Insulin-like signaling regulates developmental arrest, stress resistance and lifespan in the nematode Caenorhabditis elegans. However, the genome encodes 40 insulin-like peptides, and the regulation and function of individual peptides is largely uncharacterized. We used the nCounter platform to measure mRNA expression of all 40 insulin-like peptides as well as the insulin-like receptor daf-2, its transcriptional effector daf-16, and the daf-16 target gene sod-3. We validated the platform using 53 RNA samples previously characterized by high density oligonucleotide microarray analysis. For this set of genes and the standard nCounter protocol, sensitivity and precision were comparable between the two platforms. We optimized conditions of the nCounter assay by varying the mass of total RNA used for hybridization, thereby increasing sensitivity up to 50-fold and reducing the median coefficient of variation as much as 4-fold. We used deletion mutants to demonstrate specificity of the assay, and we used optimized conditions to assay insulin-like gene expression throughout the C. elegans life cycle. We detected expression for nearly all insulin-like genes and find that they are expressed in a variety of distinct patterns suggesting complexity of regulation and specificity of function. We identified insulin-like genes that are specifically expressed during developmental arrest, larval development, adulthood and embryogenesis. These results demonstrate that the nCounter platform provides a powerful approach to analyzing insulin-like gene expression dynamics, and they suggest hypotheses about the function of individual insulin-like genes.
Citation: Baugh LR, Kurhanewicz N, Sternberg PW (2011) Sensitive and Precise Quantification of Insulin-Like mRNA Expression in Caenorhabditis elegans. PLoS ONE 6(3): e18086. https://doi.org/10.1371/journal.pone.0018086
Editor: Simon Melov, Buck Institute for Age Research, United States of America
Received: August 20, 2010; Accepted: February 24, 2011; Published: March 22, 2011
Copyright: © 2011 Baugh et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was funded by the Howard Hughes Medical Institute (P.W.S.), the American Cancer Society (L.R.B. PF-06-028-01-DDC) and start up funds from Duke University (L.R.B.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscrip.
Competing interests: The authors have declared that no competing interests exist.
Insulin-like signaling contributes to homeostasis in multi-cellular animals by mediating physiological responses to environmental conditions through systemic signaling. In mammals, insulin signaling regulates carbohydrate metabolism, and insulin-like growth factor signaling controls growth. In invertebrates, insulin-like signaling regulates growth and metabolism as well as other aspects of developmental physiology . In the nematode C. elegans, insulin-like signaling regulates formation of a stress resistant, non-feeding developmental alternative to the third larval stage known as the dauer larva . Dauers form in conditions that are not favorable for growth and reproduction, and they serve as a dispersal mechanism. Dauer formation is triggered primarily in response to high population density but also limiting food and high temperature . Insulin-like signaling also regulates an acute form of developmental arrest that occurs in response to complete starvation (L1 arrest) , . Insulin-like signaling regulates adult lifespan in C. elegans , , , , as well as the fly Drosophila melanogaster and mammals , , .
Insects and nematodes each have several insulin-like peptides, and relatively little is known about the function of specific peptides . The C. elegans genome encodes 40 putative insulin-like peptides . The extent to which individual peptides have overlapping vs. specific function is not understood, and the complexity of the signaling network they comprise is unclear. Insulin-like gene expression is transcriptionally controlled in C. elegans , and expression analysis offers a way to infer the dynamics of the insulin-like signaling network in response to varying environmental conditions. A subset of insulin-like genes have been analyzed by transcriptional reporter genes, but conditional regulation was not investigated, expression was not quantified, and dynamics were not analyzed . Measurement of endogenous mRNA is ideal but challenging since insulin-like genes are expressed at relatively low levels in whole worms. Furthermore, microarrays produced to date measure only about half of the 40 insulin-like genes, and comprehensive QPCR analysis has not been reported.
nCounter is a commercially available platform for mRNA expression analysis . The nCounter “code set” contains a pair of ∼50 nt biotinylated DNA probes that are barcoded with different combinations of fluorescent tags. Total RNA is hybridized with the code set in solution phase, and DNA:mRNA hybrids are captured on the surface of a flow-cell, stretched by an electric field, and imaged. Fluorescent tags are optically resolved so that barcodes can be read and counted. Compared to other technologies for mRNA expression analysis, sensitivity should benefit from solution phase hybridization, whereas counting mRNA molecules should aid precision. In addition, the system has the benefit of measuring total RNA directly, avoiding biases potentially introduced by the use of enzymes or amplification. Although the approach is not genome-wide, accommodating tens to hundreds of genes per code set, the ease with which samples can be processed makes it excellent for experiments measuring expression over many time points, conditions or genotypes .
