7 Identification and Mapping of Genes Determining Longevity

Thomas E. Johnson and David R. Shook

Introduction

The world of modern biology is unified by genetics. Genetic approaches have the ability to transcend species and provide cross-links between fields for several reasons. First, is the fact that all species are evolutionarily related. Thus, distinct species have similar gene function, and DNA sequence homology can be found between even distantly related species. Indeed, DNA sequence homology is used as a metric to determine evolutionary relationships among species. Second, molecular genetic manipulation changes both the genotype and the phenotype of an organism. Such manipulations represent an extremely fine-scale tool for dissection of the underlying biochemistry, physiology, anatomy, and development of an individual species. Because virtually any gene can be manipulated at will in many species, a dedicated approach can lead to an unraveling of the relationship between genotype and phenotype for almost any gene in these species. In the study of longevity, genetic approaches can play a key role because the phenotype of longevity can be studied only at the whole-organism level; nevertheless, understanding at the molecular level could lead to accurate predictions of the dynamics of life-expectancy change. The unraveling of this genotype/phenotype relationship in determining organismic life span has only just begun. Third, and most important, is the fact that genetics has the power to reveal causality by factors that are not dependent upon investigator prejudice. Unlike a biochemical approach, which by its very nature must focus on one physiological system and on even one molecule or one part of that molecule (for example), a genetic approach can survey and identify alterations in any subsystem in the species in which the genetic alteration is detected (Botstein and Mauer, 1982). It



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 108
7 Identification and Mapping of Genes Determining Longevity Thomas E. Johnson and David R. Shook Introduction The world of modern biology is unified by genetics. Genetic approaches have the ability to transcend species and provide cross-links between fields for several reasons. First, is the fact that all species are evolutionarily related. Thus, distinct species have similar gene function, and DNA sequence homology can be found between even distantly related species. Indeed, DNA sequence homology is used as a metric to determine evolutionary relationships among species. Second, molecular genetic manipulation changes both the genotype and the phenotype of an organism. Such manipulations represent an extremely fine-scale tool for dissection of the underlying biochemistry, physiology, anatomy, and development of an individual species. Because virtually any gene can be manipulated at will in many species, a dedicated approach can lead to an unraveling of the relationship between genotype and phenotype for almost any gene in these species. In the study of longevity, genetic approaches can play a key role because the phenotype of longevity can be studied only at the whole-organism level; nevertheless, understanding at the molecular level could lead to accurate predictions of the dynamics of life-expectancy change. The unraveling of this genotype/phenotype relationship in determining organismic life span has only just begun. Third, and most important, is the fact that genetics has the power to reveal causality by factors that are not dependent upon investigator prejudice. Unlike a biochemical approach, which by its very nature must focus on one physiological system and on even one molecule or one part of that molecule (for example), a genetic approach can survey and identify alterations in any subsystem in the species in which the genetic alteration is detected (Botstein and Mauer, 1982). It

OCR for page 108
should be noted, however, that genetics alone has little hope of unraveling molecular mechanisms and that the strongest analysis results from the combined use of biochemistry, cell biology, and genetics. I will first review the concept of genetic determination of life span and life expectancy and the concept of longevity-determining genes that we call ''gerontogenes." Next, we will review relevant literature and experiments done in an effort to identify such gerontogenes. This review will focus mostly on invertebrates because few experiments in vertebrates, notably the mouse, have tried to identify gerontogenes. We will speculate as to how these gerontogenes might be identified in other species, paying careful attention to the mapping of quantitative trait loci (QTLs). This discussion will focus on identifying gerontogenes in nematodes and mice; much of the material has been selected from ongoing experiments in our own laboratory. Their subsequent extension to humans and/or the identification of gerontogenes directly in humans is the subject of another author. Finally, we will review work from our laboratory on the genetic determination of mortality rates. Many recent reviews are available on the genetics of aging (Johnson et al., 1996; Fleming and Rose, 1996; Jazwinski, 1996; Lithgow, 1996; Martin et al., 1996; Nooden and Guiamét, 1996). Concept Of Gerontogenes The gene is the basic unit of inheritance; typically a gene makes a protein. A gerontogene, then, makes a protein involved in aging and, more precisely, a protein involved in determining life span. Gerontogenes are defined functionally: they can be altered by mutation such that animals carrying them have a longer-than-normal life span. Vijg and Papaconstantinou (1990) suggested that gerontogenes might be separated into four distinct categories based upon their effects on life span and their evolutionary origin. "Deleterious" and "pleiotropic" genes are predicted by the evolutionary theory of aging (Charlesworth, 1980; Rose, 1991) and have evolved as a result of mutation accumulation or pleiotropic gene action, respectively. "Aging" genes that actively kill the organism are thought not to operate in organism senescence by evolutionary criteria; "longevity" genes promote survival, and presumably most genes that are fixed in an organism are of this type. "Longevity assurance gene," a term used by Sacher (1978) and by D'mello et al. (1994), means essentially the same as the term gerontogene. Methods For Identifying Gerontogenes Five methods have been used to identify genes involved in aging, with variable success. These approaches mirror the approaches used to identify genes involved in other biological processes. These approaches are: (1) gene association, (2) selective breeding, (3) quantitative trait locus (QTL) mapping, (4) induc-

