Quantitative Trait Loci Affecting Survival and Fertility-Related Traits in Caenorhabditis elegans Show Genotype-Environment Interactions, Pleiotropy and Epistasis

Corresponding author: Thomas E. Johnson, Campus Box 447, Institute for Behavioral Genetics, University of Colorado, Boulder, CO 80309-0447., johnsont{at}colorado.edu (E-mail)

Communicating editor: L. PARTRIDGE

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have identified, using composite interval mapping, quantitative trait loci (QTL) affecting a variety of life history traits (LHTs) in the nematode Caenorhabditis elegans. Using recombinant inbred strains assayed on the surface of agar plates, we found QTL for survival, early fertility, age of onset of sexual maturity, and population growth rate. There was no overall correlation between survival on solid media and previous measures of survival in liquid media. Of the four survival QTL found in these two environments, two have genotype-environment interactions (GEIs). Epistatic interactions between markers were detected for four traits. A multiple regression approach was used to determine which single markers and epistatic interactions best explained the phenotypic variance for each trait. The amount of phenotypic variance accounted for by genetic effects ranged from 13% (for internal hatching) to 46% (for population growth). Epistatic effects accounted for 9–11% of the phenotypic variance for three traits. Two regions containing QTL that affected more than one fertility-related trait were found. This study serves as an example of the power of QTL mapping for dissecting the genetic architecture of a suite of LHTs and indicates the potential importance of environment and GEIs in the evolution of this architecture.

LIFE history theory suggests that the evolution of genes influencing aging is strongly dependent on, and in fact may be largely an accidental by-product of, selection of other life history traits (LHTs). Such traits include age of sexual maturity and fertility (for reviews, see ROSE 1991 ; ROFF 1992 ; STEARNS 1992 ; MARTIN et al. 1996 ). The specific environment(s) in which a population evolves will determine the selective pressures on LHTs; genes that experience little or no selective pressure in the evolutionary environment may have strong genetic effects in a new environment. Alternatively, there may be trade-offs [genotype-environment interactions (GEIs)] that evolved due to the need to adapt to two environments, such that one allele is selectively favored in one environment, while a different allele is favored in a second environment (see VIA and LANDE 1987 ; GILLESPIE and TURELLI 1989 for reviews). Thus variations in environment, past or present, allow alleles that limit survival in some environments to accumulate in the population.

Selection favors alleles that allow the individual to survive long enough to reproduce competitively, as determined by other LHTs. The absence of significant selective pressure for the continued survival of the organism after reproduction allows mutations that increase the probability of death beyond this age to accumulate in the population. Alternatively, genes specifying one LHT may act pleiotropically on another LHT, such that selection of the first LHT can result in changes in the second, confounding or enhancing the effects of selection on that gene. Genes for a given LHT may also interact epistatically, confounding or enhancing the effects of selection for that trait on a given gene. Thus, an understanding of a species' LHTs, the effects of LHT genes in different environments, and the genetic architecture of these LHTs helps to clarify the evolutionary constraints limiting species survival and longevity.

Previous approaches to these issues have focused on genome-wide assessments. The existence of GEIs for fitness and the basis for their maintenance has been studied by comparing genetically distinct populations across environments (see FRY et al. 1996 , and references therein). Evidence for pleiotropy among LHTs has been sought, either by comparing the responses of one LHT to artificial selection on another (e.g., ROSE and CHARLESWORTH 1981A , ROSE and CHARLESWORTH 1981B ; LUCKINBILL et al. 1984 ; ZWAAN et al. 1995 ) or by comparing species that have evolved in different environments with presumed different selection regimes (e.g., AUSTAD 1993 ; KELLER and GENOUD 1997 ).

The approach taken here, quantitative trait locus (QTL) mapping, in the nematode Caenorhabditis elegans, allows us to address these issues at individual genetic loci. C. elegans is a useful model system for studying LHTs, and aging in particular, due to its short generation time and life span, small size, and well-characterized genetics. We use QTL mapping strategies (LANDER and BOTSTEIN 1989 ; HALEY and KNOTT 1992 ; TANKSLEY 1993 ; ZENG 1993 , ZENG 1994 ) to map loci showing allelic variation for LHTs within our test population, a series of recombinant inbred (RI) strains generated from two "wild-type" strains [Bristol (N2) and Bergerac (BO)]. Loci affecting LHTs can be identified and their GEIs or pleiotropic effects at single loci, and their epistatic interactions between loci, can be analyzed. These loci are good starting points for identifying genes important in the specification of LHTs. Each QTL represents a large genetic region containing many genes, such that GEIs, pleiotropic effects, or epistatic interactions involving specific genes must be established by further genetic characterization.

Three previous studies reported mapping QTL for survival in C. elegans. All of these studies examined survival in liquid media; the first two examined individual worms at the F₆ generation from N2-Bristol by Bergerac (EBERT et al. 1993 ) or N2-Bristol by DH424 (EBERT et al. 1996 ) crosses. A third study (SHOOK et al. 1996 ) used N2-Bristol by Bergerac (BO) RI strains, also used in the current study, and found QTL for survival in liquid media and for hermaphrodite self-fertility on solid media as well as QTL for internal hatching (bagging). NUZHDIN et al. 1997 have also used QTL mapping methods to find loci affecting longevity in Drosophila.