The objectives of this study were to determine the feasibility of using the nCounter platform for insulin-like mRNA expression analysis in C. elegans and to provide an overview of insulin-like gene expression. This goal includes validating the platform against the Affymetrix microarray platform, benchmarking and optimizing sensitivity and precision, and using the platform to measure insulin-like gene expression during the C. elegans life cycle. The results show that the platform produces reliable expression measurements with unparalleled sensitivity and precision. We also found that each of the insulin-like genes measured is expressed, and their expression patterns are largely distinct, suggesting that their regulation is complex leading to specific albeit overlapping functions.
Results and Discussion
Platform comparison between nCounter and microarray analysis
We used total RNA samples previously characterized by Affymetrix microarray analysis to test the nCounter platform . We purchased an nCounter code set from NanoString, Inc. that included probes for all 40 C. elegans insulin-like genes as well as the insulin-like receptor daf-2 , its transcriptional effector daf-16 , , and the daf-16 target gene sod-3 . The microarray contained probe sets for 22 of the insulin-like genes, as well as daf-2, daf-16, and sod-3. For microarray analysis 0.1 µg total RNA was used for biotin-labeled cRNA synthesis; for nCounter analysis 0.1 µg total RNA was used directly for hybridization. 53 RNA samples were analyzed on each platform. The experimental design included 18 groups of biological replicates, and the average of each group was compared for the 25 genes common to both platforms (Figure 1).
0.1 µg total RNA from 53 independent RNA preparations was used for nCounter analysis. The 53 samples comprise 18 groups of biological replicates (all but one with 3 replicates), and the average of each group is plotted for 25 genes common to both platforms. Each axis is on a log scale.
nCounter analysis of mRNA expression agreed well with microarray analysis (Figure 1). There was some quantitative disagreement between the platforms, but it was relatively minor and evenly distributed. The exceptions were daf-2 and daf-16. Both genes are large with multiple splice forms, and the two platforms did not target the same transcript sequences. daf-2 was not detected above background by microarray, but it was robustly detected by nCounter. daf-16 was detected with greater sensitivity by nCounter than microarray, but the expression pattern was qualitatively similar between the two (data not shown). Excluding daf-2 and daf-16, the correlation coefficient for the platform comparison was 0.85. We believe this is very good agreement considering the technical differences between the two procedures and the fact that the insulin-like genes are expressed at relatively low levels in whole worms. For this set of genes and this amount of RNA (0.1 µg) the sensitivity of the two platforms is comparable based on the number of genes detected above background. The nCounter results are modestly more reproducible with a median coefficient of variation of 17% compared to 25% for the microarray results, considering the same set of genes in each case. In summary, the nCounter platform performed well, matching or exceeding Affymetrix microarray analysis in terms of sensitivity and precision using the same mass of total RNA as starting material.
Optimization of total RNA mass used in nCounter hybridization
We performed an experiment to analyze the effects of total RNA mass used for hybridization to the nCounter code set on sensitivity and precision. The standard nCounter protocol specifies 0.1 µg total RNA for hybridization. However, our code set contained probes for only 43 genes, which is less than one-tenth of what the system can accommodate. In addition, the insulin-like genes are expressed at relatively low abundance. As a result, we typically observed less than 100,000 total counts per hybridization using 0.1 µg total RNA, though the system should allow for millions of total counts. This motivated us to try using more RNA per hybridization, and to investigate the effects of using no RNA as a control. We performed a set of technical replicates, all from a single total RNA preparation. We used no RNA, 0.1 µg, 1 µg and 10 µg total RNA in otherwise identical hybridizations. For each mass of RNA we performed 3 replicates. The code set includes a set of 10 positive controls. The positive controls include sequences from A. thaliana and D. melanogaster (File S1), and synthetic transcripts complementary to each were included in the code set at known concentrations. The standard curve resulting from positive control counts was not affected by total RNA mass (Figure 2A). We therefore used the sum of positive control counts to normalize the data across all 12 samples.