OCR for page 108
TABLE 7-1 Approaches Used to Identify Gerontogenes Allelic association Mapping quantitative trait loci Selective breeding Induction of new mutations Construction of transgenic stocks tion of new mutations, and (5) construction of transgenic stocks (Table 7-1 ). The first three approaches use genetic variation latent in the species studied. In the first approach, these genes are studied as found; this is the only approach useful for studying genes in humans because of constraints inherent in human studies. The second approach involves selective breeding using the phenotype [or more recently the genotype, as in marker-assisted selection (Tanksley, 1993)] of interest to establish a strain in which genes leading to the desired phenotypic alterations (in this case increased longevity) have been differentially accumulated in one line as compared with another. Selected lines differing in traits of interest have been used by humans since the beginning of the agricultural era. The third approach, QTL mapping, is a method for separating the overall polygenic effects of a selected line, for example, into individual components, providing estimated map positions and effect sizes for each QTL (Tanksley, 1993). Approaches 4 and 5 differ from the former two approaches in that they do not rely on preexisting genetic variation and instead involve the creation of new mutations that then become the object of study. Approach 4 involves the identification of mutants induced by mutagens; these mutants are identified solely on the basis of their effect on the phenotype of interest—i.e., longevity. The last approach targets genes of interest to test a specific hypothesis and to determine whether that gene is actually causally involved in aging. Gene Association In humans and in other vertebrates the length of the life span and other problems (ethical, technical, and budgetary) have prevented the analysis of longevity by methods other than the study of naturally occurring variation. Studies of human longevity genes have involved either the identification of syndromes leading to reduced life expectancy that in many cases are mimics of the aging process (Martin, 1978; Brown, 1987; Van Broeckhoven, 1995; Yu et al., 1996) or on gene association (Schächter et al., 1994; Takata et al., 1987). The former approach involves the identification of genetic diseases of potential relevance to aging, while the latter has most often involved the study of candidate genes for association between allelic variants in that candidate and life expectancy. In mice, a few studies have attempted to go beyond the use of candidate genes by using genetic markers spanning the whole genome (Gelman et al., 1988; Puel et

OCR for page 108
al., 1995). These studies are difficult both because of problems with maintaining strains of mice in a "clean" environment uncontaminated with pathogens for their entire life span and because of statistical problems in conducting whole-genome scans. Thus, most associations that have been discovered to date are of doubtful significance. [See Lander and Kruglyak (1995) for a discussion of the statistical problems associated with whole-genome scans and see the section on QTL mapping for a more complete description of this approach.] Selective Breeding In this approach the breeder chooses animals showing the desired characteristics and mates them; the same breeding scheme is followed for successive generations of their offspring, which leads eventually to an alteration in the trait being assessed or "selected for." This method can be as simple as "mate the best with the best and hope for the best" or can involve very sophisticated strategies wherein phenotype assessment is based on a variety of test matings and progeny testing (Falconer, 1989). Selective breeding has been conducted in Drosophila, selecting either on age of reproduction (Luckinbill et al., 1984; Rose, 1984; Partridge and Fowler, 1992) or directly on life expectancy (Zwaan et al., 1995). Most selection studies have identified a negative interaction between longevity and reproductive effort. QTL Mapping Overview When many genes are coordinately involved in the specification of a genetic trait, we can refer to the individual genes contributing to the overall phenotype as QTLs (Table 7-2). We intend to first describe how QTLs are "mapped" (localized to specific regions of the genome). Next we will address some issues in evolutionary theory and demography to which QTL mapping may be relevant. We will then describe our work on QTL mapping of life-history traits in the nematode Caenorhabditis elegans and some early work on mapping QTLs involved in specifying adult demographic characters. Finally, we will speculate as to how these studies may be extended to other demographic parameters not yet assessed in C. elegans or other species. Refer to Table 7-2 for definitions; for more background on QTL mapping see Tanksley (1993). The Basis of QTL Mapping All genetic mapping is based on the fact that genetic loci, whether QTLs or genetic markers, close together on the chromosome remain together (cosegregate) in crosses, except where recombination occurs. Genes that are close are said to

OCR for page 108
TABLE 7-2 A Lexicon of QTL Terminology Term Definition Allele A variant of a gene or genetic marker Effect size The amount of genetic variation explained by a particular QTL Genetic marker An easily-assessed DNA segment with a known genetic map position and many alleles. It should have no effect on the phenotype Interval mapping A method for QTL mapping in which one takes into account genotype information from markers on either side of the region being studied; this allows the detection of recombination events within the interval that can be used to position the QTL within the interval Quantitative trait A trait whose phenotypic variation is continuous (rather than discrete) and is determined by the segregation of many genes or QTLs QTL Quantitative trait locus; one of a suite of genes specifying any particular phenotype QTL mapping The process of determining probable sites and probable effect sizes of a QTL or QTLs RI Recombinant inbred: an inbred strain generated by 20 generations of sibling mating from the F2 of two inbred strains or by 10 generations of self-fertilization (hermaphroditic species) SDP Strain distribution pattern; the pattern of alleles at a locus across a series of RI strains derived from crosses between the same initial inbred parents be genetically linked. For 75 years, single-gene traits have been "mapped" by taking advantage of such cosegregation and the fact that the frequency of recombination events that separate linked genes is proportional to the physical distance between genes on the chromosome. The mapping of quantitative traits, especially demographic traits, is more difficult because these traits are influenced by more than one gene. However, the basis of QTL mapping has been apparent for some time. If a QTL for mean life span, for example, is genetically close (linked) to a genetic marker, then, on average, progeny that carry the marker allele from the "low" parent will have a shorter life span than progeny carrying the marker allele from the "high'' parent. The Human Genome Project has provided "dense" genetic maps of the human and mouse with highly polymorphic marker loci; as a result, the mapping of QTLs over practically the entire human genome has become possible. Nevertheless, QTL mapping in humans is difficult and involves large population sizes. Several other animal model systems are also under intense study and are of more immediate relevance for our discussion. These include other mammalian models, particularly the mouse and rat. and invertebrate models such as fruit flies and the nematode C. elegans. Mapping of QTLs involves (1) assaying the phenotype of interest in each offspring, (2) characterizing the pattern of genetic markers in each offspring, and (3) a statistical assessment of whether any part of the phenotypic variation is significantly associated with any marker. For example, a marker allele tightly