This study extends the previous results in C. elegans by localizing QTL for a suite of survival and fertility-related traits, all on solid media. This allows a comparison of our results for survival-related traits with our results for fertility-related traits in the same environment. This is important, because, as argued above, genes may have different effects in different environments, and any attempt to understand the genetic architecture of LHTs (and any pleiotropic interactions in particular) requires assessment in a common environment. We map the survival-related traits—life span and internal hatching—and the fertility-related traits—total fertility, fraction of fertility accrued at early ages, age at sexual maturity, and the Malthusian rate of population growth. HODGKIN and BARNES 1991 demonstrated that the onset or amount of early fertility can be a more important criterion for population growth (a fundamental measure of fitness) than total fertility. We have therefore assessed all these traits and consider the correlations among them, in addition to mapping QTL for each.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

General methods, media, and strains:
Both survival and fertility assays were performed on NGM agar plates (SULSTON and HODGKIN 1988 ) spotted with OP50. We examined 81 RI strains derived from two wild-type C. elegans varieties: Bristol (N2) and Bergerac (BO) as described in JOHNSON and WOOD 1982 and SHOOK et al. 1996 . Stocks were maintained at 20°, and all manipulations were done in a 20° constant temperature room. Strains were axenized (made free of bacterial and fungal contamination; EMMONS and YESNER 1984 ) and analyzed at least two generations postaxenization. Strains found to have become contaminated during the course of any phenotypic assay were not included in the results of that assay.

Survival assay:
Life span ("survival") was assessed using one cohort for each strain by picking 25 third-larval-stage (L3) worms at random from age-synchronous populations laid during a 3- to 7-hr period. Plates were inspected and worms were transferred daily during the fertile period and were inspected three times weekly thereafter. Death was determined by absence of movement, pharyngeal pumping, or touch-response, as in LITHGOW et al. 1994 . This is different from our previous study (SHOOK et al. 1996 ) in which internal decay and lack of turgor pressure were also used as indicators of death; this may have slightly biased our assessment of survival time downward, but would have done so equally for all strains. Three terminal conditions in addition to senescent death were scored: death by internal hatching of progeny ("bagging"), death by desiccation on the wall of the plate ("desiccation"), and loss during handling. All terminal events other than senescent death were calculated as the fraction of the total starting population. Only strains for which at least five worms died senescent deaths were considered for further analysis of survival.

Fertility assay:
The RI strains, N2 and BO, were assayed for age-specific fertility in four replicates, two on each of two dates. Five second-larval-stage (L2) or L3 worms were picked at random from age-synchronous populations laid during a 2-hr period by young gravid adults. Starting at 62 hr of age, observations were made every 2 hr to determine age of first reproduction ("alpha").

Age-specific fertility was measured by serial transfer of cohorts to fresh plates at 8 hr after worms with the earliest alpha started laying eggs, followed by two transfers at 8-hr intervals, then two at 12-hr intervals, and then daily until the end of each strain's fertile period. Adult survivorship was recorded at each transfer and censoring events were noted. Progeny produced was measured as the number of L2 to L3 stage larvae developed on the plate. The age-specific hourly fertility rate for a given interval was calculated as (number of progeny/number of parents at interval midpoint/interval width). Because parents that died of natural causes (senescence, bagging, or desiccation) were included, this measure gives the net fertility rate, or more specifically, the mean number of L2 to L3 progeny produced by a single worm that began the experiment as an L2 to L3.

Population growth rate (r) was determined by extrapolating the age-specific hourly fertility for several generations, assuming that each progeny from succeeding generations would have the same age-specific hourly fertility as its parent. In combination with survival data, this yields an expected number of worms alive at any time in the future. This approach is analogous to an iterated Leslie matrix (CAREY 1993 ). r was determined from the equation r = ln(N_{x + t}/N_x)/t, where N_x is the population at some time x, and N_{x + t} is the population size t hours later; for these analyses we took the average of r at several time points between 600 to 1200 hr, when the plot of population size with time began to visibly flatten (not shown, but see Figure 1). Because r was found not to be normally distributed, it was transformed as e^100r, which is normally distributed, and can be thought of as the factor by which population would increase in 100 hr (an approximate mean generation time) assuming an equilibrium age distribution and infinite resources. This value is called "population growth" in further analyses. "Early fertility" was determined by summing fertility prior to 100 hr; "late fertility" is the remainder. "Early fertility fraction" is early fertility divided by total fertility.

View larger version (17K): In this window In a new window Download PPT slide   Figure 1. Population size vs. time for N2 and BO, based on projected population growth rates. Calculations were based on age-specific fertility data.