(A) Positive control standard curve is plotted for each RNA mass. Average transcript count (3 technical replicates) is plotted against the coefficient of variation for (B) no RNA, (C) 0.1 µg, (D) 1 µg, and (E) 10 µg. (F–H) Average transcript count is plotted in 3 pair-wise comparisons of the 3 RNA masses used. Legend in B applies to B–H. The y-axis of B–E is coefficient of variation. The vertical grey line in B–E represents background. Data are normalized by positive control counts. Transcript counts are plotted on a log scale. “CV” refers to coefficient of variation. One data point is omitted from B (a target with count of 57 and CV of 120%).
Sensitivity was improved by increasing the RNA mass used in hybridization. The average number of counts obtained for target transcripts (43 genes) was approximately 10-fold higher with 1 µg compared to 0.1 µg RNA and approximately 100-fold higher with 10 µg (Table 1). However, background increased with 10 µg RNA, as indicated by an approximate doubling in the average number of counts obtained for negative controls (A. thaliana and D. melanogaster probes; Table 1 and File S1). Consistent with the number of counts increasing, we also detected more target genes with more hybridized RNA. Background was modeled for each mass of RNA as the average of negative control transcript counts plus three standard deviations. Based on this background model, none of the negative controls were detected in any of the hybridizations, and only one of the targets was detected when no RNA was used for hybridization (Table 2). Over 1.5 times as many targets were detected with 1 µg RNA compared to 0.1 µg, and all but one of the 43 targets were detected with 10 µg. However, consistent with background increasing with more hybridized RNA, the lowest abundance positive control was not detected above background with 1 µg or 10 µg RNA (Table 2). This positive control (0.1 fmol) is not always detected in any condition, and the gain in sensitivity made by using more RNA for hybridization makes up for the minor increase in background.
Precision was improved by increasing RNA mass used in hybridization. With no RNA, target probes behaved like negative control probes both in terms of transcript counts and coefficient of variation (Figure 2B). Positive and negative control probes were unaffected by no RNA vs. 1 µg, but with 10 µg RNA negative controls produced more counts and a lower coefficient of variation (Table 1; Figure 2B–E), consistent with non-specific interactions contributing to background. The major effect of increasing RNA mass was to increase the number of counts and decrease the coefficient of variation, indicating increases in sensitivity and precision. The median target coefficient of variation was 8.5%, 4.5% and 1.8% for 0.1 µg, 1 µg and 10 µg RNA, respectively. Since technical replicates of a single RNA preparation were used, this experiment captured only technical error, and biological replicates will be more variable. Nevertheless, the decrease in coefficient of variation observed indicates that, with this code set and in this system, the power to detect differential expression is greater when hybridizing more RNA.
Relative transcript abundances were comparable when different masses of RNA were used for hybridization. When comparing results of 0.1 µg and 10 µg RNA, skew in target counts is evident on the low end of transcript abundance, but there is linear concordance for moderate and high abundance transcripts (Figure 2F). This result suggests that the lowest transcript counts (∼10–100) are affected by background, including those detected above background. Minor skew in the same count range is also evident when comparing 0.1 µg and 1 µg, but it is not evident when comparing 1 µg and 10 µg (Figure 2G,H). These results suggest that with this code set and RNA preparation the linear dynamic range of the assay is between 1 µg and 10 µg total RNA, a range that is also optimal for sensitivity and precision. These results also suggest that the system does not reach saturation with 3 million total counts. We conclude that for this code set and this system using microgram quantities of total RNA for hybridization is optimal in terms of sensitivity and precision, and we used 3 µg in subsequent experiments.
We recommend that other researchers working with the nCounter platform perform a similar technical experiment varying the mass of RNA used for hybridization to optimize their results. RNA is often limiting, and such results will aid in considering trade-offs between starting material quantity and data quality. In addition, researchers should consider transcript abundances when possible while designing code sets, since the total number of counts limits sensitivity. For example, the inclusion of invariant genes as internal controls (“housekeeping” genes for normalization) is advisable, but abundantly transcribed genes limit overall sensitivity.