OCR for page 108
linked to a QTL involved with the determination of longevity will be overrepresented in the most long-lived individuals in a population in which alleles of this QTL are segregating. Exactly this method was used by Ebert et al. (1993) in mapping genes that contributed to longevity in C. elegans. Similarly, the shortest-lived individuals in the population will carry a higher proportion of the marker allele associated with the short-life span allele of the QTL. Of course, this mapping assumes a significant genetic component to the determination of life span in the particular population being studied. The degree of correlation between the marker and the phenotype is based on two things: the effect size of the QTL and the genetic distance between the QTL and the marker. If the QTL is close to the marker, then detection and mapping of the QTL can be performed easily and accurately. If the QTL is far from the marker gene, then recombination between marker and QTL will lead to a lower frequency of association, resulting in a lower observed "effect size" attributed to the QTL linked to that marker. This problem can be partially alleviated by using allelic information from markers on both sides of the QTL; this results in the determination of linkage using an "interval-mapping approach." Interval mapping is more mathematically sophisticated and more biologically precise because algorithms have been developed that can account for recombination events in the interval. Interval mapping uses these detected recombination events to suggest most probable locations and effect sizes for the QTL within the interval. Interval mapping can be done by using either regression approaches (Haley and Knott, 1992) or maximum-likelihood procedures, such as those developed by Lander and Botstein (1989) or by Zeng (1994). The interval-mapping procedure involves fitting an additive model for a putative QTL, Q, at a series of positions in the interval, d, between two markers A and B. Induced Mutants In this approach, genomes are mutagenized, and the offspring are subsequently screened for phenotypic alterations. This is the most powerful approach in that any gene in the genome can potentially be identified, even those that are essentially fixed in natural populations. A second major advantage of this approach is that it can reveal genes that were not logical candidates. This element of surprise and lack of being tied in with specific hypotheses makes mutant induction the method of choice. This approach has the major disadvantage of being time consuming and expensive and consequently unpopular in the study of aging and senescence. Only in Neurospora (Munkres and Furtek, 1984), in the nematode (Klass, 1983; Duhon et al., 1996), and in yeast (Kennedy et al., 1995) has the approach really been successfully employed. However, there is at least one reported example of an increased-longevity mutant in Drosophila (Maynard Smith, 1958).

OCR for page 108
Identification of Gerontogenes in Yeast D'Mello et al. (1994) isolated the LAG1 gerontogene based on differential times of expression by screening gene libraries for young-expressed genes and late-expressed genes. LAG1 appears to be a 411-amino acid protein that is predicted to be an integral membrane protein. Further analysis of LAG1 revealed that both null mutants, as well as dominant, C-terminal-deletion mutants in LAG1, lead to extension of yeast life span. Kennedy et al. (1995) selected for gerontogene mutants by using prolonged survival under starvation conditions. Their results offer evidence that two phenotypes, extended survival under nongrowing conditions and extended replicative life span, cosegregate. Their finding that eight mutants in four genes (UTH1-UTH4) selected on the basis of extended vegetative survival under nongrowing conditions also had increased replicative life spans argues for considerable coordinate control. One gene was studied in detail and shown to be a mutation in SIR4, a gene involved in transcriptional silencing. A number of other intriguing studies were described, including a requirement for SIR3 but not for SIR1 and evidence that a dominant negative construct of SIR4 can extend the replicative life span. Identification of Gerontogenes in C. elegans Eight gerontogenes have been identified in nematodes, and mutants in all of those tested have increased resistance to a variety of stresses (see also Lithgow et al., 1995; Murakami and Johnson, 1996; Martin et al., 1996). (1) Klass (1983) isolated the first mutant, which we subsequently called age-1 (Friedman and Johnson, 1988). (2) Mutants in spe-26 show an 80 percent increase in life expectancy of the hermaphrodite (VanVoorhies, 1992). (3) Mutants in daf-2 have consistently extended adult life expectancy under some conditions (Kenyon et al., 1993). (4) Larsen et al. (1995) demonstrated that daf-23 mutants also have a prolonged life under some conditions, and she also demonstrated that a double mutant (daf-12 daf-2) has a 4-fold life extension (daf-12 is a dauer-defective mutant, see below). (5) Wong et al. (1995) isolated a new class of mutants, and the first of these, clk-1, showed a longer adult life for several alleles. In a subsequent study, Lakowski and Hekimi (1996) showed that clk-1, clk-2, clk-3, and gro-1 are gerontogenes as well. rad-8 mutants have extended life at 16°C (Ishii et al., 1994), but this is mostly due to an extended developmental period, and thus rad-8 is not a true gerontogene. Only age-1 was identified by screening for increased longevity. All the other mutants have defects other than the altered life expectancy. spe-26 is a gene involved primarily in sperm activation and isolated on the basis of its male-specific sterility; rad-8 was isolated based on its increased sensitivity to radiation; daf-2 and daf-23 are temperature-sensitive, dauer-constitutive mutations that cause dauer formation even in the presence of food at 25°C, clk-1 mutants show a variety of