Molecular markers:
Genotype was assessed as described in SHOOK et al. 1996 ; six additional Tc1-based markers were scored in this study. Two of these (stP17 and stP128) are from the original set developed by WILLIAMS et al. 1992 , while four others are based on Tc1 polymorphisms found by the Plasterk lab [pkP5062 (I), pkP5036 (I), and pkP580-U (X); R. KORSWAGEN, personal communication] or other Tc1 inserts not previously characterized [zP589 (III)]. See Table 1 for new primer sequences. The primers for both pkP5062 and pkP5036 recognize sequences on the left flank of the Tc1 inserts, while all other markers used in this study recognize sequences on the right flank of the Tc1 inserts. The new markers were assayed as for other markers used in SHOOK et al. 1996 , with the following exceptions: pkP5062 was amplified using 1.2 mM MgCl₂ and internal Tc1 primer Tc1.L3 (which is left facing); pkP5036 was amplified using 1.6 mM MgCl₂ and internal Tc1 primer Tc1.L1 (also left facing); and zP589-L was amplified with an annealing temperature of 48°. The new markers had nonpolymorphic background bands that acted as internal controls to ensure that amplification had occurred.

Statistical analyses:
Distribution statistics for all traits were determined using SPSS 5.0 (SPSS 1992 ). The proportion of phenotypic variance for survival due to genotype was computed as V_G/(V_G + V_E), where V_G (²_B) is the component of variance between strains and V_E (²_W) is the component of variance within strains.

To detect QTL we used a composite interval mapping approach (ZENG 1993 , ZENG 1994 ), using model five of QTL Cartographer (BASTEN et al. 1996 ) with a window size of 20 cM. Model five takes into account genetic background effects due to the two most significant markers on each of the chromosomes other than the one being analyzed, as well as markers on the same chromosome farther away than the window size; this allows the genetic effect of the locus under consideration to be assayed in relative isolation. A window size of 20 was chosen to prevent markers near the locus being analyzed from being considered as background; nearby markers will have similar genetic effects due to linkage, which should not be considered as background. One thousand permutations of the strain phenotypes with respect to genotypes were considered by QTL Cartographer (BASTEN et al. 1996 ) for each trait to give experiment-wide critical significance levels, based on the method of CHURCHILL and DOERGE 1994 , and interpreted by the standards of LANDER and KRUGLYAK 1995 ; an effective P = 0.10 was considered suggestive, while an effective P = 0.05 was significant. Additive effect sizes at the position of the QTL mapped by the interval mapping approach were calculated by QTL Cartographer (BASTEN et al. 1996 ) and normalized by the standard deviation of the entire population; these estimates of QTL effect are likely to be biased upward. Positive additive effects indicate that Bergerac has the higher mean.

GEIs for survival were analyzed for each marker nearest a QTL for survival in either environment using a regression model that included a marker genotype term, an environment term (both coded as "1" or "-1") and the GEI term, which was the product of the two. We used a Bonferroni correction for the number of independent markers we tested (BELKNAP 1992 ) to evaluate the significance of the regression coefficient for the GEI term at a given marker.

Epistasis among markers for a given trait was tested for by two-factor ANOVA (KSHIRSAGAR 1983 ) of strain means by all possible marker pairs, with marker genotypes as the two factors. F-scores resulting from 10,000 permutations of strain phenotypes with respect to genotypes were generated for each trait by a C program (available on request) written by D.R.S. to determine critical experiment-wide significance levels after the methods of CHURCHILL and DOERGE 1994 . Epistatic effect (4i) was calculated as (A + D - B - C)/SD (MATHER and JINKS 1977 ), where A and D represent the means of strains with the homotypic cases (either N2 or BO alleles at both markers), B and C represent the means of the heterotypic cases (N2 allele at one marker, BO allele at the other marker), and SD is the standard deviation of the whole population. A multiple regression approach was used to find those individual markers and epistatically interacting pairs of markers that significantly accounted for the phenotypic variance observed for each trait. All 40 markers were entered into the model at the start along with those interacting pairs of markers that had been shown previously to be suggestive or significant. Markers were removed from or reentered into the model using the stepwise method for regression (SPSS 1992 ), with a regression coefficient significance cutoff that was chosen to reflect the appropriate experiment-wide significance cutoffs as determined by permutation testing for single-marker linear regression by QTL Cartographer (see above); in practice this was ~P = 0.003. The significance cutoff for the epistatically interacting marker pairs found by permutation analysis (see above) was somewhat lower than this, but we started only with those pairs that we had already shown were at least suggestive. Determination of the proportion of phenotypic variance due to epistatic interactions was determined by comparing adjusted R² scores for regressions with and without epistatic interactions. The markers that interacted epistatically were also included individually in these regressions, so that only the interaction variance from these markers was represented by the difference between adjusted R² scores. Markers were coded as 1 (for BO genotype) or -1 (for N2), and epistatic interactions were represented as the product of the two markers for each strain.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Measurements of life history traits:
Survival time was assayed in a single replicate for all 79 RI strains. The fraction of the initial population of worms dying due to bagging or desiccation was also calculated. The distribution statistics for survival, bagging, and desiccation are presented in Table 2. Age-specific fertility data were collected from four replicate assays and several fertility-related traits were subsequently derived from these data: early, late, and total fertility, early fertility fraction, and alpha. Total fertility was not significantly affected by date of assay (F = 0.12, P = 0.73) or by assayers (F = 0.10, P = 0.75), and all four replicates were highly correlated (lowest r = 0.87, 79 cases). Both alpha and population growth were significantly different between dates of assay (F = 25.14, P < 0.001; F = 14.65, P < 0.001, respectively), but not between assayers. The mean alpha for the second assay date was ~3 hr earlier, reflecting an overall systematic bias. Correlations between dates of assay for both alpha and population growth were highly significant (r = 0.75 and 0.84, respectively), and correlations among the four replicates were also high (lowest r = 0.67 for alpha, 0.79 for population growth). Thus, we used the combined mean of all four replicates for further analyses. Distribution statistics for traits based on the fertility assay are also presented in Table 2. We calculated population growth based on age-specific fertility and survival data rather than direct measurements. Examples of projected population growth curves for the first 500 hr of culture for N2 and BO are presented in Figure 1.