Specificity of nCounter hybridization
We used deletion mutants to determine the specificity of the nCounter platform with our optimized hybridization conditions (3 µg total RNA). The code set includes 8 negative control probes, but the sequences are from A. thaliana and D. melanogaster and are less likely to cross-hybridize with C. elegans transcripts than the target probes. Because our targets include a 40 gene family and we increased the mass of RNA per hybridization, cross-hybridization is a concern. We obtained deletion alleles ins-4(tm3620), ins-5(tm2560) and ins-6(tm2416) from the National BioResource Project and prepared RNA for L1 arrest. Deletions tm3620 and tm2560 eliminate all of the sequence targeted by the nCounter probes, and tm 2416 eliminates all but 25 bp. Because the nCounter relies on two probes for each target, tm2416 should be null in the assay. Each deletion resulted in a dramatic reduction in the number of detected counts, but residual counts were detected (Figure 3). Deletion reduced ins-4 expression from around 1500 to 31+/−2 counts, ins-5 was reduced from around 1500 to 124+/−55 counts, and ins-6 was reduced from around 2000 to 137+/−53 counts. Deletion of ins-4 is the only one of the three that resulted in a comparable number of counts to the A. thaliana negative control probes (Table 1). The results for ins-5 and ins-6 suggest that a higher threshold should be used to define background than the one determined from the negative control probes. We tested only three targets for specificity with deletion alleles, and we assume their behavior is representative of the other 37 insulin-like genes. Based on these results, 400 counts is an appropriate background cut-off with this code set using 3 µg RNA. This number corresponds to the max of the three counts after gene deletion plus four times the standard deviation.
Insulin-like gene expression during the C. elegans life cycle
We used the nCounter platform to measure insulin-like mRNA expression throughout the C. elegans life cycle. mRNA expression was measured from embryos, each larval stage, adults, and during L1 and dauer developmental arrest (Figure 4). The embryos spanned mid-gastrulation. The 60 hr time point includes adults as well as their early embryos. The larval time points (12–48 hr) fall near the end of each larval stage, but because larval stages vary in length, time points fall in variable places within each larval stage and molt cycle. Given the dynamics of development, the samples measured here represent discontinuous developmental stages. Nevertheless, this experiment presents an overview of insulin-like mRNA expression during the life cycle, and future work with high-resolution time series analysis at particular stages will capture true dynamics.
A schematic of the C. elegans life cycle is presented along with plots of the average and standard deviation (3 biological replicates) of transcript counts for each gene. The ins-13 probe is not specific and reports expression of acdh-2. In the plots, “emb” refers to mid-grastrulation embryos, “L1” refers to L1 arrest, and “dau” refers to dauer arrest.
The expression patterns we detected are consistent with published results. Unfortunately, the ins-13 probe is not specific since the ins-13 and acdh-2 genes overlap and share common transcript sequences, and this probe was apparently dominated by acdh-2 expression . The superoxide dismutase gene sod-3 is a direct target of DAF-16  known to be up-regulated during L1 arrest and dauer formation , , . We found that sod-3 expression is up-regulated by approximately 10-fold and 30-fold during L1 and dauer arrest, respectively, compared to developing larvae (Figure 4). In addition, daf-28 expression is greatest at the end of L1 development (12 hr), consistent with reporter gene analysis . These results extend on those presented in Figures 1 and 3 to suggest that the data are valid.
Expression of insulin-like genes that have been functionally characterized reveals correlation between function and expression, suggesting that expression patterns can guide functional analysis. For example, expression of daf-28 during late L1 development is consistent with it promoting bypass of dauer formation given that this is the critical time for the dauer decision . In contrast, ins-18 is thought to function as an antagonist of the insulin-like receptor DAF-2 , , and it is up-regulated during L1 arrest and in dauer larvae consistent with promoting developmental arrest. ins- 10, -15, -16, -17, -20, -24 have a similar expression pattern, suggesting they could also promote arrest. ins-7 functions in adults limiting lifespan , and it is up-regulated at the end of larval development and in adults (Figure 4). ins-19 appears to be expressed specifically in adults (60 hr), suggesting that it too functions in adults. However, this time point also includes early embryos that have not yet been laid (earlier embryos than those included in the mid-gastrulation “embryo” sample), and ins-19 could be expressed maternally or during early embryonic development. ins-33 is regulated by the heterochronic pathway such that it is repressed during L1 arrest and early larval development, and it promotes germline proliferation, which occurs later in larval development , . Consistent with these functional insights, ins-33 expression increased through larval development, peaking at 36 hr after L1 arrest (Figure 4). ins-9, -35 and others were also expressed during larval development, but with different timing than ins-33 or daf-28, suggesting novel larval functions. There is no known function of insulin-like signaling during C. elegans embryogenesis, though the null phenotype of the daf-2 insulin-like receptor is embryonic lethal , . Consistent with a possible function of insulin-like signaling during embryogenesis, and suggesting candidate peptides, ins-2 and ins-34 were expressed specifically in embryos (Figure 4).
