SUSCEPTIBILITY OF CAENORHABDITIS ELEGANS TO A BACTERIAL PATHOGEN IS A TYPICAL QUANTITATIVE TRAIT WITH AN ATYPICAL MUTATIONAL BIAS By VERONIQUE ETIENNE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2014
Â© 2014 Veronique Etienne
To my worms, may you RIP
4 ACKNOWLEDGMENTS I would like to express my deepest appreciation to my mother and my brother who provided immense emotional support. My sincerest gratitude goes to Florence for keeping me grounded. I also thank Dr. Charles F. Baer for being a great advisor/mentor; without his guidance, this would not have been possible. I am ever indebted to the Baer lab crew for assisting with data collection. Gramercy to Dr. Michael M. Miyamoto for continually aiding me in bettering my scientific self. Last, but certainly not least, danke Dr. Craig W. Osenber g , for being both helpful and direct, which was key in molding my approach to my time at UF.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ..................... 8 CHAPTER 1 BACKGROUND AND MOVTIVATION ................................ ................................ .... 10 Estimating V M from a Mutation Accumulation experiment ................................ ....... 12 Quantifying Genetic Variation ................................ ................................ ................. 13 Quantitative Genetics of Susceptibility of C. elegans to Mortality Due to Infection by Pseudomonas aeruginosa ................................ ................................ 14 2 MATERIALS AND METHODS ................................ ................................ ................ 18 MA Lines ................................ ................................ ................................ ................. 18 Pathogen ................................ ................................ ................................ ................ 19 Pathogen Susceptibility Assay ................................ ................................ ................ 20 Data Analysis ................................ ................................ ................................ .......... 22 LT50 ................................ ................................ ................................ ................. 22 Per ................................ ............... 22 stdLT50 ................................ ................................ ................................ ............ 23 Per generation change in the genetic variance LT50 (V M ) ............................... 23 Standing genetic variance (VG) ................................ ................................ ........ 24 3 R ESULTS AND DISCUSSION ................................ ................................ ............... 30 4 CONCLUSIONS ................................ ................................ ................................ ..... 38 APPENDIX A RECIPE ................................ ................................ ................................ .................. 39 B RELEVANT VALUES ................................ ................................ .............................. 40 LIST OF REFERENCES ................................ ................................ ............................... 43 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 49
6 LIST OF TABLES Table page 2 1 List of wild isolate worm lines ................................ ................................ ............. 27 3 1 List of trait means of mean arentheses. ....... 35 3 2 List of variances of mean ......... 36 B 1 List of releva nt values ................................ ................................ ......................... 40
7 LIST OF FIGURES Figure page 1 1 The strength of selection (as indicated by the intensity of blue) under different values of mutationa l variance (VM) and genetic variance (VG), assuming mutation selection balance (MSB). ................................ ................................ ..... 16 1 2 This figure illustrates the cumulative effect of mutations on the trait mean ................................ ................................ ................................ ................... 17 2 1 This figure illus trates the MA method/protocol. ................................ ................... 25 2 2 Each assay block consisted of 14 G250 MA lines and three ancestral G0 control "pseudolines", each replicated three times. ................................ ............ 26 3 1 This figure shows various values for the ratio of V M to V G ; the dashed line is for a neutral trait at MDE. ................................ ................................ ................... 37
8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science SUSCEPTIBILITY OF CAENORHABDITIS ELEGANS TO A BACTERIAL PATHOGEN IS A TYPICAL QUANTITATIVE TRAIT WITH AN ATYPICAL MUTATIONAL BIAS By Veronique Etienne August 2014 Chair: Charles F. Baer Major: Zoology Understanding the genetic basis of disease susceptibility is an important goal of medical ge netics and of evolutionary biology. A key first step toward understanding the genetics and evolution of any phenotypic trait is characterizing the role of mutation. Here the mutational variance (V M ) for susceptibility to one strain of the pathogenic bacte rium Pseudomonas aeruginosa in the nematode Caenorhabditis elegans is quantified . V M is a composite parameter dependent on the mutation rate, the size of the mutational target, and the magnitude of allelic effects on the trait. V M for susceptibility to P . aeruginosa was found to be on the order of, or slightly less than V M for a wide variety of traits in this strain of C. elegans , but is well within the range of reported values for those kinds of traits. Perhaps surprisingly, the average susceptibility d id not change significantly over 250 generations of mutation accumulation. The standing genetic variation (V G ) for susceptibility to P. aeruginosa in a worldwide collection of wild isolates of C. elegans is very similar to typical values of V G for life hi story and morphological traits in a variety of taxa, and comparison of V M to V G suggests an average strength of selection against mutations affecting susceptibility to P. aeruginosa of a few tenths of a percent. We conclude that, in this system, pathogen susceptibility presents a fairly
9 large mutational target under relatively weak stabilizing selection. These results should inform more realistic models of the evolutionary genetics of pathogen susceptibility.
10 CHAPTER 1 BACKGROUND AND MOVTIVATION Infecti ons by pathogens often manifest different symptoms in different individuals. Further, variation in the host's response to infection often has a genetic basis. A classic example is the sickle cell allele at the human hemoglobin locus, which confers resi stance to malaria when hetero zygou s (Bender and Hobbs 2012; Verra et al . 2009 ; Richer and Chudley 2005 , Williams et al . 2005 ). Similarly, certain alleles at the gl ucose 6 phosphate dehydrogenase locus also confer resistance to malaria in humans ( Lopez et al . 2010 ; Roth et al . 1983 ). We also see that the 32 mutation on the CCR5 recep tor confers resistance to HIV ( Galvani and Novembre 2005; . There are also numerous examples in which disease susceptibility (or resistance, depending on one's point of view) appears to be a polygenic trait, in which alleles of small to moderate effect at multiple loci affect the response to infection ( Daub et al. 2013; Lefebvre and Palloix 1996) . A comprehensive understanding of any genetically varia ble trait requires an accounting of the influence of mutation, natural selection, and random genetic drift on generating and shaping that variation. In most cases it is very difficult to separate the effects of these fo rces because they act simultaneously and different processes can lead to similar outcomes. The method of M utation A ccumulation (MA ) (Mukai 1964; Halligan and Keightley 2009) provides a powerful way to essentially isolate the effects of mutation from those of other evolutionary forces and to quantify the input of genetic variation by mutation. The principles underlying the MA method are simple. The efficiency of natural selection is inversely related to the genetic effective population size, N e ; when the product of N e and the strength of se lection (the selection coefficient, s ) is <
11 1 (technically, 4N e s <1 in diploids), genetic drift overwhelms selection and the evolutionary dynamics are effectively neutral (Kimura 62; Keightley and Caballero 1997; Kondrashov et al. 2006). Thus, by experiment ally maintaining a population at very small N e , the effect of natural selection is masked and all but the most highly deleterious mutations will accumulate in the population , as if neutral . For a trait to evolve, there must be additive genetic variance for the trait in the population (Fisher 1918, 1930 ; Lynch and Walsh 1998 ). For any trait, mutation introduces genetic variance (the "additive" will remain implicit except where noted) in to the population at rate V M per generation (Houle et al. 1996). On th e other hand, random genetic drift removes genetic variation from the population at rate equal to 1/2N e per generation ( Gillespie 1998 ). At equilib rium "mutation drift equilibrium" , the standing genetic variance for a neutral trait, V G , is proportional to N e V M ; for a random mating finite population at mutation drift equilibrium V G =2N e V M (Lynch and Hill 1986). There is an obvious and useful analogy between this relationship and the standing nucleotide sequence variation at mutation drift equilibrium, e nucleotide mutation rate ( Gillespie 1998 ). The mutational variance V M can be broken down into its biological components, the mutation rate and the effects of mutations on the phenotypic trait in question ( see Box 1 for de rivation ). Following Barton (1990), we assume 2n mutable loci in the diploid genome and a per neut ral trait z genetic variance in the trait accumulates 2 pe r generation, i.e. .