OCR for page 108
alterations in the timing of cell division and development. In addition, the dauer-defective mutant (daf-16) interacts to block the effects of daf-2, daf-23 as well as all of the other mutants tested (Murakami and Johnson, 1996; Kenyon et al., 1993: Larsen et al., 1995). The extended life span of these mutants was detected in a search for extended life expectancy in some cases and serendipitously in others. Mapping Induced Mutants The fact that genes located close together on the chromosome remain together during crosses allows the construction of genetic maps and ultimately even the identification of genes by "positional cloning." There is only one example of "mapping" a gerontogene using assessment of life expectancy as the character that is being mapped, and that is the mapping of age-1 (Friedman and Johnson, 1988; Johnson et al., 1993). More recent studies have relied on the mapping of other traits associated with the gerontogene mutant, such as resistance to stress (Duhon et al.,. 1996). Problems with directly mapping life expectancy include the lengthy period of time it takes to assess the phenotype and the influence of other genes and environmental effects, such that the mapping can become very problematic. Additional attempts to localize QTLs specifying longevity are described below. Transgenic Strategies A final approach for identifying gerontogenes has been successful in only a few instances. This involves the construction of transgenic lines carrying additional copies or altered copies of the genes of interest. Chen et al. (1990) and Sun et al. (1994) used a candidate-gene approach in that they identified genes probably involved in the determination of replicative life span based not on direct screens for longevity but on other evidence. They subsequently showed that disruption of either of two yeast candidate genes, RAS1 or RAS2 alters the replicative life span. RAS1 deletions prolong replicative life about 40 percent, while RAS2 deletions shorten it. Over-expression of RAS2 led to an increased longevity, but over-expression of RAS1 had no effect. In two cases this approach has been used in Drosophila: with EF-1α (a protein elongation factor; Shepherd et al., 1989) and superoxide dismutase/catalase, double transgenics, which both help to detoxify free radicals (Orr and Sohal, 1994). Both studies reported life extension. Subsequent studies on additional EF-1a transgenics (Shikama et al., 1994) failed to replicate the earlier findings. Although the studies on the superoxide dismutase/catalase double mutants are not without problems, they were widely replicated initially and are consistent with a large additional body of evidence showing that reactive oxidant species are involved in many aspects of aging and senescence.

OCR for page 108
Review of QTL Mapping of Longevity The most extensive genetic analysis of longevity ever conducted in a mammal was that of Gelman et al. (1988), who used the B x D (C57BL/6J x DBA/2J) series of recombinant inbred strains of mice to identify QTLs for life expectancy. They found six significant associations: however, the statistical significance of these findings is in some doubt because their stringency levels were not sufficiently high (Lander and Kruglyak, 1995); however, see Yunis and Salazar (1993). Lines of mice that have been selected for different immune response have quite different life expectancies (Covelli et al., 1989). Unfortunately, most of the variance between the lines was for decreased life span, since the lines selected for decreased immune function were very short lived. Puel et al. (1995) mapped QTLs that were associated with the different immune response and found that five QTLs explained much of the variance in immune response in the F2 population but have no data on life expectancy. Besides our work described below and those in mice just described, we are aware of only three other studies mapping or attempting to find the genes specifying longevity in animals. Ebert et al. (1993) used F2 populations of C. elegans to localize QTLs for longevity by genotyping the longest-lived members of the populations. They were not able to distinguish various aspects of life history, because these aspects were not studied independent of survival—but they were able to apply a novel approach to studying increased longevity. Using two different approaches, they found QTLs for increased survival on chromosomes II and IV and the X chromosome. These QTLs are not the same as those found by Shook et al. (1996, see below). Ebert et al. (1996) also positioned a QTL for H2O2-resistance to a site on chromosome V (stP6) near the site previously identified that is involved in longevity determination. In Drosophila, Fleming et al. (1993) used two-dimensional gel electrophoresis in a direct approach to identifying the proteins leading to long life in selected lines. These strategies identified genetic regions or gene products that were associated with increased survival, but neither study critically examined life-history traits other than survival, and neither could examine more sophisticated interactions among loci. Luckinbill et al. (1988) showed that most of the effect of life extension in their selected Drosophila lines mapped to chromosome 3 with chromosome 2 negatively affecting life span. The findings were corroborated by Arking et al. (1993) on an independent population originally selected at the same time under the same conditions. Several laboratories are currently attempting to generate QTL maps for Drosophila longevity, and several studies are under way in mice. Use of Recombinant Inbred Strains (RIs) in Mapping QTLs for Longevity and Other Life-History Traits C. elegans, a nematode worm, is a prime candidate for QTL mapping and the study of coadapted gene complexes, since it is normally inbreeding but can also

