat-map showing expression patterns of differentially up-regulated genes in different tissues of *C. roseus* analyzed in this study. The scale at the bottom represents log~2~ fold change. (PDF) ###### Click here for additional data file. ###### Graphical view showing biological process gene ontology term enrichment in up-regulated genes in (A) leaf and (B) root. The GO enrichment was performed using BiNGO. Node size is proportional to the number of genes in each category and shades represent the scale denoting significance level (white- no significant difference). (PDF) ###### Click here for additional data file. ###### Quality control and duplicate read removal statistics of C. roseus libraries. (PDF) ###### Click here for additional data file. ###### De novo assembly statistics by different assemblers at different k-mer length using total high-quality reads (a) Velvet (b) Oases (c) ABySS. (PDF) ###### Click here for additional data file. ###### De novo assembly statistics by different assemblers at different k-mer length using NR high-quality reads (a) Velvet (b) Oases (c) ABySS. (PDF) ###### Click here for additional data file. ###### Primer sequences used for real-time PCR analysis. (PDF) ###### Click here for additional data file. Authors are thankful to Dr. Rohini Garg for transcription factor analysis. [^1]: Competing Interests:The authors have declared that no competing interests exist. [^2]: Conceived and designed the experiments: MJ AKS. Performed the experiments: MV RS. Analyzed the data: MV RG RS. Contributed to the writing of the manuscript: RG MV MJ.
Introduction ============ Manganese (Mn) is an essential element for living organisms. It is the twelfth most abundant element in earth´s crust and is present in rocks, water, soil and food, normally associated with other elements (Santamaria, 2008\[[@R54]\]; Farina et al., 2013\[[@R19]\]). Environmental and occupational exposure to Mn may occur by contact with fungicides, such as Maneb, Manconzeb, methylcyclopentadienyl manganese tricarbonyl (MTT)-an anti-knock agent in gasoline, Mn-ore mining, Mn alloy production and dry alkaline battery manufacture (Mergler and Baldwin, 1997\[[@R42]\]; Mergler, 1999\[[@R41]\]). Dietary ingestion is the main source of Mn for humans and Mn absorption takes place mostly in the gastrointestinal tract where it is homeostatically controlled in the intestinal wall (Au et al., 2008\[[@R4]\]). The brain is especially susceptible to metal intoxication during embryonic development, when it is known that Mn is able to cross the placenta and to be excreted in the maternal milk (Betharia and Maher, 2012\[[@R7]\]). Mn absorption is increased during the neonatal period, when biliary excretion is poorly developed, leading to elevated concentrations of Mn in the brain and other tissues (Aschner and Aschner, 2005\[[@R2]\]). In children, Mn exposure is associated with alterations in psychomotor and cognitive development; furthermore a positive correlation exists between Mn exposure and hyperactivity (Menezes-Filho et al., 2011\[[@R40]\]; Roels et al., 2012\[[@R52]\]; Torres-Agustín et al., 2013\[[@R61]\]). Exposure to high levels of Mn can lead to pathological conditions, including neurodegeneration (Mergler et al., 1994\[[@R43]\]). The mechanisms mediating Mn toxicity are complex and not completely understood. Some of them include: \(1\) Mn accumulation in astrocytes leading to disruption of their ability to promote neuronal differentiation and decreasing glutamate uptake by astrocytes (Erikson and Aschner, 2003\[[@R16]\]; Giordano et al., 2009\[[@R21]\]); \(2\) Mn induced loss of dopaminergic neurons (Stanwood et al., 2009\[[@R57]\]); \(3\) Inhibition of respiratory chain complexes and induction of reactive oxygen species (ROS) (Zhang et al., 2004\[[@R64]\]; Sriram et al., 2010\[[@R56]\]). The use of alternative models in toxicological studies has been growing over the years. The fruit fly Drosophila melanogaster has served as a unique and powerful model for studies on human genetics and diseases. Although humans and flies are only distantly related, almost 75 % of disease related genes in humans have functional orthologs in the fly (Deepa et al., 2009\[[@R13]\]; Pandey and Nichols, 2011\[[@R46]\]). Moreover, the fast and external developmental cycle of this organism enable the study of toxicological effects of compounds during the developmental period. All these advantages make flies an appropriate model for studies related with metal toxicity (Bonilla-Ramirez et al., 2011\[[@R10]\]; Paula et al., 2012\[[@R47]\]) and human neurodegeneration (Hirth, 2010). As the embryonic development period is particularly sensitive to Mn exposure, in this paper we aimed to investigate the behavior and biochemical alterations caused by Mn exposure during the embryonic development of Drosophila melanogaster, focusing on adaptations in the antioxidant systems and MAPK signaling pathways. The levels of Mn and major essential elements were also determined. Materials and Methods ===================== Reagents -------- Anti-phospho-p38^MAPK^, anti-phospho JNK, anti-phospho ERK, anti ERK and ß-actin antibodies were purchased from Cell Signaling Technology (Danvers, MA, United States). EDTA (CAS 60-00-4), glycine (CAS 56-40-6), tris(hydroxymethyl) aminomethane (CAS 77-86-1) and ammonium persulfate (CAS 7727-54-0) were purchased from Serva (Heidelberg, Germany). L-Glutathione reduced (CAS 70-18-8), 1-chloro-2,4-dinitrobenzene (CAS 97-00-7), sodium orthovanadate (CAS 13721-39-6), manganese (II) chloride tetrahydrate (CAS 13446-34-9), ß-mercaptoethanol (CAS 60-24-2), methanol (CAS 67-56-1), tween 20 (CAS 9005-64-5), potassium phosphate dibasic (CAS 7758-11-4), potassium phosphate monobasic (CAS 7778-77-0), potassium bicarbonate (CAS 298-14-6) and anti-rabbit imunoglobulin antibody, N,N,N\',N\'-Tetramethylethylenediamine (CAS 110-18-9), quercetin (CAS 117-39-5), protease inhibitor cocktail for use with mammalian cell and tissue extracts, 5,5´dithiobis(2-nitrobenzoic acid) (CAS 69-78-3), 2´,7´-dichlorofluorescein diacetate (CAS 2044-85-1), glycerol (CAS 56-81-5), resazurin sodium salt (CAS 62758-13-8), triton x-100 (CAS 9002-93-1), sodium chloride (CAS 7647-14-5), albumin from bovine serum (CAS 9048-46-8), HEPES (CAS 7365-45-9), ß-nicotinamide adenine dinucleotide 2´-phosphate reduced tetrasodium salt were obtained from Sigma Aldrich (St. Louis, MO, United States). Bis-acrylamide, hybond nitrocellulose, acrylamide (CAS 79-06-1), sodium dodecyl sulfate (CAS 151-21-3), boric acid (CAS 10043-35-3) were purchased from GE Healthcare Bio-Sciences AB (Uppsala, Sweden). All other reagents were commercial products of the highest purity grade available. Animals ------- Drosophila melanogaster(Harwich strain) was obtained from the National Species Stock Center, Bowling Green, OH, USA. The flies were maintained at 25 °C on 12 h light/dark cycle in glass bottles containing 10 mL of standard medium (mixture of 39 % coarse and 32 % medium corn flour, 10 % wheat germ, 14 % sugar, 2 % milk powder, 1 % salt, 1 % soybean flour, 1 % rye flour, a pinch of methyl paraben (99-76-3) and lyophilized yeast. All experiments were performed with the same strain, and both genders were used at random. Animal treatment ---------------- Adults flies were placed in 10 mL of standard medium supplemented with 3 mL of a fresh solution (0.1 mM, 0.5 mM or 1 mM) of manganese chloride (MnCl~2~). In the control group the standard medium were supplemented with 3 mL of ultrapure water. After ten days laying eggs the adult flies were removed. When eggs were newly ecloded, 1 to 3 day old flies were used for all analyses. The MnCl~2~ concentrations were chosen based on previous studies (Bonilla-Ramirez et al., 2011\[[@R10]\]). Locomotor assay --------------- Locomotor activity was determined using the negative geotaxis assay as described by Bland et al. (2009\[[@R9]\]) with minor modifications. Briefly, for each assay, individual flies (1-3 days old) were immobilized on ice and placed separately in a glass tube; this method of immobilization does not affect fly neurology (Deepa et al., 2009\[[@R13]\]). After 15 minutes the flies were gently tapped to the bottom of the tube and the time required to climb up 8 cm of the tube wall was recorded. Each fly was tested 4 times at 1 minute intervals. For each experiment, the climbing mean was calculated. Metal content ------------- Two hundred flies per group were washed three times in ultrapure water and then dried on a filter paper in the incubator at 37 °C for 90 minutes. Flies were digested in closed vessels according to the procedure described previously by Bizzi et al. (2010\[[@R8]\]). Flies (\~ 70 mg) were transferred to quartz vessels with 6 mL of nitric acid 3 mol L^-1^. After closing and capping the rotor, the vessels were pressurized with 7.5 bar of oxygen using the valve originally designed for pressure release after conventional acid sample digestion. Then, the rotor was placed inside a microwave oven (Multiwave 3000 Microwave Sample Preparation System, Anton Paar, Graz, Austria). The system was equipped with eight high-pressure quartz vessels (volume of 80 mL, maximum operational temperature and pressure of 280 °C and 80 bar, respectively). Pressure was monitored in each vessel during all the runs. Microwave heating program was as follows: \(1\) 1000 W, with a ramp of 5 min; \(2\) 1000 W for 10 min; and \(3\) 0 W for 20 min (cooling step). After digestion, the pressure in each vessel was carefully released. The resulting solutions were transferred to polypropylene vials and diluted to 25 mL with water. Determination of calcium (Ca), iron (Fe), potassium (K), magnesium (Mg), manganese (Mn), sodium (Na), phosphorus (P), sulfur (S), and zinc (Zn) was performed using an inductively coupled plasma optical emission spectrometer (Optima 4300 DV, PerkinElmer, Shelton, USA) with axial view configuration. A concentric nebulizer and cyclonic spray chamber were used. Argon 99.996 % (White Martins, São Paulo, Brazil) was used for plasma generation, nebulization and as auxiliary gas. The instrumental parameters were carried out in according with previous work (Pereira et al., 2013\[[@R48]\]). Two readings were averaged to give one value per biological replicate and expressed as a mean (±) standard deviation of the mean (SD). Metals levels were expressed relative to the weight of flies used for analysis (µg metal/g of dried weight tissue). Cellular viability ------------------ Cellular viability was measured by two different methods. Firstly, cellular viability was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay as described by Sudati et al. (2013\[[@R58]\]) with minor modifications. The analysis was performed on the whole body of female flies. The flies were incubated in MTT for 60 min (37 °C), after MTT was removed the sample was incubated in DMSO for 30 min (37 °C). The absorbance from formazan dissolution by addition of DMSO was monitored in an EnsPireR multimode plate reader (PerkinElmer, USA) at 540 nm. The second method used was the resazurin reduction assay. The method uses the indicator resazurin to measure the metabolic capacity of cells. Viable cells reduce resazurin into resorufin, a fluorescent compound (Franco et al., 2009\[[@R20]\]). Groups of 40 flies were mechanically homogenized in 1 mL 20 mM Tris buffer (pH 7.0) and centrifuged at 1,000 RPM for 10 min at 4 °C. The supernatant was incubated in Elisa plates with 20 mM Tris buffer (pH 7.0) and resazurin for two hours. The fluorescence was recorded using EnsPire^R^ multimode plate reader (Perkin Elmer, USA) at~ex~579nm and ~em~584nm. DCF-DA oxidation assay ---------------------- Groups of 20 flies were mechanically homogenized in 1 mL 20 mM Tris buffer (pH 7.0), and centrifuged at 1,000 RPM for 10 min, 4 °C. The supernatant was used to quantify 2'-7'-dichlorofluorescein diacetate (DCF-DA) oxidation as a general index of oxidative stress as described by Perez-Severiano et al. (2004\[[@R49]\]). The fluorescence emission of DCF resulting from DCF-DA oxidation was monitored at regular intervals (~ex~488nm and ~em~530nm) in an EnsPire^R^ multimode plate reader (PerkinElmer, USA). Determination of gene expression by real-time quantitative PCR (qPCR) --------------------------------------------------------------------- Real-time quantitative PCR (qRT-PCR) was performed according to the method described by Paula et al. (2012\[[@R47]\]). The primers utilized are shown in Table 1[(Tab. 1)](#T1){ref-type="fig"}. All samples were analyzed as technical and biological triplicates with a negative control. Threshold and baselines were manually determined using the StepOne Software v2.0 (Applied Biosystems, NY). SYBR fluorescence was analyzed by StepOne software version 2.0 (Applied Biosystems, NY), and the CT (cycle threshold) value for each sample was calculated and reported using the 2^-ΔΔCT^ method (Livak and Schmittgen, 2001\[[@R38]\]). The GPDH gene was used as an endogenous reference showing no alterations in response to the treatment. For each well, analyzed in quadruplicate, a ΔC~T~ value was obtained by subtracting the GPDH C~T~ value from the C~T~ value of the interest gene (sequences of tested genes are represented in Table 1). The ΔC~T~ mean value obtained from the control group of each gene was used to calculate the ΔΔCT of the respective gene (2^-ΔΔCT^). Enzyme assays ------------- For enzyme activity measurements, groups of 40 flies were mechanically homogenized in 1 mL 20 mM HEPES buffer (pH 7.0), and centrifuged at 14.000 RPM for 30 min at 4 °C (Franco et al., 2009\[[@R20]\]). The supernatant was used for determination of Glutathione S-Transferase (GST), Catalase (CAT), Superoxide Dismutase (SOD) and Thioredoxin Reductase (TrxR). The GST activity was assayed following the procedure of Jakoby and Habig (1981\[[@R29]\]) using 1-chloro-2,4-dinitrobenzene (CDNB) as substrate. The assay is based on the formation of the conjugated complex of CDNB and GSH at 340 nm. The reaction was conducted in a mix consisting of 100 mM phosphate buffer (pH 7.0), 1 mM EDTA, 1 mM GSH and 2.5 mM CDNB. CAT activity was assayed following the clearance of H~2~O~2~ at 240 nm in a reaction media containing 50 mM phosphate buffer (pH 7.0), 0.5 mM EDTA, 10 mM H~2~O~2~, 0.012 % TRITON x100 as described by Aebi (1984\[[@R1]\]). SOD activity assay was performed as described by Kostyuk and Potapovich (1989\[[@R35]\]). The assay consists in the inhibition of superoxide-driven oxidation of quercetin by SOD at 406 nm. The complete reaction system consisted of 25 mM phosphate buffer (pH 10), 0.25 mM EDTA, 0.8 mM TEMED and 0.05 µM quercetin. TrxR activity was assayed as described by Holmgren and Björnstedt (1995\[[@R26]\]). The test is based on the reduction of oxidized thioredoxin (Trx-S~2~) to reduced thioredoxin \[Trx-(SH)~2~\], using NADPH at 412 nm in a reaction media containing 0.1 M phosphate buffer (pH 7.0), 10 mM EDTA, 5 mM DTNB, 0.2 mg/mL BSA, 0.2 mM NADPH. All enzyme activities were performed at room temperature (25 ± 1 °C) using a Thermo Scientific Evolution 60s UV-Vis spectrophotometer. The enzyme activities were expressed in milliunits per milligram of total protein content, which was quantified following Bradford (1976\[[@R11]\]). Western blotting ---------------- Quantification of of the phosphorylation of mitogen-activated protein kinases (MAPKs) was performed by Western blotting as described by Posser et al. (2009\[[@R50]\]) with minor modifications. Groups of 40 flies were mechanically homogenized at 4 °C in 200 µL of buffer (pH 7.0) containing 50 mM Tris, 1 mM EDTA, 20 mM Na~3~VO~4~, 100 mM sodium fluoride and protease inhibitor cocktail. The homogenate were centrifuged at 4000 RPM for 10 min at 4 °C and the supernatants collected. After protein determination following Bradford (1976\[[@R11]\]), 4 % SDS solution, ß-mercaptoethanol and glycerol was added to samples to a final concentration of 100, 8 and 25 %, respectively and the samples frozen for further analysis. Proteins were separated using SDS-PAGE with 10 % gels, and then electrotransferred to nitrocellulose membranes (Paula et al., 2012\[[@R47]\]). Membranes were washed in tris-buffered saline with Tween (100 mM tris-HCl, 0.9 % NaCl and 0.1 % Tween-20, pH 7.5) and incubated overnight at 4 °C with specific primary antibodies (anti-phospho-p38^MAPK^, anti-phospho JNK, anti-phospho ERK, anti ERK and anti ß-actin). Following incubation, membranes were washed in tris-buffered saline with Tween and incubated for 1 h at 25 °C with anti rabbit-IgG secondary specific antibodies. Antibody binding was visualized using the ECL Western Blotting substrate Kit (Promega). Band staining density was quantified using the Scion Image software (Scion Image for Windows) and expressed as the percentage (%) of the control group (mean ± standard deviation of the mean). The values were normalized using total proteins (total ERK and ß-actin). Statistical analysis -------------------- Statistical analysis was performed using one-way ANOVA followed by Tukey´s post hoc test. Pearson's correlation test was applied for detection of significant statistical differences among the metals. Differences were considered statistically significant when p \< 0.05. GraphPad Prism 5 Software was used for artwork creation. Results ======= Exposure to Mn causes hyperactive behaviors and alters metal levels in Drosophila melanogaster ---------------------------------------------------------------------------------------------- The evaluation of climbing behavior performance by negative geotaxis showed that flies exposed to 0.5 mM and 1 mM of Mn reached the limit of columns significantly (p \< 0.005) faster than controls (Figure 1[(Fig. 1)](#F1){ref-type="fig"}). Levels of Mn and other essential metals were measured in D. melanogaster exposed to Mn. At concentrations of 0.5 and 1 mM, the levels of Mn in flies increased almost two and three fold respectively when compared with the control group (Table 2[(Tab. 2)](#T2){ref-type="fig"}), whereas Ca, Cu, Na and Zn levels decreased significantly in flies treated with Mn. Statistically, a significant negative relationship between Mn uptake and levels of Ca (r = -0,7966), Fe (r = -0,6635), Cu (r = -0,8028), S (r = - 0,6018) and Zn (r = -0,9802) occurred in response to Mn treatment (Table 3[(Tab. 3)](#T3){ref-type="fig"}). Mn exposure during embryonic development decreased cell viability and increased ROS production in flies ------------------------------------------------------------------------------------------------------- Cell viability was evaluated through two different tests, MTT and Resazurin. Both procedures showed a significant drop in cell viability at higher concentrations of Mn, confirming its toxicity at the cellular level (Figure 2A and 2B[(Fig. 2)](#F2){ref-type="fig"}). Many factors have been implicated in Mn-induced neurotoxicity, among them the oxidative stress caused by dopamine oxidation, or its ability in interfering with cellular respiration (Aschner and Aschner, 2005\[[@R2]\]). In this study, exposure to Mn leads to an increase in DCF-DA oxidation, a general index of oxidative stress from 0.5 mM (Figure 3[(Fig. 3)](#F3){ref-type="fig"}). Mn increased CAT and SOD mRNA expression, without altering their enzymatic activity ----------------------------------------------------------------------------------- Expression of mRNA for Cat, Sod and Hsp83 (an homolog of mammalian HSP90) (Bandura et al., 2013\[[@R5]\]) was analyzed by qRT-PCR using specific primers (Table 1[(Tab. 1)](#T1){ref-type="fig"}) in flies treated with 1 mM Mn. CAT and SOD mRNA expression was 250 % higher in treated flies whereas Hsp83 expression was 1000 % higher (Figure 4[(Fig. 4)](#F4){ref-type="fig"}). The levels of the antioxidant enzymes activity TrxR, GST, SOD and CAT were determined TrxR and GST activity were increased at concentrations of 0.5 mM and 1 mM (Figure 5[(Fig. 5)](#F5){ref-type="fig"}), while CAT and SOD activity showed no significant differences. Mn exposure inhibited p38MAPK phosphorylation --------------------------------------------- MAPKs phosphorylation levels were investigated in flies exposed to Mn. There was a 40 % inhibition of p38^MAPK^ phosphorylation in flies exposed to Mn at 1 mM, while the phosphorylation level of extracellular signal-regulated kinases (ERK) was unaltered. C-Jun-N-terminal Kinases 2 (JNK2) phosphorylation was not statistically different from controls (Figure 6[(Fig. 6)](#F6){ref-type="fig"}). Discussion ========== In this study, we determined the biochemical and behavioral alterations in Drosophila melanogaster in response to Mn exposure during embryonic development. Mn is an essential element required in key biological processes; however, high levels of Mn are associated with neurological and neuropsychiatric disorders (Mergler, 1999\[[@R41]\]). The risk of Mn overexposure comes from both occupational and environmental sources (Mergler and Baldwin, 1997\[[@R42]\]). Mn intoxication, a syndrome known as Manganism, is characterized by an extrapyramidal dysfunction and neuropsychiatric symptomatology and is associated with prolonged occupational exposure to high concentrations of this metal. Classical symptoms include irritability, intellectual deficits, compulsive behaviors, tremors and cock-like walk (Mergler, 1999\[[@R41]\]; Roth, 2006\[[@R53]\]). In rodents, Krishna et al. (2014\[[@R36]\]) showed that adult mice exposed to Mn through the drinking water presented neurobehavioral deficits and glial activation related with Mn deposition in brain. Moreover, others studies demonstrated that Mn toxicity in rats is accompanied by increased cholesterol biosynthesis and impairments in neuronal function of the hippocampus, which is involved in learning and memory (Öner and Sentürk, 1995\[[@R45]\]; Sentürk and Öner, 1996\[[@R55]\]). It has been shown that Mn supplementation during the neonatal period of rats resulted in increased Mn concentrations in tissues leading to adverse effects on motor development and behavior (Tran et al., 2002\[[@R62]\]). Mn uptake is increased during the neonatal period as biliary excretion, which has been suggested as a pathway for Mn elimination from the body, is poorly developed at this stage (Aschner and Aschner, 2005\[[@R2]\]). Exposure to Mn during the embryonic and early postnatal periods may result in increased levels of Mn in the brain and other tissues including bone, liver, pancreas and kidney (Aschner and Aschner, 2005\[[@R2]\]; Roels et al., 2012\[[@R52]\]). Higher levels of Mn retention in utero may affect children´s psychomotor development (Takser et al., 2003\[[@R59]\]). Possible adverse effects of Mn exposure on children´s health include cognitive deficits and hyperactive behaviors (Menezes-Filho et al., 2009\[[@R39]\]; Torres-Agustín et al., 2013\[[@R61]\]). Children exposed to high levels of Mn during the fetal period were more impulsive, inattentive, agressive, defiant, disobedient, destructive and hyperactive (Ericson et al., 2007\[[@R15]\]). It is recognized that factors such as the source and the duration of exposure, as well as nutritional status, can interfere in the intensity and incidence of neurological symptoms associated with Mn exposure in humans. Chronic consumption of drinking-water containing Mn at levels ranging from 81 to 2300 µg/l was associated with progressively higher prevalence of neurological symtoms (Kondakis et al., 1989\[[@R33]\]). The concentrations used in this study were from 0.1 mM of Mn in food (corresponding to 19 mg/L in the medium). Despite the use of relatively elevated concentrations, body levels of Mn were not altered at 0.1 mM. In previous studies, adult flies were acutely exposed to Mn (0.5-20 mM) diluted in sucrose, as the only source of food and liquid, which lead to significant locomotor deficits (Bonilla-Ramirez et al., 2011\[[@R10]\]). Our study is the first where Mn was provided as a cereal based diet over all the embryonic period. Thus, more studies are necessary to understand the rate of Mn uptake from diet in flies and how it may affect neurological behaviour. In our study, flies exposed to Mn at 0.5 mM and 1 mM showed increased locomotor speed in the locomotor behavior test (assayed as negative geotaxis behavior), pointing to a hyperactive-like behavior in Drosophila melanogaster. Furthermore, Mn levels were substantially increased in treated flies, while Ca, Cu, Zn, Fe and S levels were all decreased. This relationship may be in part associated with a competition of the metals for the same mechanism of transport into the flies cells. Facilitated diffusion, active transport, divalent metal transport 1 (DMT1), ZIP8 and transferrin (Tf)-dependent transport mechanisms are all involved in cellular Mn transportation (Aschner et al., 2007\[[@R3]\]). Among these metal transport systems, DMT1 has a very broad substrate specificity and is likely to be the major transmembrane protein responsible for the uptake of a variety of divalent cations, including Mn^2+^, Cd^2+^, Zn^2+^, Co^2+^, Ni^2+^, Cu^2+^ and Pb^2+^ (Gunshin et al., 1997\[[@R22]\]). In the flies, many proteins involved in the metabolism of biometals such as ferritin, transferrin, iron regulatory proteins, divalent metal transporter are expressed (Bonilla-Ramirez et al., 2011\[[@R10]\]). In this context, Mn uptake is less frequently studied in comparison with other metals and the mechanisms related to Mn transport are considerably more complex, occurring in most part in the divalent (II) and oxide forms (Tebo et al., 2004\[[@R60]\]). Mn has the capacity to interact and /or compete with Ca (Dittman and Buchwalter, 2010\[[@R14]\]). In a study performed in the aquatic insect Hydropsyche sparna, Mn exposure decreased cadmium (Cd) and Zn accumulation. Furthermore, increased Ca concentrations significantly reduced Mn accumulation in the insect (Poteat et al., 2012\[[@R51]\]). Dittmanand and Buchwalter (2010\[[@R14]\]) suggested that Mn is also absorbed by Ca transporters in aquatic insects, where increasing ambient Ca concentrations decrease Mn accumulation. There was also a negative correlation between Mn and Fe levels. Iron deficiency has been suggested as a possible contributing cause of attention deficit and hyperactivity disorder (ADHD) in children (Konofal et al., 2008\[[@R34]\]). Concomitantly, children with ADHD showed elevated serum Mn concentrations (Konofal et al., 2008\[[@R34]\]). Recent studies have suggested that Mn accumulates in dopaminergic neurons via the presynaptic dopamine transporter (DAT) and an altered functioning of the dopaminergic system has been well established in the etiology of ADHD (Farias et al., 2010\[[@R18]\]). Our results showed decreased cell viability using two different methodologies and increased ROS generation in flies exposed to Mn during development. Previous work by our group in PC12 cells, demonstrated that Mn leads to increased production of hydrogen peroxide (H~2~O~2~) (Posser et al., 2009\[[@R50]\]). H~2~O~2~ is a highly permeable and reactive molecule being able to react with metals such as Fe, thus generating hydroxyl radicals (Jiménez Del Río M and Vélez-Pardo, 2004\[[@R30]\]; Barbosa et al., 2010\[[@R6]\]) resulting in a propagation of oxidative damage in cells. Tissues can respond to oxidative stress by modulating antioxidant defenses (Halliwell and Gutteridge, 2007\[[@R23]\]). We measured the gene expression of Hsp83, CAT and SOD in Drosophila melangaster. Earlier studies showed that cellular stress may induce heat shock proteins in parallel with ROS production (Kim et al., 2004\[[@R32]\]; Paula et al., 2012\[[@R47]\]). Our results showed a significant increase in Hsp83 mRNA levels in Mn treated flies. Previously, our group demonstrated that exposure of flies to heavy metals such as mercury causes increases in the expression of Hsp83 (Paula et al., 2012\[[@R47]\]). CAT and SOD mRNA levels were significantly increased by Mn, but the enzymatic activity of these proteins was unchanged. The antioxidant ezyme SOD converts superoxide radicals (O~2~^º-^) to H~2~O~2~ and CAT catalyzes the conversion of H~2~O~2~ to oxygen (O~2~) and water (Barbosa et al., 2010\[[@R6]\]), thus neutralizing these reactive species. Considering that both SOD and CAT are crucial in the cell defense against oxidative stress (Halliwell and Gutteridge, 2007\[[@R23]\]), it might be expected that a posttranscriptional regulation mechanisms could maintain adequate levels of these proteins, however, further studies are necessary to elucidate this. Our results also showed that the activity of TrxR and GST was enhanced in Mn exposed flies. These two enzymes play an important role in protection against oxidative stress (Mustacich and Powis, 2000\[[@R44]\]). GST is a complex group of phase II detoxifying enzymes that participate in the metabolism of electrophilic substances, including carcinogenic, mutagenic and toxic compounds (Hayes et al., 2005\[[@R24]\]). TrxR is a dimeric FAD-containing enzyme that catalyzes the NADPH-dependent reduction of the active-site disulfide in Trx-S~2~ to give a dithiol in Trx-(SH)~2~ (Zhao et al., 2002\[[@R65]\]). Thioredoxin consists in one of the major redox-regulating proteins displaying a number of biological activities, including cytoprotection against ROS, protein repairing and protein disulfide reduction and modulation of signaling pathways (Yan et al., 2012\[[@R63]\]). Our data suggest that the increase in the levels of Mn and TrxR activity could represent a response to oxidation of thioredoxin in response to Mn-induced oxidative stress. MAPKs regulate the activity of a range of transcription factors thereby controlling gene expression and cellular function. The three most-studied MAPKs are ERK1/2, JNK1/2 and p38^MAPK^ (Ichijo, 1999\[[@R27]\]). ASK1 (Apoptosis Signaling Kinase 1) is a member of mitogen activated protein kinase kinase family (MAPKK) and an upstream activator of MAPK signaling pathway (Yan et al., 2012\[[@R63]\]). The redox state of thioredoxin regulates ASK1. Under normal conditions, ASK1 is bound to and inhibited by thioredoxin and when thioredoxin is oxidized, ASK1 can be activated and apoptotic signaling through the p38 ^MAPK^/JNK1/2 MAPKs initiated (Ichijo et al., 1997\[[@R28]\]). Studies conducted by Yan et al. (2012\[[@R63]\]) in a pancreatic carcinoma cell line, showed inhibition of TrxR by indolequinones resulting in a change of Thioredoxin-1 redox state to an oxidized form and activation of p38^MAPK^/JNK1/2. Similarly, Cd treatment activated ASK1 and its downstream MAPK in neuronal cells (Kim et al., 2005\[[@R31]\]), and inhibits components of thioredoxin system (Chrestensen et al., 2000\[[@R12]\]), while that Liedhegner et al. (2011\[[@R37]\]) demonstrated that knockdown of ASK1 as well as chemical inhibition of p38^MAPK^ and JNK played protective effects against L-DOPA induced apoptosis. We show that Mn induced TrxR activity while p38^MAPK^ /JNK1/2 phosphorylation were inhibited. This suggests an involvement of thioredoxin system in the mechanism of Mn induced toxicity. Augmented TrxR activity may represent a cellular response to high levels of ROS induced by Mn exposure, thus preventing the oxidation of thioredoxin and its dissociation of ASK1. This could contribute to diminished activation of p38^MAPK^ pathway upstream kinases resulting in lower levels of phosphorylation of this MAPK thus minimizing apoptotic cell death. In summary, our study demonstrate for a first time that developmental exposure to Mn leads to hyperactive-like behavior and accumulation of this metal in Drosophila melanogaster. The observed raise in Mn levels is negatively correlated with levels of other essential metals. This result fits with previous studies showing that Mn accumulation and Fe deficiency are associated with hyperactive behavior in children (Ericson et al., 2007\[[@R15]\]; Konofal et al., 2008\[[@R34]\]). The induction of stress responsive genes and antioxidant enzyme activity associated with inhibition of p38^MAPK^ phosphorylation at higher concentrations of Mn may represent an adaptive response to oxidative stress generated by this metal, in an attempt to avoid exacerbated cellular damage. Acknowledgements ================ The authors thank to Universidade Federal do Pampa, Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq 482313/2013-7), Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (Fapergs 1954/2551/13-7), FAPERGS/PRONEM, FINEP and INCT-APA for financial support. The authors thank Professor Peter R. Dunkley for critical review of the manuscript. Conflict of interest ==================== The authors declare that they have no conflict of interest. ![Genes tested by quantitative real-time RT-PCR analysis and used forward and reverse primers](EXCLI-13-1239-t-001){#T1} ![Metal levels in flies Mn-exposed during the embryonic development](EXCLI-13-1239-t-002){#T2} ![The relationship between Mn concentration and metal levels](EXCLI-13-1239-t-003){#T3} ![Effects of exposure to Mn during the embryonic development on locomotor performance in Drosophila melanogaster. Results are expressed as mean (±) standard deviation (SD) and represent the time spent to climb up to 8cm in a glass tube. (n = 20-24). \\ indicates a significant effect of Mn in comparison to control group (p \< 0.005).](EXCLI-13-1239-g-001){#F1} ![Effect of Mn exposure on cell viability in homogenate of flies treated with metal over embryonic development performed by MTT (A) and Rezarusin (B) cell viability assays. For MTT, the average of absorbance in control group was 0.017 (n = 3-4); by Rezarusin, absorbance in control group was 229.03 (n = 4-5). Results were expressed as the percentage (%) of the control group (mean ± standard deviation)\ \* indicates a significant effect of Mn in comparison to control group (p \< 0.05).](EXCLI-13-1239-g-002){#F2} ![Analysis of ROS production in response to Mn-exposure during embryonic development in flies. The data showed the DCF-DA intensity of fluorescence in total flies homogenate, expressed as percentage of the control group (mean ± standard deviation) (n = 3-10).\ \* indicates a significant effect of Mn in comparison to control group (p \< 0.05).](EXCLI-13-1239-g-003){#F3} ![Quantitative real time PCR (qRT-PCR) analysis of CAT, SOD and HSP83 mRNA in flies Mn-exposed at 1 mM. We used qRT-PCR to quantify levels of mRNA, relative to respective controls, after exposure. The data were normalized against GPDH transcript levels and each bar represents the mean ± standard deviation expressed as percent of its respective control (n = 3-4)\ \\ and \\\* indicate a significant effect of Mn in comparison to control group (p \< 0.005 and p \< 0.0001, respectively).](EXCLI-13-1239-g-004){#F4} ![Effects observed on enzymatic activities in flies Mn-exposed during embryonic development. TrxR activity (n = 5-7) (A), GST activity (n = 3-4) (B), SOD activity (n = 6-8) (C) and CAT activity (n = 5-6) (D). The data shows the enzymatic activities in flies homogenate expressed as mean (mU/mg protein) ± standard deviation.\ \* and \\\* indicate a significant effect of Mn in comparison to control group (p \< 0.05 and p \< 0.0001, respectively).](EXCLI-13-1239-g-005){#F5} ![Modulation of MAPKs phosphorylation in response to Mn-exposed flies during the embryonic development. The upper panel is a western blot showing phosphorylated forms of p38^MAPK^, ERK and JNK ½ and total forms of ERK and ß-actin (A). The graphs are showing the quantification (percentage of control) of immunoreactive bands, each bar represents the mean ± standard deviation. ERK phosphorylation was normalized with total ERK expression (n-5) (B). p38^MAPK^ phosphorylation was normalized with ß-actin expression (n=5) (C). JNK ½ phosphorylation was normalized with ß-actin expression (n = 3-5) (D).\ \* Indicates a significant effect of Mn in comparison to control group (p \< 0.05).](EXCLI-13-1239-g-006){#F6}