Correlations among traits and estimation of V_G/V_P:
Correlations among all traits measured in this study, as well as those measured in our previous study (SHOOK et al. 1996 ), are presented in Table 3. Note that survival on agar is not correlated with earlier survival results in liquid. Comparing survival and fertility traits, only late fertility and bagging on agar are significantly correlated. Unlike the survival traits, most fertility traits were highly correlated. Early fertility, late fertility, and alpha are statistically independent and they are highly correlated (r 0.65). Early fertility is highly correlated with (r = 0.99) and is the major determinant of projected population growth rate. Alpha also shows a strong correlation (r = -0.78) with population growth, while late fertility is the least important of these three determinants (r = 0.64).

The proportion of phenotypic variance due to among-strain differences in survival was 0.33. Individual measures within strains were not available for other traits, and so proportion of phenotypic variance due to genetic effects could not be calculated using within-strain variances; however, see the section on multiple regression below.

Mapping QTL:
A composite interval mapping method (ZENG 1993 , ZENG 1994 ) using QTL Cartographer (BASTEN et al. 1996 ) was used to map QTL for all traits (Figure 2). Forty genetic markers [6 new and 34 reported previously (SHOOK et al. 1996 )] were used to localize QTL. For survival (Figure 2A) one suggestive QTL was found on the X chromosome (near stP129; see Table 4 for effect sizes and positions). No QTL for bagging were found. For total hermaphrodite self-fertility (Figure 2B) two significant QTL were found: one on II (near maP1) and one on IV (near stP51). For early fertility fraction (Figure 2C) one suggestive QTL was found on II (near stP98). For alpha (Figure 2D) two peaks were found on II (near stP19 and stP127), which probably represents only one significant QTL, and one was found on IV (near stP51) as well as a suggestive QTL on V (near stP18). For population growth (Figure 2E) suggestive QTL were found on II (near stP36 and maP1) and III (stP19), and one significant QTL was found on IV (stP51). Late fertility was found to have a QTL map very similar to that of total fertility, while the map for early fertility is nearly identical to that for population growth (data not shown).

View larger version (77K): In this window In a new window Download PPT slide   Figure 2. QTL maps. Plots are from composite interval mapping by QTL Cartographer, model 5, with a window size of 20 cM. LR, likelihood ratio. (*) Suggestive (corrected P < 0.10) and (**) significant (corrected P < 0.05) composite interval mapping peaks, as determined by permutation testing. The number of RI strains used for each QTL map is indicated in Table 3. Updated marker positions can be found in ACEDB (DURBIN and MIEG 1991 ). (A) Survival on agar (mean life span); (B) total fertility; (C) early fertility fraction (fraction of progeny produced before 100 hr); (D) alpha (time of onset of egg laying); and (E) population growth (factor of population increase over 100 hr, calculated using projected population growth rate).

Effects of environment:
Environment had a highly significant effect on survival times (t = 3.07, P = 0.003) when considered alone. This reflects the 2.7-day difference in mean survival times on agar [22.1 days (SHOOK et al. 1996 )] vs. in liquid (19.4 days, Table 2). The life spans for the parental strains N2 and BO also show a greater difference on agar than they do in liquid [13.3 and 19.5 days, respectively, on agar; 15.7 and 17.5 days in liquid (SHOOK et al. 1996 )]. The cause of this variation in life span is unknown. The extent of the differences in worm "lifestyle" experienced in liquid vs. solid is not clear, but worms show far more rapid undulation behavior in liquid culture (CROLL 1975 ; D. R. SHOOK and T. E. JOHNSON, personal observations) and have reduced food consumption (AVERY and THOMAS 1997 —also indicated by our observation of greater bagging in liquid). Other explanations may include small differences in assay technique (see MATERIALS AND METHODS) or differences in damage done to worms during transfer from plate to plate (by wire pick on solid media and by mouth pipette in liquid).