12 Obviously many phenotypic traits are not neutral. In what follows (Barton 1990) we assume a general model of pleiotropic selection wherein we have a large population in which selection contributes most to the removal of variation and the effects of drift are negligible whereby mutations affect both a phenotypic trait z and fitness ( w ), but the effects on the trait and the effects on fitness need not b e correlated. Genetic variance increases at rate V M per generation and each mutant allele multiplies the fitness of its bearer by 1 s (Barton 1990). Natural selection removes genetic variance at a rate proportional to the selection coefficient s, so at equilibrium ("mutation selection balance", MSB) the standing genetic variance in a finite population is V G M /s. Consider the intuitive example of a trait for which all the genetic variance is contributed by dominant lethal mutations; n o mutant alleles w ill make it into the next generation (because s = 1 so w = 0) and V G =V M /1, i.e. all the genetic variance is contributed by new mutations. Estimating V M from a Mutation Accumulation experiment The basic structure of a MA experiment is simple ( Figure 2 1 ). A highly homozygous starting population (i.e., an inbred line at mutation drift equilibrium) is replicated into many subpopulations ("MA lines") and each MA line is allowed to evolve for many generations at very small Ne; mutations with selective effects s <1/4N e will accumulate at the neutral rate (hence the term "mutation accumulation"). The initial genetic variance is (assumed to be) zero, although in reality the best that can be hoped for is that the starting population is at mutation drift equilibrium. As the MA experiment proceeds, mutations occur in individual lines at rate U per generation and fix with probability 1/2Ne. After t generations of MA the among line component of the phenotypic variance V L =2V M t , so V M =V L /2 t ( Lynch and Walsh 1998 ) .
13 If the average effect of mutations on a trait is non zero (i.e., there is a mutational bias), the mean of the trait averaged over the set of MA lines will differ from the mean of the ancestral population. By definition, fitness is under directional selection. It is e xtremely well documented that deleterious mutations are more common (probably much more common) than beneficial mutations, so a directional change in the mean the dire ction of lower fitness. For example, lifetime fecundity typically decreases in MA experiments whereas time to maturity often increases ( Figure 1 2 ). It is reasonable to assume that, all else equal, natural selection favors high fecundity and rapid matura tion. Quantifying Genetic Variation Genetic variance is commonly scaled in one of two ways either as a fraction of the total phenotypic variance (V P ), the heritability, or as a fraction of the trait mean, the genetic coefficient of variation, CV G or the CV 2 (Houle 1992, Houle et al. 1996). Heritability is defined in two ways; in the "broad sense", h 2 = V G /V P or in the "narrow sense", h 2 =V A /V P where V A represents additive genetic variance. Natural selection removes (additive) genetic variance for fitnes s (Fisher 1930), thus, all else equal, the more closely related a trait is to fitness, the smaller the amount of genetic variance expected at equilibrium. On the other hand, and again all else equal, the larger the fraction of the genome that contributes to a trait the "mutational target" , the greater the genetic variance. It is now well established that the (narrow sense) heritability is lower in life history traits, which are presumed to be closely related to fitness, than in morphological, physiolog ical, and behavioral traits, whose connection to fitness is less direct (Mousseau and Roff 1987). That finding is consistent with the idea that natural selection removes additive genetic variance for fitness. However, when genetic
14 variance is scaled as t he CV G , the opposite result is found; there is more additive genetic variance for life history traits than for other classes of traits (Houle 1992). A potential resolution to this apparent paradox is that (1) life history traits present a larger mutationa l target than other traits, but also (2) life history traits are more influenced by the effects of the environment such that although V G of life history traits is greater than V G of other traits, V E is even greater, so the ratio V G /V E of life history trait s is smaller than V G /V E for other traits. The idea that life history traits present a larger mutational target than other classes of traits was substantiated by Houle et al. (1996), who surveyed the existing literature and found that the mutational coeffi cient of variation, CV M , is on average greater for life history traits than for the other classes of traits. Interestingly, one of the more consistent results in evolutionary genetics is that the mutational heritability, (h 2 m = V M /V E ) for most traits in m ost organisms is rarely very different from 10 3 /generation, even for traits for which the CV M varies considerably (Pannebaker et al. 2008). There are some notable exceptions to that rule, however; for example, h m 2 for gene expression in several species i s ~10 5 /generation (Denver et al. 2005, Rifkin et al. 2005, Landry et al. 2007). Quantitative Genetics of Susceptibility of C. elegans to Mortality Due to Infection by Pseudomonas aerug i nosa There is abundant evidence that variation in the susceptibilit y (or resistance) of hosts to pathogens has a genetic basis, and that in many cases the genetic basis of susceptibility appears to be polygenic rather than due to a single locus of large effect. However, to the best of our knowledge, no one has yet attemp ted to quantify the mutational variance for susceptibility to any pathogen in any host organism. It is worth noting that this question is not merely ac ademic, because, all else equal, the larger the
15 mutational target a trait presents, the larger the numbe r of potential targets of drug acti on with respect to that trait. Here we report a study to determine the cumulative effects of spontaneous mutations (i.e., V M C. elegans resulting from infection by the (known) pathogenic bacterium Pseuodomonas aeruginosa . We integrate existing data from a set of wild isolates of C. elegans (collected by our collaborator Erik Andersen) to provide an estimate of V G for the same trait. We employed MA lines derived from the PB306 strain of C. elegans that had been propagated by single hermaphrodite descent for 250 generations s < 25% i.e., 4N e s <1 will be effectively neutral. Populations of worms were exposed to P. aeruginosa over a period of 128 hours and mortality recorded at 12 hr intervals. The time at which half the worms on a plate had died (LT50) is used as the measure of susceptibility to the pathogen ( Reddy et al. 2009).
16 Figure 1 1. The strength of selection (as i n dicated by the intensity of blue ) under different values of mutational variance (VM) and genetic variance (VG), assuming mutation selection balance (MSB). Under MSB, V M / V G =s. s=1 indicates the strongest level of selection. For a neutral trait (s=0), V G =4N e V M at mutation drift equilibrium . The ombre is imposed to illustrate that natural selection operates on a gradient an d depends upon the relative values of V M and V G.
17 Figure 1 2. This figure illustrates the cumulative effect of mutations on th deleterious mutations are allowed to accumulate, which brings the average fitness level to decrease as the number of inbred generations increase.