OCR for page 108
out-cross. Shook et al. (1996, unpublished work) used QTL mapping strategies to localize genetic regions influencing life-history traits in C. elegans. These investigators mapped QTLs for mean and maximum survival, mean and maximum fertility, temperature-sensitive fertility, and internal hatching of progeny (bagging). QTLs for mean life span were identified on linkage groups (LGs) II, IV, and the X chromosome, and QTLs for fertility were identified on LGs II. III, and IV. The QTLs for mean life span accounted for 23 percent of observed phenotypic variance (comparable to total genetic variance of 26 percent). The loci for mean fertility accounted for 46 percent of the observed phenotypic variance (genetic component of variance: 52 percent). Additional QTLs for temperature-sensitive fertility (LGs II and V) and internal hatching (LG IV) were also mapped in these crosses. The two sets of studies (Shook et al., 1996; Ebert et al., 1993) did not replicate each other. Essentially none of the markers associated in the two studies were the same. In the only marker associated in both studies (spP5 on chromosome V), the association was in the opposite direction in the two studies. There were many differences between the two studies in conduction of the experiments, and the reasons for the lack of replication are unclear. RI-based QTL mapping allows an analysis of the effects of individual genes on different traits and of interactions among genes in specifying individual traits not possible in F, populations. We found several instances where genes do not act independently and instead interact to determine mean life span and mean fertility. Negatively correlated effects between mean life expectancy and internal hatching were found linked to the stP5 genetic marker. These QTLs were mapped using RI strains described previously (Johnson and Wood, 1982; Johnson, 1987; Brooks and Johnson, 1991). The study was facilitated by a polymerase-chain-reaction method for marker assessment (Williams et al., 1992) based on the assessment of a repeated genetic sequence element (Tcl) found with different frequencies in the two wild-type parental strains that were crossed to generate the RIs; the Bergerac strain has several hundred Tcls not found in the N2 strain. The use of RIs has a tremendous advantage for the study of demographic parameters—stable unchanging genotypes that can be examined for multiple distinct phenotypes over a period of months or years. In addition to our analysis of demographic parameters, studies on behavior and sensitivity to anesthetic action in these strains are already under way. This study provides a first step toward determining the particular genes involved in specifying life-history traits and in elucidating the mechanisms underlying the interrelationships among these traits. This type of study will contribute to our understanding of the coordinated evolution of life-history traits and the genetic constraints that are placed on evolution.

OCR for page 108
TABLE 7-3 Basic Theoretical Questions to Be Addressed Number of genes specifying various aspects of the life course Location of genes specifying various aspects of the life course Size of effects specified by each gene Interactions between genes Effects of individual genes on more than one aspect of the life course Identification of and characterization of genes involved in these events Issues in Evolutionary Theory and in Demography That Can Be Addressed by QTL Mapping Gene interactions have been postulated to be important in many aspects of evolution and speciation. It is a general assumption in much evolutionary theory that most genes affect many different traits (pleiotropy). Antagonistic pleiotropy has been suggested as a mechanism for maintaining genetic variation and as a mechanism for the evolution of genes that limit life span, trading off survival with some early fitness trait. In summary, the value of QTL mapping is that it allows the assessment of effects resulting from individual loci rather than the genome as a whole. More specifically, there are a number of questions that QTL mapping can answer (Table 7-3). The genes, their location, and the amount of genetic variation explained by each can be determined by QTL mapping. Most importantly, it is possible to gain some insight into the interactions between and among these QTLs in specifying their respective phenotypes. It is also possible to identify QTLs that are pleiotropic in their action, affecting more than one stage or aspect of life history or demography. Cloning The Genes There is an "ultimate question" which one hopes to eventually answer in these mapping studies: "What is the nature of the genes responsible for these life extensions?" This same ultimate question was asked almost 50 years ago when quantitative genetics and Mendelian genetics made their great rapprochement. It is still not clear if the "mutations" identified and used to such good measure by Drosophila and nematode geneticists are stronger alleles of the same genes underlying quantitative traits. The cloning of the gerontogenes themselves should allow us to answer this question directly. The only successful cloning of gerontogenes based primarily on their extended survival has been in yeast. These studies have focused on the limited proliferative life span of individual yeast cells (Jazwinski, 1996), as well as on the limited length of viability of these cells in stationary phase (Kennedy et al., 1995). The mutants (UTH1-UTH4 ) isolated by Kennedy et al. (1995) had increased life expectancy in stationary culture. One mutant gene was cloned by complementation of its associated sterility phenotype and shown to be a mutation in SIR4, a gene involved in transcriptional silencing.

OCR for page 108
Positional Cloning of Induced Mutants Only in the nematode has positional cloning been successfully used to identify a gerontogene. The spe-26 gerontogene was cloned last year and encodes an actin-associated protein (Varkey et al., 1995). How life extension results from the altered gene is not obvious, although increased stress resistance of spe-26 has been observed for both temperature (Lithgow et al., 1995) and ultraviolet light (Murakami and Johnson, 1996). Our laboratory has been attempting to clone age-l for several years, and that effort resulted in a narrow map assignment to the center of chromosome II (Johnson et al., 1993). Recently, it was proposed that daf-23 and age- are the same gene based on noncomplementation for an unusual phenotype, dauer formation at 27°C (Malone et al., 1996) and the fact that daf-23 also maps to the center of chromosome II. We have confirmed these results and shown noncomplementation between age-1 and daf-23 alleles for several other traits as well. daf-23 has been recently cloned and shown to have homology to the phosphatidylinositol 3-kinase family (Morris et al., 1996) and thus could play a role in transducing signals of environmental stress to the worm genome. No mutations previously assigned to the age-1 gene have been found, as yet, in the daf-23 sequence, leaving open the definitive answer whether daf-23 and age-1 are the same gene. Positional Cloning of QTLs Unfortunately, cloning QTLs specifying gerontogenes will be even more problematic than cloning induced mutants leading to increased life span. For positional-cloning approaches to work, the site of the mutation must be localized to a small area of the genome. Unlike a single-gene mutant. in which a single recombination event can indicate whether the mutant is to the left or right of the recombination site, a QTL cannot be localized based on the results of any single recombinant chromosome. Instead, QTLs must be localized by making congenic strains carrying the QTL region and subsequent analysis of inbred strains carrying these regions. This approach can lead to localizing the region in which the QTL could be located to as small a region as desired but could be quite expensive; perhaps as much as $100,000 could be spent localizing the QTL to a region less than I map unit (2.1 Mb) in size, if these experiments are done in mice. This region may contain 100 or more genes. An approach successfully used in mice in my laboratory is to identify a QTL based on the effect of an additional gene added to the genome by transgenic manipulations. Both in the nematode (Murakami and Johnson, unpublished work) and in the mouse (Rikke et al., unpublished work), the construction of transgenic animals carrying additional DNA, derived from yeast artificial chromosomes (YACs) has been successfully used to show that a QTL is carried in a candidate region. Subsequent cloning of the genes involved in life expectancy and drug sensitivity, respectively, are still under way.