To test whether any of the four QTL for survival on agar (current study) and survival in liquid (SHOOK et al. 1996 ) were in fact specific to one environment, we tested for GEIs. Of the four markers most closely associated with these QTL, stp5 had a significant GEI (t = 2.86, P = 0.005), stp101 was suggestive (t = -2.24, P = 0.026), and the last two (stP129, t = 0.58, P = 0.565; and stP61, t = 1.02, P = 0.308) showed no GEI [assuming a Bonferroni correction of 3, for three independent marker groups (stP129 and stP61 are separated by only 3.6 cM)]. The QTL mapping to stP129 and stP61 are right next to each other on the X chromosome, and both influence life span in the same direction (higher for BO), and so may in fact represent the same QTL mapped to slightly different positions in the two different environments. When environment was considered as a cofactor in an ANOVA of each marker's effect on survival in both environments, only stP61 was suggestive (F = 9.35, P = 0.003).

Epistasis within traits:
All markers were tested pair-wise for each trait by two-way ANOVA to determine whether they showed any epistatic interaction (Table 5). We found one case of significant epistasis for bagging between stP124 and stP3, and one suggestive case between stP41 and stP2. For total fertility we found two cases of significant epistasis between maP1 and stP127 and between stP5 and stP6. For alpha we found one case of suggestive epistasis between stP98 and bP1, and three cases of significant epistasis between stP19 and stP17, between zP589-L and sP4, and between stP5 and stP6. For population growth we found one case of suggestive epistasis between stP5 and stP6, and one case of significant epistasis between maP1 and zP589-L.

Multiple regression models:
We used multiple regression analysis to determine the proportion of the total phenotypic variance for each trait explained by significant markers and epistatic interactions found in this study (Table 6). The amount of phenotypic variance accounted for ranged from 13% (for bagging) to 46% (for population growth). Epistatic variance accounted for ~10% of the phenotypic variance for total fertility, alpha, and population growth.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

QTL:
Over all life history traits, we found at least nine suggestive or significant QTL by composite interval mapping. Only one QTL for survival was found, while the other eight were associated with the fertility-related traits (two for total fertility, one for early fertility fraction, three for alpha, and two for population growth). The two peaks for alpha on chromosome III might reflect two QTL, but it seems more likely that mgP21 is simply a poor indicator of genotype. A similar case is seen for population growth on chromosome II (Figure 2 and Table 4). The multiple regression analysis supports the existence of only one QTL in each case, because only the stronger of the two peaks is included in the final multiple regression model for each trait (Table 6).

While a higher trait value for survival, fertility, and population growth indicates higher fitness, the interpretation for bagging is ambiguous. A lower value for alpha and for early fertility fraction is generally considered to confer higher fitness for an organism with a life style like C. elegans (HODGKIN and BARNES 1991 ; ROFF 1992 ; STEARNS 1992 ). Overall, only the QTL for survival and early fertility fraction are associated with increased fitness for the Bergerac genotype. All other fertility-related traits showed increased fitness for the N2 genotype.

Effects of environment:
We found no overall correlation between survival on agar and survival in liquid. Of the four QTL mapped for survival in the two environments, two show GEIs, indicating that these QTL affect survival differently in different environments. The existence of GEIs for LHTs has important implications in driving adaptation (and perhaps speciation) in specific environments and for maintaining genetic variation in varying environments. The existence of GEIs also suggests caution in studying pleiotropy among traits (they must be studied in the same environment for the pleiotropies to be evolutionarily relevant) and for studying the evolution of traits and their genetic architecture (they must be assessed in the environment in which evolution has or will take place for implications about the mechanisms of their evolution to be relevant). However, the assay of LHTs even in a completely artificial environment should help to elucidate the genetic pathways that determine LHTs, whether those genes played a direct role in the evolution of a given species or not.

Pleiotropy:
Among the fertility-related traits, there are two cases of QTL that affect more than one trait mapping to the same region. The QTL near stP51 affects total fertility, alpha, and population growth while that near maP1 affects both total fertility and population growth. For both QTL, the N2 genotype confers higher fitness for each of these traits. Because population growth is dependent on both alpha and total fertility, this set of positive fitness pleiotropies is not surprising. Also notable was the absence of QTL on chromosomes III and V for population growth (and early fertility), where QTL are found for alpha; this may indicate that some QTL for alpha control only the timing of the onset of fertility, while others control the number of progeny produced early.

Epistasis:
We found a total of 10 suggestive or significant epistatic interactions affecting different traits. Except for maP1 for both total fertility and population growth and zP589-L for alpha, none of the markers nearest QTL for a given trait were involved in epistatic interactions for that trait. While 7 of the interactions are unique to a specific trait, the interaction between stP5 and stP6 is pleiotropic for total fertility, alpha, and population growth. The epistatic marker pairs remaining in the multiple regression models for total fertility, alpha, and population growth account for 9–11% of the phenotypic variance. This suggests that QTL modeling and analysis approaches would do well to include the detection and analysis of epistatically interacting QTL. C. elegans is a self-fertilizing hermaphroditic species and thus normally highly inbred. This may lead to coadapted gene complexes, which are detected as epistasis when divergent strains such and N2 and BO are crossed. Because gene interactions are of particular interest in understanding the genetic pathway of a given trait, inbreeding species may prove especially valuable for QTL mapping studies.