18 CHAPTER 2 MATERIALS AND METHODS MA Lines The MA protocol has be en previously described (Vassilieva and Lynch 1999; Baer et al. 2005 ) . In this study we employed 70 MA lines derived from the PB306 wild isolate. Beginning with a single immature (L3/L4 stage) hermaphrodite, populations were initially propagated by the t ransfer of a single immature hermaphrodite at four day intervals for six generations. This demographic protocol results in an effective population size (N e ) of 1; a population is expected to reach mutation drift equilibrium after 6N e generations (Lynch an d Hill 1986). After six generations of single worm descent, the population was allowed to expand to large size (two generations of reproduction), at which time 100 replicate mutation accumulation (henceforth MA) lines were initiated from single immature h ermpaphrodites. Each MA line was subsequently propagated by transfer of a single immature hermaphrodite at four day intervals for approximately 250 generations "approximately" because some lines did not produce appropriately aged offspring in some four da y intervals and the parent was held over until the next four day interval. MA lines were maintained on 60mm NG M agar plates seeded with 0.1 mL of the OP50 strain of E. coli and incubated at 20Â°C. The common ancestor of the MA lines ("G0 control") was cry opreserved at the beginning of the MA experiment; MA lines were cryopreserved after (approximately) 250 generations of MA. During the course of the MA experiment each MA line was maintained as a "leading" individual , and the plates containing the previous two generations were maintained as backups. If a leading individual died or had not produced eggs during the four day interval it was replaced with a randomly chosen individual from the plate
19 containing the previous generation (occasionally the second pr evious generation). "Going to backup" does not affect the number of generations of MA that the line experienced because the leading worm and the backup worm are from the same generation. Going to backup DOES influence the long term N e of the line because the census size of the backup plate is approximately the number of offspring of the parental worm. When population size fluctuates in time the long term , N e is the harmonic mean census size ( Gillespie 1998), which is heavily weighted toward the smallest sizes. In the case of PB306 the long term N e was approximately 1.1 (Phillips et al. 2009 ). In a MA experiment, mutations with selective effect s < 4N e will fix at approximately the neutral rate (Kimura 1962; Keightley and Caballero 1997; Kondrashov et al . 2004), so in this experiment mutations with selective effects less than ~25% were effectively neutral. Pathogen We used the PA14 strain of Pseudomonas aeruginosa as the pathogen. P. aeruginosa is ubiquitous and is an opportunistic pathogen of a very wide spectrum of Eukaryotes, includin g plants (Adonizio et al. 2008; He et al. 2004) and humans (Pukatzk et al. 2002). The PA14 strain is known to be pathogenic to C. elegans under the condition s of this experiment (Tan et al. 1999), and the degree of suscept ibility as measured by LT50 (see below) has been shown to differ among strains of C. elegans ( Reddy et al. 2009; E. Andersen, unpublished data ). Whether C. elegans normally interacts with P. aeruginosa in its natural environment is not known, although P. aeruginosa has been identified from collections of microbes associated with C. elegans taken from nature (Buck Samuel, personal communication). Different strains of P. aeruginosa are known to kill nematodes via diff erent mechanisms ; for example, the PA01 strain induces mortality via cyanide poisoning and neuromuscular paralysis
20 (Gallagher and Manoil 2001), whereas the PA14 strain kills by means of oxidative stress mediated by phenazines (Cezairliyan 2013). Further, different mutations in PA14 cause varyin g degrees of pathogenicity (Tan et al. 1999) . Pathogen Susceptibility Assay In this section we explain the details of the pathogenesis assay with respect to a single experimental unit (i.e., a single plate of worms); in the next section we explain the detai ls of the full assay. We employ a plate based pathogenesis assay (the "slow killing assay" (SKA) of Tan et al. 1999 ) rather than the liquid culture assay used in many studies (Zaborin et al . 2009, Kirienko et al . 2013) because it is possible that hypoxia resulting from bacterial growth contributes to worm mortality in liquid culture assays (E. Andersen, personal communication). The virulent effects of the bacteria on the worm manifest themselves in one of two ways, either by distension of the intestine a nd the accumulation of outer membrane vesicles in the gut, or by "Red Death" syndrome caused by the buildup of the pathway of quorum sensing + Fe3 + , PQS+FE3 + , complex (Zaborin et al. 2009). Pathogen challenge was performed on 35mm SKA agar plates ( Appendi x A ) containing .05% of 100mg/mL filter sterilized 5 fluorodeoxyuridine , FUDR ; FUDR prevents maturation of immature worms via prevention of DNA synthesis ( Mitchell DH, et al. 1979; Angeli et al . 2013) and thus prevents the focal individuals from reproducing . Pseudomonas aeruginosa , incubated for 24 hours at 37Â° C and for an additional 24 hours at 25Â° C prior to beginning the assay. At the beginning of the SKA, approximately 30 40 immature (L3/L4 stage) hermaphrodites fr om a synchronized population (see below) were introduced onto an assay plate; this point constitutes time t =0 of the assay. Beginning at time t =32 hrs, all
21 worms on a plate were scored as live/dead at 12 hour intervals by the criterion of responsiveness t o the touch of a worm pick. Unresponsive (=dead) worms were removed from the plate. The assay was terminated at t =128 hrs, at which point all remaining worms were counted and scored as live or dead. Replicates for which at least 30 L3/L4 stage worms wer e not available at t =0 were held over for 24 hrs, at which point the replicate was begun or, if an insufficient number of worms were available, the replicate was aborted and counted as missing. Experimental d esign . The experiment was performed in two rep licated "super blocks", each of which consisted of four blocks; each MA line was present in two blocks, one in each super block. The experimental design for the lead up phase i s depicted in Figure 2 2 . An assay block consisted of 14 G250 MA lines and thr ee ancestral G0 control "pseudolines", each replicated three times. At the beginning of a block, an aliquot of cryopreserved G0 control was thawed and three worms picked singly to individual plates; each of these plates constitutes a "pseudoline". After one generation, three immature offspring from each pseudoline were picked singly to new plates; each of these plates constitutes a replicate within each pseudoline. At the same time, a cryopreserved aliquot of each MA line was thawed and three individuals picked singly to new plates, each of which constitutes a replicate within that MA line. Each replicate was then propagated by single offspring descent for an additional generation, at which point populations were allowed to expand to large size ( two gene rations of reproduction) and a small chunk of agar containing many worms was placed onto a new plate. This plate was incubated for 24 hours, at which point worms were picked individually onto