1. Introduction {#s1} =============== Consider a game in which every decision maker is faced with a finite set of choices such that one specific choice always brings him higher monetary payoff than other choices, irrespective of the choices made by other players. In this situation, the individual choice boils down to going for either a higher or a lower monetary payoff. The straightforward response of a decision maker who cares about his monetary payoff is to disregard dominated actions---i.e., actions that may only deteriorate payoff relative to other actions. This dominance principle is the most basic solution concept of game theory (Camerer, [@B12]). It becomes very powerful when embedded in a strategic reasoning as a stepwise process. In each step, the dominance principle implies that dominated strategies should be eliminated from an agent\'s strategy space. In an important class of games---known as dominance-solvable games---this iterated elimination of dominated strategies leads to a unique solution. Strikingly, the data collected from numerous experiments on dominance-solvable games raise important questions about the empirical accuracy of predictions derived from this principle. Subjects tend to display less strategic sophistication than is needed to justify many applications of iterated dominance (and related refinements) to model human decision making in strategic environments (Crawford, [@B19]). The beauty contest game is one of the textbook examples of this issue[^1^](#fn0001){ref-type="fn"}. A given set of players is asked to choose a number in the range \[0, 100\]. To win the game, a player should choose a number that is the closest to p = 2/3 of the average of all chosen numbers. Any number above 2/3 × 100 ≈ 66.7 violates first-order dominance, because the average has to be lower than 100. Knowing this, players should all choose numbers no greater than 66.7, meaning that their average may not exceed 2/3 × 66.7 ≈ 44.5. This reasoning lowers the target as the number of iterations increases, eventually leading to the unique Nash equilibrium in which all players choose 0. In many experimental studies of this game, the numbers chosen by players are used as a proxy of the depth of iterated reasoning.[^2^](#fn0002){ref-type="fn"} A well replicated stylized fact is to observe 1/3 of subjects choosing a number higher than 67, and at least 1/3---a number between 44 and 67. This paper focuses on one of the earliest and simplest example of such an empirical inaccuracy of dominance solvability, adapted from a 2 × 2 game discussed in Rosenthal ([@B43]) and first brought to the laboratory by Beard and Beil ([@B5])[^3^](#fn0003){ref-type="fn"}. The normal-form representation of this game is given in Table [1](#T1){ref-type="table"}. With L \< S \< H, m \< h, and s \< h, the game is one-step dominance solvable: the elimination of player B\'s weakly dominated strategy l immediately leads to the Pareto-Nash equilibrium (R, r)[^4^](#fn0004){ref-type="fn"}. ###### Generic form of the normal representation of Rosenthal ([@B43]) dominance solvable game. Player B -------------- ----- -------------- -------- Player A L (S; s) (S; s) R (L; m) (H; h) In line with observed behavior in other dominance solvable games, numerous studies (summarized in Table [2](#T2){ref-type="table"}) find frequent failures to achieve the Pareto-Nash equilibrium. In spite of variations in the design (described in the table), deviations from the standard theoretical predictions are systematic and sizable. First, dominance is frequently violated by player Bs. Depending on the exact experimental setup, up to 27% column players choose a strictly dominated action. Second, player As violate iterated dominance, even in those cases in which player Bs commonly obey dominance. As an example, while only 6% of player Bs violate dominance in Jacquemet and Zylbersztejn ([@B35])-ET2 and BT2, 26% of row players still contradict the predictions of dominance solvability by choosing L (and this figure may even attain 86% in other instances, see Beard, Beil -- Tr. 5 in Table [2](#T2){ref-type="table"}). As shown in the three middle columns of the table, both the absolute and the relative size of the stakes vary a great deal from one study to the other. Several lessons emerge from this accumulated evidence. First, both players react to their own monetary incentives. Second, in some cases player As also adjust their behavior to player Bs\' incentives. Finally, as shown by Jacquemet and Zylbersztejn ([@B35]), players\' inefficient behavior does not fade away with repetition and cannot be explained by inequality aversion (as framed by Fehr and Schmidt, [@B23]). ###### Overview of existing experimental evidence. Experiment Form Payoff Outcomes (%) ------------------------- ---------- --------------- ------------------ ------------ ---- ----- ----- ----- ----- Beard, Beil--Tr.1 Seq (9.75; 3.0) (10; 5.0) (3; 4.75) 66 29 6 83 --- Beard, Beil--Tr.2 Seq (9.00; 3.0) (10; 5.0) (3; 4.75) 65 35 0 100 --- Beard, Beil--Tr.3 Seq (7.00; 3.0) (10; 5.0) (3; 4.75) 20 80 0 100 --- Beard, Beil--Tr.4 Seq (9.75; 3.0) (10; 5.0) (3; 3.00) 47 53 0 100 --- Beard, Beil--Tr.5 Seq (9.75; 6.0) (10; 5.0) (3; 3.00) 86 14 0 100 --- Beard, Beil--Tr.7 Seq (58.50; 18.0) (18.0; 28.50) (60; 30.0) 67 33 0 100 --- Beard et al.--Tr.1 Seq (1450; 450) (1500; 750) (450; 700) 79 18 3 83 --- Beard et al.--Tr.2 Seq (1050; 450) (1500; 750) (450; 700) 50 32 18 64 --- Goeree, Holt--Tr.1 Ext (80; 50) (90; 70) (20; 10) 16 84 0 100 --- Goeree, Holt--Tr.2 Ext (80; 50) (90; 70) (20; 68) 52 36 12 75 --- Goeree, Holt--Tr.3 Ext (400; 250) (450; 350) (100; 348) 80 16 4 80 --- Cooper, Van Huyck--Tr.9 Str (4; 1) (6; 5) (2; 4) 27 --- --- --- 86 Cooper, Van Huyck--Tr.9 Ext (4; 1) (6; 5) (2; 4) 21 --- --- --- 84 JZ, 2014--BT1 Str (9.75; 3.0) (3.0; 4.75) (10; 5.0) 51 41 8 84 81 JZ, 2014--ET1 Str (9.75; 5.0) (5.0; 9.75) (10; 10.0) 54 33 13 72 73 JZ, 2014--ET3 Str (9.75; 5.5) (5.5; 8.50) (10; 10.0) 39 48 13 79 76 JZ, 2014--ET4 Str (8.50; 5.5) (5.5; 8.50) (10; 10.0) 25 61 14 82 82 JZ, 2014--ET2 Str (8.50; 8.5) (6.5; 8.50) (10; 10.0) 26 70 4 94 94 JZ, 2014--BT2 Str (8.50; 7.0) (6.5; 7.00) (10; 8.5) 26 70 4 94 94 For each implementation in row, the first column describes the actual design of the experiment: simultaneous-move strategic-form game (Str), simultaneous-move extensive-form game (Ext), sequential-move game (Seq). The monetary payoffsof each outcome, displayed in columns 2--4, are in USD in Beard and Beil ([@B5]) and Cooper and Van Huyck ([@B14]), in cents of USD in Goeree and Holt ([@B27]), in Yens in Beard et al. ([@B4]), and in Euros in Jacquemet and Zylbersztejn ([@B35]). The game is repeated ten times in changing pairs in Jacquemet and Zylbersztejn ([@B35]), and one-shot in all other instances. The aim of the present paper is to explore whether this empirical puzzle is related to players\' cognitive skills. In this sense, our investigation belongs to a recent and growing body of experimental studies in both psychology and economics which investigate the relationship between strategic behavior and cognitive skills[^5^](#fn0005){ref-type="fn"}. The main conclusion that can be drawn from these studies is that high cognitive skills predict strategic sophistication and efficient decision making. First, people with high cognitive skills make more accurate predictions about other people\'s intentions. Recent evidence from psychological research reveals the relationship between cognitive skills and the theory of mind. Using the "Reading the Mind in the Eyes" test (RMET, Baron-Cohen et al., [@B3]) to measure one\'s theory of mind, Ibanez et al. ([@B32]) find that people with higher cognitive skills are better at infering the internal emotional states of others[^6^](#fn0006){ref-type="fn"}. Relatedly, the results of a neuroeconomic experiment on the p-beauty contest game by Coricelli and Nagel ([@B17]) suggest that strategic thinking about other players\' thoughts and behavior is implemented by medial prefrontal cortex (mPFC) -- one of the brain areas commonly associated with theory of mind[^7^](#fn0007){ref-type="fn"}. An economic experiment by Carpenter et al. ([@B13]) also shows that people with higher cognitive ability make more accurate predictions of others\' choices in a 20-player beauty contest game. Second, people with higher cognitive skills apply more sophisticated reasoning and are more apt in strategic adaptation. Burks et al. ([@B10]) report that subjects with higher cognitive skills more accurately predict others\' behavior in a sequential prisoners\' dilemma game, and adapt their own behavior more strongly. In the context of the p-beauty contest game, subjects with higher cognitive skills are not only found to carry out more steps of reasoning on the equilibrium path (Burnham et al., [@B11]; Brañas-Garza et al., [@B8]), but also to adapt their behavior to their opponents\' cognitive skills (Gill and Prowse, [@B26]) as well as to their beliefs about their opponents\' cognitive skills (Fehr and Huck, [@B22]). Third, cognitive skills may be associated with the economic efficiency of outcomes of both individual and group activities. Corgnet et al. ([@B16]) find that higher cognitive skills predict better performance and less shirking in an experimental labor task (summing up tables of 36 numbers without using a pen). Jones ([@B36]), Al-Ubaydli et al. ([@B2]), and Proto et al. ([@B40]) report that groups with higher cognitive skills attain higher cooperation rates in repeated prisoner\'s dilemma games. On the other hand, Al-Ubaydli et al. ([@B1]) do not find a relationship between group members\' average cognitive skills and the efficiency of outcomes in a stag hunt coordination[^8^](#fn0008){ref-type="fn"}. Our contribution is two-fold. First, we provide new evidence on the relationship between strategic behavior and cognitive skills. We show that systematic mismatches between theoretical predictions and actual behavior in a classic 2 × 2 dominance-solvable game have cognitive underpinnings. Subjects with higher cognitive skills are found to be more likely to play dominant strategy and to best respond to other\'s strategy. Furthermore, cognitive skills predict strategic sophistication: only those players with sufficiently high cognitive ability are found to display sensitivity to the presence of uncertainty about others\' behavior. Our second contribution lies in experimental methodology. We extend the recent body of laboratory experiments comparing the performance of different measures of cognitive skills in predicting economic behavior. Notwithstanding the previous results (see e.g., Brañas-Garza et al., [@B8]; Corgnet et al., [@B15]), we report that the Raven\'s test score is a more general predictor of strategic behavior than the Cognitive Reflection Test score. 2. Experimental design {#s2} ====================== Our experiment is based on a 2 × 2 factorial design that varies the payoff matrix and the nature of player B. Each of the four resulting experimental treatments is implemented through a between-subject procedure---each subject participates in only one experimental condition. This data come from a large dataset, part of which has been previously used by Hanaki et al. ([@B30]). The main focus of that study is player As\' behavior under strategic uncertainty and its relation to monetary incentives and fluid intelligence. Certain elements of their design (such as the use of Human and Robot conditions and interest in players\' cognitive skills) inevitably needed to be adopted in the present study in order to address a much more general question of the empirical validity of the solution concept of dominance solvability. More precisely, we are interested in both players\' behavior (so as to measure the use of dominance by player Bs and the use of iterated dominance by player As under different information structures). We also make a methodological contribution, since in this paper we associate players\' behavior with multiple facets of cognitive skills: fluid intelligence (measured by Raven\'s test) and cognitive reflection (measured by CRT). Our first treatment variable is the size of the stakes, as represented by Game 1 and Game 2 in Table [3](#T3){ref-type="table"}. Although they have the same strategic properties, these two game matrices differ in terms of the saliency of monetary incentives to use (iterated) dominance. In Game 2, player As may earn a surplus of only 0.25 when moving from L to (R, r) (with payoff going from 9.75 to 10), while ending up in (R, l) is relatively costly (yielding only 3). In Game 1, the potential gains and losses from action R relative to L are more balanced: the gain from moving from L to (R, r) increases to 1.5 (with payoff moving from 8.5 to 10), while the outcome (R, l) becomes less costly (now yielding 6.5). The incentives of player Bs, in turn, go in the opposite direction: the gain from using the dominant strategy r (and conditional on player As\' choice R) is lower in Game 1 \[with payoff increasing from 4.75 to 5 between (R, l) and (R, r)\] than in Game 2 (where payoff increases from 8.5 to 10). In line with Jacquemet and Zylbersztejn ([@B35]) and Hanaki et al. ([@B30]) (who report that both players only react to their own monetary incentives) and as discussed in Section 3.1, each of these games generates sizable yet diverse empirical violations of dominance solvability. These two games together thus provide a wide range of monetary incentives to use dominance solvability within a common strategic environment[^9^](#fn0009){ref-type="fn"}. ###### The experimental games. GAME 1 ------------ ----- --------------- ----------------- A L (8.50 ; 3.00) (8.50 ; 3.00) R (6.50 ; 4.75) (10.00 ; 5.00) GAME 2 B **l r A** L (9.75 ; 8.50) (9.75 ; 8.50) R (3.00 ; 8.50) (10.00 ; 10.00) Our second treatment variable is related to the nature of player B (the column player) who may be represented either by a human subject (Human condition) or a pre-programmed computer (Robot condition). The Human condition enables us to capture two cardinal breaches of dominance solvability: the failure to use the dominant strategy (player Bs\' behavior) and the failure to best respond to others\' dominant actions (player As\' behavior). However, the latter behavior occurs under strategic uncertainty and thus might stem from two distinct sources: bounded rationality and rational behavior under uncertainty. More precisely, player As may simply have a limited capability of best responding to dominant strategy, but may also intentionally refrain from best responding when in doubt about player Bs\' use of dominant strategy. To separate these two effects, we introduce the Robot condition in which a human subject acting as player A interacts with a computerized player B who is pre-programmed to always choose r. We clearly inform the subjects in the Robot condition that they are interacting with a pre-programmed computer: "the computer chooses r at each round, without exception" (bold in the original instruction sheet). This is the only difference in the rules and procedures between Human and Robot conditions[^10^](#fn0010){ref-type="fn"}. Thus, the key property of the Robot condition as compared to the Human condition is neutralizing strategic uncertainty player As face, while maintaining space for boundedly rational behavior. The design of the experiment is otherwise the same in all four experimental conditions. We explore whether behavior is sensitive to learning by considering ten uniform, one-shot interactions. In order to homogenize incentives across rounds, the following rules are implemented: all games are played in strict anonymity, roles are fixed, and subjects\' payoffs are computed based one randomly drawn round. In the Human condition, players are matched into pairs using a perfect stranger, round-robin scheme, which guarantees that subjects are involved in a series of one-shot interactions despite the repetition of the game[^11^](#fn0011){ref-type="fn"}. Our control variables also include two measures of cognitive skills. Both of them are introduced as part of a post-experimental supplementary task. Subjects\' participation is rewarded with extra five Euros; otherwise, their answers are not incentivized[^12^](#fn0012){ref-type="fn"}. The supplementary task starts with a debriefing question, where subjects are asked to "report any information they find relevant about how their decisions has been made." Then, we implement the following measures of cognitive skills. The first task is the standard Cognitive Reflection Test based on Frederick ([@B25]) which "measures cognitive reflectiveness or impulsiveness, respondents\' automatic response versus more elaborate and deliberative thought" (Brañas-Garza et al., [@B8], p. 255). It contains three questions: 1. A notebook and a pencil cost 1.10 Euros in total. The notebook costs 1 Euro more than the pencil. How much does the pencil cost? 2. If it takes 5 machines 5 min to make 5 widgets, how long would it take 100 machines to make 100 widgets? 3. In a lake, there is a patch of lily pads. Every day, the patch doubles in size. If it takes 48 days for the patch to cover the entire lake, how long would it take for the patch to cover half of the lake? Subjects are informed that this set of three questions should be answered within 30 s (although we allow them to provide answers even after this time has elapsed). In this way, subjects can be classified according to their overall score (that is, the total number of correct answers) which can range from 0 to 3. The second task is Raven\'s progressive matrix test (often called Raven\'s test), a picture based, non-verbal measure of fluid intelligence, that is "the capacity to think logically, analyze and solve novel problems, independent of background knowledge" (Mullainathan and Shafir, [@B38], p. 48). It is widely used by, e.g., psychologists, educators, and the military (Raven, [@B41]). It consists of a series of tasks to be solved within a fixed amount of time. In each task, a subject should pick a single element (among eight options) that best fits a set of eight pictures. The level of difficulty increases from one question to the other[^13^](#fn0013){ref-type="fn"}. In our experiment, each participant is given a series of 16 tasks to be solved within 10 min. Individual scores in Raven\' test are computed as the number of correct answers to the 16 items of the test. 2.1. Experimental procedures ---------------------------- For each game matrix, we run three Human sessions (involving 20 subjects per session: 10 player As interacting with 10 player Bs), and two Robot sessions (involving 20 player As per session interacting with automated player Bs). Subjects are given a fixed fee equal to five euros to compensate participation to the experiment. Upon arrival to the laboratory, participants are randomly assigned to their computers and asked to fill in a short administrative questionnaire containing basic questions about their age, gender, education, etc. Experimental instructions are then read aloud: subjects are informed that they will play multiple rounds of the same game, each round with a different partner, and that their own role will remain unchanged throughout the experiment. Finally, subjects are asked to answer a short comprehension quiz. Once the quiz and any questions from participants are answered, the experiment begins. After each of the ten rounds of the game, subjects are only informed of their own payoffs. Information about past choices and payoffs is updated after each round and displayed at the bottom of the screen. Take-home earnings correspond to the outcome of a single round that is randomly drawn at the end of each experimental session. In addition, the experimental game is followed by supplementary tasks. An additional five euros fee is paid to each subject for completing this part. Immediately after the end of the experimental game, participants are provided with a brief round-by-round summary of their decisions and outcomes, and are asked to provide in a blank space on their computer screens any relevant comments in particular about what might have affected their decisions during the experiment. Subjects are also asked to solve the CRT test and a reduced-form Raven\'s test described above. All the sessions were conducted in February and March 2014. Out of the 200 participants (94 males), 155 were students with various fields of specialization. The majority of subjects (65%) had already taken part in economic experiments. Participants\' average age was 25.6 (st. dev. is 7.5). All sessions took place at the Laboratoire d\'Economie Experimentale de Paris (LEEP) at Paris School of Economics. Subjects were recruited via an on-line registration system based on O[rsee]{.smallcaps} (Greiner, [@B28]) and the experiment was computerized through software developed under R[egate]{.smallcaps} (Zeiliger, [@B47]) and z-Tree (Fischbacher, [@B24]). Sessions lasted about 45--60 min, with an average payoff of roughly 18.83 euros (including a five euros show-up fee and five euros for completing the post-experimental tasks). 3. Results {#s3} ========== Our main experimental results can be summarized as follows. First, in line with the existing literature, we observe systematic and sizable deviations from standard predictions based on the principle of dominance solvability. This phenomenon persists across game matrices and despite repetition. Second, we associate strategic behavior with cognitive skills. We find that Raven\'s test score is a more reliable predictor of strategic behavior than CRT score: whenever the latter predicts behavior, the former does too, but not vice versa. Subjects with higher Raven\'s test scores are more likely to use the dominant strategy and to best respond to other player\'s dominant strategy. Unlike those with low Raven\'s test score, they also react to the presence of strategic uncertainty. 3.1. Aggregate behavior in experimental games --------------------------------------------- Table [4](#T4){ref-type="table"} outlines the main patterns of behavior in our experimental games. The statistical significance of the changes observed in this table is tested by Models 1--3 in Table [5](#T5){ref-type="table"}. We first focus on the aggregate frequency of Pareto-Nash equilibrium (R, r) -- the sole outcome that survives the iterated elimination of (weakly) dominated strategies---found in the Human condition. In both games, we observe substantial deviations from the predictions of this solution concept: overall, players attain the (R, r) outcome 58% of times in Game 1 and 43% in Game 2 (Model 1, H~0~ : β~1~ = 0, p = 0.318). We also observe that efficiency increases over time: in both games, we observe the lowest frequency of (R, r) in the initial round (0.333 in Game 1 and 0.200 in Game 2), whereas the highest frequency of (R, r) occurs in the final round (0.700 in Game 1 and 0.533 in Game 2). ###### Aggregate results. Round Overall --------------------------------------------------------------- ----------- ------------- ------- ------- ------- ------- ------- ------- ------- ------- ------- *Pr* (*R,r*) in the Human condition Game 1 0.333 0.600 0.667 0.700 0.567 0.600 0.433 0.633 0.567 0.700 0.580 Game 2 0.200 0.333 0.400 0.400 0.433 0.500 0.500 0.433 0.467 0.533 0.420 *Pr* (*r*) by player B in the Human condition Game 1 0.767 0.800 0.867 0.900 0.800 0.800 0.700 0.833 0.867 0.800 0.813 Game 2 0.833 0.933 0.900 0.933 1.000 0.933 0.933 0.900 0.900 0.933 0.920 *Pr* (*R*) by player A in the Human condition Game 1 0.500 0.733 0.700 0.767 0.767 0.800 0.700 0.767 0.700 0.867 0.730 Game 2 0.300 0.333 0.400 0.400 0.433 0.533 0.533 0.500 0.500 0.533 0.447 *Pr* (*R*) by player A in the Robot condition Game 1 0.700 0.750 0.750 0.725 0.800 0.800 0.800 0.825 0.800 0.775 0.773 Game 2 0.500 0.575 0.725 0.575 0.800 0.700 0.700 0.775 0.775 0.775 0.690 Columns 1--10 summarize the frequencies of outcomes (defined in rows) as % of all outcomes observed in each round of a given experimental treatment. The last column provides overall results. ###### Aggregate results: statistical support. Model 1 Model 2 Model 3 ------------------------------------- -------------------------------------------- -------------------------------------------- -------------------------------------------- Constant (β~0~) 0.580[^\\\^](#TN1){ref-type="table-fn"} 0.813[^\\\^](#TN1){ref-type="table-fn"} 0.730[^\\\^](#TN1){ref-type="table-fn"} (0.144) (0.032) (0.084) 1\[Game* 2\] (β~1~) --0.160 0.107[^\^](#TN1){ref-type="table-fn"} --0.283[^\\^](#TN1){ref-type="table-fn"} (0.103) (0.044) (0.140) 1\[Robot\] (β~2~) 0.043 (0.103) 1\[Robot\] × 1\[Game* 2\] (β~3~) 0.201 (0.161) N 600 600 1400 R^2^ 0.026 0.025 0.066 Estimates of linear probability models on outcome (R, r) (Model 1), decision r by player B (Model 2) and decision R by player A (Model 3). Standard errors (in parentheses) are clustered at the session level in Human treatments (three clusters per game matrix, six in total) and individual level in the Robot condition (40 clusters per game matrix, 80 in total) and computed using the delete-one jackknife procedure. All models contain a dummy variable set to 1 for game matrix 2 (and 0 for game matrix 1). In Model 3, we also introduce an additional dummy variable set to 1 for Robot condition (and 0 for Human condition) and well as the interaction between these two variables. indicate significance at the 10/5/1% level. To further explore the roots of these deviations, we turn to the aggregate patterns of both players\' behavior in Human and Robot conditions. We focus on three behavioral dimensions of dominance solvability: the use of dominant strategy (captured by player Bs\' behavior in the Human condition) and the ability to best respond to other player\'s dominant action with and without bearing the uncertainty about the latter (which is captured by player As\' behavior in the Human and Robot conditions, respectively). Inefficiency is caused by both players, although their roles differ from one game to another: the scope of inefficient behavior is similar for both players in Game 1, and highly asymmetric in Game 2. Overall, player As select action R with probability 0.730 in Game 1 and 0.447 in Game 2 (Model 3, H~0~ : β~1~ = 0, p = 0.047). However, player As\' behavior happens to be misaligned with player Bs\' actual decisions which follow the opposite trend: the total frequency of action r increases from 0.813 in Game 1 to 0.920 in Game 2 (Model 2, H~0~ : β~1~ = 0, p = 0.060). Importantly, the data from Robot sessions suggest that the uncertainty about player Bs\' behavior is not the only driver of player As\' choices. Player As frequently and systematically fail to best respond to player Bs\' dominant action even when the latter comes with certainty in the Robot condition, although their willingness to select action R increases in both games as compared to the Human condition (to 0.773 in Game 1 and 0.690 in Game 2)[^14^](#fn0014){ref-type="fn"}. The fact that inefficient actions from player As prevail in the absence of strategic uncertainty may suggest that at least some of them are boundedly rational decision makers. In the next section, we analyze how these three behavioral components of dominance solvability vary as a function of players\' cognitive skills. 3.2. Cognitive skills and strategic behavior -------------------------------------------- The average score in Raven\'s test (CRT) is 8.679 out of 16 with SD 3.117 (0.479 out of 3 with SD 0.852). Our experimental sample is properly randomized across treatments regarding both measures. We do not reject the null hypothesis that Raven\'s test scores have the same distributions in all treatments (p = 0.275, Kruskal-Wallis test). A Kruskal-Wallis test applied to the CRT scores leads to the same conclusion (p = 0.502). We also replicate several results from previous studies combining Raven\'s test and CRT regarding the relationship between both scores as well as gender differences (Brañas-Garza et al., [@B8]; Corgnet et al., [@B15]). There is a moderate, yet highly significant correlation between Raven and CRT scores (Spearman\'s ρ = 0.306, p \< 0.001) which suggests that they may have a common source, but do not capture the same cognitive skills. Furthermore, the average score of males is significantly higher than the average score of females (Raven\'s test: 9.382 with SD 0.341 vs. 8.014 with SD 0.384, p = 0.009; CRT: 0.676 with SD 0.111 vs. 0.291 with SD 0.087, p = 0.007; two-sided t-tests)[^15^](#fn0015){ref-type="fn"}. We also observe that many subjects (70%) of our 200 participants fail to provide at least one correct answer in our standard CRT. 16% provide exactly one, 8% -- two, and 6% -- three correct answers. This stands in line with Brañas-Garza et al. ([@B8]) who report the respective frequencies of 67, 23, 9, and 1% for a similar sample size (N = 191), and echoes the scores in the least performant sample reported in a seminal study by Frederick ([@B25]): out of 138 students of the University of Toledo, 64% provide no correct answer, 21% provide one, 10% provide two, and 5% provide three corrects answers. ### 3.2.1. Cognitive predictors of strategic behavior: aggregate results In this part, we study the cognitive correlates of strategic behavior. Figures [1](#F1){ref-type="fig"}, [2](#F2){ref-type="fig"} present the aggregate evolution of behavior as a function of cognitive skills, measured either by CRT score or by Raven\'s test score across roles (player A or player B) and experimental conditions (Human or Robot). ![CRT score and aggregate behavior across rounds and treatments.](fpsyg-07-01188-g0001){#F1} ![Raven\'s test score and aggregate behavior across rounds and treatments.](fpsyg-07-01188-g0002){#F2} In Figure [1](#F1){ref-type="fig"}, the sample is divided into two subsamples: subjects who provided at least one correct answer to CRT (referred to as CRT \> 0) and those who did not (referred to as CRT = 0). The aggregate patterns of behavior weakly differ between the two subsamples. Bootstrap proportion tests fail to reject the null hypothesis that the overall proportions of decision R are the same for both CRT categories in the Human condition (p = 0.126) and in the Robot condition (p = 0.235)[^16^](#fn0016){ref-type="fn"}. The aggregate proportions of decision r, in turn, are found to be statistically different (p = 0.037), subjects with a CRT score zero being less likely to play r than subjects who gave at least one correct answer. In Figure [2](#F2){ref-type="fig"}, we split our sample into three subsamples based on Raven\'s test score (1st tertile: less than 8 correct answers, 2nd tertile: between 8 and 10 correct answers, 3rd tertile: more than 10 correct answers). Although, bootstrap proportion tests suggest that player As\' behavior in the Human condition does not vary significantly between these three subsamples (1st tertile vs. 2nd tertile: p = 0.255, 2nd vs. 3rd: p = 0.580, 1st vs. 3rd: p = 0.565), significant differences arise for both player As in the Robot condition (p = 0.001, p = 0.735, p \< 0.001, respectively) and for player Bs (p = 0.064, p = 0.057, p \< 0.001, respectively). Raven\'s test score seems to have a more systematic association with players\' behavior than CRT score, although both measures fail to predict behavior under strategic uncertainty. ### 3.2.2. Cognitive skills and dominance solvability: regression analysis In what follows, we provide further econometric insights into these preliminary results. Following Brañas-Garza et al. ([@B8]); Corgnet et al. ([@B15]), we use three individual characteristics discussed in the previous section -- gender, Raven\'s test score and CRT score (kept as a dummy variable with value 1 if the subject gave at least one correct answer at the CRT test and 0 otherwise) -- to explain behavior in our experimental games[^17^](#fn0017){ref-type="fn"}. The econometric specification is based on the linear probability model and the estimation procedure is outlined in Jacquemet and Zylbersztejn ([@B35]). We also control for payoff scheme and repetition effects by including game matrix and round dummies. We consider three different outcome variables: player As\' behavior in the Human and the Robot treatment, and player Bs\' behavior in the Human treatment. Given the correlation between CRT and Raven\'s test scores, including both variables in the model might result in multicollinearity and lead to the under-rejection of the nullity of respective coefficients. For each outcome, we first include these two measures separately in Models 1 and 2, while Model 3 includes both variables. This evidence is summarized in Table [6](#T6){ref-type="table"}. ###### Cognitive predictors of strategic behavior: regression analysis. ****Pr** (******R****) by player A Pr (****r****) by player B --------------- ---------------------------------------------------- ------------------------------------------ ------------------------------------------ -------------------------------------------- --------------------------------------------- --------------------------------------------- -------------------------------------------- -------------------------------------------- -------------------------------------------- Const. 