Comparisons with previous QTL mapping studies in C. elegans:
The QTL of EBERT et al. 1993 , EBERT et al. 1996 for survival in liquid, which were at least partially in agreement with our previous results (see SHOOK et al. 1996 for discussion), are not replicated in our results for survival on solid media. This study did confirm two QTL for total fertility in the same regions (near maP1 and stP51) as detected previously (SHOOK et al. 1996 ), supporting the replicability of the results, despite small differences in methods of assay. However, in the earlier study a QTL associated with stP19 was found, which was not replicated here.

Conclusions:
We have taken a first step in unraveling the genetic architecture of life history traits in a simple model organism, C. elegans; further studies using this system may eventually help us to understand the complex functional and genetic relationships among these traits as well as their evolution. These studies provide evidence for several QTL for survival and other life history traits. More QTL are suggested by analysis of epistatic interactions among traits; these interactions make up a substantial part of the genetic variance for these traits. We found GEIs for two survival QTL, indicating the importance of environment in considering the evolution of LHTs.

	FOOTNOTES

¹ Present address: Department of Biology, Gilmer Hall, University of Virginia, Charlottesville, VA 22903.

	ACKNOWLEDGMENTS

We thank Steve Wilson and Doug Hinerfeld for technical assistance with fertility and survival assays and strain genotyping. Robin Corley assisted in determining the proportion of variance due to epistatic effects in the multiple regression models and John DeFries made helpful suggestions throughout the project. John Thaden, Rik Korswagen, and the Plasterk lab made the new Tc1 markers available to us. This work was supported by research grants from the National Institutes of Health (NIH; K02-AA00195, P01AG08761, RO1-AG08322, and RO1 AG10248) to T.E.J., from the Glenn Foundation, and by other private gifts to the University of Colorado. D.R.S. was supported by NIH training grant 5T32-MH16880.

Manuscript received June 3, 1998; Accepted for publication July 8, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AUSTAD, S. N., 1993  Retarded senescence in an insular population of Virginia opossums (Dilelphis virginia). J. Zool. 229:695-708.

AVERY, L., and J. H. THOMAS, 1997 Feeding and defecation, pp. 679–716 in C. elegans II, edited by D. L. RIDDLE, T. BLUMENTHAL, B. J. MEYER and J. R. PRIESS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

BASTEN, C. J., B. S. WEIR and Z. ZENG, 1996 QTL Cartographer: a reference manual and tutorial for QTL mapping. North Carolina State University.

BELKNAP, J. K., 1992  Empirical estimates of Bonferroni corrections for use in chromosome mapping studies with the BXD recombinant inbred strains. Behav. Genet. 22:677-684[Medline].

CAREY, J. R., 1993 Applied Demography for Biologists, with Special Emphasis on Insects. Oxford University Press, Oxford.

CHURCHILL, G. A. and R. W. DOERGE, 1994  Empirical threshold values for quantitative trait mapping. Genetics 138:963-971[Abstract].

CROLL, N. A., 1975  Components and patterns in the behavior of the nematode Caenorhabditis elegans.. J. Zool. 176:159-176.

DURBIN, R., and J. T. MIEG, 1991 A C. elegans database. Documentation, code and data available from anonymous FTP servers at lirmm.lirmm.fr, cele.mrc-lmb.cam.ac.uk and ncbi.nlm.nih.gov.

EBERT, R. H., V. A. CHERKASOVA, R. A. DENNIS, J. H. WU, and S. RUGGLES et al., 1993  Longevity-determining genes in Caenorhabditis elegans: chromosomal mapping of multiple noninteractive loci. Genetics 135:1003-1010[Abstract].

EBERT, R. H., II, M. A. SHAMMAS, B. H. SOHAL, R. S. SOHAL, and J. K. EGILMEZ et al., 1996  Defining genes that govern longevity in Caenorhabditis elegans.. Dev. Genet. 18:131-143[Medline].

EMMONS, S. W. and L. YESNER, 1984  High-frequency excision of transposable element Tc1 in the nematode Caenorhabditis elegans is limited to somatic cells. Cell 36:599-605[Medline].

FRY, J. D., S. HEINSOHN, and T. F. C. MACKAY, 1996  The contribution of new mutations to genotype-environment interaction for fitness in Drosophila melanogaster. Evolution 50:2316-2327.

GILLESPIE, J. H. and M. TURELLI, 1989  Genotype-environment interactions and the maintenance of polygenic variation. Genetics 121:129-138[Abstract/Free Full Text].

HALEY, C. S. and S. A. KNOTT, 1992  A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69:315-324[Medline].