22 the SKA plate. During this lead up phase worms were kept under MA conditions (NGM plates, OP50 food, and incubation at 20Â° C). Data Anal ysis LT50 We use mortality as a proxy for pathogen susceptibility. Ideally we would have run a parallel control of worms kept identically except fed on their normal bacterial foo d (i.e., the OP50 strain of E. coli ) rather than the pathogen. However, maintaining a parallel control would have required a concomitant reduction in the number of lines that could be assessed for survivorship in the pathogen treatment. A pilot study wit h several wild isolates showed that almost all individuals lived well past the termination time of the SKA (E. And ersen, personal communication), so it is reasonable to think that most mortality in the SKA was due to the effects of the pathogen rather than extraneous causes. We use an estimate of the median time of death (LT50, for "time of 50% lethality") as our quantitative measure of s usceptibility since we expect the survivorship curves to be sigmoid (Reddy et al. 2009 ) .We used nonlinear least squares t o fit a logistic regression, P t = 1 1 / (1+e B (G X l n( t)) ), where P t is the proportion alive, t time units after the start of the assay, and then estimated LT50 as e B/G . Per generation c The question of interest is: at what r ate does LT50 change per generation of mutation accumulation? The change in LT50 is a function both of the MA phenotype and the G0 phenotype, both of which are subject to several sources of estimation error. We first divided each data point (LT50 value) by the mean LT50 of the G0 control for the block. Doing this makes the trait dimensionless and enables comparisons between
23 traits measured on different scales (Hansen and Houle 2008 ) among and across various organisms . We then analyzed the linear model (s ee below) stdLT50 stdLT50=Gmax + Block + Block*Line(Treatment) + Line(Treatment) + Error, where stdLT50 is the G 0 mean standardized LT50, Gmax is generations of MA (0 or 250) and Treatment is MA or G 0 . Note that Gmax and Treatment are continuous and discr ete representations of the number of generations of mutation accumulation. Block and Line are modeled as random effects, Treatment is a fixed effect. Variance components of all random effects are assumed to be heterogeneous across treatments and were est imated independently for each treatment group. Degrees of freedom were determined by the Kenward Roger method ( Kenward and Roger 1997 ). The slope of the regression of stdLT50 on Gmax , estimated from the full model, generation change in the (mean standardized) LT50 with MA. Per generation c hange in the genetic variance LT50 ( V M ) The mutational variance V M is equal to half the difference between the among line component of variance (V L ) in the MA lines and V L in the G 0 lines, di vided by the number of generations of MA, i.e., , where t is the number of generations of MA ( Lynch and Walsh 1998) . The among line variances of the MA lines and the G 0 ancestor were estimated by restricted maximum l ikelihood (REML) separately for the MA lines and G 0 pseudolines, from the linear model stdLT50= Block + Line + Block*Line + Error . Statistical significance of the among line component of variance (V L ) within each group was assessed by Likelihood Ratio Tes t. The likelihood of the model with the among line variance included was compared to the likelihood of
24 the model without the among line term; twice the difference in the (log)likelihoods is chi square distributed with degrees of freedom equal to the diffe rence in the number of parameters estimated (one, in this case). The REML estimate of V L for the G 0 pseudolines was zero so all downstream calculations were based on the MA lines only. Standing genetic variance (VG) Erik Andersen previously collected d ata on LT50 under exposure to the PA14 strain of P. aeruginosa on a collection of 20 wild isolates of C. elegans . The SKA protocol was the same as ours except mortality data were collected at 8 hour intervals rather than 12 hour intervals. The wild isola tes represent a worldwide collection ( Table 2 1 ). There is little global population structure in C. elegans (Cutter 2006 and BarriÃ¨re A, and FÃ©lix 2005 ), but there is presumably a small among population component of variance. Wild isolates of C. elegans are almost always highly homozygous, and each wild isolate underwent an additional few generations of self fertilization prior to preservation. Thus, we treat the wild isolates as a set of homozygous lines, and assume the genetic (= genotypic) component o f phenotypic variance , V G , is V L /2, where V L is the among line component of variance (Lynch and Walsh 1998 ). Each line was present in three replicates in each of three assay blocks. Variance components were estimated from the linear model stdLT50 = block + line + block*line + error , using the same REML methodology as described for the MA lines.
25 Figure 2 1. This figure illustrates the MA method/protocol. Imposing severe bottlenecks allows drift to swamp the effects of natural selection, thus allo wing (deleterious) mutations (indicated by various colors) to accumulate . There is some random chance as to whether or not a new mutation will be arbitrarily selected to propagate. The ancestral controls are maintained by freezing these worms.
26 Figure 2 2. Each assay block consisted of 14 G250 MA lines and three ancestral G0 control "pseudolines", each replicated three times. After one generation, three immature offspring from each pseudoline/replicate were picked singly to new plates. Each pseudolin e/replicate was then propagated by single offspring descent for an additional generation, at which point populations were allowed to expand to large size ( two generations of reproduction). 30 40 individual L3 worms from these populations were segregated a nd placed onto population specific SKA plates (indicated in orange).