0.423[^\\\^](#TN2){ref-type="table-fn"} 0.552[^\\^](#TN2){ref-type="table-fn"} 0.563[^\\^](#TN2){ref-type="table-fn"} 0.573[^\\\^](#TN2){ref-type="table-fn"} 0.242[^\^](#TN2){ref-type="table-fn"} 0.240[^\^](#TN2){ref-type="table-fn"} 0.705[^\\\^](#TN2){ref-type="table-fn"} 0.430[^\\\^](#TN2){ref-type="table-fn"} 0.444[^\\\^](#TN2){ref-type="table-fn"} (0.080) (0.176) (0.197) (0.088) (0.135) (0.135) (0.027) (0.103) (0.099) 1\[CRT\>0\] 0.131 0.152 0.062 (0.024) 0.109[^\^](#TN2){ref-type="table-fn"} 0.046 (0.095) (0.121) (0.102) (0.102) (0.047) (0.036) Raven 0.013 0.018 0.0426[^\\\^](#TN2){ref-type="table-fn"} 0.0434[^\\\^](#TN2){ref-type="table-fn"} 0.0313[^\\^](#TN2){ref-type="table-fn"} 0.0287[^\\^](#TN2){ref-type="table-fn"} (0.025) (0.027) (0.013) (0.012) (0.009) (0.008) 1\[Game 2\] −0.270[^\^](#TN2){ref-type="table-fn"} 0.263 0.266 0.068 0.054 0.056 0.100 0.132[^\^](#TN2){ref-type="table-fn"} 0.129[^\^](#TN2){ref-type="table-fn"} (0.129) (0.139) (0.136) (0.083) (0.076) (0.079) (0.052) (0.056) (0.056) 1\[Male\] 0.132 0.187 0.158 0.096 0.072 0.077 0.025 0.024 0.017 (0.126) (0.100) (0.107) (0.090) (0.076) (0.089) (0.046) (0.046) (0.047) Round: 2 0.133 0.133 0.133 0.063 0.063 0.063 0.067 0.067 0.067 (0.092) (0.092) (0.092) (0.048) (0.048) (0.048) (0.042) (0.042) (0.042) 3 0.150 0.150 0.150 0.138[^\\\^](#TN2){ref-type="table-fn"} 0.138[^\\\^](#TN2){ref-type="table-fn"} 0.138[^\\\^](#TN2){ref-type="table-fn"} 0.083 0.083 0.083 (0.109) (0.109) (0.109) (0.050) (0.050) (0.050) (0.048) (0.048) (0.048) 4 0.183[^\\^](#TN2){ref-type="table-fn"} 0.183[^\\^](#TN2){ref-type="table-fn"} 0.183[^\\^](#TN2){ref-type="table-fn"} 0.050 0.050 0.050 0.117[^\^](#TN2){ref-type="table-fn"} 0.117[^\^](#TN2){ref-type="table-fn"} 0.117[^\^](#TN2){ref-type="table-fn"} (0.070) (0.070) (0.070) (0.047) (0.047) (0.047) (0.048) (0.048) (0.048) 5 0.200[^\^](#TN2){ref-type="table-fn"} 0.200[^\^](#TN2){ref-type="table-fn"} 0.200[^\^](#TN2){ref-type="table-fn"} 0.200[^\\\^](#TN2){ref-type="table-fn"} 0.200[^\\\^](#TN2){ref-type="table-fn"} 0.200[^\\\^](#TN2){ref-type="table-fn"} 0.100 0.100 0.100 (0.089) (0.089) (0.089) (0.045) (0.045) (0.045) (0.052) (0.052) (0.052) 6 0.267[^\\^](#TN2){ref-type="table-fn"} 0.267[^\\^](#TN2){ref-type="table-fn"} 0.267[^\\^](#TN2){ref-type="table-fn"} 0.150[^\\\^](#TN2){ref-type="table-fn"} 0.150[^\\\^](#TN2){ref-type="table-fn"} 0.150[^\\\^](#TN2){ref-type="table-fn"} 0.067 0.067 0.067 (0.088) (0.088) (0.088) (0.054) (0.054) (0.054) (0.049) (0.049) (0.049) 7 0.217[^\^](#TN2){ref-type="table-fn"} 0.217[^\^](#TN2){ref-type="table-fn"} 0.217[^\^](#TN2){ref-type="table-fn"} 0.150[^\\\^](#TN2){ref-type="table-fn"} 0.150[^\\\^](#TN2){ref-type="table-fn"} 0.150[^\\\^](#TN2){ref-type="table-fn"} 0.017 0.017 0.017 (0.098) (0.098) (0.098) (0.047) (0.047) (0.047) (0.048) (0.048) (0.048) 8 0.233[^\\^](#TN2){ref-type="table-fn"} 0.233[^\\^](#TN2){ref-type="table-fn"} 0.233[^\\^](#TN2){ref-type="table-fn"} 0.200[^\\\^](#TN2){ref-type="table-fn"} 0.200[^\\\^](#TN2){ref-type="table-fn"} 0.200[^\\\^](#TN2){ref-type="table-fn"} 0.067 0.067 0.067 (0.088) (0.088) (0.088) (0.057) (0.057) (0.057) (0.056) (0.056) (0.056) 9 0.200 0.200 0.200 0.188[^\\\^](#TN2){ref-type="table-fn"} 0.188[^\\\^](#TN2){ref-type="table-fn"} 0.188[^\\\^](#TN2){ref-type="table-fn"} 0.083 0.083 0.083 (0.113) (0.113) (0.113) (0.047) (0.047) (0.047) (0.060) (0.060) (0.060) 10 0.300[^\\^](#TN2){ref-type="table-fn"} 0.300[^\\^](#TN2){ref-type="table-fn"} 0.300[^\\^](#TN2){ref-type="table-fn"} 0.175[^\\\^](#TN2){ref-type="table-fn"} 0.175[^\\\^](#TN2){ref-type="table-fn"} 0.175[^\\\^](#TN2){ref-type="table-fn"} 0.067 0.067 0.067 (0.115) (0.115) (0.115) (0.056) (0.056) (0.056) (0.049) (0.049) (0.049) R^2^ 0.151 0.141 0.160 0.050 0.139 0.140 0.060 0.108 0.111 Estimates of linear probability models explaining the likelihood of decision R by player A and decision r by player B. Standard errors (in parantheses) are clustered at the session level in the Human condition (three clusters per game matrix, six in total) and individual level in the Robot condition (40 clusters per game matrix, 80 in total) and computed using the delete-one jackknife procedure. Models 1 and 2 include a single measure of cognitive skills (a dummy set to 1 for a positive CRT score, or Raven\'s test score), while Model 3 combines both variables. Other independent variables include gender, game matrix and round dummies. The number of observations is N = 600 for Human and N = 800 for Robot conditions. indicate significance at the 10/5/1% level. We first turn to player Bs\' behavior. Models 1 and 2 suggest that both the coefficient of CRT* \> 0 dummy and the coefficient of Raven\'s test score are positive and significant (p = 0.067 for CRT \> 0 and p = 0.015 for Raven). In Model 3, the coefficient of Raven\'s test score remains highly significant (p = 0.014), while the coefficient of CRT becomes insignificant (p = 0.253). Their joint significance (p = 0.034) implies that cognitive skills predict the use of dominant strategy. We now turn to player As\' behavior in the Human condition. Notwithstanding the previous set of results, cognitive skills are not found to explain player As\' choices. The coefficient of CRT \> 0 dummy is insignificant (p = 0.226) in Model 1, and so is the coefficient of Raven\'s test score (p = 0.633) in Model 2. If we account for both, Model 3 reveals that the coefficients of both scores are neither individually (p = 0.226 for CRT \> 0 and p = 0.550 for Raven\'s test score) nor jointly significant (p = 0.503). Finally, the behavior of player As in the Robot condition is only predicted by Raven\'s test score: unlike CRT \> 0 dummy, its coefficient remains positive and highly significant across models (p ≤ 0.001). Unsurprisingly, the joint insignificance of both coefficients in Model 3 is also rejected (p = 0.003). Altogether, the results presented in Table [6](#T6){ref-type="table"} suggest that cognitive skills predict certain components of strategic behavior: the use of dominant strategy (reflected in player Bs\' behavior), as well as the ability to best respond to other player\'s dominant strategy (reflected in player As\' behavior in the Robot condition). Moreover, in both cases Raven\'s test score is a more reliable predictor of behavior than CRT score. However, we also observe that Raven\'s test score fails to predict player As\' behavior once player Bs\' behavior becomes uncertain, that is once we move from Robot to Human condition. This, in turn, points toward an interplay between the degree of strategic uncertainty, behavior in the experimental games, and individual cognitive skills. Importantly, the existence of such an interplay is also supported by Figure [2](#F2){ref-type="fig"} which shows that the aggregate levels of efficiency shift upwards between the Human condition and the Robot condition for the 2nd and 3rd Raven\'s score tertile, but not the 1st tertile. In order to formally test this conjecture, we now look at the reaction of player As with different cognitive skills to the disappearance of strategic uncertainty. Splitting the data according to Raven\'s score tertile, for each of the three subsamples we compare player As\' behavior in the Human condition to their behavior in the Robot condition by regressing player As\' choice on the Robot dummy (set to 1 for the Robot and to 0 for the Human condition). We also include the previous set of independent variables (except for Raven\'s test score itself). These results are summarized in Table [7](#T7){ref-type="table"}. The coefficient of the Robot dummy captures the effect of eliminating strategic uncertainty on player As\' behavior for each of the three subsamples. This suggests that only player As with high enough cognitive skills are sensitive to the uncertainty about player Bs\' behavior. The behavior of players with low Raven\'s test score (1st tertile) is unresponsive to the degree of strategic uncertainty: the coefficient of the Robot dummy is close to zero and insignificant (p = 0.822). For players with medium scores (2nd tertile), we find a positive yet weakly significant effect (p = 0.087) which becomes amplified and highly significant for those player As whose Raven\'s test score belongs to the 3rd tertile of the experimental sample (p = 0.012). ###### The effect of strategic uncertainty and cognitive skills: evidence from player As\' behavior in Human and Robot conditions. Raven\'s test score tertile --------------- ---------------------------------------- -------------------------------------------- ------------------------------------------ Constant 0.277[^\^](#TN3){ref-type="table-fn"} 0.592[^\\\^](#TN3){ref-type="table-fn"} 0.330[^\\^](#TN3){ref-type="table-fn"} (0.147) (0.065) (0.135) 1\[Robot\] 0.044 0.158[^\^](#TN3){ref-type="table-fn"} 0.428[^\\^](#TN3){ref-type="table-fn"} (0.195) (0.090) (0.155) 1\[CRT\>0\] 0.002 0.038 0.016 (0.262) (0.066) (0.188) 1\[Male\] 0.144 0.033 0.212 (0.145) (0.063) (0.179) 1\[Game 2\] 0.034 −0.245[^\\^](#TN3){ref-type="table-fn"} −0.176 (0.146) (0.092) (0.155) Round dummies Yes Yes Yes N* 480 610 310 R^2^ 0.048 0.173 0.298 Estimates of linear probability models on decision R by player A. Standard errors (in parentheses) are clustered at the session level in the Human condition (three clusters per game matrix, six in total) and individual level in the Robot condition (40 clusters per game matrix, 80 in total) and computed using the delete-one jackknife procedure. Data from Human and Robot conditions are pooled and split into three subsamples based on Raven\'s test score tertiles. Other independent variables include a dummy set to 1 for a positive CRT score, as well as gender, game matrix and round dummies (omitted from the table). indicate significance at the 10/5/1% level. Finally, it is also worth noting that player As\' reaction to the payoff scheme also varies as a function of Raven\'s test score. The coefficient of the Game 2 dummy is close to zero and highly insignificant in the 1st tertile regression (p = 0.890). Then, it becomes negative in 2nd and 3rd tertile models (although it is only statistically significant in the former with p = 0.012 and p = 0.271, respectively). This, in turn, stands in line with the previous finding that player As\' willingness to play R increases as the safe choice L becomes less attractive relative to outcome (R, r). It also seems that the magnitude of this effect is mediated by player As\' cognitive skills, although not in a monotone way. 4. Conclusion {#s4} ============= This paper studies the relationship between strategic behavior and cognitive skills---cognitive reflection and fluid intelligence---in a classic 2 × 2 dominance-solvable game. Our results show that subjects with higher fluid intelligence (measured by Raven\'s progressive matrices test) are more likely to play dominant strategy, and also more likely to best respond to other\'s strategy. Furthermore, fluid intelligence predicts strategic sophistication: only those players with sufficiently high Raven\'s test score are found to display sensitivity to the presence of uncertainty about others\' behavior. Cognitive reflection (measured by CRT), in turn, lacks the power to predict behavior in our experimental setting. We see three main conclusions that stem from these findings. First, these results contribute to the ongoing debate on the relationship between rationality and intelligence (see Stanovich, [@B45], for a critical review). For instance, Stanovich and West ([@B46]) distinguish between two aspects of rational behavior: instrumental rationality which is understood as the "ability to take appropriate action given one\'s goals and beliefs," and epistemic rationality which enables agents to hold "beliefs that are commensurate with available evidence." In the strategic environment investigated in this paper, instrumental rationality can be associated with the ability to solve the game, while epistemic rationality---with the ability to play it with others. Our experimental data suggest an important relationship between fluid intelligence (rather than reflective thinking) and both of these facets of rationality in strategic settings. Both the ability to use dominance and iterated dominance to efficiently solve the game, as well as the responsiveness to the availability of strategic information, is found to be predicted by Raven\'s test score (but not by CRT score). The second contribution is related to the experimental methodology. Despite the fact that CRT and Raven\'s test are both commonly used to measure cognitive skills in experimental subject pools, still very little is known about their relative performance in predicting different types of behavior. Therefore, the choice of one test over the other may happen to be at least as intuitive as evidence-based. As mentioned before, to the best of our knowledge only two experiments address this issue. Brañas-Garza et al. ([@B8]) do so in a strategic environment (p-beauty contest game), while Corgnet et al. ([@B15])---in a non-strategic one (individual choices on wealth distribution). Both studies find that CRT performs better than Raven\'s test in predicting subjects\' behavior. The result of the present experiment points the to the opposite conclusion. We believe that this difference is driven by the very nature of the experimental tasks which may involve different types of cognitive effort. In our view, this issue deserves attention in future research. Finally, although we find evidence that behaving in accordance with dominance solvability is positively correlated with cognitive skills, we also substantiate that most of the variance in individual decision making cannot be explained by such skills. Thus, exploring factors alongside cognitive skills that generate strategic behavior remains an open and important empirical question. An interesting avenue is to disentangle individual determinants, e.g., personal characteristics (such as cognitive skills) that are associated with appropriate behavior, from environmental determinants, that is, those features of the decision making environment that lead decision makers to take certain types of actions. 5. Author contributions {#s5} ======================= NH, NJ, SL, and AZ all contributed equally to this work. Authors are listed in an alphabetical order. 6. Funding ========== This project has received funding from JSPS-ANR bilateral research grant BECOA (ANR-11-FRJA-0002), as well as the LABEX CORTEX (ANR-11-LABX-0042) of Université de Lyon, and LABEX OSE of the Paris School of Economics (ANR-10-LABX_93-01), both within the program "Investissements d\'Avenir" (ANR-11-IDEX-007) operated by the French National Research Agency (ANR). Ivan Ouss provided efficient research assistance. We thank Juergen Bracht, Colin Camerer, Guillaume Fréchette, Haoran He, Asen Ivanov, Frédéric Koessler, Rosemarie Nagel, Ariel Rubinstein, Jason F. Shogren, Jean-Marc Tallon, Antoine Terracol, and Marie Claire Villeval for their comments. NH and NJ gratefully acknowledge the Institut Universitaire de France. SL thanks the School of Business at the University of Western Australia for hospitality and support. A major part of this work was conducted while NH was affiliated with Aix-Marseille University (Aix-Marseille School of Economics, AMSE) and NJ was affiliated with Université de Lorraine (BETA). NH and NJ thank both institutions for their various supports. Conflict of interest statement ------------------------------ The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. ^1^This class of games has been first introduced by Moulin ([@B37]) as the p−beauty contest games, where p (often equal 2/3) stands for the target fraction of all numbers\' average. ^2^See Nagel ([@B39]) and Ho et al. ([@B31]) for early evidence from the laboratory, Costa-Gomes and Crawford ([@B18]) for a laboratory experiment supporting a behavioral model of bounded rationality, and Bosch-Domenech et al. ([@B7]) for related evidence from the field. ^3^Both Camerer ([@B12]) and Crawford ([@B19]) consider this game as a basic example of a dominance-solvable game, and a glaring case of a mismatch between theoretical predictions and actual behavior. ^4^If the game is played sequentially (so that player A moves first), the same solution can be obtained through backward induction. Note that if s \> h, the solution does not change (since l remains player B\'s weakly dominated strategy), but the outcomes are no longer Pareto-rankable. Beard and Beil ([@B5]), Schotter et al. ([@B44]), and Goeree and Holt ([@B27]) find that this environment also generates important violations of standard theoretical predictions. ^5^Cognitive skills are often measured using (amongst others) the Cognitive Reflection Test (CRT, Frederick, [@B25]), the Raven\'s progressive matrices test (Raven, [@B42]), or both (like in this study). The details of these two measures are presented in Section 2. ^6^RMET consists of a series of photos of the area of the face involving the eyes. Subjects are asked to choose one of the four words that best describes what the person in the photo is thinking or feeling. ^7^See Hampton et al. ([@B29]) for related evidence. ^8^Al-Ubaydli et al. ([@B1], [@B2]) also report that individual cognitive skills do not predict individual willingness to reach efficient outcomes in these two game. ^9^Herein, we restrict our design to these two game matrices and do not seek to further investigate the effects of monetary incentives on both players\' behavior. These effects are analyzed in detail in Jacquemet and Zylbersztejn ([@B35]) and Hanaki et al. ([@B30]). ^10^An English translation of the original instructions in French is provided as supplementary material. ^11^See Jacquemet and Zylbersztejn ([@B34]) for a detailed motivation and description of this design. ^12^Absence of monetary incentives for providing corrects answers is a standard procedure for both CRT and Raven\'s tests. Recent evidence on both tests suggests that monetary incentives do not per se affect people\'s performance. See Brañas Garza et al. ([@B9]) for a metastudy on the determinants of CRT scores and Eckartz et al. ([@B21]) and Dessi and Rustichini ([@B20]) for experimental evidence on the role of monetary incentives in Raven\'s test. ^13^See Raven ([@B42]) for an overview. ^14^Model 3 suggests that these two proportions are not significantly different: testing H~0~ : β~1~ + β~3~ = 0 yields p = 0.303. The increase in the proportion of decisions R between Human and Robot conditions is insignificant for Game 1 (H~0~ : β~2~ = 0, p = 0.679) and significant for Game 2 (H~0~ : β~2~ + β~3~ = 0, p = 0.054). ^15^See also Frederick ([@B25]) and Bosch-Domènech et al. ([@B6]) for related evidence. ^16^We test the difference in proportion of a given outcome between two experimental conditions by carrying out a bootstrap proportion test that accounts for within-subject correlation, i.e., the fact that the same individual takes 10 decisions. The procedure consists of bootstrapping subjects and their corresponding decisions over all 10 rounds instead of bootstrapping decisions as independent observations (see e.g., Jacquemet et al., [@B33], for a detailed description of the procedure). ^17^Given that most CRT scores in our sample are null and the higher the score, the less frequent it gets, dichotomizing the CRT score variable limits the impact of the outliers on the overall results. Supplementary material {#s6} ====================== The Supplementary Material for this article can be found online at: <http://journal.frontiersin.org/article/10.3389/fpsyg.2016.01188> ###### Click here for additional data file. [^1]: Edited by: Nikolaos Georgantzis, University of Reading, UK [^2]: Reviewed by: Adriana Breaban, Tilburg University, Netherlands; Gerardo Sabater-Grande, Jaume I University, Spain [^3]: This article was submitted to Personality and Social Psychology, a section of the journal Frontiers in Psychology
*Background.* VRE is an endemic hospital-acquired organism in the US. There are no proven methods or means of decreasing or eliminating carriage of VRE which is a risk factor for invasive disease. Two small clinical studies showed reduction or elimination of VRE colonization using the probiotic LGG but neither study cultured stool for the presence of LGG, raising questions about whether effects on VRE colonization were due to LGG or other factors. *Methods.* To determine whether LGG can eliminate colonization by VRE, we undertook a randomized, double blind, placebo controlled trial in which subjects with VRE received 2 weeks of LGG or placebo. Stool samples were collected and cultured for the presence of VRE and LGG at baseline and days 7, 14, 21, 28 and 56. Presence of LGG was also assessed by PCR on day 14 samples from a subset of 7 subjects. Subjects were monitored closely for adverse events. *Results.* Of 694 patients screened, 11 were randomized. Five subjects received LGG and 6 received placebo. VRE was not eliminated during or after treatment in any subject. No significant decreases in VRE colony counts occurred in the subjects receiving LGG (table). LGG was recovered from stool in 2/5 subjects who received LGG. These 2 subjects were among 3/5 who received LGG and were also treated with antibiotics. LGG was detected by PCR in 4/4 samples from subjects who received LGG and 0/3 subjects in the placebo group. Adverse events were similar in the 2 groups. Median Stool VRE Counts by Study Day -------------------------------------- ------ ------ ------ ----------- ------ -------- ------ Day 0 6.97 4.48 7.72 Day 0 7.18 4.36 9.53 Day 7^a^ 7.49 5.72 8.46 Day 7^b^ 8.11 1.04 8.65 Day 14 6.26 4.51 9.54 Day 14 7.86 2.79 8.56 Day 21 6.90 4.60 7.43 Day 21^b^ 7.28 2.54 8.45 Day 28 6.59 4.18 7.38 Day 28 8.28 6.08 8.83 Day 56^b^ 5.26 3.81 7.52 Day 56^b^ 6.74 \<1^c^ 7.20 ^a^ one positive swab included, no CFU available ^b^ one sample not received ^c^ lowest level of detection based on 1ml plated *Conclusion.* Although the probiotic LGG is safe in patients with comorbid conditions, we were unable to demonstrate that it resulted in elimination of VRE in an RCT, likely because of continued administration of antibiotics in this patient population. *Disclosures.* All authors: No reported disclosures. [^1]: Session: 108. Clinical - Enteric Infections [^2]: Friday, October 10, 2014: 12:30 PM
A stressful life event may affect prognosis of breast cancer directly through stress-induced alterations of the immune and neuroendocrine system and indirectly through changes in health behaviour, such as physical activity, consumption of alcohol, compliance to therapy and coping with disease. In all, six previous studies have addressed the association between stressful life events and breast cancer prognosis: two of which found a significantly increased risk for recurrence ([@bib12]; [@bib10]), two studies found no association ([@bib6]; [@bib8]), whereas two studies found, contrary to what was expected, a significantly lower risk of recurrence ([@bib2]; [@bib5]). These inconsistent findings may be due to the use of different measures of exposure and methodological weaknesses including self-reported measure of exposure, small samples (N=94--665) and selection or recall bias (refer to web appendix: [Supplementary Table A1](#sup1){ref-type="supplementary-material"}). In this large population-based study, of more than 20 000 breast cancer cases, we use objective information from population-based registers and clinical databases to examine the association between a single life event stressor, loss of a partner and breast cancer prognosis. Loss of a partner is a common and also very stressful life event, implying considerable changes to everyday life ([@bib7]). Materials and methods ===================== Linkage of registry data ------------------------ Information on sex, date of birth, current and historical addresses, emigration, disappearance and death with date of these incidences were obtained from the Central Population Register in which all Danish residents since 1968 have been registered with a personal identification number allowing linkage of information between national registers ([@bib11]). Breast cancer ------------- We obtained information on date of breast cancer diagnosis (defined as date of primary surgery), date of recurrence (which is reported up to 10 years after diagnosis), tumour size (in mm), number of tumour-positive lymph nodes, malignancy grade, hormone receptor status and menopausal status, from Danish Breast Cancer Cooperative Group, which contain information on nearly 95% of all breast cancer cases in Denmark since 1977 ([@bib9]). Other cancers ------------- Information on first primary cancer, excluding nonmelanoma skin cancer, was obtained from the Danish Cancer Registry, which since 1943 have registered all cases of cancer (ICD-7) in Denmark ([@bib13]). Stressful life event -------------------- A stressful life event was defined as the death of a cohabiting partner either in the 4 years before breast cancer diagnosis or in subsequent years. 'Cohabitation\' was defined as two persons of the opposite sex over the age of 16, with a maximum age difference of 15 years, living at the same address with no other adults in the residence. Socioeconomic status and comorbidity ------------------------------------ Information on educational level and disposable income (categorised in [Table 1](#tbl1){ref-type="table"}) was obtained 2 years before breast cancer diagnosis from the population-based Integrated Database for Labour Market Research in Statistics Denmark with data on sociodemographic factors in Denmark since 1980 ([@bib14]). Information on comorbidity categorised according to the Charlson comorbidity index (scores 0, 1 and ⩾2), excluding cancers ([@bib3]) was obtained from the Danish National Patient Register with information on all somatic diseases leading to hospitalisation since 1977, and from 1995 also information on all outpatient visits ([@bib1]). Analysed cohort --------------- To attain accurate information from the included registers our base population is restricted to all 3.4 million Danish residents born between 1925 and 1973 who resided in Denmark 1994--2006 and entered the cohort at age 30 (for more details [@bib4]). From this population we identified 22 366 women diagnosed with breast cancer between 1 January 1994 and 31 December 2006 with no previous history of cancer (except nonmelanoma skin cancer), who resided in Denmark 2 years before diagnosis, and had a cohabiting partner up to 4 years before their breast cancer diagnosis. We excluded 1153 women with missing values on one or more covariates or who had resided in Denmark for less than 2 years. In all, 21 213 eligible women were followed for recurrence and death. Statistical analyses -------------------- We used Cox regression to assess hazard ratios (HRs) with 95% confidence intervals (CIs) for all-cause mortality and breast cancer recurrence, respectively, according to the vital status of the partner. For all-cause mortality follow-up time was counted from the date of diagnosis until death, emigration or 31 December 2010, whichever came first. For breast cancer recurrence follow-up time was counted from the date of diagnosis until death, emigration, 10 years of follow-up or 31 December 2006, whichever came first. The exposure, death of partner (after diagnosis), was included as a time-dependent variable, so that person--time before the death of the partner was counted as unexposed, whereas person--time after the date of death of the partner was counted as exposed. In all analyses, time since breast cancer diagnosis was used as the underlying time scale, and baseline hazards were allowed to vary across age at breast cancer diagnosis in 1-year intervals. The HRs were first adjusted for educational level and income, both considered as potential confounders. Subsequently, we adjusted for comorbidity, period of diagnosis, tumour size, number of positive lymph nodes, receptor status and malignancy grade (I--IV), as these factors are strongly associated with outcome, however, not obviously associated with the exposure. We estimated HRs in the intervals \[0--1\],\]1--2\],\]3--4\] years for exposure before diagnosis and for latencies of: \[0--2\],\] 2--5\],\]5--17\] years for exposure after diagnosis. We investigated whether a change in cohabitation status influenced the estimated association, by censoring at 1 January in the year in which the cohabitation status changed. Further, we estimated the association for women diagnosed with hormone-receptor-positive breast cancer and post-menopausal women only. Results ======= In the analysis of the association between loss of partner and all-cause mortality 172 773 person--years of follow-up were accrued, with a median follow-up of 7.7 years (ranging 0--17). During follow-up, 5660 women died, 762 lost their partner in the 4 years before their breast cancer diagnosis and 2259 lost their partner during follow-up. As expected, women who lost their partner were older and had a lower education and income compared with those who did not ([Table 1](#tbl1){ref-type="table"}). After adjustment for education and income as well as period of diagnosis, comorbidity and severity of breast cancer, women who had lost a partner were not at a significantly higher risk for recurrence or all-cause death from that of women who did not lose a partner, no matter if the event happened in the 4 years before diagnosis or in subsequent years ([Table 2](#tbl2){ref-type="table"}) or at different latencies ([Table 3](#tbl3){ref-type="table"}). We found only minor changes to the estimates when censoring at change in cohabitation status and when measuring the association only among women diagnosed with hormone-receptor-positive breast cancer or post-menopausal women (results not shown). Discussion ========== Our results do not support the concern that experiencing a major stressful life event, loss of a partner, negatively affects breast cancer recurrence or all-cause mortality, whether the event occurs in the 4 years before or 0--17 years after diagnosis. Of six previous studies four support this finding ([@bib6]; [@bib2]; [@bib8]; [@bib5]). Two previous studies found a higher risk of recurrence among women reporting stressful life events, however, both were of retrospective design and the observed estimate may reflect differential recall and reporting of events (Ramirez et al, 1989; [@bib10]). We addressed several methodological limitations of the previous studies. First, we included more than 30 times as many cancer patients as the largest study published so far. Second, the use of national registers and databases to identify the study population ensures high representativeness and minimal risk for selection bias. Third, the exposure (death of partner) was measured independently of the participants, eliminating recall bias. Fourth, the register-based cohort design ensures temporality and minimises loss to follow-up and misclassification of outcomes. Finally, we estimated the effect of the death of partner before and after diagnosis on both recurrence and all-cause mortality. Still, loss of partner is relatively rare among newly diagnosed breast cancer patients, and 73% of the cohort was alive and recurrence free at exit, resulting in small number of events in certain subgroups. The observed estimates of all-cause mortality may represent overestimates of breast cancer-specific mortality. We adjusted, however, the analyses for comorbidity and investigated also recurrence as outcome. Finally, an association may be present among women with poor coping resources or accumulated stressful life events. We were unable to take these into account. The death of a partner is a common major life event, and our finding of no association with breast cancer recurrence or all-cause mortality may provide reassurance for women confronting breast cancer. This work was funded by the Health Insurance Foundation; and the Danish Cancer Society. For their support in data management we thank Visti Birk Larsen, MD, and Marianne Steding-Jessen MSci, Survivorship, Danish Cancer Society Research Center. Author contributions All authors have contributed to the conception and design or analysis and interpretation of data and approved the final version of the report. [Supplementary Information](#sup1){ref-type="supplementary-material"} accompanies the paper on British Journal of Cancer website (http://www.nature.com/bjc) The authors declare no conflict of interest. Supplementary Material {#sup1} ====================== ###### Click here for additional data file. ###### Descriptive characteristics at entry of 21 213 women with breast cancer diagnosed in 1994--2006, Denmark, by vital-status of the partner at exit *Partner\'s death before diagnosis (n=762), No. (%)* *Partner\'s death after diagnosis (n=2259), No. (%)* *Partner alive at exit (n=18 192), No. (%)* -------------------------------------------------------- ---------------------------------------------------------- ---------------------------------------------------------- ------------------------------------------------- Sociodemographic characteristics Age at time of diagnosis (years) 30--39 2 (1) 25 (1) 1266 (7) 40--49 40 (5) 219 (10) 4640 (26) 50--59 154 (20) 670 (30) 6926 (38) 60--69 385 (51) 1061 (47) 4501 (25) ⩾70 181 (24) 284 (13) 859 (5) Mean/median 64/65 60/61 54/54 Level of education[a](#t1-fn1){ref-type="fn"} Basic or high school 440 (58) 1204 (53) 7127 (39) Vocational education 227 (30) 703 (31) 6551 (36) Higher education 95 (12) 352 (16) 4514 (25) Disposable income[b](#t1-fn2){ref-type="fn"} Lowest (1st quartile) 340 (45) 536 (24) 2489 (14) Middle (2nd--3rd quartile) 308 (40) 1107 (49) 8468 (47) Highest (4th quartile) 114 (15) 616 (27) 7235 (40) Medical characteristics Tumour size (mm) 0--10 98 (13) 415 (18) 2851 (16) 11--20 297 (39) 963 (43) 7352 (40) 21--50 311 (41) 780 (35) 6560 (36) ⩾51 22 (3) 53 (2) 737 (4) Unknown 34 (4) 48 (2) 692 (4) No. of positive lymph nodes 0 393 (52) 1323 (59) 8979 (49) 1--3 204 (27) 644 (29) 5380 (30) ⩾4 140 (18) 274 (12) 3434 (19) Unknown 25 (3) 18 (1) 399 (2) Malignancy grade Grade I 186 (24) 660 (29) 4197 (23) Grade II 261 (34) 738 (33) 6248 (34) Grade III 130 (17) 321 (14) 3795 (21) Grade IV 49 (6) 85 (4) 861 (5) Unknown 136 (18) 455 (20) 3091 (17) ER and PR Status[c](#t1-fn3){ref-type="fn"} Negative 129 (17) 380 (17) 3882 (21) Positive 588 (77) 1733 (77) 13 254 (73) Unknown 45 (6) 146 (6) 1056 (6) Period of diagnosis 1994--1996 49 (6) 608 (27) 3100 (17) 1997--1999 134 (18) 557 (25) 3792 (21) 2000--2002 217 (28) 580 (26) 4527 (25) 2003--2006 362 (48) 514 (23) 6773 (37) Charlson comorbidity index[d](#t1-fn4){ref-type="fn"} 0 603 (79) 1976 (87) 16 334 (90) 1 100 (13) 176 (8) 1201 (7) ⩾2 59 (8) 107 (5) 657 (4) Highest-education attained 2 years before diagnosis: basic school/high school (7--12 years of primary, secondary and grammar school); vocational training (10--12 years of education); higher education (⩾13 years of education). Disposable income extracted 2 years before diagnosis; income quartiles calculated according to the entire population in the investigated age group, after taxation and interest per person, adjusted for number of people in the household and deflated according to the 2000 value of the Danish crown, as: deflated household income/(no. of person in household 0.6). Negative indicates that individual was estrogen receptor (ER) and progesterone receptor (PR) negative. Positive indicates ER-positive or PR-positive. Accumulated value of all hospital contacts from 1978 to date of diagnosis; scores weighted by level of severity assigned to 19 conditions (excluding cancers) and grouped on the basis of the cumulated sum of scores of 0, 1 and ⩾2. ###### Hazard ratios (HRs) for death and recurrence with 95% confidence intervals (CIs) according to partner\'s vital status among 21 213 women with breast cancer diagnosed in 1994--2006, Denmark Crude[a](#t2-fn1){ref-type="fn"} Adjusted HR[b](#t2-fn2){ref-type="fn"} Adjusted HR[c](#t2-fn3){ref-type="fn"} ------------------------------------------ -------- --------- ------ -------------------------------------- -------------------------------------------- -------------------------------------------- -------------- ------ -------------- All-cause mortality Partner alive at exit (ref.) 18 192 144 410 4977 1.00 1.00 1.00 Exposed in the 4 years before diagnosis 762 5216 243 1.14 (1.00, 1.30) 1.07 (0.94, 1.22) 1.10 (0.95, 1.27) Exposed 0--17 years after diagnosis 2259 23 147 440 1.12 (1.01, 1.25) 1.09 (0.98, 1.20) 1.09 (0.98, 1.22) Recurrence Partner alive at exit (ref.) 19 312 86 453 2635 1.00 1.00 1.00 Exposed in the 4 years before diagnosis 762 2839 59 0.83 (0.64, 1.08) 0.82 (0.63, 1.06) 0.82 (0.61, 1.09) Exposed 0--17 years after diagnosis 1139 7581 85 0.98 (0.78, 1.22) 0.97 (0.78, 1.21) 0.93 (0.73, 1.18) Stratified by age at diagnosis in 1-year intervals. Adjusted for the highest-attained educational level, disposable income, and stratified by age at diagnosis in 1-year intervals (defined in [Table 1](#tbl1){ref-type="table"}). Adjusted for the highest-attained educational level, disposable income, period of diagnosis, comorbidity, tumour size, no. of tumour-positive lymph-nodes, hormone receptor status, malignancy grade and stratified by age at diagnosis in 1-year intervals (N=19 186) (defined in [Table 1](#tbl1){ref-type="table"}). ###### Hazard ratios (HRs) for death and their 95% confidence intervals (CIs) according to partner\'s vital status and time between diagnosis, death and death of partner among 21 213 women with breast cancer diagnosed in 1994--2006, Denmark Crude[a](#t3-fn1){ref-type="fn"} Adjusted HR[b](#t3-fn2){ref-type="fn"} Adjusted HR[c](#t3-fn3){ref-type="fn"} ---------------------------------------------- ----- ------- ----- -------------------------------------- -------------------------------------------- -------------------------------------------- -------------- ------ -------------- All-cause mortality Exposure before diagnosis Time from bereavement to diagnosis (years) \[0--1\] 228 1610 76 1.17 (0.93, 1.48) 1.15 (0.91, 1.44) 1.04 (0.82, 1.34) \]1--2\] 195 1391 65 1.14 (0.89, 1.46) 1.04 (0.81, 1.34) 1.05 (0.80, 1.38) \]2--4\] 339 2215 102 1.11 (0.91, 1.36) 1.04 (0.85, 1.26) 1.19 (0.96, 1.47) Exposure after diagnosis Time from bereavement to death (years) \[0--1\] 647 5768 173 1.20 (0.98, 1.48) 1.16 (0.95, 1.43) 1.11 (0.89, 1.39) \[1--5\] 766 7237 156 1.11 (0.97, 1.28) 1.08 (0.94, 1.23) 1.09 (0.94, 1.26) \>5 846 10142 111 1.07 (0.88, 1.31) 1.04 (0.85, 1.27) 1.08 (0.87, 1.34) Recurrence Exposure before diagnosis Time from bereavement to diagnosis (years) \[0--1\] 228 904 15 0.66 (0.40, 1.10) 0.66 (0.39, 1.09) 0.57 (0.32, 1.01) \[1--2\] 195 800 14 0.71 (0.42, 1.20) 0.69 (0.41, 1.17) 0.74 (0.42, 1.32) \[2--4\] 339 1135 30 1.05 (0.73, 1.51) 1.03 (0.72, 1.49) 1.11 (0.74, 1.65) Exposure after diagnosis Time from bereavement to death (years) \[0--1\] 467 2439 48 0.90 (0.61, 1.34) 0.90 (0.60, 1.33) 0.73 (0.46, 1.17) \[1--5\] 412 2837 28 1.01 (0.76, 1.34) 1.00 (0.75, 1.33) 1.01 (0.75, 1.36) \>5 260 2306 9 1.02 (0.52, 2.00) 1.02 (0.52, 2.00) 1.16 (0.57, 2.37) Stratified by age at diagnosis in 1-year intervals. Adjusted for the highest-attained educational level, disposable income and stratified by age at diagnosis in 1-year intervals. Adjusted for the highest-attained educational level, disposable income, period of diagnosis, comorbidity, tumour size, no. of tumour-positive lymph-nodes, hormone receptor status, malignancy grade and stratified by age at diagnosis in 1-year intervals (N=19 186) (defined in [Table 1](#tbl1){ref-type="table"}). The reference category (HR=1) are not experiencing the death of a cohabiting partner in the given year-interval.