OCR for page 108
Genetics of age-specific mortality in c. Elegans Mortality increases with chronological age in C. elegans just as in many other species. The standard laboratory wild-type strain, N2, shows age-specific mortality rates (the fraction of worms that die in a 24-hour period) that are highly correlated with age. The normal mortality doubling time is about 4-6 days (Johnson, 1987 and 1990; Brooks et al., 1994). Mortality of Gerontogene Mutants The mortality doubling time of strains carrying mutations in age-1 (a gene that can be mutated to increase mean life span 70 percent) has been increased about 3-fold, and this increase is recessive to the normal allele in that individuals with one normal and one mutant allele have mortality doubling times no different from wild type (Johnson, 1990). Males have shorter life spans and a more rapid rate of mortality increase than hermaphrodites (Johnson, 1990). Males carrying age-1 mutations also show a lengthening of life and a slowing of the acceleration rate of mortality, although not as much as hermaphrodites, and age-1 mutant males still have significantly shorter life spans than do hermaphrodites of the same genotype. A major limitation of almost all studies on age-specific mortality rate is the lack of analysis of large populations of nematodes. Only small numbers (200) of worms have been studied, and these results may be modified after larger populations are analyzed. RI Strains Both long-lived mutants and RI strains also have been analyzed to see how age-specific mortality rate is altered. Long-lived R1 strains, some having mean and maximum life spans up to 70 percent longer than wild type, still showed exponential rates of increase of mortality with chronological age, as did both wild-type progenitor strains (Johnson, 1987). Longer life results from a slowing of the characteristic increase in mortality rate that had been thought typical of aging populations in most, if not all, species. The lengths of the developmental and reproductive period were unrelated to increased life span but instead were under independent genetic control. Lengthened life resulted entirely from an increase in postreproductive life span. General motor activity decayed linearly with chronological age in all genotypes. The loss of general motor activity was both correlated with and a predictor of life expectancy, suggesting that both share at least one common, rate-determining component (Johnson, 1987). Mortality of Pure Strains Because increased mean life expectancy could result from lower initial mortality rate, or from a slower rate of increase in mortality (or to alterations in both),

OCR for page 108
we asked how these components varied in the various R1 strains. The slopes of the age-specific mortality rates vary among the R1 strains (Johnson, 1987; Brooks et al., 1994). The longest-lived strain had a consistent increase in both mean and maximum life span and a highly significant, 1.5-fold decrease in the rate of change of age-specific mortality rate. Another short-lived strain had a significant increase in the rate of change of age-specific mortality rate. No significant change in initial vulnerability was seen for either (Johnson, 1987). Mortality of Quartiles, Clustered by Life Expectancy When the combined survival data from these 79 RIs were assessed (Brooks et al., 1994), we found mortality kinetics of the combined population to be similar to that observed in the medfly (Carey et al., 1992) and in Drosophila (Curtsinger et al., 1992). In other words, the survival curve had a long tail, and mortality rates approached 5 percent daily mortality by 17 days of age and remained at that level for several weeks. The experiment was replicated with essentially the same observations. In contrast, when we then divided the set of RIs into quartiles based on life expectancy (Brooks et al., 1994), we observed quite distinct mean survival times and survival curves for each quartile. Not surprisingly, each had significantly different life expectancies (p < .001) from the other. Now, within each quartile, age-specific mortality rate increased almost exponentially with chronological age until considerably later in life. The differences among quartiles was modeled using a Gompertz model and was due almost entirely to differences in the rate of increase of the exponent with essentially no significant difference in initial mortality rate. There still was evidence of a linear period late in life when mortality rates stopped increasing exponentially with increasing age. as demonstrated by a model-comparison approach (Brooks et al., 1994). However, the detailed interpretation of the exact kinetics await analyses of larger populations. Survival and mortality kinetics of the combined R1 populations closely approximate that previously observed by Carey et al. (1992) and Curtsinger et al. (1992). We also observed periods of several weeks at the end of life when the mortality rate did not increase with chronological age. We conclude that heterogeneity within populations (genetic or environmental) will lead to a deceleration of the age-specific mortality curve as soon as the most at-risk population has expired. Exploration of the effects of this genetic heterogeneity on age-specific mortalities is a primary focus of our future studies in which we characterize and genetically map QTLs specifying various aspects of the demography of mortality, and it is also being explored by Curtsinger and his colleagues. Although it is clear that genetic and phenotypic heterogeneity can produce mortality deceleration, there may also be some inherent slowing of mortality with age. Even when both environment and genotype have been kept constant, variability in life span or in other quantitative variables are still observed (Brooks et al., 1994). In this study, variability was a function of average life span, as shown