27 Table 2 1. List of wild isolate worm lines Isotype Source Lab GPS Latitude GPS Longitude Isolation Date AB1 CGC 34.93 138.59 1983 AB4 CGC 34.93 138.59 1983 CB4851 CGC 44.85 0.4 8 pre 1949 CB4852 CGC NA NA pre 1966 CB4853 CGC 34.189 118.131 1974/05 CB4854 CGC 34.189 118.131 1974/05 CB4856 CGC 21.33 157.86 1972/08 CB4857 CB 34.096 117.719 1972/11 CB4858 CGC NA NA 1973 CB4932 CGC 51.02 3.1 pre 1991 CX11262 CX 34.12946 118.10987 2003/09 CX11264 CX 34.12946 118.10987 2003/09 CX11271 CX 34.13712 118.12532 2003/09 CX11276 CX 34.20111 118.21198 2003/09 CX11285 CX 34.14331 118.05496 2003/09 CX11292 CX 34.13531 118.30582 2004/02 CX11307 CX 34.12946 118.10987 2003/0 9 CX11314 CX 34.12946 118.10987 2003/09 CX11315 CX 34.12946 118.10987 2003/09 DL200 DL 9.03 38.74 2007/12 DL226 DL 44.5633 123.2821 2007 DL238 DL 19.22 155.82 2008/07/15 ED3005 VX 55.94 3.36 2004/10/25 ED3011 VX 55.92 3.19 2004/11/26 ED3012 V X 55.92 3.19 2004/11/26 ED3017 VX 55.92 3.19 2004/12/03 ED3040 VX 26.1 28.01 2006/03 ED3046 VX 33.22 19.19 2006/03 ED3048 VX 33.22 19.19 2006/03 ED3049 VX 33.22 19.19 2006/03 ED3052 VX 33.22 19.19 2006/03 ED3073 VX 1.05 36.39 2006/03 ED3077 VX 1.19 36.48 2006/03 EG4347 EG 44.04789 123.07108 2006/10 EG4349 EG 40.77167 111.87316 2006/10 EG4724 EG 41.628771 8.347617 2007/03 EG4725 EG 41.628771 8.347617 2007/03 EG4946 EG 40.72596 111.82184 2007/09/27 JT11398 JT 47.763944 122.275484 2003/12 JU258 JU 32.73 16.89 2001/10 JU310 JU 46.63 1.06 2002/08/25 JU311 JU 44.42 4.4 2002/09/08
28 Table 2 1. Continued. Isotype Source Lab GPS Latitude GPS Longitude Isolation Date JU323 JU 44.42 4.4 2002/09/08 JU346 JU 44.42 4.4 2002/09/08 JU360 JU 48.98 2.23 2002/09/02 JU363 JU 48.98 2.23 2002/09/16 JU367 JU 48.98 2.23 2002/09/16 JU393 JU 49.28 0.32 2002/09 JU394 JU 49.28 0.32 2002/09 JU397 JU 49.28 0.32 2002/09 JU406 JU 49.28 0.32 2002/12/30 JU440 JU 48.715 1.56 2003/09/12 JU561 JU 4 8.71 3.81 2004/10/03 JU642 JU 48.84 2.5 2004/12/14 JU751 JU 48.84 2.5 2005/06/08 JU774 JU 38.683 9.34 2005/07/10 JU775 JU 38.7175 9.1486 2005/07/10 JU778 JU 38.719 9.1491 2005/07/10 JU782 JU 38.7191 9.1503 2005/07/10 JU792 JU 43.06 0.24 2005/08 /31 JU830 JU 48.52 9.05 2005/09/28 JU847 JU 48.46 7.461 2005/10/03 JU1088 JU 34.7613 138.0149 2007/03/14 JU1172 JU 36.87 73.04 2007/04 JU1200 JU 55.577 4.6 2007/08/01 JU1212 JU 48.71 3.81 2007/09/24 JU1213 JU 48.71 3.81 2007/09/24 JU1242 JU 49 .1269 1.9595 2007/10/14 JU1246 JU 49.12618 1.96152 2007/10/14 JU1395 JU 47.2199 0.04619 2008/03/01 JU1400 JU 37.3845 5.988 2008/03 JU1409 JU 37.468 5.637 2008/03/31 JU1440 JU 41.41307 2.15231 2008/06/09 JU1491 JU 46.63 1.06 2008/08/17 JU1530 JU 48 .7015 2.1725 2008/09/09 JU1568 JU 48.8092 2.3862 2008/10/05 JU1580 JU 48.7015 2.1725 2008/10/06 JU1581 JU 48.7015 2.1725 2008/10/23 JU1586 JU 46.63 1.06 2008/11/03 JU1652 JU 34.86 56.19 2009 JU1896 JU 37.999722 23.749673 2010/01/02 KR314 CGC 49.28 123.13 1984/05 LKC34 CGC 18 46 2005/06/17 LSJ1 CGC 51.45 2.59 1951 MY1 CGC 52.54 7.31 2002/07 MY10 CGC 51.96 7.53 2002/07
29 Table 2 1. Continued. Isotype Source Lab GPS Latitude GPS Longitude Isolation Date MY16 CGC 51.93 7.57 2002/07 MY18 CGC 51. 96 7.53 2002/07 MY23 CGC 51.96 7.53 2002/07 PB303 CGC NA NA 1998/11/14 PB306 CGC NA NA 1998/11/28 PS2025 CGC 34.19 118.13 Unknown PX179 CGC 44.035 123.058 2001/10/02 QX1211 QX 37.7502 122.4331 2007/11/26 QX1233 QX 37.8804 122.2838 2007/11/24 RC 301 CGC 47.99 7.84 1983 WN2002 WN 51.975285 5.694834 2007/11/20
30 CHAPTER 3 RESULTS AND DISCUSSION Two important primary results emerge from this study. First, mean LT50 of C. elegans in the presence of pathogenic Pseudomonas aeruginosa d id not change significantly over 250 generations of mutation accumulation ( 1.1x10 4 /generation, p > 0.28). This result stands in contrast to many other traits we have investigated in this set of MA lines, for which trait means change significantly in the directio n of (assumed) lower fitness ( Table 3 1 ). Second, there was significant mutational variance for LT50 (V M = 1.8x10 5 /generation, p<0.006). Taken together, the lack of change in mean LT50 coupled with the significant mutational variance indicates that LT50 presents a non trivial mutational target ( Table 3 2 ) and that there is no consistent relationship between LT50 in the presence of P. aeruginosa and fitness. Because we do not have a parallel control grown in the absence of the pathogen, the possibility ex ists that the V M in LT50 simply reflects underlying variation in natural lifespan rather than variation in susceptibility to P. aeurginosa . In addition to the known pathogenicity of the PA14 strain , two additional lines of evidence suggest that most morta lity was due to the pathogenic effects of P. aeruginosa as opposed to normal variation in life span . First, most dead worms possessed the vesicles associated with infection by P. aeruginosa , and many worms exhibited symptoms of "red death" syndrome, also associated with infection by P. aeruginosa ( Williams and CÃ¡mara 2009; Zaborin et al. 2009). Second, reanalysis of data on longevity in the PB306 MA lines assayed under MA conditions (fed OP50 strain of E. coli on NGM plates, incubated at 20Â°; see Joyner Ma tos et al. 2009 ) indicates a LT50 much greater than observed here ;
31 Table 3 1 ) indicating that normal variation in lifespan cannot be attributed to this level of mortality . Additionally, w e e xpect the LT50 under non pathogenic conditions to be ~18 19 days (Leiser et al. 2011) it to be anywhere near as low as 85 hours. Given that significant mutational variance accumulated, it is of interest to compare V M for pathogen susceptibility (i.e., L T50 in this experiment) to other traits to assess the relative size of mutational targets presented by different kinds of traits. LT50 under (benign) MA conditions presents an obvious comparison. Whereas LT50 under pathogen exposure provides a measure of pathogen susceptibility, LT50 under benign conditions provides a measure of average lifespan. V M for LT50 under MA conditions is 5.9x10 5 /generation, approximately 3X greater than for pathogen susceptibility. Given the magnitude of the sampling variatio n associated with these estimates of V M , it is reasonable to conclude that pathogen susceptibility probably presents a somewhat smaller mutational target than LT50 under benign conditions, but that equivalently sized mutational targets for the two traits a re plausible. Interestingly, for LT50 under MA conditions is about six fold greater than LT50 under pathogen exposure ( = 6.2 x 10 4 /generation ( p< 0.002) vs. 1.1x10 4 /generation). This result is counterintuitive, because the average lifespan of th e worms in benign conditions is well past the time at which worms cease reproducing. We would expect mutations that affect the ability to survive infection by a pathogen and some worms did survive the full 128 hr SKA to be acted upon by natural selection, whereas mutations that affect lifespan past the reproductive period would not. Apparently, mutations that reduce LT50 under benign conditions have deleterious pleiotropic effects on other traits that are directly related to fitness, whereas mutations tha t increase