"INTRODUCTION {#sec1-1}\n============\n\nHeat is a physical hazard that could create problems almost(...TRUNCATED)
"1. Introduction {#sec1}\n===============\n\nThe nervous system has two parts: (1) the CNS and (2) t(...TRUNCATED)
"Introduction {#s1}\n============\n\nThe sequencing of the human genome \\[[@pgen-0030099-b001],[@pg(...TRUNCATED)
"Introduction\n============\n\nBipolar disorder (BP) is among the more reliably diagnosed^[@bib1]^ a(...TRUNCATED)
"Introduction\n============\n\nThe metabolic properties of cancer cells are different from those of (...TRUNCATED)

The authors confirm that all data underlying the findings are fully available without restriction. The sequence data generated in this study are available in the Gene Expression Omnibus database at NCBI under the series accession number GSE57326. Whole C. roseus transcriptome sequence along with functional annotation, expression profiling and identified SSRs are available at the Catharanthus Transcriptome web page (<http://nipgr.res.in/mjain.html?page=catharanthus>). Introduction {#s1} ============ *Catharanthus roseus*, popularly known as Madagascar periwinkle, is a medicinal plant which belongs to family Apocyanaceae. The plant is diploid (2n = 16) and native to islands of Madagascar, but now grown in many tropical countries as ornamental plant [@pone.0103583-Magnotta1]. *C. roseus* is well known for its pharmacological importance as it produces more than 130 terpenoid indole alkaloids (TIAs) including vinblastine and vincristine, which are widely used in anti-cancer chemotherapies [@pone.0103583-VanderHeijden1], [@pone.0103583-Guimaraes1]. Most tissues of *C. roseus* are known to produce alkaloids and no other single plant is known to produce such a wide spectrum of alkaloids [@pone.0103583-Blasko1]. The plant is known to treat diabetes also, due to hypoglycemic properties in its tissue extracts [@pone.0103583-Nammi1]. Moreover, roots of *C. roseus* are known to accumulate ajmalicine and serpentine which help controlling blood pressure and cardio-vascular disorders [@pone.0103583-Svoboda1]. Alkaloid biosynthetic pathways are highly branched and complex, with wide differences in alkaloid composition between underground and aerial tissues. TIAs have high commercial value because they are produced by plants in very low amounts and its infusion is very difficult. The common precursor of TIAs, strictosidine, is the central intermediate formed by the condensation of tryptamine (product of shikimate pathway) and secologanin (product of non-mevalonate pathway) involving strictosidine synthase (STR). Pharmacologically important alkaloids, vinblastine and vincristine (found only in aerial tissues) are synthesized *in vivo* by the condensation of vindoline and catharanthine, both of which are obtained from branch-point intermediate cathenamine. Biochemical pathway resulting in formation of vindoline is specifically present in well differentiated aerial tissues of the plant, but not in roots and cell cultures, thereby marking the presence of tissue-specific TIA pathway in *C. roseus* [@pone.0103583-Rischer1]--[@pone.0103583-Jaggi1]. In recent years, next generation sequencing has become the method of choice for fast and cost-effective transcriptome characterization for non-model plants [@pone.0103583-Morozova1]--[@pone.0103583-Garg1]. Earlier, a large-scale transcriptomic resource from three medicinal plants (*Camptotheca acuminate*, *Catharanthus roseus* and *Rauvolfia serpentina*) have been developed for elucidating monoterpene indole alkaloid (MIA) pathways [@pone.0103583-GongoraCastillo1]. Recently, Van Moerkercke et al. [@pone.0103583-VanMoerkercke1] used RNA-seq approach to construct metabolic pathway database, CathaCyc, for *C. roseus*. Such databases can facilitate identification of key regulator(s) for metabolic pathway engineering. To add on to existing resources, in this study, we generated *C. roseus* transcriptome by assembling RNA-seq data generated from *C. roseus* tissues (leaf, flower and root) and merged it with previously reported *C. roseus* transcripts. The updated comprehensive *C. roseus* transcriptome was screened for simple sequence repeats (SSRs) which might be helpful in development of functional molecular markers. We also identified transcription factor (TF) encoding transcripts in *C. roseus* transcriptome as only few TFs are known, which regulate TIA pathway genes. Expression analysis of genes involved in TIA pathways was also undertaken to reveal their tissue-specific expression. Gene ontology (GO) enrichment analysis highlighted the tissue-preferential/specific expression of transcripts in various biological processes. These data will provide a framework for further functional analysis of genes involved in biosynthesis of important alkaloids. Results {#s2} ======= Transcriptome sequencing and preprocessing of data {#s2a} -------------------------------------------------- To generate the transcriptome of *C. roseus*, three tissue samples (leaf, root and flower) were subjected to next generation sequencing using Illumina platform. Of the total ∼347 million reads generated from all the tissue samples, about 343 million reads were found to be of high quality after filtering with NGS QC Toolkit ([Table S1](#pone.0103583.s008){ref-type="supplementary-material"}), having average Phred quality score of at least 30 at each base position. As the short reads obtained may be redundant (due to PCR amplification at library preparation step) and their assembly needs a high-end server with high random access memory (RAM). Therefore, duplicate reads from each sample were removed and about 230 million non-redundant (NR) reads were obtained ([Table S1](#pone.0103583.s008){ref-type="supplementary-material"}). Optimization and validation of transcriptome assembly {#s2b} ----------------------------------------------------- To generate an optimal transcriptome assembly of *C. roseus*, we systematically compared the performance of various *de novo* short read assembly tools, including Velvet, Oases and ABySS. *De novo* assembly of total (343,384,084) and NR (230,715,698) high quality reads was performed employing a two-step approach. In the first step, primary assembly (best k-mer assembly) was generated using Velvet, Oases and ABySS at different k-mer lengths ranging from 31 to 95 ([Table S2](#pone.0103583.s009){ref-type="supplementary-material"} and [S3](#pone.0103583.s010){ref-type="supplementary-material"}). On the basis of several parameters described earlier [@pone.0103583-Garg2], assemblies obtained from respective assemblers at different k-mers were compared. Assemblies generated by Velvet showed a gradual increase in N50 and average read length with the k-mer length, best being at k-93 ([Table S2A and S3A](#pone.0103583.s009){ref-type="supplementary-material"}). Similarly, assembly at k-93 (from total reads; [Table S2C](#pone.0103583.s009){ref-type="supplementary-material"}) and k-87 (from NR reads; [Table S3C](#pone.0103583.s010){ref-type="supplementary-material"}) had higher N50 and average read lengths, and were considered to be the best assembly generated from ABySS. On the other hand, choosing the best assembly generated from Oases (using NR dataset) was a tricky task, as there was not much difference in N50 and average lengths at different k-mers ([Table S2B](#pone.0103583.s009){ref-type="supplementary-material"}). Finally, assembly at k-57 was selected, which had an optimal assembly size (number of contigs; [Table S3B](#pone.0103583.s010){ref-type="supplementary-material"}) and minimum redundant unigenes. Whereas, assembly of Oases at k-61 generated from total dataset had higher N50 and average lengths ([Table S2B](#pone.0103583.s009){ref-type="supplementary-material"}). By taking all the assembly parameters into consideration along with BLAST results, assembly generated by Oases at k-57 using NR short read dataset (NR-Oases k-57), was considered to be the best, which generated 42909 contigs (≥250 bp) of 1161 bp average length and 1990 bp N50 length ([Table 1](#pone-0103583-t001){ref-type="table"}). 10.1371/journal.pone.0103583.t001 ###### Assembly optimization/validation of Illumina data of *C. roseus*. ![](pone.0103583.t001){#pone-0103583-t001-1} Total high quality reads (best k-mer) Non-redundant high quality reads (best k-mer) MPGR assembly Merged assembly[2](#nt102){ref-type="table-fn"} -------------------------------------------------------------------------- --------------------------------------- ----------------------------------------------- --------------- ------------------------------------------------- ------- ------- ------- -------- -------- -------- **Number of contigs** 133650 39010 128716 106736 71190 39276 53017 42909 86726 59220 **Total size (Mb)** 72.78 25.38 165.01 160.92 57.62 25.62 51.86 49.81 107.79 76.03 **Minimum length (bp)** 100 185 100 250 100 185 107 250 251 250 **Maximum length (bp)** 15524 10893 17071 17071 15524 10913 17112 17112 12048 17141 **Average length (bp)** 544.6 650.7 1282 1507.8 809.5 652.3 978.2 1160.8 1243 1283.9 **N50 length (bp)** 1087 848 2161 2205 1400 841 1911 1990 1683 2115 **Contigs with significant similarity** [1](#nt101){ref-type="table-fn"} 63706 28142 76866 73292 41643 28477 22593 21275 66788 32666 Similarity search was done against TAIR10 proteome. Assembly generated by merging best k-mer with MPGR transcriptome using TGICL. Gongora-Castillo et al. [@pone.0103583-GongoraCastillo1] followed a robust approach to generate a comprehensive *C. roseus* transcriptome from different tissues and treatments. Our second step involved merging of primary assembly (best k-mer; NR-Oases k-57) with previously existing *C. roseus* assembly (MPGR de novo assembly; [@pone.0103583-GongoraCastillo1]). Our earlier studies have shown that TGICL software generates optimal merged assemblies [@pone.0103583-Garg1], [@pone.0103583-Agarwal1]. Unigenes from both the assemblies were size selected (≥250 bp) and transcript isoforms from MPGR assembly were removed (retaining the longest isoform) before subjecting to assembly using TGICL program. The merged assembly resulted in a total of 59220 contigs with improved average length (1284 bp) and increased N50 read length (2115 bp; [Table 1](#pone-0103583-t001){ref-type="table"}). Overall, the merged assembly was found to be much better than those reported previously [@pone.0103583-GongoraCastillo1], [@pone.0103583-Kumar2]. To assess the quality of *C. roseus* transcriptome thus obtained, we checked for the presence of publicly available *C. roseus* sequences in recently assembled transcriptome. Out of 287 full-length protein sequences (downloaded from NCBI), 220 (77%) were found to be present in the assembled transcriptome. Moreover, we also checked for the genes involved in various biochemical pathways and found that all the 108 genes previously reported by Van Moerkercke et al. [@pone.0103583-VanMoerkercke1] were represented in our transcriptome data. Similarly, all the 30 enzymes (genes) known to be involved in TIA bio-synthesis [@pone.0103583-VanMoerkercke1], were also present in the *C. roseus* transcriptome generated in this study. As compared with earlier reported transcriptome assemblies of *C. roseus*, we obtained nearly 19% (MPGR assembly) [@pone.0103583-GongoraCastillo1] and 42% (CathaCyc) [@pone.0103583-VanMoerkercke1] novel transcripts in our assembly. *C. roseus* belongs to the clade Asterids, and genome of three plant *species* of this clade, including *Solanum tuberosum* (potato), *Solanum lycopersicum* (tomato) and *Sesamum indicum* (sesame) have been sequenced so far. A BLAST analysis of *C. roseus* transcriptome against proteomes of tomato, potato, cucumber, grapevine and Arabidopsis, and transcriptomes of six known alkaloid producing plants (*Atropa belladonna*, *C. acuminata*, *Cannabis sativa*, *R. serpentine*, *Rosmarinus officinalis* and *Valeriana officinalis*) revealed higher similarity of *C. roseus* transcripts with *R. officinalis* (60.9%) followed by tomato (58.7%), potato (56.3%), cucumber (56.2%) and grapevine (56.1%) ([Fig. S1](#pone.0103583.s001){ref-type="supplementary-material"}). Further, reciprocal BLAST analysis with annotated protein sequences from closest reference genomes (tomato, potato, cucumber and grapevine) and Arabidopsis showed that, although Arabidopsis happens to be distantly related to *C. roseus* in phylogenetic tree but had the highest number of orthologs (15252) as compared to tomato (12118) and potato (11263), which belong to the same clade (Asterids) as that of *C. roseus* ([Fig. S2](#pone.0103583.s002){ref-type="supplementary-material"}). This may be due to availability of better genome annotation of the model plant Arabidopsis. Cucumber and grapevine had the least number of orthologs in *C. roseus*, because they belong to different clades. *C. roseus* transcripts generated above were designated as *C. roseus* tentative consensus (Cr_TC) and were assigned a unique identifier number from Cr_TC00001 to Cr_TC59220. The whole transcriptome sequence is available at Catharanthus Transcriptome Sequence web page (<http://nipgr.res.in/mjain.html?page=catharanthus>). The total size of transcriptome is ∼76 Mb with nearly 65% of the transcripts longer than 500 bp and more than 40% transcripts larger than 1000 bp ([Fig. S3](#pone.0103583.s003){ref-type="supplementary-material"}). Average GC content of *C. roseus* transcriptome was little lower (40.65%) than Arabidopsis (42.5%; [Fig. S4](#pone.0103583.s004){ref-type="supplementary-material"}), and comparable with that of soybean (40.9%) and chickpea (40.3%) [@pone.0103583-Garg2]. Whereas, average GC content of rice was much higher (55%) with respect to *C. roseus* and other dicot plant species analyzed. Functional annotation of *C. roseus* transcriptome {#s2c} -------------------------------------------------- For comprehensive annotation of *C. roseus* transcripts, similarity search was performed against several public databases sequentially. We were able to annotate 38380 (65%) unigenes with confidence (e-value≤1E-05), while others were considered to be *C. roseus* - specific which may be involved in various important biochemical pathways, whose intermediates and enzymes involved have not been catalogued in public repositories as of now. The putative function assigned to the transcripts is available at Catharanthus Transcriptome Sequence web page. Based on their similarity with Arabidopsis genes, *C. roseus* transcripts were assigned GOSlim terms under biological process, molecular function and cellular components categories ([Fig. S5A](#pone.0103583.s005){ref-type="supplementary-material"}). Among the biological process category, maximum number of *C. roseus* transcripts were assigned with the specific term, protein metabolism (13.17%) followed by response to stress (12.02%). GOSlim terms, nucleotide binding (9.81%), hydrolase activity (8.32%), transferase activity (8.03%) and protein binding (7.36%) were most represented under molecular function category. Among the cellular component category, nucleus (19%) followed by other cytoplasmic component (18.8%) were most represented ([Fig. S5A](#pone.0103583.s005){ref-type="supplementary-material"}). Based on COG (cluster of orthologous groups) classification, at least 33422 (56.43%) transcripts could be classified into 25 COG categories. Among the 25 COG categories, the cluster for general function prediction represented the largest group (7124; 21.31%), followed by post-translational modification, protein turnover, chaperones (3809; 11.40%) and signal transduction mechanisms (3380; 10.11%). In addition, 1941 (5.8%) of *C. roseus* transcripts were assigned into the cluster of unknown function ([Fig. S5B](#pone.0103583.s005){ref-type="supplementary-material"}). To elucidate various biochemical pathways represented in the transcriptome, *C. roseus* transcripts were searched against Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database, which aids in studying the role of gene(s) in complex metabolic pathways. In total, 4436 genes (4738 transcripts) were found to be involved in one of the 318 different KEGG metabolic pathways. Few of the metabolic pathways represented with higher number of genes were, ribosome (117 genes), spliceosome (98 genes), biosynthesis of amino acids (96 genes), RNA transport (87 genes), purine metabolism (82 genes), carbon metabolism (80 genes), oxidative phosphorylation (73 genes), pyrimidine metabolism (71 genes) and protein processing in endoplasmic reticulum (71 genes). Along with plant hormone signal transduction (38 genes) pathway, we found genes involved in various alkaloid biosynthesis pathways also, like terpenoid backbone biosynthesis (26 genes), phenylalanine tyrosine and tryptophan biosynthesis (23 genes), ubiquinone and other terpenoid-quinone biosynthesis (17 genes), tryptophan metabolism (10 genes), tropane piperidine and pyridine alkaloid biosynthesis (eight genes), diterpenoid biosynthesis (11 genes), isoquinoline alkaloid biosynthesis (seven genes), sesquiterpenoid and triterpenoid biosynthesis (five genes), monoterpenoid biosynthesis (two genes) and indole alkaloid biosynthesis (one gene). We also identified TF-encoding genes in *C. roseus* transcriptome and found 1820 transcripts representing 73 TF families. Among the 73 families represented, the MYB-domain (116) family TFs were the most abundant followed by AP2-EREBP- (106), WRKY- (78), bHLH- (75) and HB- (75) domain TFs ([Fig. 1](#pone-0103583-g001){ref-type="fig"}). TFs play an important role in secondary metabolite and TIA accumulation [@pone.0103583-Jaggi1], [@pone.0103583-VanMoerkercke1]. TFs known to be involved in TIA pathway like ORCA2, ORCA3, WRKY, MYC2 and zinc finger DNA-binding protein 1 and 2, were represented in the assembled transcriptome of *C. roseus*. ![Number of transcripts representing different transcription factor families in *C. roseus* transcriptome.](pone.0103583.g001){#pone-0103583-g001} Identification of simple sequence repeats {#s2d} ----------------------------------------- Transcriptome resources have been harnessed for mining of SSRs in several plant species. EST-SSRs provide an insight on the density of SSRs in the transcribed region of genome and have higher rates of transferability across species [@pone.0103583-Varshney1]. *C. roseus* transcriptome was screened for SSRs using MISA search tool and a total of 11620 SSRs were identified in 8644 (14.6%) *C. roseus* transcripts. Among the identified EST-SSRs, di-nucleotide repeats were most represented (55.55%), followed by tri-nucleotide repeats with 41.67% (4842; [Fig. 2A](#pone-0103583-g002){ref-type="fig"}). Among di-nucleotide repeats, AG/CT showed highest occurrence (60.84%), followed by AT/AT (31.73%), AC/GT (7.41%) and CG/CG (0.03%). In case of tri-nucleotide repeats, occurrence of various motifs was uniform except for AAG/CTT, which showed highest frequency (34%) and CCG/CGG being the least abundant (2.15%) ([Fig. 2B](#pone-0103583-g002){ref-type="fig"}). Further, we developed a comprehensive SSR marker resource for *C. roseus* by designing forward and reverse primers from their flanking sequences. In total, we could design primers for 7158 (61.6%) SSR repeat-motifs identified, which can be used for the generation of functionally relevant markers in *C. roseus*. The complete list of SSRs identified in *C. roseus* along with primer sequences are available at Catharanthus Transcriptome Sequence web page. ![Identification of simple sequence repeats (SSRs) in *C. roseus* transcriptome.\ (A) Distribution of SSRs in different classes (B) Frequency of most common SSRs motifs is shown by bar graph.](pone.0103583.g002){#pone-0103583-g002} Differential gene expression and gene ontology enrichment analysis {#s2e} ------------------------------------------------------------------ RNA-Seq has been considered to be the method of choice for differential gene expression studies at whole genome level [@pone.0103583-Ozsolak1], [@pone.0103583-Jain2]. In total, approximately 87--90% short reads mapped onto *C. roseus* transcriptome and nearly 84--87% mapped uniquely ([Table 2](#pone-0103583-t002){ref-type="table"}). DESeq package, was used to identify the genes differential expressed in different tissue samples [@pone.0103583-Anders1]. We identified differentially expressed genes among different tissues of *C. roseus* via pairwise comparisons ([Fig. 3A](#pone-0103583-g003){ref-type="fig"}). Leaf and root being the primary site for alkaloid production, differential gene expression in these tissues as compared to flower was analyzed in more detail. A total of 2443 and 2153 genes were differentially expressed in leaf and root, respectively, as compared to flower tissue ([Fig. 3A, B](#pone-0103583-g003){ref-type="fig"}). In leaves, higher number of genes (1635) were down-regulated than 808 up-regulated genes, whereas in roots, there were nearly equal number of up- (1125) and down-regulated (1028) genes ([Fig. 3A](#pone-0103583-g003){ref-type="fig"}). Out of the total 4596 differentially expressed genes in leaf and roots, 679 were common. Among 679 genes commonly differentially expressed in roots and leaves, 72 and 552 genes were up- and down-regulated, respectively ([Fig. 3B](#pone-0103583-g003){ref-type="fig"}). Fifty five genes were found to be up-regulated in one tissue and down-regulated in another tissue. The genes related to photosynthesis were up-regulated in leaf, whereas genes annotated as DNA binding proteins (TFs), disease-resistant proteins and wound response proteins were up-regulated in roots. A heat-map of 1861 up-regulated genes in at least one tissue is represented in [Fig. S6](#pone.0103583.s006){ref-type="supplementary-material"}. Among 1861 up-regulated genes, at least 148 were found to encode for TFs, which are key regulatory proteins. TFs are known to play an important role in accumulation of secondary metabolites in plants [@pone.0103583-Broun1]--[@pone.0103583-Patra1]. Among 148 TFs exhibiting significant differential expression in leaf and root tissues, majority of the transcription factors belonging to AP2-EREBP, HB, MYB, NAC, Tify and WRKY families were up-regulated in root tissues ([Fig. 3C](#pone-0103583-g003){ref-type="fig"}). ![Differential expression analysis of *C. roseus* transcriptome.\ (A) Number of differentially expressed genes in different tissues in pairwise comparisons. Number of up-regulated genes are in bold, while down-regulated are in normal font. (B) Venn diagram showing number of up- and down-regulated (in parentheses) genes in leaf and root tissues as compared to flower. Asterisk represents genes up-regulated in one tissue and down-regulated in another tissue. (C) Heat-map showing expression patterns of differentially up-regulated TF encoding genes in different tissues. The scale at the bottom represents log~2~ fold change. (D) Graphical view showing GO terms associated with biological process enriched in up-regulated genes of leaf. The GO enrichment was performed using BiNGO. Node size is proportional to the number of genes in each category and shades represent the scale denoting significance level (white- no significant difference).](pone.0103583.g003){#pone-0103583-g003} 10.1371/journal.pone.0103583.t002 ###### Mapping of non-redundant high-quality reads on *C. roseus* transcriptome. ![](pone.0103583.t002){#pone-0103583-t002-2} Tissue samples High quality reads Total mapped reads (%) Uniquely mapped reads (%) ---------------- -------------------- ------------------------ --------------------------- **Leaf** 79025564 70974531 (89.81) 68635709 (86.85) **Flower** 78728416 69955703 (88.86) 67774853 (86.09) **Root** 72961718 63571877 (87.13) 61747179 (84.63) We performed GO enrichment analysis to explore the major functional categories in up-regulated genes in leaves and roots. GO terms associated with various biological processes, such as metabolic process, nucleic acid metabolic process and cellular metabolic process were found to be enriched in up-regulated genes of leaf ([Fig. S7A](#pone.0103583.s007){ref-type="supplementary-material"}). Leaves being actively participating in photosynthesis, GO terms associated with photosynthetic process were also significantly enriched in leaves. Apart from these, biological process GO terms, like cellular response to jasmonic acid stimulus, ion homeostasis, carotenoid, isoprenoid and tertraterpenoid metabolic process were significantly enriched in genes up-regulated in leaf ([Fig. 3D](#pone-0103583-g003){ref-type="fig"}). Likewise, response to jasmonic acid stimulus was also significantly enriched in up-regulated genes of root. Thus, supporting the previous remarks that plant hormone jasmonic acid is one of the main drivers of TIA synthesis in *C. roseus* and plant secondary metabolism in general [@pone.0103583-Rischer1], [@pone.0103583-DeGeyter1]. Roots being an underground tissue is subjected to various biotic stresses present in the rhizosphere. GO terms like response to stress, response to biotic stimulus, defense response and response to fungus were significantly enriched in up-regulated root genes under biological process category. Regulation of transcription, cellular amino-acid derivative metabolic process, jasmonic acid biosynthesis, response to chemical stimulus, response to endogenous stimulus and response to salicylic acid were few other biological process GO terms, which were significantly enriched in up-regulated genes of roots ([Fig. S7B](#pone.0103583.s007){ref-type="supplementary-material"}). Expression profiling and validation of genes involved in TIA pathway {#s2f} -------------------------------------------------------------------- For expression analysis, we mapped the short reads of individual sample from our study and previous study [@pone.0103583-GongoraCastillo1] onto our *C. roseus* transcriptome and analyzed expression profile using DESeq software. Complete expression analysis of genes is available at Catharanthus Transcriptome Sequence web page. This resource can be utilized by researchers to look for expression profile of their gene(s) of interest. *C. roseus* (L.) var. Prabal is well known for its high alkaloid content [@pone.0103583-Dwivedi1] and we were interested in expression analysis of important alkaloids (TIA) biosynthesis pathway genes. We used CathaCyc database [@pone.0103583-VanMoerkercke1] for the analysis of these pathways. The common precursor of TIAs, strictosidine, is the central intermediate formed by the coupling of tryptamine (shikimate pathway) and monoterpene secologanin (methyl erythritol phosphate pathway) as shown in [Fig. 4A](#pone-0103583-g004){ref-type="fig"}. The alkaloid, vinblastine, is synthesized by coupling of vindoline and catharanthine, both of which are obtained from branch-point intermediate cathenamine ([Fig. 4A](#pone-0103583-g004){ref-type="fig"}). We identified *C. roseus* transcripts encoding for most of the enzymes catalyzing different reactions involved in these pathways ([Fig. 4A](#pone-0103583-g004){ref-type="fig"}). ![Expression patterns of transcripts involved in TIA biosynthesis.