OCR for page 108
by the constancy of the coefficient of variation, consistent with the hypothesis that those inbred populations that die late in life do not fundamentally differ from those that die early. The biphasic mortality rate seen in the genetically heterogeneous populations of nematodes results from the genetic heterogeneity in the population. Genetic heterogeneity could explain similar mortality kinetics in human populations (Vaupel et al., 1979) in which, beyond 85 years, the mortality rate stops increasing exponentially and becomes constant, or actually decreases. Genetic heterogeneity could play a large role in the long period of ''flat" mortality rate increases separated by Carey et al. (1992). Mortality Rates in Large Populations The relationship between age-specific mortality rate and chronological age on a population of 180,000 worms has been modeled using an exponential function of chronological age (Brooks et al., 1994). Vaupel et al. (1994) used a two-stage Gompertz model and showed that two curves (each with exponential rates of increase) provided a significantly better fit. The break in the curves occurred at day 8, about the time the wild type stops reproducing. Clearly the "best" model for mortality in large populations of C. elegans does not fit a two-parameter Gompertz model. Conclusions It is clear that gerontogenes exist. Identification of these genes has been achieved in lower eukaryotes and in one metazoan, C. elegans. The most powerful approach, inducing single-gene mutants, may be too expensive to ever be pursued in mammals. QTL mapping offers a useful alternative to localizing genes, but cloning of the genes underlying these QTLs will be problematic. Mortality alterations that result from changes in these genes can be studied, and results from the Johnson laboratory suggest that much of the flattening of mortality rates at later ages in humans could result from genetic heterogeneity among individuals. Clearly such heterogeneity leads to an extended period of nonexponentially increases in age-specific mortality at the end of life. References Arking, R.. S.P. Dudas, and G.T. Baker III 1993  Genetic and environmental factors regulating the expression of an extended longevity phenotype in a long-lived strain of Drosophila. Genetica 91:127-142. Botstein, D., and R. Mauer 1982  Genetic approaches to the analysis of microbial development. Annual Review of Genetics 16:61-83. Brooks, A., and T.E. Johnson 1991  Genetic specification of life span and self-fertility in recombinant-inbred strains of Caenorhabditis elegans. Heredity 67:19-28.

OCR for page 108
Brooks, A., G.J. Lithgow, and T.E. Johnson 1994   Mortality rates in a genetically heterogeneous population of Caenorhabditis elegans. Science 263:668-671. Brown, W.T. 1987   Genetic aspects of aging in humans. Review of Biological Research in Aging 3:77-91. Carey, J.R., P. Liedo, D. Orozco, and J.W. Vaupel 1992   Slowing of mortality rates at older ages in large medfly cohorts. Science 258:457-461. Charlesworth, B. 1980   Evolution in Age-Structured Populations. London: Cambridge University Press. Chen, J.B., J. Sun, and S.M. Jazwinski 1990   Prolongation of the yeast life span by the v-HA-RAS oncogene. Molecular Microbiology 4:2081-2086. Covelli. V., D. Mouton, V. Di Majo, Y. Bouthillier, C. Bangrazi, J.-C. Mevel. S. Rebessi, G. Doria, and G. Biozzi 1989   Inheritance of immune responsiveness, life span, and disease incidence in interline crosses of mice selected for high or low multispecific antibody production. Journal of Immunology 142:1224-1234. Curtsinger, J.W., H.H. Fukui, D.R. Townsend, and J.W. Vaupel 1992   Demography of genotypes: Failure of the limited life span paradigm in Drosophila melanogaster. Science 258:461-463. D'mello, N.P.. A.M. Childress, D.S. Franklin, S.P. Kale, C. Pinswasdi, and S.M. Jazwinski 1994   Cloning and characterization of LAG1, a longevity-assurance gene in yeast. Journal of Biological Chemistry 269:15451-15459. Duhon, S.A., S. Murakami. and T.E. Johnson 1996   Direct isolation of longevity mutants in the nematode Caenolrhabditis elegans. Developmental Genetics 18:144-153. Ebert, R.H., V.A. Cherkasova, R.A. Dennis, J.H. Wu, S. Ruggles, T.E. Perrin, and R.J.S. Reis 1993   Longevity-determining genes in Caenorhabditis elegans: Chromosomal mapping of multiple noninteractive loci. Genetics 135:1003-1010. Ebert, R.H., M.A. Shammas, B.H. Sohal, R.S. Sohal, N.K. Egilmez, S. Ruggles. and R.J.S. Reis 1996   Defining genes that govern longevity in Caenorhabditis elegans. Developmental Genetics 18:131-143. Falconer, D.S. 1989   Introduction to Quantitative Genetics, 2nd ed. London: Longman. Fleming, J.E., and M.R. Rose 1996   Genetics of aging in Drosophila. Pp. 74-93 in J.W. Rowe and E.L. Schneider, eds., Handbook of the Biology of Aging, 4th ed. New York: Academic Press. Fleming, J.E., G.S. Spicer, R.C. Garrison, and M.R. Rose 1993   Two-dimensional protein electrophoretic analysis of postponed aging in Drosophila. Genetica 91:183-195. Friedman, D.B., and T.E. Johnson 1988   A mutation in the age-1 gene in Caenorhabditis elegans lengthens life and reduces hermaphrodite fertility. Genetics 118:75-86. Gelman, R., A. Watson, R. Bronson, and E. Yunis 1988   Murine chromosomal regions correlated with longevity. Genetics 118:693-704. Haley, C.S., and S.A. Knott 1992   A simple regression methods for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69:315-324. Ishii, N.. N. Suzuki, P.S. Hartman, and K. Suzuki 1994   The effects of temperature on the longevity of a radiation-sensitive mutant rad-8 of the nematode Caenorhabditis elegans. Journal of Gerontology and Biological Science, 49:B1 17-B120.