32 susceptibility to P. aeruginosa (i.e., reduce LT50) have at most weak pleiotropic effects on fitness in the absence of infection by P. aeruginosa . Body size presents another illuminating comparison. Reanalysi s of data in Ostrow et al. (2007 ) sh ows that V M for body volume at 72 hrs (~ time of maturity) i s 1.2 x 10 4 /generation, an order of magnitude greater than for LT50 under pathogen exposure (1.8x10 5 ) . It is not surprising that body size presents a larger mutational target than LT50 under pathogen exposure, because presumably many loci affect organis mal growth. Similarly, body volume declines significantly with MA; 7.0 x 10 3 /generation (p<0.003), consistent with mutations that affect body size either being under direct ional selection (i.e., all else equal, larger worms have higher fitness t han smaller worms) or having deleterious pleiotropic effects on fitness. Although LT50 under pathogen exposure presents a smaller mutational target than lifespan and body size (and also relative fitness; Baer et al. 2006), there are other traits that have similar or smaller V M than LT50 under pathogenesis. For example, a suite of eight traits associated with embryo morphology and/or the first mitotic cell division have a median V M of ~ 1x10 6 /generation in the same set of PB306 MA lines (Far hadifar, Needle man , and B aer unpublished data). Thus, we conclude tha t LT50 under pathogen exposure pathogen susceptibility presents a rather typical mutational target rather than an unusually small target. That a trait related to pathogen susceptibility presents a su bstantial mutational target was not a foregone conclusion, for two reasons. First, many host pathogen relationships are known to be mediated by a small number of genes in the host genome ( Wilfert and Schmid Hempel 2008 ). More particularly, several patho gen related traits in
33 C. elegans are known to be mediated by large effect mutations at one or a few loci ; i n particular , npr 1 encodes a neuropeptide receptor that is known to mediate numerous aspects of worm behavior , including pathogen avoidance (Reddy e t al. 2009 ) . It is possible that variation in LT50 in our assay is mediated by some component of behavior, although we have no evidence that it is or is not. Mutational variance can also be scaled relative to the environmental variance, V E , i.e., = V M /V E . In general, mean standardized genetic variances are preferable to heritabilities because they permit an unambiguous comparison of the genetic variance between different traits and taxa that is unbiased by variation in environmental circums tances (Houle 1992). However, heritability has one specific useful interpretation, which is that it quantifies the genetic variation immediately available to respond to natural selection in the particular environment in which the trait is measured and can used to meaningfully compare traits amongst various types of organisms . The mutational heritability f or LT50 under pathogen exposure is = 1.1x10 3 /generation ( Table 3 2 ). Unsurprisingly in hindsight, this value is almost exactly the "typical" value of observed in many studies of many disparate traits (Houle et al. 1996; Pannebakker et al. 2008; Halligan and Keightley 2009 ). A primary motivation of this study is to provide an estimate of V M for susceptibility to P. aeruginosa to compare to a previous study of the same trait in a collection of 20 wild isolates of C. elegans (Erik Andersen, unpublished data). We employed th e same analytical methodology to determine the standing genetic (=genotypic) variance (V G ) for LT50 under pathogen exposure. As described in the Introduction, with some reasonable assumptions, for a trait under selection (directional or stabilizing), at
34 m utation selection balance the ratio of V M to V G is expected to be proportional to the strength of selection acting on new mutations (the selection coefficient, s ). At the other end of the spectrum, V G for a neutral trait is expected to be equal to 2N e V M a t mutation drift equilibrium (4N e V M in a predominantly selfing group such as C. elegans ( Lynch and Hill 1986 ). The ratio V M /V G = 0.0028 /generation, suggesting that the selection coefficient acting on mutations affecting susceptibility to P. aeruginosa is 3 %, which is similar to other life history traits (Figure 3 1) .
35 Table 3 1 . List of trait means of mean standardized traits Trait Trait mean (G 0 ) Trait Mean (MA) std mean (G 0 ) std mean (MA) LT50Pa (hrs) 85.2 7 (2.83) 83.50 (2.57) 1.007 (0.013) 0.976 (0.022) 0.00012 x 10 3 (0.10 x 10 3 ) > 0.25 Fitness at 25Â°C (# offspring) 90.89 (22.34) 74.78 (14.60) 1.001 (0.057) 0.818 (0.055) 0.83 x 10 3 (0.36 x 10 3 ) 0.024 Survivorship (proportion) 0.71 (0.03) 0.61 (0.02) 0.989 (0.027) 0.854 (0.027) 0.54 x 10 3 (0.15 x 10 3 ) 0.0005 LT50MA (hrs) 407.82 (43.46) 339.36 (14.35) 0.998 (0.030) 0.843 (0.123) 0.62 x 10 3 (0.51 x 10 3 ) 0.0017 Size (mm 3 ) 1.60 x 10 3 (0.48 x 10 3 ) 1.38 x 10 3 (0. 91 x 10 3 ) 0.998 (0.023) 0.857 (0.043) 0.70 x 10 3 (0.25 x 10 3 ) 0.0022
36 Table 3 2 . List of v ariance s of mean standardized traits Trait V L, MA V L, G0 V E, MA V E, G0 V M P(V M >0) LT50Pa (hrs) 8.97 x 10 3 (.00326) 0 1.62 x 10 2 ( .01 56) 1.45 x 10 2 (.0222) 1.79 x 10 5 (.65210 5 ) 1.17 x 10 3 (2.99 x 10 5 ) 0.006 Fitness at 25Â°C (# offspring) 99.11 x 10 3 (.03903) 9.71 x 10 3 (.03396) 34.63 x 10 2 ( .0 381) 48.01 x 10 2 (.0622) 20.32 x 10 5 (3.85 x 10 5 ) 0.49 x 10 3 (2 .65 x 10 5 ) > 0.09 Survivorship (proportion) 16.72 x 10 3 (.01032) 0 16.58 x 10 2 (.01 50) 12.62 x 10 2 (.0138) 3.39 x 10 5 (2.06 x 10 5 ) 0.23 x 10 3 (.336 x 10 5 ) > 0.09 LT50MA (hrs) 33.37 x 10 3 (9.43 x 10 3 ) 3.91 x 10 3 (.00449) 2.52 x 10 2 ( . 0040 ) 1.42 x 10 2 (.0462) 5.72 x 10 5 (1.66 x 10 5 ) 2.91 x 10 3 (3.89 x 10 5 ) 0.019 Size (mm 3 ) 54.15 x 10 3 (.01553) 1.08 x 10 3 (.00366) 36.93 x 10 2 (.0732 ) 37.63 x 10 2 (.0722) 13.35 x 10 5 (3.02 x 10 5 ) 3.58 x 10 3 (2.61 x 10 5 ) <0.0 001
37 Figure 3 1. This figure shows various values for the ratio of V M to V G ; the dashed line is for a neutral trait at MDE. The distance from the line allows one to see how strong selection is acting on a given trait. We see here that s for susceptibility to P. aeruginosa is comparable to several other life history traits .
38 CHAPTER 4 CONCLUSIONS N e for C. elegans has been estimated to between 10,000 and 50,000 (Andersen et al. 2012). As stated previously, the selection coefficient for susce ptibility to P. ecall that for a trait to be neutral while in mutation drift equilibrium V G =4N e V M . Our estimation for V G was .081 and was 1.8 x 10 5 for V M , which would require an effective population size of 1125. However, this is infeasible give the observed V E value; this indicates that the trait is indeed under weak (stabilizing) selection. On the whole we see that the mutational variance and the standing genetic variance for susceptibility to P. aerunginosa is similar to that fo r a variety of traits in C. elegans and in other organisms (Houle et al. 1996). One possibility is that C. elegans rarely encounters P. aerunginosa , even when found in close proximity in the wild, in which case susceptibility to this particular pathogen wou ld be a neutral trait in C. elegans . Alternatively, the high standing genetic variance may be due to a high number of m utations affecting this trait a large mutational target. The fact that the mutational variance is similar to other life history and morph ological traits adds credence to this possibility.