HODGKIN, J. and T. M. BARNES, 1991  More is not better: brood size and population growth in a self-fertilizing nematode. Proc. R. Soc. Lond. Ser. B 246:19-24[Medline].

JOHNSON, T. E. and W. B. WOOD, 1982  Genetic analysis of life-span in Caenorhabditis elegans.. Proc. Natl. Acad. Sci. USA 79:6603-6607[Abstract/Free Full Text].

KELLER, L. and M. GENOUD, 1997  Extraordinary lifespans in ants: a test of evolutionary theories of ageing. Nature 389:958-960.

KSHIRSAGAR, A. M., 1983 A Course in Linear Models. Marcel Dekker, Inc., New York.

LANDER, E. S. and D. BOTSTEIN, 1989  Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185-199[Abstract/Free Full Text].

LANDER, E. and L. KRUGLYAK, 1995  Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat. Genet. 11:241-247[Medline].

LITHGOW, G. J., T. M. WHITE, D. A. HINERFELD, and T. E. JOHNSON, 1994  Thermotolerance of a long-lived mutant of Caenorhabditis elegans.. J. Gerontol. Biol. Sci. 49:B270-B276.

LUCKINBILL, L. S., R. ARKING, M. J. CLARE, W. C. CIROCCO, and S. A. MUCK, 1984  Selection for delayed senescence in Drosophila melanogaster.. Evolution 38:996-1003.

MANLY, K. F., and R. CUDMORE, 1994 Map Manager: A Program for Genetic Mapping. Version 2.6.

MARTIN, G. M., S. A. AUSTAD, and T. E. JOHNSON, 1996  Genetic analysis of ageing: role of oxidative damage and environmental stresses. Nat. Genet. 13:25-34[Medline].

MATHER, K., and J. L. JINKS, 1977 Introduction to Biometrical Genetics. Cornell University Press, Ithaca, NY.

NUZHDIN, S. V., E. G. PASYUKOVA, C. L. DILDA, Z. B. ZENG, and T. F. C. MACKAY, 1997  Sex-specific quantitative trait loci affecting longevity in Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 94:9734-9739[Abstract/Free Full Text].

ROFF, D. A., 1992 The Evolution of Life Histories: Theory and Analysis. Chapman and Hall, New York.

ROSE, M. R., 1991 Evolutionary Biology of Aging. Oxford University Press, New York.

ROSE, M. R. and B. CHARLESWORTH, 1981a  Genetics of life history in Drosophila melanogaster. I. Sib analysis of adult females. Genetics 97:173-186[Abstract/Free Full Text].

ROSE, M. R. and B. CHARLESWORTH, 1981b  Genetics of life history in Drosophila melanogaster. II. Exploratory selection experiments. Genetics 97:187-196[Abstract/Free Full Text].

SHOOK, D. R., A. BROOKS, and T. E. JOHNSON, 1996  Mapping quantitative trait loci affecting life history traits in the nematode Caenorhabditis elegans.. Genetics 142:801-817[Abstract].

SPSS, 1992 SPSS Base System Syntax Reference Guide Release 5.0. SPSS Inc., Chicago.

STEARNS, S. C., 1992 The Evolution of Life Histories. Oxford University Press, Oxford.

SULSTON, J. E., and J. HODGKIN, 1988 Methods, pp. 587–606 in The Nematode: Caenorhabditis elegans, edited by W. B. WOOD. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

TANKSLEY, S. D., 1993  Mapping polygenes. Annu. Rev. Genet. 27:205-233[Medline].

VIA, S. and R. LANDE, 1987  Evolution of genetic variability in a spatially heterogeneous environment: effects of genotype-environment interaction. Genet. Res. 49:147-156[Medline].

WILLIAMS, B. D., B. SCHRANK, C. HUYNH, R. SHOWNKEEN, and R. H. WATERSTON, 1992  A genetic mapping system in Caenorhabditis elegans based on polymorphic sequence-tagged-sites. Genetics 131:609-624[Abstract].

ZENG, Z., 1993  Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proc. Natl. Acad. Sci. USA 90:10972-10976[Abstract/Free Full Text].

ZENG, Z., 1994  Precision mapping of quantitative trait loci. Genetics 136:1457-1468[Abstract].

ZWAAN, B., R. BIJLSMA, and R. F. HOEKSTRA, 1995  Direct selection on life span in Drosophila melanogaster.. Evolution 49:649-659.