\ (A) Vindoline biosynthetic pathway showing important enzymes involved in different reactions. The IDs of *C. roseus* transcript encoding for the respective enzymes are also indicated. The important intermediates have been highlighted in bold font. (B) Heat-map showing expression patterns of TIA genes in different tissues and treatment. The scale at the bottom of each study represents log~2~ value of RPKM. Transcript IDs are given at left side and their putative annotation is on right side. (C) The correlation of gene expression results obtained from RNA-seq and real time RT-PCR analysis (D) Heat-map showing expression pattern of TF encoding genes in different tissues and treatment. The scale at the bottom of each study represents log~2~ value of RPKM. Transcript IDs are given at left side and their putative annotation is on right side.](pone.0103583.g004){#pone-0103583-g004} We identified 30 genes well-known to be involved in TIA biosynthetic pathways. The BLAST analysis showed that all of these genes were conserved in the sequenced genomes from Asterid clade (tomato and potato) and Arabidopsis at the protein level. However, only 17 (∼57%) and 22 (∼73%) of them exhibited significant similarity with annotated coding region sequences of Arabidopsis and Asterids (tomato and potato), respectively, at the nucleotide level. Further, we analyzed the expression of TIA biosynthetic pathway genes using RNA-Seq data in different tissue samples and treatments reported in our study and previous studies [@pone.0103583-GongoraCastillo1], [@pone.0103583-VanMoerkercke1]. As shown in the [Fig. 4B](#pone-0103583-g004){ref-type="fig"}, majority of the genes of TIA pathway were up-regulated in leaf and root tissues implying that these alkaloids are synthesized mainly in leaves and roots. Similar pattern of expression is also visible in developmental tissues used in the study by Gongora-Castillo et al. [@pone.0103583-GongoraCastillo1]. Both leaf and root tissues share more or less a common gene expression pattern for most TIA pathway genes, except for tabersonine 16-hydroxylase (Cr_TC35206) and deacetylvindoline 4-O-acetyltransferase (Cr_TC35622), which are highly down-regulated in roots and tabersonine 19-hydroxylase (Cr_TC04217), which is highly down-regulated in leaves ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). Apart from tabersonine 16-hydroxylase and deacetylvindoline 4-O-acetyltransferase, decreased expression of desacetoxyvindoline 4-hydroxylase was seen in root and hairy root cultures ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). Down-regulation of these enzymes in root tissues and hairy root cultures is in agreement with previous studies [@pone.0103583-Shukla1], [@pone.0103583-StPierre1], as they participate in terminal reactions for vindoline biosynthesis, which is restricted to aerial tissues. On the other hand, tabersonine 19-hydroxylase, which was found to be up-regulated in roots, further endorses previous finding that it helps to operate an alternate mechanism for tabersonine metabolism in roots by side-chain hydroxylation [@pone.0103583-Giddings1]. Gene expression analysis using RNA-seq data revealed that many of the genes involved in TIA pathway were differentially expressed in root and leaf tissues. To validate these findings, quantitative RT-PCR was performed for at least 10 genes of TIA pathway detected to be differentially expressed in root and leaf tissues. Real-time RT-PCR analysis revealed similar expression patterns of all the selected genes as observed in RNA-seq data. Moreover, the statistical analysis also showed a very good correspondence (correlation coefficient of 0.80) among the results of real time RT-PCR and RNA-seq data analysis as shown in [Fig. 4C](#pone-0103583-g004){ref-type="fig"}. Suspension culture supplemented with yeast extract does not seem to be an attractive approach for TIA production due to lower gene expression as seen in gene expression profile ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). On the other hand, seedlings treated with MeJa showed increased expression, after exposure for longer duration i.e. 5 and 12 days. Expression of TIA genes was relatively higher in hairy root cultures, however it escalated when subjected to MeJa treatment ([Fig. 4B](#pone-0103583-g004){ref-type="fig"}). Expression profiling of TFs known to be involved in TIA pathway across different tissues and treatments revealed that their expression is highly up-regulated in hairy root culture, stem and root ([Fig. 4D](#pone-0103583-g004){ref-type="fig"}). Similar to earlier observations, there was a growth related decrease in TIA transcripts in *C. roseus* and accumulation of bisindole alkaloid content depends on tissue maturity [@pone.0103583-Shukla1], [@pone.0103583-StPierre1], [@pone.0103583-Naaranlahti1], we also found that the expression of TIA genes diminished in the mature leaf tissue. Discussion {#s3} ========== *C. roseus* is widely known for its pharmaceutical potential and has become one of the extensively studied medicinal plants. It is considered to be single biological source of the anti-cancer compounds, vinblastine and vincristine [@pone.0103583-VanderHeijden1], [@pone.0103583-ElSayed1]. Although many good efforts have been made to elucidate the complete pathway of TIA biosynthesis, but few complex steps and intermediate compounds are still unknown. Transcriptome studies, with the advent of next generation sequencing technologies, can help addressing few of these problems via gene discovery. Here, we performed high-throughput sequencing of transcriptome from different tissues of *C. roseus* and used short read assembly tools (Velvet, Oases and ABySS) for *de novo* assembly optimization. A two-step strategy involving merging of best k-mer assembly (from out data) with earlier reported MPGR assembly [@pone.0103583-GongoraCastillo1] was employed to obtain a robust *C. roseus* transcriptome. Based upon various parameters [@pone.0103583-Garg2], assembly generated from Oases at k-mer length of 57 taking NR short reads was considered to be the best. This is in agreement with a previous study by Ghangal et al. [@pone.0103583-Ghangal1] who also reported that assembly generated from NR reads was better than total reads. Merging of best k-mer assembly (NR-Oases-k-57) with MPGR assembly using TGICL further improved the assembly assessment parameters, such as N50 length (2115 bp), average transcript length (1283 bp) and sequence similarity with closely related species. As sequence similarity also marks the completeness of the transcriptome, 77% of the known full-length *C. roseus* proteins were found to be present in our *C. roseus* transcriptome. We also found all the previously reported [@pone.0103583-VanMoerkercke1] genes involved in TIA bio-synthesis (30 genes) represented in our assembled transcriptome. The presence of already reported important alkaloid biosynthetic genes and full-length proteins marks the quality of *C. roseus* transcriptome. Moreover, BLAST analysis with earlier transcriptome sequences of *C. roseus* [@pone.0103583-GongoraCastillo1], [@pone.0103583-VanMoerkercke1] and other related plant proteome and transcriptome sequences revealed a better transcriptome assembly presented in our study. For comprehensive annotation, *C. roseus* transcriptome was subjected to similarity search against various known protein databases. We were able to annotate about 65% of *C. roseus* transcripts. Recently discovered genes encoding geraniol synthase (GES) [@pone.0103583-Simkin1] and iridoid synthase (IS) [@pone.0103583-GeuFlores1], known to be involved in biosynthesis of secologanin (a monoterpenoid alkaloid) from geranyl pyrophosphate were also present in our *C. roseus* transcriptome. Overall, more than 56% of transcripts were classified into 25 COG categories, which is quite higher than other studies [@pone.0103583-Lai1]--[@pone.0103583-Wei1]. We found the category "general function prediction" to be the most represented in COG classification accounting for its need for basic physiological and metabolic functions. The GC content of *C. roseus* transcriptome was found to be very much similar to other dicot plants. Our results concord with earlier findings that there is only marginal variation in average GC content between dicots like Arabidopsis, soybean, tomato, potato, pea and tobacco [@pone.0103583-Carels1]. Many studies have been undertaken to characterize and differentiate different *C. roseus* cultivars using various molecular markers, such as AFLP [@pone.0103583-Kim1], [@pone.0103583-ElDomyati1], RAPD [@pone.0103583-Kim1], [@pone.0103583-Shaw1], ISSR [@pone.0103583-ElDomyati1] and SSRs [@pone.0103583-Shokeen1]--[@pone.0103583-Mishra1]. Microsatellites are co-dominant molecular markers used for marker-assisted selection studies and their identification from high-throughput transcriptome studies have been reported in large number of plant species. A total of 11620 SSRs of 2--6 nucleotides were predicted in *C. roseus* transcripts with di-nucleotides repeats being most abundant followed by tri-nucleotide repeat. This is in accordance with previous studies on *C. roseus*, who also observed more di-nucleotide repeats than tri-nucleotide repeats in the EST datasets [@pone.0103583-Mishra1]. The availability of a large number of SSRs with primer sequences can help large-scale genotyping studies for various applications. Existence of genetic diversity in *C. roseus* have been demonstrated by developing STMS markers [@pone.0103583-Shokeen1], [@pone.0103583-Shokeen2]. Thus, availability of transcriptome for screening of SSRs hold an immense potential for high-throughput genotyping applications in *C. roseus*. TFs are key regulators that can alter the gene expression of several target genes, thereby can regulate metabolic flux. Members of some TF families, such as MYB, AP2-EREBP, WRKY, MYB-related and bHLH, are known to regulate secondary metabolism in plants [@pone.0103583-Kato1]--[@pone.0103583-Zhang1]. Members of the plant-specific AP2-EREBP TF family, namely octadecanoid-derivative responsive Catharanthus AP2-domain protein (ORCA2 and ORCA3), which are known to activate expression of several genes (enzymes) involved in TIA biosynthesis, were identified in our *C. roseus* transcriptome. Recently, Suttipanta et al. [@pone.0103583-Suttipanta1] characterized CrWRKY1 and reported its involvement in the transcriptional regulation of TIA pathway in *C. roseus*. Apart from AP2-domain and WRKY proteins, we also found MYC2, zinc-finger DNA binding protein 1 and 2 in the *C. roseus* transcriptome, which were reported by Van Moerkercke et al. [@pone.0103583-VanMoerkercke1] to be involved in regulation of TIA biosynthesis. Digital expression profiling, a powerful and efficient approach for *in-silico* analysis of gene expression, was employed to determine the expression of genes involved in TIA biosynthesis. When compared with flower, genes involved in TIA biosynthesis were highly active in leaf and root tissues. Except few genes, majority of the TIA pathway genes were up-regulated in leaves and roots, which are the prime source of anticancer and antihypertensive alkaloids, respectively. Our study further confirms previous findings that some of the enzymes involved in late reactions of vindoline biosynthesis are not expressed in cell cultures or in tissues unable to produce vindoline [@pone.0103583-DeLuca1], [@pone.0103583-De1]. As reported earlier by Goklany et al. [@pone.0103583-Goklany1], TIA pathway genes are up-regulated in hairy root cultures elicited with MeJa [@pone.0103583-Kumar1], [@pone.0103583-Raina1], we also observed an increase in gene expression of TIA pathway genes in MeJa induced hairy root cultures. Elicitor, like MeJa are compounds, which induce plant stress response and thereby increasing gene expression and alkaloid biosynthesis. Looking at overall digital expression profiling of TIA pathway genes, we conclude that expression of genes are dependent on plant maturity and are highly expressed in MeJa-elicited hairy root cultures. Nearly equal number of genes were differentially expressed in leaf and roots. Enrichment of GO terms, performed on differentially expressed genes showed that photosynthesis related genes were up-regulated in leaves. GO terms, like jasmonic acid biosynthesis, response to jasmonic acid stimulus, isoprenoid and tetraprenoid metabolic process were also enriched in differentially expressed genes. This further adds on to the findings that MeJa induces expression of TIA pathway genes. In conclusion, we assembled and annotated *C. roseus* transcriptome. Many transcripts harboring microsatellite repeats were identified, which can be used for marker-assisted breeding in *C. roseus*. Differential gene expression and GO enrichment analyses revealed the enrichment of genes involved in secondary metabolite production in leaf tissues, which is the prime source of bisindole alkaloids. Further, expression profiling of TIA genes determined that vindoline exclusively accumulates in aerial tissue of *C. roseus* and exposure to MeJa increases its production. However, deeper understanding of regulatory network governing TIA biosynthesis could help in successful metabolic engineering of alkaloid biosynthesis. The transcriptome resource generated in this study can facilitate understanding of regulatory and metabolic pathways underlying the biosynthesis of alkaloids. Materials and Methods {#s4} ===================== RNA isolation, sequencing and quality filtering {#s4a} ----------------------------------------------- Leaf, root and flower tissues of *C. roseus* L. var. Prabal were harvested from the adult plants grown in field. The tissues were harvested from the plants grown under natural environmental conditions in the experimentation field (28°31′55.3′′N 77°09′54.9′′E) of the National Institute of Plant Genome Research, New Delhi. The field experiments conducted in this study did not involve endangered or protected species and no specific permission was required for these location/activities. The tissues were snap frozen in liquid nitrogen and stored in −80°C until further use. RNA was isolated from tissue samples using TRI reagent (Sigma Life Science, USA). Quantity and quality of RNA samples were measured using Nanodrop (Thermo Fisher Scientific) and Agilent Bioanalyzer (Agilent technologies, Singapore). Sequencing was performed using HiSeq 2000 platform generating paired-end reads of 100 bp length. Stringent quality check was performed on short read datasets by using NGS QC Toolkit v2.3 [@pone.0103583-Patel1] to remove the low quality reads and those having primer/adaptor contamination. Duplicate reads from the dataset were removed using CLC Genomic Workbench (v4.7.2, <http://maq.sourceforge.net/index.shtml>) to obtain NR dataset. *De novo* short read assembly and validation {#s4b} -------------------------------------------- *De novo* transcriptome assembly was performed by using three commonly used short read assemblers, Velvet (v1.2.01) [@pone.0103583-Zerbino1], Oases (v0.2.04) [@pone.0103583-Schulz1] and ABySS (v1.2.6) [@pone.0103583-Simpson1]. All the three assemblers used in this study were run at different k-mer lengths, ranging from 31--95. We employed a two-step approach, the best k-mer assembly obtained from different assemblers was merged with MPGR *C. roseus* transcriptome [@pone.0103583-GongoraCastillo1] and subjected to second round of assembly using TGICL (v2.0) [@pone.0103583-Pertea1] with minimum and maximum overlap length of 40 and 90, respectively. Various assembly parameters kept in consideration for marking the best assembly has been described previously by Garg et al. [@pone.0103583-Garg2]. GC content analysis was done using in-house perl script. To validate the quality of assembled transcriptome, we performed simple and reciprocal BLAST searches at an *E*-value cut-off of ≤1e-05 for identification of best significant match. The proteome sequences of tomato, potato, cucumber, grapevine and Arabidopsis were downloaded from phytozome v9.1 ([www.phytozome.net](http://www.phytozome.net)) and transcriptome sequences of alkaloid producing plants (*A. belladonna*, *C. acuminata*, *C. sativa*, *R. serpentine*, *R. officinalis* and *V. officinalis*) were downloaded from Medicinal plant genomics resource ([www.medicinalplantgenomics.msu.edu](http://www.medicinalplantgenomics.msu.edu)). Functional annotation {#s4c} --------------------- One of the most common approach for annotating transcriptome assembly is similarity search via BLAST. *C. roseus* transcripts were searched against TAIR10 proteome, Uniref90, Uniref100 and non-redundant protein (NCBI-nr) data sets at an *E*-value cut-off of ≤1e-05 for identification of best significant match. GOSlim terms for molecular function, biological process and cellular component were assigned to each *C. roseus* transcripts on the basis of their best match Arabidopsis protein. Similarity search against COG database classified *C. roseus* transcripts among different categories of COG classification system. To look for the genes involved in various pathways, assignment of KEGG Orthology (KO) terms and KEGG pathway construction was performed using KAAS (KEGG Automatic Annotation Server) [@pone.0103583-Moriya1] at default parameters. Read mapping and gene expression analysis {#s4d} ----------------------------------------- For gene expression analysis, high-quality short reads were mapped on to *C. roseus* transcriptome assembly using RNA-seq analysis utility of CLC Genomics Workbench. A maximum of two mismatches were permitted for alignments. Unique read counts for each tissue sample were normalized by calculating the read per kilo-base per million (RPKM) for each transcript. DESeq (v1.10.1) [@pone.0103583-Anders1], a software of R package, was used for differential gene expression analysis. It measures gene expression based on the negative binomial distribution with variance and mean linked by local regression. We calculated the size factor for each sample for normalization of read count data using DESeq. A p-value cut-off of ≤0.05 and at least two-fold change in gene expression was used to identify differentially expressed genes. RPKM values were log~2~ transformed and heat-map showing expression profiles for genes involved in TIA pathway were generated using MultiExperiment Viewer (MeV, v4.8). Hierarchical clustering was performed using Pearson correlation metrics and average linkage rule using MeV. Real-time PCR analysis {#s4e} ---------------------- For real-time PCR analysis, gene-specific primers ([Table S4](#pone.0103583.s011){ref-type="supplementary-material"}) were designed using Primer Express (v3.0) software (Applied Biosystems, USA). Actin was used as an internal control. At least three independent biological replicates with three technical replicates of each biological replicate for each tissue sample were used for analysis. Real-time PCR reactions were carried out essentially following the protocol described previously [@pone.0103583-Garg3]. The correlation between expression profiles of selected genes obtained from real-time RT-PCR and RNA-seq data analysis was determined in R program. Identification of SSR and transcription factors {#s4f} ----------------------------------------------- *C. roseus* transcriptome was screened for the presence of microsatellites (SSRs) using MISA [@pone.0103583-Thiel1]. The number of repeating units considered in this study was, six for di-nucleotides, and five for tri-, tetra-, penta- and hexa-nucleotides. We did not consider mono-nucleotide repeats in this study. Primers for all the identified SSRs were designed using BatchPrimer3 v1.0 (probes.pw.usda.gov/batchprimer3). TFs encoding *C. roseus* transcripts were identified based on the Hidden Markov Model (HMM) profile search of conserved domain present in each TF family as described earlier [@pone.0103583-Garg2]. GO enrichment analysis {#s4g} ---------------------- For GO enrichment analysis, similarity search (BLASTX) was carried out against Arabidopsis proteome and the best hit corresponding to each *C. roseus* transcripts was identified. GO enrichment of different sets of genes was performed using BiNGO tool [@pone.0103583-Maere1] as described previously [@pone.0103583-Singh1]. Supporting Information {#s5} ====================== ###### **Number of** ***C. roseus*** **transcripts showing significant similarity with proteome/transcriptome sequences of closely related/alkaloid producing plants.** (PDF) ###### Click here for additional data file. ###### **Reciprocal BLAST analysis of** ***C. roseus*** **transcripts showing number of orthologous genes in closely related plant species.** (PDF) ###### Click here for additional data file. ###### **Length distribution of transcripts in the** ***C. roseus*** **transcriptome.** (PDF) ###### Click here for additional data file. ###### **GC content distribution in the** ***C. roseus*** **and** ***A. thaliana*** **transcripts.** (PDF) ###### Click here for additional data file. ###### **Functional annotation of** ***C. roseus*** **transcripts.** (A) GOSlim term assignment to the *C. roseus* transcripts in different categories of biological process, molecular function and cellular component. (B) COG function classification of *C. roseus* transcripts. (PDF) ###### Click here for additional data file. ###### **Heat-map showing expression patterns of differentially up-regulated genes in different tissues of** ***C. roseus*** **analyzed in this study.** The scale at the bottom represents log~2~ fold change. (PDF) ###### Click here for additional data file. ###### **Graphical view showing biological process gene ontology term enrichment in up-regulated genes in (A) leaf and (B) root.** The GO enrichment was performed using BiNGO. Node size is proportional to the number of genes in each category and shades represent the scale denoting significance level (white- no significant difference). (PDF) ###### Click here for additional data file. ###### Quality control and duplicate read removal statistics of *C. roseus* libraries. (PDF) ###### Click here for additional data file. ###### *De novo* assembly statistics by different assemblers at different k-mer length using total high-quality reads (a) Velvet (b) Oases (c) ABySS. (PDF) ###### Click here for additional data file. ###### *De novo* assembly statistics by different assemblers at different k-mer length using NR high-quality reads (a) Velvet (b) Oases (c) ABySS. (PDF) ###### Click here for additional data file. ###### Primer sequences used for real-time PCR analysis. (PDF) ###### Click here for additional data file. Authors are thankful to Dr. Rohini Garg for transcription factor analysis. [^1]: **Competing Interests:**The authors have declared that no competing interests exist. [^2]: Conceived and designed the experiments: MJ AKS. Performed the experiments: MV RS. Analyzed the data: MV RG RS. Contributed to the writing of the manuscript: RG MV MJ.