Expression of a few insulin-like genes is very close to the background inferred from deletion analysis (∼400 counts, Figure 3) and should be treated with caution. We used QPCR on the remaining RNA to generate similar expression profiles (excluding dauer since no RNA remained) for a few insulin-like genes with low signal (ins-8, -21, -25, -32, -36 and -39) as well as daf-28 and sod-3. sod-3 and daf-28 agree remarkably well between platforms (Figure 5). Much of the expression of the 6 low abundance genes is at or below background, but where they are above background there is generally good agreement with QPCR. Nevertheless, there are a couple of examples of what appears to be differential expression between embryo and L1 arrest on one platform not captured on the other. These are at most 2–3-fold differences in expression. Given the low abundance of these transcripts and the modesty of these differences this result does not undermine our nCounter analysis, but it does highlight the need for caution in interpreting results at or near background.
Average and standard deviation (3 biological replicates) of transcript abundance determined by QPCR is plotted for ins-8, -21, -25, -32, -36, -39, daf-28 and sod-3.
In summary, expression of nearly all 40 insulin-like genes was convincingly detected above background, and expression of most of them was modulated during the life cycle (Figure 4). We did not observe any correlation between expression pattern and classification of peptides as either α, β, or γ based on predicted structural features . Widespread modulation of expression is consistent with extensive developmental and physiological regulation of insulin-like gene expression. The expression patterns are largely distinct, suggesting that regulation of insulin-like gene expression is complex. Furthermore, distinct expression patterns suggest specificity in insulin-like gene function that merits functional analysis.
Nematode culture and sample preparation
RNA samples used for validation were prepared as described , and the same RNA was used here for nCounter analysis after 1.5 yr storage at −80°C. The same general procedure was used to prepare RNA for optimization and life cycle analysis, but with less precise staging. Nematode cultures were maintained, passaged and collected at 20°C. A starved 5 cm plate was used to inoculate a 10 ml liquid culture (S-complete medium plus 40 mg/ml E. coli HB101), and the liquid culture was incubated for 65 hr at 180 rpm and then bleached to produce a clean preparation of embryos . 500,000 embryos were suspended in 85 ml of S-complete and they were cultured for 31 hr at 180 rpm allowing them to hatch and enter L1 arrest. Cultures were fed by adding HB101 to a final concentration of 40 mg/ml and 5 larvae/µl. Cultures were bleached 60 hr after feeding. Embryos were suspended at 5 eggs/µl in S-complete with no food. The embryo sample was aged 3 hr (to mid-gastrulation) and collected. The remainder of the culture was incubated so that the animals hatched and entered L1 arrest. The L1 arrest sample was collected 24 hr after bleaching, and then the remainder of the culture was fed with 40 mg/ml HB101 to initiate post-embryonic development. Samples were collected 12, 24, 36, 48 and 60 hr later. For the dauer culture, embryos were also suspended at 5 eggs/µl and allowed to hatch and enter L1 arrest. 24 hr after bleaching they were fed with 1 mg/ml HB101 to initiate recovery with limiting food. Dauers were collected after 5 d; cultures had at least 99% dauers based on visual inspection. The entire procedure was repeated 3 times to produce 3 biological replicates. Samples were flash frozen in liquid nitrogen, and RNA was prepared using TRIzol (Invitrogen) plus sand. RNA concentration was determined by UV absorbance and RNA integrity was confirmed by agarose gel electrophoresis. This project has been reviewed and approved by the Duke University Institutional Biosafety Committee (IBC registration# 09-6093-01).