39 APPENDIX A RECIPE SKA Plates The plates were made by adding 3.0g NaCl, 17g BactoAgar, 3.5g Peptone, and filling up with DI water to 1L. The solution was then autoclaved and allowed to cool to 55 Â° C. T hen, 1mL each of the following was added: cholesterol (5mg/mL in EtOH), 1M CaCl 2 , and 1 M MgSO 4 . 500ÂµL of FUDR (filter sterilized, 100mg/mL) and 1M KH 2 PO 4 (pH6) was also added. To each 35x10mm plate, 4mL of the above solution was added.
40 APPENDIX B R ELEVANT VALUES Table B 1. List of relevant values Term Abbrev. Definition Our value Among line variance V L 2V M t V L,LT50Pa N/A Average mutational effect the absolute value of the average mutational effect on a given trait (I M /U)1/2 N/A W25 Broad sense Heritability H 2 Proportion of V P that is contributed by the V G . V G /V P 23% WSORT ::28% W20::7% Siz e::5% Coefficient of Variation CV Mean standardized std. deviation LT50PaMA::.976 SizeMA::.818 Environmental variance V E The residual component of variance due to non genetic factors .016 N/A Genetic effective population size N e Number of individuals contributing (genetically) to the next generation/ size of population that will lose heterozygosity C. elegans 10 4 5x10 4 Lethality time 50% LT50 The amount of time it takes for half of the population to die LT50PaMA ::83.5hrs LT50PaG0:: 85hrs LT50MA:: 407.8hrs
41 Table B 1. Continued Term Abbrev. Definition Our value Muta tion accumulation MA The Ne is kept very low so as to negate natural selection, allowing deleterious mutations to accumulate as if neutral N/A N/A Mutation Selection Balance MSB Natural selection removes deleterious traits at the same rate mutation introd uces them N/A N/A Mutational bias The change in the trait mean under MA conditions 1.1 x 10 4 /generation, P>0.28, F test MA conditions:: 6.2 x 10 4 /generation (p< 0.002) Mutational heritability h 2 V M /V E Allows for comparison of traits measure on different scales 3 /generat ion 2 /generation to 5 /generation Mutational Target Q Z fraction of the genome with the potential to affect trait Z if mutated N/A Q W25 = 0.5 Mutational Variance V M The per generation VM=UQ Z 2 (QZ is the mutational target) 1.8x10 5 /generation, p <0.006 MA conditions:: 5.9x10 5 /generation Body volume 4 /generation Selection coefficient S The strength of selection V M /V G 3% LT50MA::1 2% Size::
42 Table B 1. Continued Term Abbrev. Definition Our value Squared coefficient of variation CV 2 Mean standardized variance MA::1.8 x10 5 /generation (P<0.006 LT50MA:: (5.2 x 10 5 , P<0.04) Standing Genetic Variance V G If we were to pool the worldwide population, what is the genetic variance that we would see .08 1 LT50MA::.117 Size:: .117
43 LIST OF REFERENCES 1. Adonizio A, Leal SM, Ausubel FM, and Mathee K (2008) Attenuation of Pseudomonas aeruginosa virulence by medicinal plants in a Caenorhabditis elegans m odel system. Med Microbiol 57: 809 813. 2. Alatalo RV , Gusta fsson L, Lunberg A (1990) Phenotypic selection on heritable size traits: environmental variance and genetic response . Th e American Naturalist 135(3): 464 471. 3. Andersen EC, Gerke JP, Shapiro JA, Crissman JR, et al . (2012) Chromosome scale selective sweeps s hape Caenorhabditis elegans genomic diversity. Nature Genetics 44:285 283. 4. Angeli S, Klang I, Sivapatham R, Mark K , et al. (2013) A DNA synthesis inhibitor is protective against proteotoxic stressors via modulation of fertility pathways in Caenorhabditis elegans . Aging 5 (10): 759 769. 5. Baer CF, Phillips N, Ostrow D, Avalos A, et al . (2006) Cumulative effects of spontaneous mutations for fitness in Caenorhabditis : role of genotype, environ ment and stress. Genetics 174: 1387 1395. 6. Baer CF, Sha w F, Sted ing C, Baumgartner M, et al . ( 2005 ) Comparative evolutionary genetics of spontaneous mutations affecting fitness in rhabditid nematodes. PNAS 102:5785 5790. 7. BarriÃ¨re A, FÃ©lix MA (2005) High local genetic diversity and low outcrossing rate in Caen orhabditis elegans natural populations. Curr Biol 15: 1176 1184. 8. Barton NH ( 1990 ) Pleiotropic models of quantitative variation. Genetics 124:773 782. 9. Bender MA, Hobbs W (2012) Sickle cell d isease. GeneReviews Web . 10. Cezairliyan B , Vinayavekhin N, Grenfe ll Lee D, Yuen GJ, et al . (2013) Identification of Pseudomonas aeruginosa p henazines that k ill Caenorhabditis elegans . PLOS 9(1): e1003101 . 11. Cutter AD (2006) Nucleotide polymorphism and linkage disequilibrium in wild populations of the partial selfer Caenor habditis elegans . Genetics 172: 171 184.
44 12. Daub JT , Hofer T, Cutivet E, Dupanloup I , et al . (2013) Evidence for Polygenic Adaptation to Pathogens in the Human Genome. Mol Biol Evol 30: 1544 1558. 13. Denver DR , Morris K, Streelman JT, Kim SK , et al. (2005) The tr anscriptional consequences of mutation and natural selection in Caenorhabditis elegans . Nat Genet 37:544 548 . 14. Fisher RA (1930) The genetical theory of natural selection . Oxford University Press, Oxford . 15. Fisher, RA (1918) The correlation between relatives on the supposition of mendelian inheritance . Transactions of the Royal Society of Edinburgh 52: 399 433. 16. Gallagher LA, Manoil C (2001) Pseudomonas aeruginosa PAO1 k ills Caenorhabditis elegans by cyanide p o isoning. J Bacteriol 183(21): 6207 6214. 17. Ga lvani A P, Novembre P (2005) The evolutionary history of the CCR5 resistance mutation . Microbes Infect 2: 302 309. 18. Gillespie JH (1998) Population g enetics: a concise guide. The Johns Hopkins University Press, Maryland . 19. Hall igan DL, Keightley PD (2009) Spontaneous mutation acc umulation studies in evolutionary genetics . A nnu Rev Ecol Evol Syst 40: 151 172. 20. Hansen TF, Houle D (2008) Measuring and comparing evolvability and constraint in multivariate characters. J Evol Biol 21:1201 1219. 21. He J, Baldini RL, DÃ©ziel E, Saucier M , et al . (2004) The broad host range pathogen Pseudomonas aeruginosa strain PA14 carries two pathogenicity islands harboring plant and an imal virulence genes. PNAS 101(8): 2530 2535 . 22. Houle D (1992) Comparing Evolvability and variability of quantitative traits . Genetics 130 : 195 204. 23. Houle D, Morikawa B, Lynch M (1996) Comparing mutational variabilities . Genetics 143: 1467 1483.