This article has been cited by other articles:

				J. Leips, P. Gilligan, and T. F. C. Mackay Quantitative Trait Loci With Age-Specific Effects on Fecundity in Drosophila melanogaster Genetics, March 1, 2006; 172(3): 1595 - 1605. [Abstract] [Full Text] [PDF]

				W. R. Swindell and J. L. Bouzat Inbreeding Depression and Male Survivorship in Drosophila: Implications for Senescence Theory Genetics, January 1, 2006; 172(1): 317 - 327. [Abstract] [Full Text] [PDF]

				R. L. Malmberg, S. Held, A. Waits, and R. Mauricio Epistasis for Fitness-Related Quantitative Traits in Arabidopsis thaliana Grown in the Field and in the Greenhouse Genetics, December 1, 2005; 171(4): 2013 - 2027. [Abstract] [Full Text] [PDF]

				T. E. Johnson Genes, Phenes, and Dreams of Immortality: The 2003 Kleemeier Award Lecture J. Gerontol. A Biol. Sci. Med. Sci., June 1, 2005; 60(6): 680 - 687. [Abstract] [Full Text] [PDF]

				S. Estes, B. C. Ajie, M. Lynch, and P. C. Phillips Spontaneous Mutational Correlations for Life-History, Morphological and Behavioral Characters in Caenorhabditis elegans Genetics, June 1, 2005; 170(2): 645 - 653. [Abstract] [Full Text] [PDF]

				A. L. Caicedo, J. R. Stinchcombe, K. M. Olsen, J. Schmitt, and M. D. Purugganan Epistatic interaction between Arabidopsis FRI and FLC flowering time genes generates a latitudinal cline in a life history trait PNAS, November 2, 2004; 101(44): 15670 - 15675. [Abstract] [Full Text] [PDF]

				M. Turelli and N. H. Barton Polygenic Variation Maintained by Balancing Selection: Pleiotropy, Sex-Dependent Allelic Effects and G x E Interactions Genetics, February 1, 2004; 166(2): 1053 - 1079. [Abstract] [Full Text] [PDF]

				D. Zhong, A. Pai, and G. Yan Quantitative Trait Loci for Susceptibility to Tapeworm Infection in the Red Flour Beetle Genetics, November 1, 2003; 165(3): 1307 - 1315. [Abstract] [Full Text] [PDF]

				K. L. Montooth, J. H. Marden, and A. G. Clark Mapping Determinants of Variation in Energy Metabolism, Respiration and Flight in Drosophila Genetics, October 1, 2003; 165(2): 623 - 635. [Abstract] [Full Text] [PDF]

				C. Weinig, L. A. Dorn, N. C. Kane, Z. M. German, S. S. Halldorsdottir, M. C. Ungerer, Y. Toyonaga, T. F. C. Mackay, M. D. Purugganan, and J. Schmitt Heterogeneous Selection at Specific Loci in Natural Environments in Arabidopsis thaliana Genetics, September 1, 2003; 165(1): 321 - 329. [Abstract] [Full Text] [PDF]

				M. C. Ungerer, S. S. Halldorsdottir, M. D. Purugganan, and T. F. C. Mackay Genotype-Environment Interactions at Quantitative Trait Loci Affecting Inflorescence Development in Arabidopsis thaliana Genetics, September 1, 2003; 165(1): 353 - 365. [Abstract] [Full Text] [PDF]

				V. Le Corre and A. Kremer Genetic Variability at Neutral Markers, Quantitative Trait Loci and Trait in a Subdivided Population Under Selection Genetics, July 1, 2003; 164(3): 1205 - 1219. [Abstract] [Full Text] [PDF]

				S. Ayyadevara, R. Ayyadevara, A. Vertino, A. Galecki, J. J. Thaden, and R. J. S. Reis Genetic Loci Modulating Fitness and Life Span in Caenorhabditis elegans: Categorical Trait Interval Mapping in CL2a x Bergerac-BO Recombinant-Inbred Worms Genetics, February 1, 2003; 163(2): 557 - 570. [Abstract] [Full Text] [PDF]

				M. C. Ungerer, C. R. Linder, and L. H. Rieseberg Effects of Genetic Background on Response to Selection in Experimental Populations of Arabidopsis thaliana Genetics, January 1, 2003; 163(1): 277 - 286. [Abstract] [Full Text] [PDF]

				J. W. Curtsinger Sex Specificity, Life-Span QTLs, and Statistical Power J. Gerontol. A Biol. Sci. Med. Sci., December 1, 2002; 57(12): B409 - 414. [Abstract] [Full Text] [PDF]

				J. O. Borevitz, J. N. Maloof, J. Lutes, T. Dabi, J. L. Redfern, G. T. Trainer, J. D. Werner, T. Asami, C. C. Berry, D. Weigel, et al. Quantitative Trait Loci Controlling Light and Hormone Response in Two Accessions of Arabidopsis thaliana Genetics, February 1, 2002; 160(2): 683 - 696. [Abstract] [Full Text] [PDF]

				A. U. Jackson, A. T. Galecki, D. T. Burke, and R. A. Miller Mouse Loci Associated With Life Span Exhibit Sex-Specific and Epistatic Effects J. Gerontol. A Biol. Sci. Med. Sci., January 1, 2002; 57(1): B9 - 15. [Abstract] [Full Text] [PDF]

				R. A. Miller A Position Paper on Longevity Genes Sci. Aging Knowl. Environ., November 28, 2001; 2001(9): vp6 - 6. [Full Text]

T. E. Johnson, D. Wu, P. Tedesco, S. Dames, and J. W. Vaupel Age-Specific Demographic Profiles of Longevity Mutants in Caenorhabditis elegans Show Segmental Effects