The nCounter code set was designed by NanoString, Inc. (Seattle, WA USA; http://www.NanoString.com/). It includes a pair of approximately 50 nt probes complementary to adjacent sequences in each target transcript. Probes were designed to be specific to the target transcript and to have a uniform melting temperature . Transcript sequences targeted are available in File S1. The ins-13 probe is not specific, and it appears to report expression of acdh-2 based on microarray analysis . The code set also includes probes for 10 positive control targets, and those target transcripts were included directly in the code set at known concentrations. The code set also includes 8 probes for negative control genes (from A. thaliana). For validation (Figure 1) 0.1 µg total was used for hybridization. For optimization (Figure 2), varying amounts of a single total RNA preparation was used for hybridization (no RNA, 0.1 µg, 1 µg or 10 µg); replicates were performed on a single RNA preparation (technical as opposed to biological). For life cycle analysis, 3 µg total RNA was used per hybridization, and three biological replicates were performed. Hybridization, flow cell preparation and scanning were performed according to the standard nCounter protocol. Transcript counts were normalized between samples in a particular experiment by the positive control transcript counts. For validation and life cycle analysis, target transcript counts were further normalized by the sum of target transcript counts. Ideally, the code set should include genes with invariant expression as internal controls for normalization, but the code set used here does not. The complete data set of insulin-like gene expression during the C. elegans life cycle is available in File S2. Microarray analysis was performed as described  and the complete data set is available from the Gene Expression Omnibus (GSE11055). Our results are MIAME compliant.
QPCR was performed using the Fast Start Universal SYBR Green Master (Roche) on an Eppendorf MasterCycler QPCR machine according to the manufacturer's instructions. Analysis of QPCR products by gel electrophoresis and melting curves is consistent with amplification of a single, specific product. Genomic DNA was used as template for standard curves for each primer pair, and the standard curve was used to convert cycle thresholds to copy number. cDNA was prepared from total RNA using oligo-(dT) primer and SuperScript III (Invitrogen), and the product was divided between QPCR reactions so that each 20 µl reaction had 20 ng-equivalents of total RNA as template. rpl-12 and rpl-19 were used as standards for normalization between RNA/cDNA preparations. Primer sequences used for QPCR are available in File S3.
Complete code set design, including genes, accession numbers, targeted region and target sequence for each gene (including positive and negative controls).
Complete data set for analysis of insulin-like mRNA expression through the C. elegans life cycle. The file contains three sheets: one with raw data, one with the averages of replicates after normalization, and one with the corresponding standard deviations.
We would like to thank Eric Davidson for use of the nCounter instrument, and Stefan Materna for advice and help processing samples for nCounter analysis. We would also like to thank the Mitani Lab and the National BioResource Project for providing deletion alleles of insulin-like genes.
Conceived and designed the experiments: LRB. Performed the experiments: LRB NK. Analyzed the data: LRB. Contributed reagents/materials/analysis tools: PWS. Wrote the paper: LRB PWS.
- 1. Wu Q, Brown MR (2006) Signaling and function of insulin-like peptides in insects. Annu Rev Entomol 51: 1–24.
- 2. Hu PJ (2007) Dauer. WormBook 1–19.
- 3. Golden JW, Riddle DL (1984) The Caenorhabditis elegans dauer larva: developmental effects of pheromone, food, and temperature. Dev Biol 102: 368–378.
- 4. Baugh LR, Sternberg PW (2006) DAF-16/FOXO regulates transcription of cki-1/Cip/Kip and repression of lin-4 during C. elegans L1 arrest. Curr Biol 16: 780–785.
- 5. Kao G, Nordenson C, Still M, Ronnlund A, Tuck S, et al. (2007) ASNA-1 positively regulates insulin secretion in C. elegans and mammalian cells. Cell 128: 577–587.
- 6. Kenyon C, Chang J, Gensch E, Rudner A, Tabtiang R (1993) A C. elegans mutant that lives twice as long as wild type. Nature 366: 461–464.
- 7. Kimura KD, Tissenbaum HA, Liu Y, Ruvkun G (1997) daf-2, an insulin receptor-like gene that regulates longevity and diapause in Caenorhabditis elegans. Science 277: 942–946.
- 8. Lin K, Dorman JB, Rodan A, Kenyon C (1997) daf-16: An HNF-3/forkhead family member that can function to double the life-span of Caenorhabditis elegans. Science 278: 1319–1322.
- 9. Ogg S, Paradis S, Gottlieb S, Patterson GI, Lee L, et al. (1997) The Fork head transcription factor DAF-16 transduces insulin-like metabolic and longevity signals in C. elegans. Nature 389: 994–999.