Introduction ============ Manganese (Mn) is an essential element for living organisms. It is the twelfth most abundant element in earth´s crust and is present in rocks, water, soil and food, normally associated with other elements (Santamaria, 2008\[[@R54]\]; Farina et al., 2013\[[@R19]\]). Environmental and occupational exposure to Mn may occur by contact with fungicides, such as Maneb, Manconzeb, methylcyclopentadienyl manganese tricarbonyl (MTT)-an anti-knock agent in gasoline, Mn-ore mining, Mn alloy production and dry alkaline battery manufacture (Mergler and Baldwin, 1997\[[@R42]\]; Mergler, 1999\[[@R41]\]). Dietary ingestion is the main source of Mn for humans and Mn absorption takes place mostly in the gastrointestinal tract where it is homeostatically controlled in the intestinal wall (Au et al., 2008\[[@R4]\]). The brain is especially susceptible to metal intoxication during embryonic development, when it is known that Mn is able to cross the placenta and to be excreted in the maternal milk (Betharia and Maher, 2012\[[@R7]\]). Mn absorption is increased during the neonatal period, when biliary excretion is poorly developed, leading to elevated concentrations of Mn in the brain and other tissues (Aschner and Aschner, 2005\[[@R2]\]). In children, Mn exposure is associated with alterations in psychomotor and cognitive development; furthermore a positive correlation exists between Mn exposure and hyperactivity (Menezes-Filho et al., 2011\[[@R40]\]; Roels et al., 2012\[[@R52]\]; Torres-Agustín et al., 2013\[[@R61]\]). Exposure to high levels of Mn can lead to pathological conditions, including neurodegeneration (Mergler et al., 1994\[[@R43]\]). The mechanisms mediating Mn toxicity are complex and not completely understood. Some of them include: \(1\) Mn accumulation in astrocytes leading to disruption of their ability to promote neuronal differentiation and decreasing glutamate uptake by astrocytes (Erikson and Aschner, 2003\[[@R16]\]; Giordano et al., 2009\[[@R21]\]); \(2\) Mn induced loss of dopaminergic neurons (Stanwood et al., 2009\[[@R57]\]); \(3\) Inhibition of respiratory chain complexes and induction of reactive oxygen species (ROS) (Zhang et al., 2004\[[@R64]\]; Sriram et al., 2010\[[@R56]\]). The use of alternative models in toxicological studies has been growing over the years. The fruit fly *Drosophila melanogaster* has served as a unique and powerful model for studies on human genetics and diseases. Although humans and flies are only distantly related, almost 75 % of disease related genes in humans have functional orthologs in the fly (Deepa et al., 2009\[[@R13]\]; Pandey and Nichols, 2011\[[@R46]\]). Moreover, the fast and external developmental cycle of this organism enable the study of toxicological effects of compounds during the developmental period. All these advantages make flies an appropriate model for studies related with metal toxicity (Bonilla-Ramirez et al., 2011\[[@R10]\]; Paula et al., 2012\[[@R47]\]) and human neurodegeneration (Hirth, 2010). As the embryonic development period is particularly sensitive to Mn exposure, in this paper we aimed to investigate the behavior and biochemical alterations caused by Mn exposure during the embryonic development of *Drosophila melanogaster*, focusing on adaptations in the antioxidant systems and MAPK signaling pathways. The levels of Mn and major essential elements were also determined. Materials and Methods ===================== Reagents -------- Anti-phospho-p38^MAPK^, anti-phospho JNK, anti-phospho ERK, anti ERK and ß-actin antibodies were purchased from Cell Signaling Technology (Danvers, MA, United States). EDTA (CAS 60-00-4), glycine (CAS 56-40-6), tris(hydroxymethyl) aminomethane (CAS 77-86-1) and ammonium persulfate (CAS 7727-54-0) were purchased from Serva (Heidelberg, Germany). L-Glutathione reduced (CAS 70-18-8), 1-chloro-2,4-dinitrobenzene (CAS 97-00-7), sodium orthovanadate (CAS 13721-39-6), manganese (II) chloride tetrahydrate (CAS 13446-34-9), ß-mercaptoethanol (CAS 60-24-2), methanol (CAS 67-56-1), tween 20 (CAS 9005-64-5), potassium phosphate dibasic (CAS 7758-11-4), potassium phosphate monobasic (CAS 7778-77-0), potassium bicarbonate (CAS 298-14-6) and anti-rabbit imunoglobulin antibody, *N,N,N\',N\'*-Tetramethylethylenediamine (CAS 110-18-9), quercetin (CAS 117-39-5), protease inhibitor cocktail for use with mammalian cell and tissue extracts, 5,5´dithiobis(2-nitrobenzoic acid) (CAS 69-78-3), 2´,7´-dichlorofluorescein diacetate (CAS 2044-85-1), glycerol (CAS 56-81-5), resazurin sodium salt (CAS 62758-13-8), triton x-100 (CAS 9002-93-1), sodium chloride (CAS 7647-14-5), albumin from bovine serum (CAS 9048-46-8), HEPES (CAS 7365-45-9), ß-nicotinamide adenine dinucleotide 2´-phosphate reduced tetrasodium salt were obtained from Sigma Aldrich (St. Louis, MO, United States). Bis-acrylamide, hybond nitrocellulose, acrylamide (CAS 79-06-1), sodium dodecyl sulfate (CAS 151-21-3), boric acid (CAS 10043-35-3) were purchased from GE Healthcare Bio-Sciences AB (Uppsala, Sweden). All other reagents were commercial products of the highest purity grade available. Animals ------- *Drosophila melanogaster*(Harwich strain) was obtained from the National Species Stock Center, Bowling Green, OH, USA. The flies were maintained at 25 °C on 12 h light/dark cycle in glass bottles containing 10 mL of standard medium (mixture of 39 % coarse and 32 % medium corn flour, 10 % wheat germ, 14 % sugar, 2 % milk powder, 1 % salt, 1 % soybean flour, 1 % rye flour, a pinch of methyl paraben (99-76-3) and lyophilized yeast. All experiments were performed with the same strain, and both genders were used at random. Animal treatment ---------------- Adults flies were placed in 10 mL of standard medium supplemented with 3 mL of a fresh solution (0.1 mM, 0.5 mM or 1 mM) of manganese chloride (MnCl~2~). In the control group the standard medium were supplemented with 3 mL of ultrapure water. After ten days laying eggs the adult flies were removed. When eggs were newly ecloded, 1 to 3 day old flies were used for all analyses. The MnCl~2~ concentrations were chosen based on previous studies (Bonilla-Ramirez et al., 2011\[[@R10]\]). Locomotor assay --------------- Locomotor activity was determined using the negative geotaxis assay as described by Bland et al. (2009\[[@R9]\]) with minor modifications. Briefly, for each assay, individual flies (1-3 days old) were immobilized on ice and placed separately in a glass tube; this method of immobilization does not affect fly neurology (Deepa et al., 2009\[[@R13]\]). After 15 minutes the flies were gently tapped to the bottom of the tube and the time required to climb up 8 cm of the tube wall was recorded. Each fly was tested 4 times at 1 minute intervals. For each experiment, the climbing mean was calculated. Metal content ------------- Two hundred flies per group were washed three times in ultrapure water and then dried on a filter paper in the incubator at 37 °C for 90 minutes. Flies were digested in closed vessels according to the procedure described previously by Bizzi et al. (2010\[[@R8]\]). Flies (\~ 70 mg) were transferred to quartz vessels with 6 mL of nitric acid 3 mol L^-1^. After closing and capping the rotor, the vessels were pressurized with 7.5 bar of oxygen using the valve originally designed for pressure release after conventional acid sample digestion. Then, the rotor was placed inside a microwave oven (Multiwave 3000 Microwave Sample Preparation System, Anton Paar, Graz, Austria). The system was equipped with eight high-pressure quartz vessels (volume of 80 mL, maximum operational temperature and pressure of 280 °C and 80 bar, respectively). Pressure was monitored in each vessel during all the runs. Microwave heating program was as follows: \(1\) 1000 W, with a ramp of 5 min; \(2\) 1000 W for 10 min; and \(3\) 0 W for 20 min (cooling step). After digestion, the pressure in each vessel was carefully released. The resulting solutions were transferred to polypropylene vials and diluted to 25 mL with water. Determination of calcium (Ca), iron (Fe), potassium (K), magnesium (Mg), manganese (Mn), sodium (Na), phosphorus (P), sulfur (S), and zinc (Zn) was performed using an inductively coupled plasma optical emission spectrometer (Optima 4300 DV, PerkinElmer, Shelton, USA) with axial view configuration. A concentric nebulizer and cyclonic spray chamber were used. Argon 99.996 % (White Martins, São Paulo, Brazil) was used for plasma generation, nebulization and as auxiliary gas. The instrumental parameters were carried out in according with previous work (Pereira et al., 2013\[[@R48]\]). Two readings were averaged to give one value per biological replicate and expressed as a mean (±) standard deviation of the mean (SD). Metals levels were expressed relative to the weight of flies used for analysis (µg metal/g of dried weight tissue). Cellular viability ------------------ Cellular viability was measured by two different methods. Firstly, cellular viability was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay as described by Sudati et al. (2013\[[@R58]\]) with minor modifications. The analysis was performed on the whole body of female flies. The flies were incubated in MTT for 60 min (37 °C), after MTT was removed the sample was incubated in DMSO for 30 min (37 °C). The absorbance from formazan dissolution by addition of DMSO was monitored in an EnsPireR multimode plate reader (PerkinElmer, USA) at 540 nm. The second method used was the resazurin reduction assay. The method uses the indicator resazurin to measure the metabolic capacity of cells. Viable cells reduce resazurin into resorufin, a fluorescent compound (Franco et al., 2009\[[@R20]\]). Groups of 40 flies were mechanically homogenized in 1 mL 20 mM Tris buffer (pH 7.0) and centrifuged at 1,000 RPM for 10 min at 4 °C. The supernatant was incubated in Elisa plates with 20 mM Tris buffer (pH 7.0) and resazurin for two hours. The fluorescence was recorded using EnsPire^R^ multimode plate reader (Perkin Elmer, USA) at~ex~579nm and ~em~584nm. DCF-DA oxidation assay ---------------------- Groups of 20 flies were mechanically homogenized in 1 mL 20 mM Tris buffer (pH 7.0), and centrifuged at 1,000 RPM for 10 min, 4 °C. The supernatant was used to quantify 2'-7'-dichlorofluorescein diacetate (DCF-DA) oxidation as a general index of oxidative stress as described by Perez-Severiano et al. (2004\[[@R49]\]). The fluorescence emission of DCF resulting from DCF-DA oxidation was monitored at regular intervals (~ex~488nm and ~em~530nm) in an EnsPire^R^ multimode plate reader (PerkinElmer, USA). Determination of gene expression by real-time quantitative PCR (qPCR) --------------------------------------------------------------------- Real-time quantitative PCR (qRT-PCR) was performed according to the method described by Paula et al. (2012\[[@R47]\]). The primers utilized are shown in Table 1[(Tab. 1)](#T1){ref-type="fig"}. All samples were analyzed as technical and biological triplicates with a negative control. Threshold and baselines were manually determined using the StepOne Software v2.0 (Applied Biosystems, NY). SYBR fluorescence was analyzed by StepOne software version 2.0 (Applied Biosystems, NY), and the CT (cycle threshold) value for each sample was calculated and reported using the 2^-ΔΔCT^ method (Livak and Schmittgen, 2001\[[@R38]\]). The GPDH gene was used as an endogenous reference showing no alterations in response to the treatment. For each well, analyzed in quadruplicate, a ΔC~T~ value was obtained by subtracting the GPDH C~T~ value from the C~T~ value of the interest gene (sequences of tested genes are represented in Table 1). The ΔC~T~ mean value obtained from the control group of each gene was used to calculate the ΔΔCT of the respective gene (2^-ΔΔCT^). Enzyme assays ------------- For enzyme activity measurements, groups of 40 flies were mechanically homogenized in 1 mL 20 mM HEPES buffer (pH 7.0), and centrifuged at 14.000 RPM for 30 min at 4 °C (Franco et al., 2009\[[@R20]\]). The supernatant was used for determination of Glutathione S-Transferase (GST), Catalase (CAT), Superoxide Dismutase (SOD) and Thioredoxin Reductase (TrxR). The GST activity was assayed following the procedure of Jakoby and Habig (1981\[[@R29]\]) using 1-chloro-2,4-dinitrobenzene (CDNB) as substrate. The assay is based on the formation of the conjugated complex of CDNB and GSH at 340 nm. The reaction was conducted in a mix consisting of 100 mM phosphate buffer (pH 7.0), 1 mM EDTA, 1 mM GSH and 2.5 mM CDNB. CAT activity was assayed following the clearance of H~2~O~2~ at 240 nm in a reaction media containing 50 mM phosphate buffer (pH 7.0), 0.5 mM EDTA, 10 mM H~2~O~2~, 0.012 % TRITON x100 as described by Aebi (1984\[[@R1]\]). SOD activity assay was performed as described by Kostyuk and Potapovich (1989\[[@R35]\]). The assay consists in the inhibition of superoxide-driven oxidation of quercetin by SOD at 406 nm. The complete reaction system consisted of 25 mM phosphate buffer (pH 10), 0.25 mM EDTA, 0.8 mM TEMED and 0.05 µM quercetin. TrxR activity was assayed as described by Holmgren and Björnstedt (1995\[[@R26]\]). The test is based on the reduction of oxidized thioredoxin (Trx-S~2~) to reduced thioredoxin \[Trx-(SH)~2~\], using NADPH at 412 nm in a reaction media containing 0.1 M phosphate buffer (pH 7.0), 10 mM EDTA, 5 mM DTNB, 0.2 mg/mL BSA, 0.2 mM NADPH. All enzyme activities were performed at room temperature (25 ± 1 °C) using a Thermo Scientific Evolution 60s UV-Vis spectrophotometer. The enzyme activities were expressed in milliunits per milligram of total protein content, which was quantified following Bradford (1976\[[@R11]\]). Western blotting ---------------- Quantification of of the phosphorylation of mitogen-activated protein kinases (MAPKs) was performed by Western blotting as described by Posser et al. (2009\[[@R50]\]) with minor modifications. Groups of 40 flies were mechanically homogenized at 4 °C in 200 µL of buffer (pH 7.0) containing 50 mM Tris, 1 mM EDTA, 20 mM Na~3~VO~4~, 100 mM sodium fluoride and protease inhibitor cocktail. The homogenate were centrifuged at 4000 RPM for 10 min at 4 °C and the supernatants collected. After protein determination following Bradford (1976\[[@R11]\]), 4 % SDS solution, ß-mercaptoethanol and glycerol was added to samples to a final concentration of 100, 8 and 25 %, respectively and the samples frozen for further analysis. Proteins were separated using SDS-PAGE with 10 % gels, and then electrotransferred to nitrocellulose membranes (Paula et al., 2012\[[@R47]\]). Membranes were washed in tris-buffered saline with Tween (100 mM tris-HCl, 0.9 % NaCl and 0.1 % Tween-20, pH 7.5) and incubated overnight at 4 °C with specific primary antibodies (anti-phospho-p38^MAPK^, anti-phospho JNK, anti-phospho ERK, anti ERK and anti ß-actin). Following incubation, membranes were washed in tris-buffered saline with Tween and incubated for 1 h at 25 °C with anti rabbit-IgG secondary specific antibodies. Antibody binding was visualized using the ECL Western Blotting substrate Kit (Promega). Band staining density was quantified using the Scion Image software (Scion Image for Windows) and expressed as the percentage (%) of the control group (mean ± standard deviation of the mean). The values were normalized using total proteins (total ERK and ß-actin). Statistical analysis -------------------- Statistical analysis was performed using one-way ANOVA followed by Tukey´s post hoc test. Pearson's correlation test was applied for detection of significant statistical differences among the metals. Differences were considered statistically significant when p \< 0.05. GraphPad Prism 5 Software was used for artwork creation. Results ======= Exposure to Mn causes hyperactive behaviors and alters metal levels in Drosophila melanogaster ---------------------------------------------------------------------------------------------- The evaluation of climbing behavior performance by negative geotaxis showed that flies exposed to 0.5 mM and 1 mM of Mn reached the limit of columns significantly (p \< 0.005) faster than controls (Figure 1[(Fig. 1)](#F1){ref-type="fig"}). Levels of Mn and other essential metals were measured in *D. melanogaster* exposed to Mn. At concentrations of 0.5 and 1 mM, the levels of Mn in flies increased almost two and three fold respectively when compared with the control group (Table 2[(Tab. 2)](#T2){ref-type="fig"}), whereas Ca, Cu, Na and Zn levels decreased significantly in flies treated with Mn. Statistically, a significant negative relationship between Mn uptake and levels of Ca (r = -0,7966), Fe (r = -0,6635), Cu (r = -0,8028), S (r = - 0,6018) and Zn (r = -0,9802) occurred in response to Mn treatment (Table 3[(Tab. 3)](#T3){ref-type="fig"}). Mn exposure during embryonic development decreased cell viability and increased ROS production in flies ------------------------------------------------------------------------------------------------------- Cell viability was evaluated through two different tests, MTT and Resazurin. Both procedures showed a significant drop in cell viability at higher concentrations of Mn, confirming its toxicity at the cellular level (Figure 2A and 2B[(Fig. 2)](#F2){ref-type="fig"}). Many factors have been implicated in Mn-induced neurotoxicity, among them the oxidative stress caused by dopamine oxidation, or its ability in interfering with cellular respiration (Aschner and Aschner, 2005\[[@R2]\]). In this study, exposure to Mn leads to an increase in DCF-DA oxidation, a general index of oxidative stress from 0.5 mM (Figure 3[(Fig. 3)](#F3){ref-type="fig"}). Mn increased CAT and SOD mRNA expression, without altering their enzymatic activity ----------------------------------------------------------------------------------- Expression of mRNA for Cat, Sod and Hsp83 (an homolog of mammalian HSP90) (Bandura et al., 2013\[[@R5]\]) was analyzed by qRT-PCR using specific primers (Table 1[(Tab. 1)](#T1){ref-type="fig"}) in flies treated with 1 mM Mn. CAT and SOD mRNA expression was 250 % higher in treated flies whereas Hsp83 expression was 1000 % higher (Figure 4[(Fig. 4)](#F4){ref-type="fig"}). The levels of the antioxidant enzymes activity TrxR, GST, SOD and CAT were determined TrxR and GST activity were increased at concentrations of 0.5 mM and 1 mM (Figure 5[(Fig. 5)](#F5){ref-type="fig"}), while CAT and SOD activity showed no significant differences. Mn exposure inhibited p38MAPK phosphorylation --------------------------------------------- MAPKs phosphorylation levels were investigated in flies exposed to Mn. There was a 40 % inhibition of p38^MAPK^ phosphorylation in flies exposed to Mn at 1 mM, while the phosphorylation level of extracellular signal-regulated kinases (ERK) was unaltered. C-Jun-N-terminal Kinases 2 (JNK2) phosphorylation was not statistically different from controls (Figure 6[(Fig. 6)](#F6){ref-type="fig"}). Discussion ========== In this study, we determined the biochemical and behavioral alterations in *Drosophila melanogaster* in response to Mn exposure during embryonic development. Mn is an essential element required in key biological processes; however, high levels of Mn are associated with neurological and neuropsychiatric disorders (Mergler, 1999\[[@R41]\]). The risk of Mn overexposure comes from both occupational and environmental sources (Mergler and Baldwin, 1997\[[@R42]\]). Mn intoxication, a syndrome known as Manganism, is characterized by an extrapyramidal dysfunction and neuropsychiatric symptomatology and is associated with prolonged occupational exposure to high concentrations of this metal. Classical symptoms include irritability, intellectual deficits, compulsive behaviors, tremors and cock-like walk (Mergler, 1999\[[@R41]\]; Roth, 2006\[[@R53]\]). In rodents, Krishna et al. (2014\[[@R36]\]) showed that adult mice exposed to Mn through the drinking water presented neurobehavioral deficits and glial activation related with Mn deposition in brain. Moreover, others studies demonstrated that Mn toxicity in rats is accompanied by increased cholesterol biosynthesis and impairments in neuronal function of the hippocampus, which is involved in learning and memory (Öner and Sentürk, 1995\[[@R45]\]; Sentürk and Öner, 1996\[[@R55]\]). It has been shown that Mn supplementation during the neonatal period of rats resulted in increased Mn concentrations in tissues leading to adverse effects on motor development and behavior (Tran et al., 2002\[[@R62]\]). Mn uptake is increased during the neonatal period as biliary excretion, which has been suggested as a pathway for Mn elimination from the body, is poorly developed at this stage (Aschner and Aschner, 2005\[[@R2]\]). Exposure to Mn during the embryonic and early postnatal periods may result in increased levels of Mn in the brain and other tissues including bone, liver, pancreas and kidney (Aschner and Aschner, 2005\[[@R2]\]; Roels et al., 2012\[[@R52]\]). Higher levels of Mn retention *in utero* may affect children´s psychomotor development (Takser et al., 2003\[[@R59]\]). Possible adverse effects of Mn exposure on children´s health include cognitive deficits and hyperactive behaviors (Menezes-Filho et al., 2009\[[@R39]\]; Torres-Agustín et al., 2013\[[@R61]\]). Children exposed to high levels of Mn during the fetal period were more impulsive, inattentive, agressive, defiant, disobedient, destructive and hyperactive (Ericson et al., 2007\[[@R15]\]). It is recognized that factors such as the source and the duration of exposure, as well as nutritional status, can interfere in the intensity and incidence of neurological symptoms associated with Mn exposure in humans. Chronic consumption of drinking-water containing Mn at levels ranging from 81 to 2300 µg/l was associated with progressively higher prevalence of neurological symtoms (Kondakis et al., 1989\[[@R33]\]). The concentrations used in this study were from 0.1 mM of Mn in food (corresponding to 19 mg/L in the medium). Despite the use of relatively elevated concentrations, body levels of Mn were not altered at 0.1 mM. In previous studies, adult flies were acutely exposed to Mn (0.5-20 mM) diluted in sucrose, as the only source of food and liquid, which lead to significant locomotor deficits (Bonilla-Ramirez et al., 2011\[[@R10]\]). Our study is the first where Mn was provided as a cereal based diet over all the embryonic period. Thus, more studies are necessary to understand the rate of Mn uptake from diet in flies and how it may affect neurological behaviour. In our study, flies exposed to Mn at 0.5 mM and 1 mM showed increased locomotor speed in the locomotor behavior test (assayed as negative geotaxis behavior), pointing to a hyperactive-like behavior in *Drosophila melanogaster*. Furthermore, Mn levels were substantially increased in treated flies, while Ca, Cu, Zn, Fe and S levels were all decreased. This relationship may be in part associated with a competition of the metals for the same mechanism of transport into the flies cells. Facilitated diffusion, active transport, divalent metal transport 1 (DMT1), ZIP8 and transferrin (Tf)-dependent transport mechanisms are all involved in cellular Mn transportation (Aschner et al., 2007\[[@R3]\]). Among these metal transport systems, DMT1 has a very broad substrate specificity and is likely to be the major transmembrane protein responsible for the uptake of a variety of divalent cations, including Mn^2+^, Cd^2+^, Zn^2+^, Co^2+^, Ni^2+^, Cu^2+^ and Pb^2+^ (Gunshin et al., 1997\[[@R22]\]). In the flies, many proteins involved in the metabolism of biometals such as ferritin, transferrin, iron regulatory proteins, divalent metal transporter are expressed (Bonilla-Ramirez et al., 2011\[[@R10]\]). In this context, Mn uptake is less frequently studied in comparison with other metals and the mechanisms related to Mn transport are considerably more complex, occurring in most part in the divalent (II) and oxide forms (Tebo et al., 2004\[[@R60]\]). Mn has the capacity to interact and /or compete with Ca (Dittman and Buchwalter, 2010\[[@R14]\]). In a study performed in the aquatic insect *Hydropsyche sparna*, Mn exposure decreased cadmium (Cd) and Zn accumulation. Furthermore, increased Ca concentrations significantly reduced Mn accumulation in the insect (Poteat et al., 2012\[[@R51]\]). Dittmanand and Buchwalter (2010\[[@R14]\]) suggested that Mn is also absorbed by Ca transporters in aquatic insects, where increasing ambient Ca concentrations decrease Mn accumulation. There was also a negative correlation between Mn and Fe levels. Iron deficiency has been suggested as a possible contributing cause of attention deficit and hyperactivity disorder (ADHD) in children (Konofal et al., 2008\[[@R34]\]). Concomitantly, children with ADHD showed elevated serum Mn concentrations (Konofal et al., 2008\[[@R34]\]). Recent studies have suggested that Mn accumulates in dopaminergic neurons via the presynaptic dopamine transporter (DAT) and an altered functioning of the dopaminergic system has been well established in the etiology of ADHD (Farias et al., 2010\[[@R18]\]). Our results showed decreased cell viability using two different methodologies and increased ROS generation in flies exposed to Mn during development. Previous work by our group in PC12 cells, demonstrated that Mn leads to increased production of hydrogen peroxide (H~2~O~2~) (Posser et al., 2009\[[@R50]\]). H~2~O~2~ is a highly permeable and reactive molecule being able to react with metals such as Fe, thus generating hydroxyl radicals (Jiménez Del Río M and Vélez-Pardo, 2004\[[@R30]\]; Barbosa et al., 2010\[[@R6]\]) resulting in a propagation of oxidative damage in cells. Tissues can respond to oxidative stress by modulating antioxidant defenses (Halliwell and Gutteridge, 2007\[[@R23]\]). We measured the gene expression of Hsp83, CAT and SOD in *Drosophila melangaster*. Earlier studies showed that cellular stress may induce heat shock proteins in parallel with ROS production (Kim et al., 2004\[[@R32]\]; Paula et al., 2012\[[@R47]\]). Our results showed a significant increase in Hsp83 mRNA levels in Mn treated flies. Previously, our group demonstrated that exposure of flies to heavy metals such as mercury causes increases in the expression of Hsp83 (Paula et al., 2012\[[@R47]\]). CAT and SOD mRNA levels were significantly increased by Mn, but the enzymatic activity of these proteins was unchanged. The antioxidant ezyme SOD converts superoxide radicals (O~2~^º-^) to H~2~O~2~ and CAT catalyzes the conversion of H~2~O~2~ to oxygen (O~2~) and water (Barbosa et al., 2010\[[@R6]\]), thus neutralizing these reactive species. Considering that both SOD and CAT are crucial in the cell defense against oxidative stress (Halliwell and Gutteridge, 2007\[[@R23]\]), it might be expected that a posttranscriptional regulation mechanisms could maintain adequate levels of these proteins, however, further studies are necessary to elucidate this. Our results also showed that the activity of TrxR and GST was enhanced in Mn exposed flies. These two enzymes play an important role in protection against oxidative stress (Mustacich and Powis, 2000\[[@R44]\]). GST is a complex group of phase II detoxifying enzymes that participate in the metabolism of electrophilic substances, including carcinogenic, mutagenic and toxic compounds (Hayes et al., 2005\[[@R24]\]). TrxR is a dimeric FAD-containing enzyme that catalyzes the NADPH-dependent reduction of the active-site disulfide in Trx-S~2~ to give a dithiol in Trx-(SH)~2~ (Zhao et al., 2002\[[@R65]\]). Thioredoxin consists in one of the major redox-regulating proteins displaying a number of biological activities, including cytoprotection against ROS, protein repairing and protein disulfide reduction and modulation of signaling pathways (Yan et al., 2012\[[@R63]\]). Our data suggest that the increase in the levels of Mn and TrxR activity could represent a response to oxidation of thioredoxin in response to Mn-induced oxidative stress. MAPKs regulate the activity of a range of transcription factors thereby controlling gene expression and cellular function. The three most-studied MAPKs are ERK1/2, JNK1/2 and p38^MAPK^ (Ichijo, 1999\[[@R27]\]). ASK1 (Apoptosis Signaling Kinase 1) is a member of mitogen activated protein kinase kinase family (MAPKK) and an upstream activator of MAPK signaling pathway (Yan et al., 2012\[[@R63]\]). The redox state of thioredoxin regulates ASK1. Under normal conditions, ASK1 is bound to and inhibited by thioredoxin and when thioredoxin is oxidized, ASK1 can be activated and apoptotic signaling through the p38 ^MAPK^/JNK1/2 MAPKs initiated (Ichijo et al., 1997\[[@R28]\]). Studies conducted by Yan et al. (2012\[[@R63]\]) in a pancreatic carcinoma cell line, showed inhibition of TrxR by indolequinones resulting in a change of Thioredoxin-1 redox state to an oxidized form and activation of p38^MAPK^/JNK1/2. Similarly, Cd treatment activated ASK1 and its downstream MAPK in neuronal cells (Kim et al., 2005\[[@R31]\]), and inhibits components of thioredoxin system (Chrestensen et al., 2000\[[@R12]\]), while that Liedhegner et al. (2011\[[@R37]\]) demonstrated that knockdown of ASK1 as well as chemical inhibition of p38^MAPK^ and JNK played protective effects against L-DOPA induced apoptosis. We show that Mn induced TrxR activity while p38^MAPK^ /JNK1/2 phosphorylation were inhibited. This suggests an involvement of thioredoxin system in the mechanism of Mn induced toxicity. Augmented TrxR activity may represent a cellular response to high levels of ROS induced by Mn exposure, thus preventing the oxidation of thioredoxin and its dissociation of ASK1. This could contribute to diminished activation of p38^MAPK^ pathway upstream kinases resulting in lower levels of phosphorylation of this MAPK thus minimizing apoptotic cell death. In summary, our study demonstrate for a first time that developmental exposure to Mn leads to hyperactive-like behavior and accumulation of this metal in *Drosophila melanogaster*. The observed raise in Mn levels is negatively correlated with levels of other essential metals. This result fits with previous studies showing that Mn accumulation and Fe deficiency are associated with hyperactive behavior in children (Ericson et al., 2007\[[@R15]\]; Konofal et al., 2008\[[@R34]\]). The induction of stress responsive genes and antioxidant enzyme activity associated with inhibition of p38^MAPK^ phosphorylation at higher concentrations of Mn may represent an adaptive response to oxidative stress generated by this metal, in an attempt to avoid exacerbated cellular damage. Acknowledgements ================ The authors thank to Universidade Federal do Pampa, Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq 482313/2013-7), Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (Fapergs 1954/2551/13-7), FAPERGS/PRONEM, FINEP and INCT-APA for financial support. The authors thank Professor Peter R. Dunkley for critical review of the manuscript. Conflict of interest ==================== The authors declare that they have no conflict of interest. ![Genes tested by quantitative real-time RT-PCR analysis and used forward and reverse primers](EXCLI-13-1239-t-001){#T1} ![Metal levels in flies Mn-exposed during the embryonic development](EXCLI-13-1239-t-002){#T2} ![The relationship between Mn concentration and metal levels](EXCLI-13-1239-t-003){#T3} ![Effects of exposure to Mn during the embryonic development on locomotor performance in *Drosophila melanogaster*. Results are expressed as mean (±) standard deviation (SD) and represent the time spent to climb up to 8cm in a glass tube. (n = 20-24). \*\* indicates a significant effect of Mn in comparison to control group (p \< 0.005).](EXCLI-13-1239-g-001){#F1} ![Effect of Mn exposure on cell viability in homogenate of flies treated with metal over embryonic development performed by MTT (A) and Rezarusin (B) cell viability assays. For MTT, the average of absorbance in control group was 0.017 (n = 3-4); by Rezarusin, absorbance in control group was 229.03 (n = 4-5). Results were expressed as the percentage (%) of the control group (mean ± standard deviation)\ \* indicates a significant effect of Mn in comparison to control group (p \< 0.05).](EXCLI-13-1239-g-002){#F2} ![Analysis of ROS production in response to Mn-exposure during embryonic development in flies. The data showed the DCF-DA intensity of fluorescence in total flies homogenate, expressed as percentage of the control group (mean ± standard deviation) (n = 3-10).\ \* indicates a significant effect of Mn in comparison to control group (p \< 0.05).](EXCLI-13-1239-g-003){#F3} ![Quantitative real time PCR (qRT-PCR) analysis of CAT, SOD and HSP83 mRNA in flies Mn-exposed at 1 mM. We used qRT-PCR to quantify levels of mRNA, relative to respective controls, after exposure. The data were normalized against GPDH transcript levels and each bar represents the mean ± standard deviation expressed as percent of its respective control (n = 3-4)\ \*\* and \*\*\* indicate a significant effect of Mn in comparison to control group (p \< 0.005 and p \< 0.0001, respectively).](EXCLI-13-1239-g-004){#F4} ![Effects observed on enzymatic activities in flies Mn-exposed during embryonic development. TrxR activity (n = 5-7) (A), GST activity (n = 3-4) (B), SOD activity (n = 6-8) (C) and CAT activity (n = 5-6) (D). The data shows the enzymatic activities in flies homogenate expressed as mean (mU/mg protein) ± standard deviation.\ \* and \*\*\* indicate a significant effect of Mn in comparison to control group (p \< 0.05 and p \< 0.0001, respectively).](EXCLI-13-1239-g-005){#F5} ![Modulation of MAPKs phosphorylation in response to Mn-exposed flies during the embryonic development. The upper panel is a western blot showing phosphorylated forms of p38^MAPK^, ERK and JNK ½ and total forms of ERK and ß-actin (A). The graphs are showing the quantification (percentage of control) of immunoreactive bands, each bar represents the mean ± standard deviation. ERK phosphorylation was normalized with total ERK expression (n-5) (B). p38^MAPK^ phosphorylation was normalized with ß-actin expression (n=5) (C). JNK ½ phosphorylation was normalized with ß-actin expression (n = 3-5) (D).\ \* Indicates a significant effect of Mn in comparison to control group (p \< 0.05).](EXCLI-13-1239-g-006){#F6}