45 24. Irazoqui JE, Troemel ER, Feinbaum RL, Luhachack LG, et al . (2010) Distinct pathogenesis and host responses during infection of C. elegan s by P. aeruginosa and S. aureus . PLoS Pathogens 6(7): e1000982. 25. Joyner Matos J , Upadhyay A, Salomon MP, Grigaltchik V, Baer CF ( 2009 ) Genetic (co)variation for life span in rhabditid nematodes: role of mutation, selection, and history. J Gerontol A Biol S ci Med Sci 64:1134 1145 . 26. Keightly P, Caballero A ( 1997) Genomic mutation rates for lifetime reproductive output and lifespan in Caenorhabditis elegans . PNAS 94: 3823 3827. 27. Kenward M G , Roger JH (1997) Small sample inference for fixed effects from restricte d max imum likelihood. Biometrics 53: 983 997. 28. Kimura M (1962) On th e probability of fixation of mutant genes in a population . Genetics 47(6): 713 719. 29. Kirienko NV, Kirienko DR, Larkins Ford J, WÃ¤hlby C , et al . (2013) Pseudomonas aeruginosa disrupts Caenorh abditis elegans iron homeostasis, causing a hypoxic response and death . C ell Host Microbe 13( 4 ): 406 4 41 . 30. Kondrashov FA, Ogurtsov AY, Kondrashov AS (2004) Bioinformatical assay of human ge ne morbidity. Nucleic Acids Res 32:1731 1737. 31. Kondrashov FA , Ogurts ov AY, Kondrashov AS (2006) Selection in favor of nucleotides G and C diversifies evolution rates and levels of polymorphism at mammalian synony mous sites. J Theor Biol 240: 616 626. 32. Landry CR , Lemos B, Rifkin SA, Dickinson WJ, Hartl DL (2007) Genetic prop erties influencing the evolvability o f gene expression. Science 317: 118 121 33. Lefebvre A, Palloix A (1996) Both epistatic and additive effects of QTLs are involved in polygenic induced resistance to disease: A case study, the interaction pepper Phytophtho ra capsici Leonian . Theor Appl Genet 93: 503 511 . 34. Leiser, SF, Begun A, Kaeberlein M ( 2011 ) HIF 1 modulates longevity and healthspan in a temperature dependent manner. Aging Cell 10:318 326.
46 35. Lopez C , Saravia C, Gomez A, Hoebeke J, Patarroyo MA (2010) Mecha nisms of genetically based resistance to malaria . Gene 467(1 2): 1 12. 36. Lynch M, Hill WG (1986) Phenotypic evolution by neutral mutation . Evolution 40 (5):915 935. 37. Lynch M, Walsh B (1998) Genetics and Analysis of Quantitative Traits. Sinauer Associates, Inc ., Massachusetts. 38. Mitchell DH, Stiles JW, Santelli J, Sanadi DR (1979) Synchronous growth and aging of Caenorhabditis elegans eoxyuridine. J Gerontol 34: 28 36. 39. Mousseu TA, Roff DA (1987) Natural selection and the heritability of fitness components. Heredity 59: 181 197 . 40. Mukai T (1964) The genetic structure of natural populations of Drosophila melanogaste r . I. Spontaneous mutation rate of polygenes controlling viability. Genetics 50(1): 1 19. 41. Moore JP (2000) The effect of genetic variation in chemokines and their receptors on HIV transmission and progress ion to AIDS. Immunol Rev 177: 99 111. 42. Os trow D, Phillips N, Avalos A, Blanton D , e t al. (2007) Mutational bias for body size in rhab ditid nematodes. Genetics 176: 1653 1661. 43. Pannebakker, Halligan DL, Reynolds KT, Ballantyne GA , et al. (2008) Effects of spontaneous mutation accumulation on sex ra tio traits in a parasitoid wasp. Evolution 62 8: 1921 1935 44. Pukatzk S, Kessin RH, Mekalanos JJ ( 2002) The human pathogen Pseudomonas aeruginosa utilizes conserved virulence pathways to infect the social amoeba Dictyos telium discoideum . PNAS 99(5): 3159 3164 . 45. Reddy KC, Andersen EC, Kruglyak L, Kim DH (2009) A p olymorphism in npr 1 i s a beh avioral determinant of pathogen susceptibility in C. elegans . Science 323(5912): 382 384.
47 46. Richer J, Chudley AE (2005) The hemog lobinopathies and malaria. Clin Genet 68:332 3 36. 47. Rifkin SA, Houle D, Kim J, White KP (2005) A mutation accumulation assay reveals a broad capacity for rapid evolution of gene expression. Nature 438 : 220 223. 48. Roth EF Jr, Ravento s Suarez C, Rinaldi A, Nagel RL (1983) Glucose 6 phosphate dehydrogenase deficiency inhibits in vitro growth of Plasmodium falciparum . PNAS 80: 298 299 . 49. Smith MP, Laws TR, Atkins TP, Oyston PC, et al. (2002) A liquid based method for the assessment of bacterial pathogenicity using the nematode Caenorhabditis elegans . FEMS Micro biol Lett 210: 181 185. 50. Tan MW (1999) Killing of Caenorhabditis elegans by Pseudomonas aeruginosa used to model mammalian bacterial pathogenesis. PNAS 96(2):715 20. 51. Tan MW, Rahme LG, Sternberg JA, Tompkins RG, Ausubel FM (1999) Pseudomonas aeruginosa kill ing of Caenorhabditis elegans used to identify P. aeruginosa virulence factors. PNAS 96( 151 ): 2408 2413. 52. Vas silieva L, Lynch M (1999) Accumulation of spontaneous mutations in Caenorhabditis elegans . Genetics 151: 119 129 53. Verra F, M angano VD, Modiano D (200 9) Genetics of susceptibility to Plasmodium falciparum : from classical malaria resistance genes towards genome wide association studies . Parasite Immunol 31(5): 234 253. 54. Wilfert L , Schmid Hempe l P (2008) The genetic architecture of susceptibility to parasi tes. BMC Evol Biol 8 :187 . 55. Williams TN, Mwangi TW, Wambua S, Alexander ND , et al . (2005) Sickle cell trait and the risk of Plasmodium falciparum malaria and other childhood diseases. J Infect. Dis 192: 178 186 56. Williams WP, CÃ¡mara M (2009) Quorum sensing an d environmental adaptation in Pseudomonas aeruginosa : a tale of regulatory networks and multifunctional signal molecules . Curr Opin Microbiol 1 2 (2): 182 191
48 57. Zaborin A, Romanowski K, Gerdes S, Holbrook C , et al . (2009) Red death in Caenorhabditis elegans ca used by Pseudomonas aeruginosa P AO1. PNAS 106(15): 6327 6332.
49 BIOGRAPHICAL SKETCH Veronique Etienne, a Florida native, has always been interested in the sciences. As a youth she was involved in workshops and spent much of her time exploring the outdoor s. A experience. Whilst an undergraduate she worked in a variety of research settings, including plant ecology which allowed her to realize that research was her desired career path and public health aspects of prostate cancer. She received her Bac helor of Science from the University of South Florida; while at USF she completed an undergraduate thesis that focused on the macro parasites of Cuban tree frogs. She then matriculated to the University of Florida, where she received her Master of Science . While at UF, her interest in antagonistic relationships, i.e. host parasite/pathogen interactions, was further cultivated by courses in virology and bacterial patho gen in the nematode C. elegans ; she examined how this trait is evolving as well as how selection is operating on it.