- 10. Bartke A (2005) Minireview: role of the growth hormone/insulin-like growth factor system in mammalian aging. Endocrinology 146: 3718–3723.
- 11. Clancy DJ, Gems D, Harshman LG, Oldham S, Stocker H, et al. (2001) Extension of life-span by loss of CHICO, a Drosophila insulin receptor substrate protein. Science 292: 104–106.
- 12. Tatar M, Kopelman A, Epstein D, Tu MP, Yin CM, et al. (2001) A mutant Drosophila insulin receptor homolog that extends life-span and impairs neuroendocrine function. Science 292: 107–110.
- 13. Pierce SB, Costa M, Wisotzkey R, Devadhar S, Homburger SA, et al. (2001) Regulation of DAF-2 receptor signaling by human insulin and ins-1, a member of the unusually large and diverse C. elegans insulin gene family. Genes Dev 15: 672–686.
- 14. Murphy CT, McCarroll SA, Bargmann CI, Fraser A, Kamath RS, et al. (2003) Genes that act downstream of DAF-16 to influence the lifespan of Caenorhabditis elegans. Nature 424: 277–283.
- 15. Malkov VA, Serikawa KA, Balantac N, Watters J, Geiss G, et al. (2009) Multiplexed measurements of gene signatures in different analytes using the Nanostring nCounter Assay System. BMC Res Notes 2: 80.
- 16. Materna SC, Nam J, Davidson EH (2010) High accuracy, high-resolution prevalence measurement for the majority of locally expressed regulatory genes in early sea urchin development. Gene Expr Patterns 10: 177–184.
- 17. Baugh LR, Demodena J, Sternberg PW (2009) RNA Pol II Accumulates at Promoters of Growth Genes During Developmental Arrest. Science.
- 18. Oh SW, Mukhopadhyay A, Dixit BL, Raha T, Green MR, et al. (2006) Identification of direct DAF-16 targets controlling longevity, metabolism and diapause by chromatin immunoprecipitation. Nat Genet 38: 251–257.
- 19. Honda Y, Honda S (1999) The daf-2 gene network for longevity regulates oxidative stress resistance and Mn-superoxide dismutase gene expression in Caenorhabditis elegans. FASEB J 13: 1385–1393.
- 20. McElwee JJ, Schuster E, Blanc E, Thomas JH, Gems D (2004) Shared transcriptional signature in Caenorhabditis elegans Dauer larvae and long-lived daf-2 mutants implicates detoxification system in longevity assurance. J Biol Chem 279: 44533–44543.
- 21. Li W, Kennedy SG, Ruvkun G (2003) daf-28 encodes a C. elegans insulin superfamily member that is regulated by environmental cues and acts in the DAF-2 signaling pathway. Genes Dev 17: 844–858.
- 22. Swanson MM, Riddle DL (1981) Critical periods in the development of the Caenorhabditis elegans dauer larva. Dev Biol 84: 27–40.
- 23. Hristova M, Birse D, Hong Y, Ambros V (2005) The Caenorhabditis elegans heterochronic regulator LIN-14 is a novel transcription factor that controls the developmental timing of transcription from the insulin/insulin-like growth factor gene ins-33 by direct DNA binding. Mol Cell Biol 25: 11059–11072.
- 24. Michaelson D, Korta DZ, Capua Y, Hubbard EJ (2010) Insulin signaling promotes germline proliferation in C. elegans. Development 137: 671–680.
- 25. Gems D, Sutton AJ, Sundermeyer ML, Albert PS, King KV, et al. (1998) Two pleiotropic classes of daf-2 mutation affect larval arrest, adult behavior, reproduction and longevity in Caenorhabditis elegans. Genetics 150: 129–155.
- 26. Patel DS, Garza-Garcia A, Nanji M, McElwee JJ, Ackerman D, et al. (2008) Clustering of Genetically Defined Allele Classes in the Caenorhabditis elegans DAF-2 Insulin/IGF-1 Receptor. Genetics 178: 931–946.
- 27. Lewis JA, and Fleming , John T (1995) Basic Culture Methods. In: Epstein HF, Shakes DC, editors. Caenorhabditis elegans: Modern biological analysis of an organism. San Diego: Academic Press. pp. 4–27.