1. Introduction {#s1} =============== Consider a game in which every decision maker is faced with a finite set of choices such that one specific choice always brings him higher monetary payoff than other choices, irrespective of the choices made by other players. In this situation, the individual choice boils down to going for either a higher or a lower monetary payoff. The straightforward response of a decision maker who cares about his monetary payoff is to disregard dominated actions---i.e., actions that may only deteriorate payoff relative to other actions. This dominance principle is the most basic solution concept of game theory (Camerer, [@B12]). It becomes very powerful when embedded in a strategic reasoning as a stepwise process. In each step, the dominance principle implies that dominated strategies should be eliminated from an agent\'s strategy space. In an important class of games---known as dominance-solvable games---this iterated elimination of dominated strategies leads to a unique solution. Strikingly, the data collected from numerous experiments on dominance-solvable games raise important questions about the empirical accuracy of predictions derived from this principle. Subjects tend to display less strategic sophistication than is needed to justify many applications of iterated dominance (and related refinements) to model human decision making in strategic environments (Crawford, [@B19]). The beauty contest game is one of the textbook examples of this issue[^1^](#fn0001){ref-type="fn"}. A given set of players is asked to choose a number in the range \[0, 100\]. To win the game, a player should choose a number that is the closest to *p* = 2/3 of the average of all chosen numbers. Any number above 2/3 × 100 ≈ 66.7 violates first-order dominance, because the average has to be lower than 100. Knowing this, players should all choose numbers no greater than 66.7, meaning that their average may not exceed 2/3 × 66.7 ≈ 44.5. This reasoning lowers the target as the number of iterations increases, eventually leading to the unique Nash equilibrium in which all players choose 0. In many experimental studies of this game, the numbers chosen by players are used as a proxy of the depth of iterated reasoning.[^2^](#fn0002){ref-type="fn"} A well replicated stylized fact is to observe 1/3 of subjects choosing a number higher than 67, and at least 1/3---a number between 44 and 67. This paper focuses on one of the earliest and simplest example of such an empirical inaccuracy of dominance solvability, adapted from a 2 × 2 game discussed in Rosenthal ([@B43]) and first brought to the laboratory by Beard and Beil ([@B5])[^3^](#fn0003){ref-type="fn"}. The normal-form representation of this game is given in Table [1](#T1){ref-type="table"}. With *L* \< *S* \< *H*, *m* \< *h*, and *s* \< *h*, the game is one-step dominance solvable: the elimination of player B\'s weakly dominated strategy *l* immediately leads to the Pareto-Nash equilibrium (*R, r*)[^4^](#fn0004){ref-type="fn"}. ###### **Generic form of the normal representation of Rosenthal ([@B43]) dominance solvable game**. **Player B** -------------- ----- -------------- -------- **Player A** *L* (S; s) (S; s) *R* (L; m) (H; h) In line with observed behavior in other dominance solvable games, numerous studies (summarized in Table [2](#T2){ref-type="table"}) find frequent failures to achieve the Pareto-Nash equilibrium. In spite of variations in the design (described in the table), deviations from the standard theoretical predictions are systematic and sizable. First, dominance is frequently violated by player Bs. Depending on the exact experimental setup, up to 27% column players choose a strictly dominated action. Second, player As violate iterated dominance, even in those cases in which player Bs commonly obey dominance. As an example, while only 6% of player Bs violate dominance in Jacquemet and Zylbersztejn ([@B35])-ET2 and BT2, 26% of row players still contradict the predictions of dominance solvability by choosing *L* (and this figure may even attain 86% in other instances, see Beard, Beil -- Tr. 5 in Table [2](#T2){ref-type="table"}). As shown in the three middle columns of the table, both the absolute and the relative size of the stakes vary a great deal from one study to the other. Several lessons emerge from this accumulated evidence. First, both players react to their own monetary incentives. Second, in some cases player As also adjust their behavior to player Bs\' incentives. Finally, as shown by Jacquemet and Zylbersztejn ([@B35]), players\' inefficient behavior does not fade away with repetition and cannot be explained by inequality aversion (as framed by Fehr and Schmidt, [@B23]). ###### **Overview of existing experimental evidence**. **Experiment** **Form** **Payoff** **Outcomes (%)** ------------------------- ---------- --------------- ------------------ ------------ ---- ----- ----- ----- ----- Beard, Beil--Tr.1 Seq (9.75; 3.0) (10; 5.0) (3; 4.75) 66 29 6 83 --- Beard, Beil--Tr.2 Seq (9.00; 3.0) (10; 5.0) (3; 4.75) 65 35 0 100 --- Beard, Beil--Tr.3 Seq (7.00; 3.0) (10; 5.0) (3; 4.75) 20 80 0 100 --- Beard, Beil--Tr.4 Seq (9.75; 3.0) (10; 5.0) (3; 3.00) 47 53 0 100 --- Beard, Beil--Tr.5 Seq (9.75; 6.0) (10; 5.0) (3; 3.00) 86 14 0 100 --- Beard, Beil--Tr.7 Seq (58.50; 18.0) (18.0; 28.50) (60; 30.0) 67 33 0 100 --- Beard et al.--Tr.1 Seq (1450; 450) (1500; 750) (450; 700) 79 18 3 83 --- Beard et al.--Tr.2 Seq (1050; 450) (1500; 750) (450; 700) 50 32 18 64 --- Goeree, Holt--Tr.1 Ext (80; 50) (90; 70) (20; 10) 16 84 0 100 --- Goeree, Holt--Tr.2 Ext (80; 50) (90; 70) (20; 68) 52 36 12 75 --- Goeree, Holt--Tr.3 Ext (400; 250) (450; 350) (100; 348) 80 16 4 80 --- Cooper, Van Huyck--Tr.9 Str (4; 1) (6; 5) (2; 4) 27 --- --- --- 86 Cooper, Van Huyck--Tr.9 Ext (4; 1) (6; 5) (2; 4) 21 --- --- --- 84 JZ, 2014--BT1 Str (9.75; 3.0) (3.0; 4.75) (10; 5.0) 51 41 8 84 81 JZ, 2014--ET1 Str (9.75; 5.0) (5.0; 9.75) (10; 10.0) 54 33 13 72 73 JZ, 2014--ET3 Str (9.75; 5.5) (5.5; 8.50) (10; 10.0) 39 48 13 79 76 JZ, 2014--ET4 Str (8.50; 5.5) (5.5; 8.50) (10; 10.0) 25 61 14 82 82 JZ, 2014--ET2 Str (8.50; 8.5) (6.5; 8.50) (10; 10.0) 26 70 4 94 94 JZ, 2014--BT2 Str (8.50; 7.0) (6.5; 7.00) (10; 8.5) 26 70 4 94 94 *For each implementation in row, the first column describes the actual design of the experiment: simultaneous-move strategic-form game (Str), simultaneous-move extensive-form game (Ext), sequential-move game (Seq). The monetary payoffsof each outcome, displayed in columns 2--4, are in USD in Beard and Beil ([@B5]) and Cooper and Van Huyck ([@B14]), in cents of USD in Goeree and Holt ([@B27]), in Yens in Beard et al. ([@B4]), and in Euros in Jacquemet and Zylbersztejn ([@B35]). The game is repeated ten times in changing pairs in Jacquemet and Zylbersztejn ([@B35]), and one-shot in all other instances*. The aim of the present paper is to explore whether this empirical puzzle is related to players\' cognitive skills. In this sense, our investigation belongs to a recent and growing body of experimental studies in both psychology and economics which investigate the relationship between strategic behavior and cognitive skills[^5^](#fn0005){ref-type="fn"}. The main conclusion that can be drawn from these studies is that high cognitive skills predict strategic sophistication and efficient decision making. First, people with high cognitive skills make more accurate predictions about other people\'s intentions. Recent evidence from psychological research reveals the relationship between cognitive skills and the theory of mind. Using the "Reading the Mind in the Eyes" test (RMET, Baron-Cohen et al., [@B3]) to measure one\'s theory of mind, Ibanez et al. ([@B32]) find that people with higher cognitive skills are better at infering the internal emotional states of others[^6^](#fn0006){ref-type="fn"}. Relatedly, the results of a neuroeconomic experiment on the *p*-beauty contest game by Coricelli and Nagel ([@B17]) suggest that strategic thinking about other players\' thoughts and behavior is implemented by medial prefrontal cortex (mPFC) -- one of the brain areas commonly associated with theory of mind[^7^](#fn0007){ref-type="fn"}. An economic experiment by Carpenter et al. ([@B13]) also shows that people with higher cognitive ability make more accurate predictions of others\' choices in a 20-player beauty contest game. Second, people with higher cognitive skills apply more sophisticated reasoning and are more apt in strategic adaptation. Burks et al. ([@B10]) report that subjects with higher cognitive skills more accurately predict others\' behavior in a sequential prisoners\' dilemma game, and adapt their own behavior more strongly. In the context of the *p*-beauty contest game, subjects with higher cognitive skills are not only found to carry out more steps of reasoning on the equilibrium path (Burnham et al., [@B11]; Brañas-Garza et al., [@B8]), but also to adapt their behavior to their opponents\' cognitive skills (Gill and Prowse, [@B26]) as well as to their beliefs about their opponents\' cognitive skills (Fehr and Huck, [@B22]). Third, cognitive skills may be associated with the economic efficiency of outcomes of both individual and group activities. Corgnet et al. ([@B16]) find that higher cognitive skills predict better performance and less shirking in an experimental labor task (summing up tables of 36 numbers without using a pen). Jones ([@B36]), Al-Ubaydli et al. ([@B2]), and Proto et al. ([@B40]) report that groups with higher cognitive skills attain higher cooperation rates in repeated prisoner\'s dilemma games. On the other hand, Al-Ubaydli et al. ([@B1]) do not find a relationship between group members\' average cognitive skills and the efficiency of outcomes in a stag hunt coordination[^8^](#fn0008){ref-type="fn"}. Our contribution is two-fold. First, we provide new evidence on the relationship between strategic behavior and cognitive skills. We show that systematic mismatches between theoretical predictions and actual behavior in a classic 2 × 2 dominance-solvable game have cognitive underpinnings. Subjects with higher cognitive skills are found to be more likely to play dominant strategy and to best respond to other\'s strategy. Furthermore, cognitive skills predict strategic sophistication: only those players with sufficiently high cognitive ability are found to display sensitivity to the presence of uncertainty about others\' behavior. Our second contribution lies in experimental methodology. We extend the recent body of laboratory experiments comparing the performance of different measures of cognitive skills in predicting economic behavior. Notwithstanding the previous results (see e.g., Brañas-Garza et al., [@B8]; Corgnet et al., [@B15]), we report that the Raven\'s test score is a more general predictor of strategic behavior than the Cognitive Reflection Test score. 2. Experimental design {#s2} ====================== Our experiment is based on a 2 × 2 factorial design that varies the payoff matrix and the nature of player B. Each of the four resulting experimental treatments is implemented through a between-subject procedure---each subject participates in only one experimental condition. This data come from a large dataset, part of which has been previously used by Hanaki et al. ([@B30]). The main focus of that study is player As\' behavior under strategic uncertainty and its relation to monetary incentives and fluid intelligence. Certain elements of their design (such as the use of Human and Robot conditions and interest in players\' cognitive skills) inevitably needed to be adopted in the present study in order to address a much more general question of the empirical validity of the solution concept of dominance solvability. More precisely, we are interested in both players\' behavior (so as to measure the use of dominance by player Bs and the use of iterated dominance by player As under different information structures). We also make a methodological contribution, since in this paper we associate players\' behavior with multiple facets of cognitive skills: fluid intelligence (measured by Raven\'s test) and cognitive reflection (measured by CRT). Our first treatment variable is the size of the stakes, as represented by Game 1 and Game 2 in Table [3](#T3){ref-type="table"}. Although they have the same strategic properties, these two game matrices differ in terms of the saliency of monetary incentives to use (iterated) dominance. In Game 2, player As may earn a surplus of only 0.25 when moving from *L* to (*R, r*) (with payoff going from 9.75 to 10), while ending up in (*R, l*) is relatively costly (yielding only 3). In Game 1, the potential gains and losses from action *R* relative to *L* are more balanced: the gain from moving from *L* to (*R, r*) increases to 1.5 (with payoff moving from 8.5 to 10), while the outcome (*R, l*) becomes less costly (now yielding 6.5). The incentives of player Bs, in turn, go in the opposite direction: the gain from using the dominant strategy *r* (and conditional on player As\' choice *R*) is lower in Game 1 \[with payoff increasing from 4.75 to 5 between (*R, l*) and (*R, r*)\] than in Game 2 (where payoff increases from 8.5 to 10). In line with Jacquemet and Zylbersztejn ([@B35]) and Hanaki et al. ([@B30]) (who report that both players only react to their own monetary incentives) and as discussed in Section 3.1, each of these games generates sizable yet diverse empirical violations of dominance solvability. These two games together thus provide a wide range of monetary incentives to use dominance solvability within a common strategic environment[^9^](#fn0009){ref-type="fn"}. ###### **The experimental games**. **GAME 1** ------------ ----- --------------- ----------------- **A** *L* (8.50 ; 3.00) (8.50 ; 3.00) *R* (6.50 ; 4.75) (10.00 ; 5.00) **GAME 2** **B** ****l**** ****r**** **A** *L* (9.75 ; 8.50) (9.75 ; 8.50) *R* (3.00 ; 8.50) (10.00 ; 10.00) Our second treatment variable is related to the nature of player B (the column player) who may be represented either by a human subject (Human condition) or a pre-programmed computer (Robot condition). The Human condition enables us to capture two cardinal breaches of dominance solvability: the failure to use the dominant strategy (player Bs\' behavior) and the failure to best respond to others\' dominant actions (player As\' behavior). However, the latter behavior occurs under strategic uncertainty and thus might stem from two distinct sources: bounded rationality and rational behavior under uncertainty. More precisely, player As may simply have a limited capability of best responding to dominant strategy, but may also intentionally refrain from best responding when in doubt about player Bs\' use of dominant strategy. To separate these two effects, we introduce the Robot condition in which a human subject acting as player A interacts with a computerized player B who is pre-programmed to always choose *r*. We clearly inform the subjects in the Robot condition that they are interacting with a pre-programmed computer: "**the computer chooses** *r* **at each round, without exception**" (bold in the original instruction sheet). This is the only difference in the rules and procedures between Human and Robot conditions[^10^](#fn0010){ref-type="fn"}. Thus, the key property of the Robot condition as compared to the Human condition is neutralizing strategic uncertainty player As face, while maintaining space for boundedly rational behavior. The design of the experiment is otherwise the same in all four experimental conditions. We explore whether behavior is sensitive to learning by considering ten uniform, one-shot interactions. In order to homogenize incentives across rounds, the following rules are implemented: all games are played in strict anonymity, roles are fixed, and subjects\' payoffs are computed based one randomly drawn round. In the Human condition, players are matched into pairs using a perfect stranger, round-robin scheme, which guarantees that subjects are involved in a series of one-shot interactions despite the repetition of the game[^11^](#fn0011){ref-type="fn"}. Our control variables also include two measures of cognitive skills. Both of them are introduced as part of a post-experimental supplementary task. Subjects\' participation is rewarded with extra five Euros; otherwise, their answers are not incentivized[^12^](#fn0012){ref-type="fn"}. The supplementary task starts with a debriefing question, where subjects are asked to "report any information they find relevant about how their decisions has been made." Then, we implement the following measures of cognitive skills. The first task is the standard Cognitive Reflection Test based on Frederick ([@B25]) which "*measures cognitive reflectiveness or impulsiveness, respondents\' automatic response versus more elaborate and deliberative thought*" (Brañas-Garza et al., [@B8], p. 255). It contains three questions: 1. A notebook and a pencil cost 1.10 Euros in total. The notebook costs 1 Euro more than the pencil. How much does the pencil cost? 2. If it takes 5 machines 5 min to make 5 widgets, how long would it take 100 machines to make 100 widgets? 3. In a lake, there is a patch of lily pads. Every day, the patch doubles in size. If it takes 48 days for the patch to cover the entire lake, how long would it take for the patch to cover half of the lake? Subjects are informed that this set of three questions should be answered within 30 s (although we allow them to provide answers even after this time has elapsed). In this way, subjects can be classified according to their overall score (that is, the total number of correct answers) which can range from 0 to 3. The second task is Raven\'s progressive matrix test (often called Raven\'s test), a picture based, non-verbal measure of fluid intelligence, that is "*the capacity to think logically, analyze and solve novel problems, independent of background knowledge*" (Mullainathan and Shafir, [@B38], p. 48). It is widely used by, e.g., psychologists, educators, and the military (Raven, [@B41]). It consists of a series of tasks to be solved within a fixed amount of time. In each task, a subject should pick a single element (among eight options) that best fits a set of eight pictures. The level of difficulty increases from one question to the other[^13^](#fn0013){ref-type="fn"}. In our experiment, each participant is given a series of 16 tasks to be solved within 10 min. Individual scores in Raven\' test are computed as the number of correct answers to the 16 items of the test. 2.1. Experimental procedures ---------------------------- For each game matrix, we run three Human sessions (involving 20 subjects per session: 10 player As interacting with 10 player Bs), and two Robot sessions (involving 20 player As per session interacting with automated player Bs). Subjects are given a fixed fee equal to five euros to compensate participation to the experiment. Upon arrival to the laboratory, participants are randomly assigned to their computers and asked to fill in a short administrative questionnaire containing basic questions about their age, gender, education, etc. Experimental instructions are then read aloud: subjects are informed that they will play multiple rounds of the same game, each round with a different partner, and that their own role will remain unchanged throughout the experiment. Finally, subjects are asked to answer a short comprehension quiz. Once the quiz and any questions from participants are answered, the experiment begins. After each of the ten rounds of the game, subjects are only informed of their own payoffs. Information about past choices and payoffs is updated after each round and displayed at the bottom of the screen. Take-home earnings correspond to the outcome of a single round that is randomly drawn at the end of each experimental session. In addition, the experimental game is followed by supplementary tasks. An additional five euros fee is paid to each subject for completing this part. Immediately after the end of the experimental game, participants are provided with a brief round-by-round summary of their decisions and outcomes, and are asked to provide in a blank space on their computer screens any relevant comments in particular about what might have affected their decisions during the experiment. Subjects are also asked to solve the CRT test and a reduced-form Raven\'s test described above. All the sessions were conducted in February and March 2014. Out of the 200 participants (94 males), 155 were students with various fields of specialization. The majority of subjects (65%) had already taken part in economic experiments. Participants\' average age was 25.6 (st. dev. is 7.5). All sessions took place at the *Laboratoire d\'Economie Experimentale de Paris* (LEEP) at Paris School of Economics. Subjects were recruited via an on-line registration system based on O[rsee]{.smallcaps} (Greiner, [@B28]) and the experiment was computerized through software developed under R[egate]{.smallcaps} (Zeiliger, [@B47]) and z-Tree (Fischbacher, [@B24]). Sessions lasted about 45--60 min, with an average payoff of roughly 18.83 euros (including a five euros show-up fee and five euros for completing the post-experimental tasks). 3. Results {#s3} ========== Our main experimental results can be summarized as follows. First, in line with the existing literature, we observe systematic and sizable deviations from standard predictions based on the principle of dominance solvability. This phenomenon persists across game matrices and despite repetition. Second, we associate strategic behavior with cognitive skills. We find that Raven\'s test score is a more reliable predictor of strategic behavior than CRT score: whenever the latter predicts behavior, the former does too, but not *vice versa*. Subjects with higher Raven\'s test scores are more likely to use the dominant strategy and to best respond to other player\'s dominant strategy. Unlike those with low Raven\'s test score, they also react to the presence of strategic uncertainty. 3.1. Aggregate behavior in experimental games --------------------------------------------- Table [4](#T4){ref-type="table"} outlines the main patterns of behavior in our experimental games. The statistical significance of the changes observed in this table is tested by Models 1--3 in Table [5](#T5){ref-type="table"}. We first focus on the aggregate frequency of Pareto-Nash equilibrium (*R, r*) -- the sole outcome that survives the iterated elimination of (weakly) dominated strategies---found in the Human condition. In both games, we observe substantial deviations from the predictions of this solution concept: overall, players attain the (*R, r*) outcome 58% of times in Game 1 and 43% in Game 2 (Model 1, *H*~0~ : β~1~ = 0, *p* = 0.318). We also observe that efficiency increases over time: in both games, we observe the lowest frequency of (*R, r*) in the initial round (0.333 in Game 1 and 0.200 in Game 2), whereas the highest frequency of (*R, r*) occurs in the final round (0.700 in Game 1 and 0.533 in Game 2). ###### **Aggregate results**. **Round** **Overall** --------------------------------------------------------------- ----------- ------------- ------- ------- ------- ------- ------- ------- ------- ------- ------- ***Pr*** **(*****R,r*****) in the Human condition** Game 1 0.333 0.600 0.667 0.700 0.567 0.600 0.433 0.633 0.567 0.700 0.580 Game 2 0.200 0.333 0.400 0.400 0.433 0.500 0.500 0.433 0.467 0.533 0.420 ***Pr*** **(*****r*****) by player B in the Human condition** Game 1 0.767 0.800 0.867 0.900 0.800 0.800 0.700 0.833 0.867 0.800 0.813 Game 2 0.833 0.933 0.900 0.933 1.000 0.933 0.933 0.900 0.900 0.933 0.920 ***Pr*** **(*****R*****) by player A in the Human condition** Game 1 0.500 0.733 0.700 0.767 0.767 0.800 0.700 0.767 0.700 0.867 0.730 Game 2 0.300 0.333 0.400 0.400 0.433 0.533 0.533 0.500 0.500 0.533 0.447 ***Pr*** **(*****R*****) by player A in the Robot condition** Game 1 0.700 0.750 0.750 0.725 0.800 0.800 0.800 0.825 0.800 0.775 0.773 Game 2 0.500 0.575 0.725 0.575 0.800 0.700 0.700 0.775 0.775 0.775 0.690 *Columns 1--10 summarize the frequencies of outcomes (defined in rows) as % of all outcomes observed in each round of a given experimental treatment. The last column provides overall results*. ###### **Aggregate results: statistical support**. **Model 1** **Model 2** **Model 3** ------------------------------------- -------------------------------------------- -------------------------------------------- -------------------------------------------- Constant (β~0~) 0.580[^\*\*\*^](#TN1){ref-type="table-fn"} 0.813[^\*\*\*^](#TN1){ref-type="table-fn"} 0.730[^\*\*\*^](#TN1){ref-type="table-fn"} (0.144) (0.032) (0.084) 1\[*Game* 2\] (β~1~) --0.160 0.107[^\*^](#TN1){ref-type="table-fn"} --0.283[^\*\*^](#TN1){ref-type="table-fn"} (0.103) (0.044) (0.140) 1\[*Robot*\] (β~2~) 0.043 (0.103) 1\[*Robot*\] × 1\[*Game* 2\] (β~3~) 0.201 (0.161) *N* 600 600 1400 *R*^2^ 0.026 0.025 0.066 Estimates of linear probability models on outcome (R, r) (Model 1), decision r by player B (Model 2) and decision R by player A (Model 3). Standard errors (in parentheses) are clustered at the session level in Human treatments (three clusters per game matrix, six in total) and individual level in the Robot condition (40 clusters per game matrix, 80 in total) and computed using the delete-one jackknife procedure. All models contain a dummy variable set to 1 for game matrix 2 (and 0 for game matrix 1). In Model 3, we also introduce an additional dummy variable set to 1 for Robot condition (and 0 for Human condition) and well as the interaction between these two variables. *indicate significance at the 10/5/1% level*. To further explore the roots of these deviations, we turn to the aggregate patterns of both players\' behavior in Human and Robot conditions. We focus on three behavioral dimensions of dominance solvability: the use of dominant strategy (captured by player Bs\' behavior in the Human condition) and the ability to best respond to other player\'s dominant action with and without bearing the uncertainty about the latter (which is captured by player As\' behavior in the Human and Robot conditions, respectively). Inefficiency is caused by both players, although their roles differ from one game to another: the scope of inefficient behavior is similar for both players in Game 1, and highly asymmetric in Game 2. Overall, player As select action *R* with probability 0.730 in Game 1 and 0.447 in Game 2 (Model 3, *H*~0~ : β~1~ = 0, *p* = 0.047). However, player As\' behavior happens to be misaligned with player Bs\' actual decisions which follow the opposite trend: the total frequency of action *r* increases from 0.813 in Game 1 to 0.920 in Game 2 (Model 2, *H*~0~ : β~1~ = 0, *p* = 0.060). Importantly, the data from Robot sessions suggest that the uncertainty about player Bs\' behavior is not the only driver of player As\' choices. Player As frequently and systematically fail to best respond to player Bs\' dominant action even when the latter comes with certainty in the Robot condition, although their willingness to select action *R* increases in both games as compared to the Human condition (to 0.773 in Game 1 and 0.690 in Game 2)[^14^](#fn0014){ref-type="fn"}. The fact that inefficient actions from player As prevail in the absence of strategic uncertainty may suggest that at least some of them are boundedly rational decision makers. In the next section, we analyze how these three behavioral components of dominance solvability vary as a function of players\' cognitive skills. 3.2. Cognitive skills and strategic behavior -------------------------------------------- The average score in Raven\'s test (CRT) is 8.679 out of 16 with SD 3.117 (0.479 out of 3 with SD 0.852). Our experimental sample is properly randomized across treatments regarding both measures. We do not reject the null hypothesis that Raven\'s test scores have the same distributions in all treatments (*p* = 0.275, Kruskal-Wallis test). A Kruskal-Wallis test applied to the CRT scores leads to the same conclusion (*p* = 0.502). We also replicate several results from previous studies combining Raven\'s test and CRT regarding the relationship between both scores as well as gender differences (Brañas-Garza et al., [@B8]; Corgnet et al., [@B15]). There is a moderate, yet highly significant correlation between Raven and CRT scores (Spearman\'s ρ = 0.306, *p* \< 0.001) which suggests that they may have a common source, but do not capture the same cognitive skills. Furthermore, the average score of males is significantly higher than the average score of females (Raven\'s test: 9.382 with SD 0.341 vs. 8.014 with SD 0.384, *p* = 0.009; CRT: 0.676 with SD 0.111 vs. 0.291 with SD 0.087, *p* = 0.007; two-sided *t*-tests)[^15^](#fn0015){ref-type="fn"}. We also observe that many subjects (70%) of our 200 participants fail to provide at least one correct answer in our standard CRT. 16% provide exactly one, 8% -- two, and 6% -- three correct answers. This stands in line with Brañas-Garza et al. ([@B8]) who report the respective frequencies of 67, 23, 9, and 1% for a similar sample size (*N* = 191), and echoes the scores in the least performant sample reported in a seminal study by Frederick ([@B25]): out of 138 students of the University of Toledo, 64% provide no correct answer, 21% provide one, 10% provide two, and 5% provide three corrects answers. ### 3.2.1. Cognitive predictors of strategic behavior: aggregate results In this part, we study the cognitive correlates of strategic behavior. Figures [1](#F1){ref-type="fig"}, [2](#F2){ref-type="fig"} present the aggregate evolution of behavior as a function of cognitive skills, measured either by CRT score or by Raven\'s test score across roles (player A or player B) and experimental conditions (Human or Robot). ![**CRT score and aggregate behavior across rounds and treatments**.](fpsyg-07-01188-g0001){#F1} ![**Raven\'s test score and aggregate behavior across rounds and treatments**.](fpsyg-07-01188-g0002){#F2} In Figure [1](#F1){ref-type="fig"}, the sample is divided into two subsamples: subjects who provided at least one correct answer to CRT (referred to as *CRT* \> 0) and those who did not (referred to as *CRT* = 0). The aggregate patterns of behavior weakly differ between the two subsamples. Bootstrap proportion tests fail to reject the null hypothesis that the overall proportions of decision *R* are the same for both CRT categories in the Human condition (*p* = 0.126) and in the Robot condition (*p* = 0.235)[^16^](#fn0016){ref-type="fn"}. The aggregate proportions of decision *r*, in turn, are found to be statistically different (*p* = 0.037), subjects with a CRT score zero being less likely to play *r* than subjects who gave at least one correct answer. In Figure [2](#F2){ref-type="fig"}, we split our sample into three subsamples based on Raven\'s test score (1st tertile: less than 8 correct answers, 2nd tertile: between 8 and 10 correct answers, 3rd tertile: more than 10 correct answers). Although, bootstrap proportion tests suggest that player As\' behavior in the Human condition does not vary significantly between these three subsamples (1st tertile vs. 2nd tertile: *p* = 0.255, 2nd vs. 3rd: *p* = 0.580, 1st vs. 3rd: *p* = 0.565), significant differences arise for both player As in the Robot condition (*p* = 0.001, *p* = 0.735, *p* \< 0.001, respectively) and for player Bs (*p* = 0.064, *p* = 0.057, *p* \< 0.001, respectively). Raven\'s test score seems to have a more systematic association with players\' behavior than CRT score, although both measures fail to predict behavior under strategic uncertainty. ### 3.2.2. Cognitive skills and dominance solvability: regression analysis In what follows, we provide further econometric insights into these preliminary results. Following Brañas-Garza et al. ([@B8]); Corgnet et al. ([@B15]), we use three individual characteristics discussed in the previous section -- gender, Raven\'s test score and CRT score (kept as a dummy variable with value 1 if the subject gave at least one correct answer at the CRT test and 0 otherwise) -- to explain behavior in our experimental games[^17^](#fn0017){ref-type="fn"}. The econometric specification is based on the linear probability model and the estimation procedure is outlined in Jacquemet and Zylbersztejn ([@B35]). We also control for payoff scheme and repetition effects by including game matrix and round dummies. We consider three different outcome variables: player As\' behavior in the Human and the Robot treatment, and player Bs\' behavior in the Human treatment. Given the correlation between CRT and Raven\'s test scores, including both variables in the model might result in multicollinearity and lead to the under-rejection of the nullity of respective coefficients. For each outcome, we first include these two measures separately in Models 1 and 2, while Model 3 includes both variables. This evidence is summarized in Table [6](#T6){ref-type="table"}. ###### **Cognitive predictors of strategic behavior: regression analysis**. ******Pr****** **(********R********) by player A** **Pr (********r********) by player B** --------------- ---------------------------------------------------- ------------------------------------------ ------------------------------------------ -------------------------------------------- --------------------------------------------- --------------------------------------------- -------------------------------------------- -------------------------------------------- -------------------------------------------- Const. 0.423[^\*\*\*^](#TN2){ref-type="table-fn"} 0.552[^\*\*^](#TN2){ref-type="table-fn"} 0.563[^\*\*^](#TN2){ref-type="table-fn"} 0.573[^\*\*\*^](#TN2){ref-type="table-fn"} 0.242[^\*^](#TN2){ref-type="table-fn"} 0.240[^\*^](#TN2){ref-type="table-fn"} 0.705[^\*\*\*^](#TN2){ref-type="table-fn"} 0.430[^\*\*\*^](#TN2){ref-type="table-fn"} 0.444[^\*\*\*^](#TN2){ref-type="table-fn"} (0.080) (0.176) (0.197) (0.088) (0.135) (0.135) (0.027) (0.103) (0.099) 1\[CRT\>0\] 0.131 0.152 0.062 (0.024) 0.109[^\*^](#TN2){ref-type="table-fn"} 0.046 (0.095) (0.121) (0.102) (0.102) (0.047) (0.036) Raven 0.013 0.018 0.0426[^\*\*\*^](#TN2){ref-type="table-fn"} 0.0434[^\*\*\*^](#TN2){ref-type="table-fn"} 0.0313[^\*\*^](#TN2){ref-type="table-fn"} 0.0287[^\*\*^](#TN2){ref-type="table-fn"} (0.025) (0.027) (0.013) (0.012) (0.009) (0.008) 1\[*Game* 2\] −0.270[^\*^](#TN2){ref-type="table-fn"} 0.263 0.266 0.068 0.054 0.056 0.100 0.132[^\*^](#TN2){ref-type="table-fn"} 0.129[^\*^](#TN2){ref-type="table-fn"} (0.129) (0.139) (0.136) (0.083) (0.076) (0.079) (0.052) (0.056) (0.056) 1\[Male\] 0.132 0.187 0.158 0.096 0.072 0.077 0.025 0.024 0.017 (0.126) (0.100) (0.107) (0.090) (0.076) (0.089) (0.046) (0.046) (0.047) Round: 2 0.133 0.133 0.133 0.063 0.063 0.063 0.067 0.067 0.067 (0.092) (0.092) (0.092) (0.048) (0.048) (0.048) (0.042) (0.042) (0.042) 3 0.150 0.150 0.150 0.138[^\*\*\*^](#TN2){ref-type="table-fn"} 0.138[^\*\*\*^](#TN2){ref-type="table-fn"} 0.138[^\*\*\*^](#TN2){ref-type="table-fn"} 0.083 0.083 0.083 (0.109) (0.109) (0.109) (0.050) (0.050) (0.050) (0.048) (0.048) (0.048) 4 0.183[^\*\*^](#TN2){ref-type="table-fn"} 0.183[^\*\*^](#TN2){ref-type="table-fn"} 0.183[^\*\*^](#TN2){ref-type="table-fn"} 0.050 0.050 0.050 0.117[^\*^](#TN2){ref-type="table-fn"} 0.117[^\*^](#TN2){ref-type="table-fn"} 0.117[^\*^](#TN2){ref-type="table-fn"} (0.070) (0.070) (0.070) (0.047) (0.047) (0.047) (0.048) (0.048) (0.048) 5 0.200[^\*^](#TN2){ref-type="table-fn"} 0.200[^\*^](#TN2){ref-type="table-fn"} 0.200[^\*^](#TN2){ref-type="table-fn"} 0.200[^\*\*\*^](#TN2){ref-type="table-fn"} 0.200[^\*\*\*^](#TN2){ref-type="table-fn"} 0.200[^\*\*\*^](#TN2){ref-type="table-fn"} 0.100 0.100 0.100 (0.089) (0.089) (0.089) (0.045) (0.045) (0.045) (0.052) (0.052) (0.052) 6 0.267[^\*\*^](#TN2){ref-type="table-fn"} 0.267[^\*\*^](#TN2){ref-type="table-fn"} 0.267[^\*\*^](#TN2){ref-type="table-fn"} 0.150[^\*\*\*^](#TN2){ref-type="table-fn"} 0.150[^\*\*\*^](#TN2){ref-type="table-fn"} 0.150[^\*\*\*^](#TN2){ref-type="table-fn"} 0.067 0.067 0.067 (0.088) (0.088) (0.088) (0.054) (0.054) (0.054) (0.049) (0.049) (0.049) 7 0.217[^\*^](#TN2){ref-type="table-fn"} 0.217[^\*^](#TN2){ref-type="table-fn"} 0.217[^\*^](#TN2){ref-type="table-fn"} 0.150[^\*\*\*^](#TN2){ref-type="table-fn"} 0.150[^\*\*\*^](#TN2){ref-type="table-fn"} 0.150[^\*\*\*^](#TN2){ref-type="table-fn"} 0.017 0.017 0.017 (0.098) (0.098) (0.098) (0.047) (0.047) (0.047) (0.048) (0.048) (0.048) 8 0.233[^\*\*^](#TN2){ref-type="table-fn"} 0.233[^\*\*^](#TN2){ref-type="table-fn"} 0.233[^\*\*^](#TN2){ref-type="table-fn"} 0.200[^\*\*\*^](#TN2){ref-type="table-fn"} 0.200[^\*\*\*^](#TN2){ref-type="table-fn"} 0.200[^\*\*\*^](#TN2){ref-type="table-fn"} 0.067 0.067 0.067 (0.088) (0.088) (0.088) (0.057) (0.057) (0.057) (0.056) (0.056) (0.056) 9 0.200 0.200 0.200 0.188[^\*\*\*^](#TN2){ref-type="table-fn"} 0.188[^\*\*\*^](#TN2){ref-type="table-fn"} 0.188[^\*\*\*^](#TN2){ref-type="table-fn"} 0.083 0.083 0.083 (0.113) (0.113) (0.113) (0.047) (0.047) (0.047) (0.060) (0.060) (0.060) 10 0.300[^\*\*^](#TN2){ref-type="table-fn"} 0.300[^\*\*^](#TN2){ref-type="table-fn"} 0.300[^\*\*^](#TN2){ref-type="table-fn"} 0.175[^\*\*\*^](#TN2){ref-type="table-fn"} 0.175[^\*\*\*^](#TN2){ref-type="table-fn"} 0.175[^\*\*\*^](#TN2){ref-type="table-fn"} 0.067 0.067 0.067 (0.115) (0.115) (0.115) (0.056) (0.056) (0.056) (0.049) (0.049) (0.049) *R*^2^ 0.151 0.141 0.160 0.050 0.139 0.140 0.060 0.108 0.111 Estimates of linear probability models explaining the likelihood of decision R by player A and decision r by player B. Standard errors (in parantheses) are clustered at the session level in the Human condition (three clusters per game matrix, six in total) and individual level in the Robot condition (40 clusters per game matrix, 80 in total) and computed using the delete-one jackknife procedure. Models 1 and 2 include a single measure of cognitive skills (a dummy set to 1 for a positive CRT score, or Raven\'s test score), while Model 3 combines both variables. Other independent variables include gender, game matrix and round dummies. The number of observations is N = 600 for Human and N = 800 for Robot conditions. *indicate significance at the 10/5/1% level*. We first turn to player Bs\' behavior. Models 1 and 2 suggest that both the coefficient of *CRT* \> 0 dummy and the coefficient of Raven\'s test score are positive and significant (*p* = 0.067 for *CRT* \> 0 and *p* = 0.015 for Raven). In Model 3, the coefficient of Raven\'s test score remains highly significant (*p* = 0.014), while the coefficient of CRT becomes insignificant (*p* = 0.253). Their joint significance (*p* = 0.034) implies that cognitive skills predict the use of dominant strategy. We now turn to player As\' behavior in the Human condition. Notwithstanding the previous set of results, cognitive skills are not found to explain player As\' choices. The coefficient of *CRT* \> 0 dummy is insignificant (*p* = 0.226) in Model 1, and so is the coefficient of Raven\'s test score (*p* = 0.633) in Model 2. If we account for both, Model 3 reveals that the coefficients of both scores are neither individually (*p* = 0.226 for *CRT* \> 0 and *p* = 0.550 for Raven\'s test score) nor jointly significant (*p* = 0.503). Finally, the behavior of player As in the Robot condition is only predicted by Raven\'s test score: unlike *CRT* \> 0 dummy, its coefficient remains positive and highly significant across models (*p* ≤ 0.001). Unsurprisingly, the joint insignificance of both coefficients in Model 3 is also rejected (*p* = 0.003). Altogether, the results presented in Table [6](#T6){ref-type="table"} suggest that cognitive skills predict certain components of strategic behavior: the use of dominant strategy (reflected in player Bs\' behavior), as well as the ability to best respond to other player\'s dominant strategy (reflected in player As\' behavior in the Robot condition). Moreover, in both cases Raven\'s test score is a more reliable predictor of behavior than CRT score. However, we also observe that Raven\'s test score fails to predict player As\' behavior once player Bs\' behavior becomes uncertain, that is once we move from Robot to Human condition. This, in turn, points toward an interplay between the degree of strategic uncertainty, behavior in the experimental games, and individual cognitive skills. Importantly, the existence of such an interplay is also supported by Figure [2](#F2){ref-type="fig"} which shows that the aggregate levels of efficiency shift upwards between the Human condition and the Robot condition for the 2nd and 3rd Raven\'s score tertile, but not the 1st tertile. In order to formally test this conjecture, we now look at the reaction of player As with different cognitive skills to the disappearance of strategic uncertainty. Splitting the data according to Raven\'s score tertile, for each of the three subsamples we compare player As\' behavior in the Human condition to their behavior in the Robot condition by regressing player As\' choice on the Robot dummy (set to 1 for the Robot and to 0 for the Human condition). We also include the previous set of independent variables (except for Raven\'s test score itself). These results are summarized in Table [7](#T7){ref-type="table"}. The coefficient of the Robot dummy captures the effect of eliminating strategic uncertainty on player As\' behavior for each of the three subsamples. This suggests that only player As with high enough cognitive skills are sensitive to the uncertainty about player Bs\' behavior. The behavior of players with low Raven\'s test score (1st tertile) is unresponsive to the degree of strategic uncertainty: the coefficient of the Robot dummy is close to zero and insignificant (*p* = 0.822). For players with medium scores (2nd tertile), we find a positive yet weakly significant effect (*p* = 0.087) which becomes amplified and highly significant for those player As whose Raven\'s test score belongs to the 3rd tertile of the experimental sample (*p* = 0.012). ###### **The effect of strategic uncertainty and cognitive skills: evidence from player As\' behavior in Human and Robot conditions**. **Raven\'s test score tertile** --------------- ---------------------------------------- -------------------------------------------- ------------------------------------------ Constant 0.277[^\*^](#TN3){ref-type="table-fn"} 0.592[^\*\*\*^](#TN3){ref-type="table-fn"} 0.330[^\*\*^](#TN3){ref-type="table-fn"} (0.147) (0.065) (0.135) 1\[Robot\] 0.044 0.158[^\*^](#TN3){ref-type="table-fn"} 0.428[^\*\*^](#TN3){ref-type="table-fn"} (0.195) (0.090) (0.155) 1\[CRT\>0\] 0.002 0.038 0.016 (0.262) (0.066) (0.188) 1\[Male\] 0.144 0.033 0.212 (0.145) (0.063) (0.179) 1\[Game 2\] 0.034 −0.245[^\*\*^](#TN3){ref-type="table-fn"} −0.176 (0.146) (0.092) (0.155) Round dummies Yes Yes Yes *N* 480 610 310 *R*^2^ 0.048 0.173 0.298 Estimates of linear probability models on decision R by player A. Standard errors (in parentheses) are clustered at the session level in the Human condition (three clusters per game matrix, six in total) and individual level in the Robot condition (40 clusters per game matrix, 80 in total) and computed using the delete-one jackknife procedure. Data from Human and Robot conditions are pooled and split into three subsamples based on Raven\'s test score tertiles. Other independent variables include a dummy set to 1 for a positive CRT score, as well as gender, game matrix and round dummies (omitted from the table). *indicate significance at the 10/5/1% level*. Finally, it is also worth noting that player As\' reaction to the payoff scheme also varies as a function of Raven\'s test score. The coefficient of the Game 2 dummy is close to zero and highly insignificant in the 1st tertile regression (*p* = 0.890). Then, it becomes negative in 2nd and 3rd tertile models (although it is only statistically significant in the former with *p* = 0.012 and *p* = 0.271, respectively). This, in turn, stands in line with the previous finding that player As\' willingness to play *R* increases as the safe choice *L* becomes less attractive relative to outcome (*R, r*). It also seems that the magnitude of this effect is mediated by player As\' cognitive skills, although not in a monotone way. 4. Conclusion {#s4} ============= This paper studies the relationship between strategic behavior and cognitive skills---cognitive reflection and fluid intelligence---in a classic 2 × 2 dominance-solvable game. Our results show that subjects with higher fluid intelligence (measured by Raven\'s progressive matrices test) are more likely to play dominant strategy, and also more likely to best respond to other\'s strategy. Furthermore, fluid intelligence predicts strategic sophistication: only those players with sufficiently high Raven\'s test score are found to display sensitivity to the presence of uncertainty about others\' behavior. Cognitive reflection (measured by CRT), in turn, lacks the power to predict behavior in our experimental setting. We see three main conclusions that stem from these findings. First, these results contribute to the ongoing debate on the relationship between rationality and intelligence (see Stanovich, [@B45], for a critical review). For instance, Stanovich and West ([@B46]) distinguish between two aspects of rational behavior: instrumental rationality which is understood as the "ability to take appropriate action given one\'s goals and beliefs," and epistemic rationality which enables agents to hold "beliefs that are commensurate with available evidence." In the strategic environment investigated in this paper, instrumental rationality can be associated with the ability to solve the game, while epistemic rationality---with the ability to play it with others. Our experimental data suggest an important relationship between fluid intelligence (rather than reflective thinking) and both of these facets of rationality in strategic settings. Both the ability to use dominance and iterated dominance to efficiently solve the game, as well as the responsiveness to the availability of strategic information, is found to be predicted by Raven\'s test score (but not by CRT score). The second contribution is related to the experimental methodology. Despite the fact that CRT and Raven\'s test are both commonly used to measure cognitive skills in experimental subject pools, still very little is known about their relative performance in predicting different types of behavior. Therefore, the choice of one test over the other may happen to be at least as intuitive as evidence-based. As mentioned before, to the best of our knowledge only two experiments address this issue. Brañas-Garza et al. ([@B8]) do so in a strategic environment (*p*-beauty contest game), while Corgnet et al. ([@B15])---in a non-strategic one (individual choices on wealth distribution). Both studies find that CRT performs better than Raven\'s test in predicting subjects\' behavior. The result of the present experiment points the to the opposite conclusion. We believe that this difference is driven by the very nature of the experimental tasks which may involve different types of cognitive effort. In our view, this issue deserves attention in future research. Finally, although we find evidence that behaving in accordance with dominance solvability is positively correlated with cognitive skills, we also substantiate that most of the variance in individual decision making cannot be explained by such skills. Thus, exploring factors alongside cognitive skills that generate strategic behavior remains an open and important empirical question. An interesting avenue is to disentangle individual determinants, e.g., personal characteristics (such as cognitive skills) that are associated with appropriate behavior, from environmental determinants, that is, those features of the decision making environment that lead decision makers to take certain types of actions. 5. Author contributions {#s5} ======================= NH, NJ, SL, and AZ all contributed equally to this work. Authors are listed in an alphabetical order. 6. Funding ========== This project has received funding from JSPS-ANR bilateral research grant BECOA (ANR-11-FRJA-0002), as well as the LABEX CORTEX (ANR-11-LABX-0042) of Université de Lyon, and LABEX OSE of the Paris School of Economics (ANR-10-LABX_93-01), both within the program "Investissements d\'Avenir" (ANR-11-IDEX-007) operated by the French National Research Agency (ANR). Ivan Ouss provided efficient research assistance. We thank Juergen Bracht, Colin Camerer, Guillaume Fréchette, Haoran He, Asen Ivanov, Frédéric Koessler, Rosemarie Nagel, Ariel Rubinstein, Jason F. Shogren, Jean-Marc Tallon, Antoine Terracol, and Marie Claire Villeval for their comments. NH and NJ gratefully acknowledge the *Institut Universitaire de France*. SL thanks the School of Business at the University of Western Australia for hospitality and support. A major part of this work was conducted while NH was affiliated with Aix-Marseille University (Aix-Marseille School of Economics, AMSE) and NJ was affiliated with Université de Lorraine (BETA). NH and NJ thank both institutions for their various supports. Conflict of interest statement ------------------------------ The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. ^1^This class of games has been first introduced by Moulin ([@B37]) as the *p*−beauty contest games, where *p* (often equal 2/3) stands for the target fraction of all numbers\' average. ^2^See Nagel ([@B39]) and Ho et al. ([@B31]) for early evidence from the laboratory, Costa-Gomes and Crawford ([@B18]) for a laboratory experiment supporting a behavioral model of bounded rationality, and Bosch-Domenech et al. ([@B7]) for related evidence from the field. ^3^Both Camerer ([@B12]) and Crawford ([@B19]) consider this game as a basic example of a dominance-solvable game, and a glaring case of a mismatch between theoretical predictions and actual behavior. ^4^If the game is played sequentially (so that player A moves first), the same solution can be obtained through backward induction. Note that if *s* \> *h*, the solution does not change (since *l* remains player B\'s weakly dominated strategy), but the outcomes are no longer Pareto-rankable. Beard and Beil ([@B5]), Schotter et al. ([@B44]), and Goeree and Holt ([@B27]) find that this environment also generates important violations of standard theoretical predictions. ^5^Cognitive skills are often measured using (amongst others) the Cognitive Reflection Test (CRT, Frederick, [@B25]), the Raven\'s progressive matrices test (Raven, [@B42]), or both (like in this study). The details of these two measures are presented in Section 2. ^6^RMET consists of a series of photos of the area of the face involving the eyes. Subjects are asked to choose one of the four words that best describes what the person in the photo is thinking or feeling. ^7^See Hampton et al. ([@B29]) for related evidence. ^8^Al-Ubaydli et al. ([@B1], [@B2]) also report that individual cognitive skills do not predict individual willingness to reach efficient outcomes in these two game. ^9^Herein, we restrict our design to these two game matrices and do not seek to further investigate the effects of monetary incentives on both players\' behavior. These effects are analyzed in detail in Jacquemet and Zylbersztejn ([@B35]) and Hanaki et al. ([@B30]). ^10^An English translation of the original instructions in French is provided as supplementary material. ^11^See Jacquemet and Zylbersztejn ([@B34]) for a detailed motivation and description of this design. ^12^Absence of monetary incentives for providing corrects answers is a standard procedure for both CRT and Raven\'s tests. Recent evidence on both tests suggests that monetary incentives do not *per se* affect people\'s performance. See Brañas Garza et al. ([@B9]) for a metastudy on the determinants of CRT scores and Eckartz et al. ([@B21]) and Dessi and Rustichini ([@B20]) for experimental evidence on the role of monetary incentives in Raven\'s test. ^13^See Raven ([@B42]) for an overview. ^14^Model 3 suggests that these two proportions are not significantly different: testing *H*~0~ : β~1~ + β~3~ = 0 yields *p* = 0.303. The increase in the proportion of decisions *R* between Human and Robot conditions is insignificant for Game 1 (*H*~0~ : β~2~ = 0, *p* = 0.679) and significant for Game 2 (*H*~0~ : β~2~ + β~3~ = 0, *p* = 0.054). ^15^See also Frederick ([@B25]) and Bosch-Domènech et al. ([@B6]) for related evidence. ^16^We test the difference in proportion of a given outcome between two experimental conditions by carrying out a bootstrap proportion test that accounts for within-subject correlation, i.e., the fact that the same individual takes 10 decisions. The procedure consists of bootstrapping subjects and their corresponding decisions over all 10 rounds instead of bootstrapping decisions as independent observations (see e.g., Jacquemet et al., [@B33], for a detailed description of the procedure). ^17^Given that most CRT scores in our sample are null and the higher the score, the less frequent it gets, dichotomizing the CRT score variable limits the impact of the outliers on the overall results. Supplementary material {#s6} ====================== The Supplementary Material for this article can be found online at: <http://journal.frontiersin.org/article/10.3389/fpsyg.2016.01188> ###### Click here for additional data file. [^1]: Edited by: Nikolaos Georgantzis, University of Reading, UK [^2]: Reviewed by: Adriana Breaban, Tilburg University, Netherlands; Gerardo Sabater-Grande, Jaume I University, Spain [^3]: This article was submitted to Personality and Social Psychology, a section of the journal Frontiers in Psychology

***Background.*** VRE is an endemic hospital-acquired organism in the US. There are no proven methods or means of decreasing or eliminating carriage of VRE which is a risk factor for invasive disease. Two small clinical studies showed reduction or elimination of VRE colonization using the probiotic LGG but neither study cultured stool for the presence of LGG, raising questions about whether effects on VRE colonization were due to LGG or other factors. ***Methods.*** To determine whether LGG can eliminate colonization by VRE, we undertook a randomized, double blind, placebo controlled trial in which subjects with VRE received 2 weeks of LGG or placebo. Stool samples were collected and cultured for the presence of VRE and LGG at baseline and days 7, 14, 21, 28 and 56. Presence of LGG was also assessed by PCR on day 14 samples from a subset of 7 subjects. Subjects were monitored closely for adverse events. ***Results.*** Of 694 patients screened, 11 were randomized. Five subjects received LGG and 6 received placebo. VRE was not eliminated during or after treatment in any subject. No significant decreases in VRE colony counts occurred in the subjects receiving LGG (table). LGG was recovered from stool in 2/5 subjects who received LGG. These 2 subjects were among 3/5 who received LGG and were also treated with antibiotics. LGG was detected by PCR in 4/4 samples from subjects who received LGG and 0/3 subjects in the placebo group. Adverse events were similar in the 2 groups. Median Stool VRE Counts by Study Day -------------------------------------- ------ ------ ------ ----------- ------ -------- ------ Day 0 6.97 4.48 7.72 Day 0 7.18 4.36 9.53 Day 7^a^ 7.49 5.72 8.46 Day 7^b^ 8.11 1.04 8.65 Day 14 6.26 4.51 9.54 Day 14 7.86 2.79 8.56 Day 21 6.90 4.60 7.43 Day 21^b^ 7.28 2.54 8.45 Day 28 6.59 4.18 7.38 Day 28 8.28 6.08 8.83 Day 56^b^ 5.26 3.81 7.52 Day 56^b^ 6.74 \<1^c^ 7.20 ^a^ one positive swab included, no CFU available ^b^ one sample not received ^c^ lowest level of detection based on 1ml plated ***Conclusion.*** Although the probiotic LGG is safe in patients with comorbid conditions, we were unable to demonstrate that it resulted in elimination of VRE in an RCT, likely because of continued administration of antibiotics in this patient population. ***Disclosures.*** **All authors:** No reported disclosures. [^1]: **Session:** 108. Clinical - Enteric Infections [^2]: Friday, October 10, 2014: 12:30 PM

A stressful life event may affect prognosis of breast cancer directly through stress-induced alterations of the immune and neuroendocrine system and indirectly through changes in health behaviour, such as physical activity, consumption of alcohol, compliance to therapy and coping with disease. In all, six previous studies have addressed the association between stressful life events and breast cancer prognosis: two of which found a significantly increased risk for recurrence ([@bib12]; [@bib10]), two studies found no association ([@bib6]; [@bib8]), whereas two studies found, contrary to what was expected, a significantly lower risk of recurrence ([@bib2]; [@bib5]). These inconsistent findings may be due to the use of different measures of exposure and methodological weaknesses including self-reported measure of exposure, small samples (*N*=94--665) and selection or recall bias (refer to web appendix: [Supplementary Table A1](#sup1){ref-type="supplementary-material"}). In this large population-based study, of more than 20 000 breast cancer cases, we use objective information from population-based registers and clinical databases to examine the association between a single life event stressor, loss of a partner and breast cancer prognosis. Loss of a partner is a common and also very stressful life event, implying considerable changes to everyday life ([@bib7]). Materials and methods ===================== Linkage of registry data ------------------------ Information on sex, date of birth, current and historical addresses, emigration, disappearance and death with date of these incidences were obtained from the Central Population Register in which all Danish residents since 1968 have been registered with a personal identification number allowing linkage of information between national registers ([@bib11]). Breast cancer ------------- We obtained information on date of breast cancer diagnosis (defined as date of primary surgery), date of recurrence (which is reported up to 10 years after diagnosis), tumour size (in mm), number of tumour-positive lymph nodes, malignancy grade, hormone receptor status and menopausal status, from Danish Breast Cancer Cooperative Group, which contain information on nearly 95% of all breast cancer cases in Denmark since 1977 ([@bib9]). Other cancers ------------- Information on first primary cancer, excluding nonmelanoma skin cancer, was obtained from the Danish Cancer Registry, which since 1943 have registered all cases of cancer (ICD-7) in Denmark ([@bib13]). Stressful life event -------------------- A stressful life event was defined as the death of a cohabiting partner either in the 4 years before breast cancer diagnosis or in subsequent years. 'Cohabitation\' was defined as two persons of the opposite sex over the age of 16, with a maximum age difference of 15 years, living at the same address with no other adults in the residence. Socioeconomic status and comorbidity ------------------------------------ Information on educational level and disposable income (categorised in [Table 1](#tbl1){ref-type="table"}) was obtained 2 years before breast cancer diagnosis from the population-based Integrated Database for Labour Market Research in Statistics Denmark with data on sociodemographic factors in Denmark since 1980 ([@bib14]). Information on comorbidity categorised according to the Charlson comorbidity index (scores 0, 1 and ⩾2), excluding cancers ([@bib3]) was obtained from the Danish National Patient Register with information on all somatic diseases leading to hospitalisation since 1977, and from 1995 also information on all outpatient visits ([@bib1]). Analysed cohort --------------- To attain accurate information from the included registers our base population is restricted to all 3.4 million Danish residents born between 1925 and 1973 who resided in Denmark 1994--2006 and entered the cohort at age 30 (for more details [@bib4]). From this population we identified 22 366 women diagnosed with breast cancer between 1 January 1994 and 31 December 2006 with no previous history of cancer (except nonmelanoma skin cancer), who resided in Denmark 2 years before diagnosis, and had a cohabiting partner up to 4 years before their breast cancer diagnosis. We excluded 1153 women with missing values on one or more covariates or who had resided in Denmark for less than 2 years. In all, 21 213 eligible women were followed for recurrence and death. Statistical analyses -------------------- We used Cox regression to assess hazard ratios (HRs) with 95% confidence intervals (CIs) for all-cause mortality and breast cancer recurrence, respectively, according to the vital status of the partner. For all-cause mortality follow-up time was counted from the date of diagnosis until death, emigration or 31 December 2010, whichever came first. For breast cancer recurrence follow-up time was counted from the date of diagnosis until death, emigration, 10 years of follow-up or 31 December 2006, whichever came first. The exposure, death of partner (after diagnosis), was included as a time-dependent variable, so that person--time before the death of the partner was counted as unexposed, whereas person--time after the date of death of the partner was counted as exposed. In all analyses, time since breast cancer diagnosis was used as the underlying time scale, and baseline hazards were allowed to vary across age at breast cancer diagnosis in 1-year intervals. The HRs were first adjusted for educational level and income, both considered as potential confounders. Subsequently, we adjusted for comorbidity, period of diagnosis, tumour size, number of positive lymph nodes, receptor status and malignancy grade (I--IV), as these factors are strongly associated with outcome, however, not obviously associated with the exposure. We estimated HRs in the intervals \[0--1\],\]1--2\],\]3--4\] years for exposure before diagnosis and for latencies of: \[0--2\],\] 2--5\],\]5--17\] years for exposure after diagnosis. We investigated whether a change in cohabitation status influenced the estimated association, by censoring at 1 January in the year in which the cohabitation status changed. Further, we estimated the association for women diagnosed with hormone-receptor-positive breast cancer and post-menopausal women only. Results ======= In the analysis of the association between loss of partner and all-cause mortality 172 773 person--years of follow-up were accrued, with a median follow-up of 7.7 years (ranging 0--17). During follow-up, 5660 women died, 762 lost their partner in the 4 years before their breast cancer diagnosis and 2259 lost their partner during follow-up. As expected, women who lost their partner were older and had a lower education and income compared with those who did not ([Table 1](#tbl1){ref-type="table"}). After adjustment for education and income as well as period of diagnosis, comorbidity and severity of breast cancer, women who had lost a partner were not at a significantly higher risk for recurrence or all-cause death from that of women who did not lose a partner, no matter if the event happened in the 4 years before diagnosis or in subsequent years ([Table 2](#tbl2){ref-type="table"}) or at different latencies ([Table 3](#tbl3){ref-type="table"}). We found only minor changes to the estimates when censoring at change in cohabitation status and when measuring the association only among women diagnosed with hormone-receptor-positive breast cancer or post-menopausal women (results not shown). Discussion ========== Our results do not support the concern that experiencing a major stressful life event, loss of a partner, negatively affects breast cancer recurrence or all-cause mortality, whether the event occurs in the 4 years before or 0--17 years after diagnosis. Of six previous studies four support this finding ([@bib6]; [@bib2]; [@bib8]; [@bib5]). Two previous studies found a higher risk of recurrence among women reporting stressful life events, however, both were of retrospective design and the observed estimate may reflect differential recall and reporting of events (Ramirez *et al*, 1989; [@bib10]). We addressed several methodological limitations of the previous studies. First, we included more than 30 times as many cancer patients as the largest study published so far. Second, the use of national registers and databases to identify the study population ensures high representativeness and minimal risk for selection bias. Third, the exposure (death of partner) was measured independently of the participants, eliminating recall bias. Fourth, the register-based cohort design ensures temporality and minimises loss to follow-up and misclassification of outcomes. Finally, we estimated the effect of the death of partner before and after diagnosis on both recurrence and all-cause mortality. Still, loss of partner is relatively rare among newly diagnosed breast cancer patients, and 73% of the cohort was alive and recurrence free at exit, resulting in small number of events in certain subgroups. The observed estimates of all-cause mortality may represent overestimates of breast cancer-specific mortality. We adjusted, however, the analyses for comorbidity and investigated also recurrence as outcome. Finally, an association may be present among women with poor coping resources or accumulated stressful life events. We were unable to take these into account. The death of a partner is a common major life event, and our finding of no association with breast cancer recurrence or all-cause mortality may provide reassurance for women confronting breast cancer. This work was funded by the Health Insurance Foundation; and the Danish Cancer Society. For their support in data management we thank Visti Birk Larsen, MD, and Marianne Steding-Jessen MSci, Survivorship, Danish Cancer Society Research Center. **Author contributions** All authors have contributed to the conception and design or analysis and interpretation of data and approved the final version of the report. [Supplementary Information](#sup1){ref-type="supplementary-material"} accompanies the paper on British Journal of Cancer website (http://www.nature.com/bjc) The authors declare no conflict of interest. Supplementary Material {#sup1} ====================== ###### Click here for additional data file. ###### Descriptive characteristics at entry of 21 213 women with breast cancer diagnosed in 1994--2006, Denmark, by vital-status of the partner at exit **Partner\'s death before diagnosis (*n*=762), No. (%)** **Partner\'s death after diagnosis (*n*=2259), No. (%)** **Partner alive at exit (*n*=18 192), No. (%)** -------------------------------------------------------- ---------------------------------------------------------- ---------------------------------------------------------- ------------------------------------------------- *Sociodemographic characteristics* Age at time of diagnosis (years) 30--39 2 (1) 25 (1) 1266 (7) 40--49 40 (5) 219 (10) 4640 (26) 50--59 154 (20) 670 (30) 6926 (38) 60--69 385 (51) 1061 (47) 4501 (25) ⩾70 181 (24) 284 (13) 859 (5) Mean/median 64/65 60/61 54/54 Level of education[a](#t1-fn1){ref-type="fn"} Basic or high school 440 (58) 1204 (53) 7127 (39) Vocational education 227 (30) 703 (31) 6551 (36) Higher education 95 (12) 352 (16) 4514 (25) Disposable income[b](#t1-fn2){ref-type="fn"} Lowest (1st quartile) 340 (45) 536 (24) 2489 (14) Middle (2nd--3rd quartile) 308 (40) 1107 (49) 8468 (47) Highest (4th quartile) 114 (15) 616 (27) 7235 (40) *Medical characteristics* Tumour size (mm) 0--10 98 (13) 415 (18) 2851 (16) 11--20 297 (39) 963 (43) 7352 (40) 21--50 311 (41) 780 (35) 6560 (36) ⩾51 22 (3) 53 (2) 737 (4) Unknown 34 (4) 48 (2) 692 (4) No. of positive lymph nodes 0 393 (52) 1323 (59) 8979 (49) 1--3 204 (27) 644 (29) 5380 (30) ⩾4 140 (18) 274 (12) 3434 (19) Unknown 25 (3) 18 (1) 399 (2) Malignancy grade Grade I 186 (24) 660 (29) 4197 (23) Grade II 261 (34) 738 (33) 6248 (34) Grade III 130 (17) 321 (14) 3795 (21) Grade IV 49 (6) 85 (4) 861 (5) Unknown 136 (18) 455 (20) 3091 (17) ER and PR Status[c](#t1-fn3){ref-type="fn"} Negative 129 (17) 380 (17) 3882 (21) Positive 588 (77) 1733 (77) 13 254 (73) Unknown 45 (6) 146 (6) 1056 (6) Period of diagnosis 1994--1996 49 (6) 608 (27) 3100 (17) 1997--1999 134 (18) 557 (25) 3792 (21) 2000--2002 217 (28) 580 (26) 4527 (25) 2003--2006 362 (48) 514 (23) 6773 (37) Charlson comorbidity index[d](#t1-fn4){ref-type="fn"} 0 603 (79) 1976 (87) 16 334 (90) 1 100 (13) 176 (8) 1201 (7) ⩾2 59 (8) 107 (5) 657 (4) Highest-education attained 2 years before diagnosis: basic school/high school (7--12 years of primary, secondary and grammar school); vocational training (10--12 years of education); higher education (⩾13 years of education). Disposable income extracted 2 years before diagnosis; income quartiles calculated according to the entire population in the investigated age group, after taxation and interest per person, adjusted for number of people in the household and deflated according to the 2000 value of the Danish crown, as: deflated household income/(no. of person in household 0.6). Negative indicates that individual was estrogen receptor (ER) and progesterone receptor (PR) negative. Positive indicates ER-positive or PR-positive. Accumulated value of all hospital contacts from 1978 to date of diagnosis; scores weighted by level of severity assigned to 19 conditions (excluding cancers) and grouped on the basis of the cumulated sum of scores of 0, 1 and ⩾2. ###### Hazard ratios (HRs) for death and recurrence with 95% confidence intervals (CIs) according to partner\'s vital status among 21 213 women with breast cancer diagnosed in 1994--2006, Denmark **Crude**[a](#t2-fn1){ref-type="fn"} **Adjusted HR**[b](#t2-fn2){ref-type="fn"} **Adjusted HR**[c](#t2-fn3){ref-type="fn"} ------------------------------------------ -------- --------- ------ -------------------------------------- -------------------------------------------- -------------------------------------------- -------------- ------ -------------- *All-cause mortality* Partner alive at exit (ref.) 18 192 144 410 4977 1.00 1.00 1.00 Exposed in the 4 years before diagnosis 762 5216 243 1.14 (1.00, 1.30) 1.07 (0.94, 1.22) 1.10 (0.95, 1.27) Exposed 0--17 years after diagnosis 2259 23 147 440 1.12 (1.01, 1.25) 1.09 (0.98, 1.20) 1.09 (0.98, 1.22) *Recurrence* Partner alive at exit (ref.) 19 312 86 453 2635 1.00 1.00 1.00 Exposed in the 4 years before diagnosis 762 2839 59 0.83 (0.64, 1.08) 0.82 (0.63, 1.06) 0.82 (0.61, 1.09) Exposed 0--17 years after diagnosis 1139 7581 85 0.98 (0.78, 1.22) 0.97 (0.78, 1.21) 0.93 (0.73, 1.18) Stratified by age at diagnosis in 1-year intervals. Adjusted for the highest-attained educational level, disposable income, and stratified by age at diagnosis in 1-year intervals (defined in [Table 1](#tbl1){ref-type="table"}). Adjusted for the highest-attained educational level, disposable income, period of diagnosis, comorbidity, tumour size, no. of tumour-positive lymph-nodes, hormone receptor status, malignancy grade and stratified by age at diagnosis in 1-year intervals (*N*=19 186) (defined in [Table 1](#tbl1){ref-type="table"}). ###### Hazard ratios (HRs) for death and their 95% confidence intervals (CIs) according to partner\'s vital status and time between diagnosis, death and death of partner among 21 213 women with breast cancer diagnosed in 1994--2006, Denmark **Crude**[a](#t3-fn1){ref-type="fn"} **Adjusted HR**[b](#t3-fn2){ref-type="fn"} **Adjusted HR**[c](#t3-fn3){ref-type="fn"} ---------------------------------------------- ----- ------- ----- -------------------------------------- -------------------------------------------- -------------------------------------------- -------------- ------ -------------- *All-cause mortality* Exposure before diagnosis Time from bereavement to diagnosis (years) \[0--1\] 228 1610 76 1.17 (0.93, 1.48) 1.15 (0.91, 1.44) 1.04 (0.82, 1.34) \]1--2\] 195 1391 65 1.14 (0.89, 1.46) 1.04 (0.81, 1.34) 1.05 (0.80, 1.38) \]2--4\] 339 2215 102 1.11 (0.91, 1.36) 1.04 (0.85, 1.26) 1.19 (0.96, 1.47) Exposure after diagnosis Time from bereavement to death (years) \[0--1\] 647 5768 173 1.20 (0.98, 1.48) 1.16 (0.95, 1.43) 1.11 (0.89, 1.39) \[1--5\] 766 7237 156 1.11 (0.97, 1.28) 1.08 (0.94, 1.23) 1.09 (0.94, 1.26) \>5 846 10142 111 1.07 (0.88, 1.31) 1.04 (0.85, 1.27) 1.08 (0.87, 1.34) *Recurrence* Exposure before diagnosis Time from bereavement to diagnosis (years) \[0--1\] 228 904 15 0.66 (0.40, 1.10) 0.66 (0.39, 1.09) 0.57 (0.32, 1.01) \[1--2\] 195 800 14 0.71 (0.42, 1.20) 0.69 (0.41, 1.17) 0.74 (0.42, 1.32) \[2--4\] 339 1135 30 1.05 (0.73, 1.51) 1.03 (0.72, 1.49) 1.11 (0.74, 1.65) Exposure after diagnosis Time from bereavement to death (years) \[0--1\] 467 2439 48 0.90 (0.61, 1.34) 0.90 (0.60, 1.33) 0.73 (0.46, 1.17) \[1--5\] 412 2837 28 1.01 (0.76, 1.34) 1.00 (0.75, 1.33) 1.01 (0.75, 1.36) \>5 260 2306 9 1.02 (0.52, 2.00) 1.02 (0.52, 2.00) 1.16 (0.57, 2.37) Stratified by age at diagnosis in 1-year intervals. Adjusted for the highest-attained educational level, disposable income and stratified by age at diagnosis in 1-year intervals. Adjusted for the highest-attained educational level, disposable income, period of diagnosis, comorbidity, tumour size, no. of tumour-positive lymph-nodes, hormone receptor status, malignancy grade and stratified by age at diagnosis in 1-year intervals (*N*=19 186) (defined in [Table 1](#tbl1){ref-type="table"}). The reference category (HR=1) are not experiencing the death of a cohabiting partner in the given year-interval.

"INTRODUCTION {#sec1-1}\n============\n\nHeat is a physical hazard that could create problems almost(...TRUNCATED)

"1. Introduction {#sec1}\n===============\n\nThe nervous system has two parts: (1) the CNS and (2) t(...TRUNCATED)

"Introduction {#s1}\n============\n\nThe sequencing of the human genome \\[[@pgen-0030099-b001],[@pg(...TRUNCATED)

"Introduction\n============\n\nBipolar disorder (BP) is among the more reliably diagnosed^[@bib1]^ a(...TRUNCATED)

"Introduction\n============\n\nThe metabolic properties of cancer cells are different from those of (...TRUNCATED)