OCR for page 108
Jazwinski. S.M. 1996   Longevity-assurance genes and mitochondrial DNA alterations: Yeast and filamentous fungi. Pp. 39-54 in J.W. Rowe and E.L. Schneider, eds., Handbook of the Biology of Aging, 4th ed. New York: Academic Press. Johnson, T.E. 1987   Aging can he genetically dissected into component processes using long-lived lines of Caenorhabditis elegans. Proceedings of the National Academy of Sciences U.S.A. 84:3777-3781. 1990   The increased life span of age-1 mutants in Caenorhabditis elegans results from lowering the Gompertz rate of aging. Science 249:908-912. Johnson, T.E., G.J. Lithgow, S. Murakami, S.A. Duhon, and D.R. Shook 1996   Genetics of aging and longevity in lower organisms. Pp. 1-17 in N. Holbrook, G.R. Martin, and R.A. Lockshin. eds., Cellular Aging and Cell Death. New York: John Wiley and Sons. Johnson, T.E., P.M. Tedesco, and G.J. Lithgow 1993   Comparing mutants, selective breeding, and transgenics in the dissection of aging processes of Caenorhabditis elegans. Genetica 91:65-77. Johnson, T.E., and W.B. Wood 1982   Genetic analysis of the life span of Caenorhabditis elegans. Proceedings of the National Academy of Sciences U.S.A. 79:6603-6607. Kennedy, B.K., N.R. Austriaco, Jr., J. Zhang, and L. Guarante 1995  Mutation in the silencing gene SIR4 can delay aging in C. cerevisiae. Cell 80:485-496. Kenyon, C., J. Chang, E. Gensch, A. Rudner, and R. Tabtiang 1993   A C. elegans mutant that lives twice as long as wild type. Nature 366:461-464. Klass, M.R. 1983   A method for the isolation of longevity mutants in the nematode Caenorhabditis elegans and initial results. Mechanisms of Ageing and Development 22:279-286. Lakowski, B., and S. Hekimi 1996   Determination of life span in Caenorhabditis elegans by four clock genes. Nature Genetics 272:1010-1013. Lander, E.S., and D. Botstein 1989   Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps . Genetics 121:185-99. Lander, E.S., and L. Kruglyak 1995   Genetic dissection of complex traits: Guidelines for interpreting and reporting linkage results. Nature Genetics 11:241-247. Larsen. P.L., P.S. Albert, and D.L. Riddle 1995   Genes that regulate both development and longevity in Caenorhabditis elegans. Genetics 139:1567-1583. Lithgow, G.J. 1996   The molecular genetics of Caenorhabditis elegans aging. Pp. 55-73 in J.W. Rowe and E.L. Schneider, eds., Handbook of the Biology of Aging, 4th ed. New York: Academic Press. Lithgow. G.J., T.M. White, S. Melov, and T.E. Johnson 1995   Thermotolerance and extended life span conferred by single-gene mutations and induced by thermal stress. Proceedings of the National Academy of Sciences U.S.A. 92:7540-7544. 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. Luckinbill, L.S., J.L. Graves, A.H. Reed, and S. Koetsawang 1988   Localizing genes that defer senescence in Drosophila melanogaster. Heredity 60: 367-374.

OCR for page 108
Malone, E.A., T. Inoue, and J.H. Thomas 1996   Analysis of the roles of daf-28 and age-1 in regulating Caenorhabditis elegans dauer formation. Genetics 143:1193-1205. Martin. G.M. 1978   Genetic syndromes in man with potential relevance to the pathobiology of aging. Pp. 539 in D. Bergsma and D.E. Harrison, eds., Genetic Effects on Aging. The National Foundation March of Dimes Birth Defects: Original Article series Vol. 15. New York: Alan R. Liss. Martin. G.M., S.N. Austad, and T.E. Johnson 1996   Genetic analysis of aging: Role of oxidative damage and environmental stresses. Nature Genetics 13:25-34. Maynard Smith, J. 1958   The effects of temperature and of egg-laying on the longevity of Drosophila suhobscura. Journal of Experimental Biology 35:832-842. Morris, J.Z., H.A. Tissenbaum, and G. Ruvkun 1996   A phosphatidylinositol-3-OH kinase family member regulating longevity and diapause in Caenorhabditis elegans. Nature 382:536-539. Munkres, K.D., and C.A. Furtek 1984   Selection of conidial longevity mutants of Neurospora crassa Mechanisms of Ageing and Development 25:47-62. Murakami, S., and T.E. Johnson 1996   A genetic pathway conferring life extension and resistance to UV stress in Caenorhabditis elegans. Genetics 143:1207-1218. Nooden, L.D., and J.J. Guiamét 1996   Genetic control of senescence and aging in plants. Pp. 94-118 in J.W. Rowe and E.L. Schneider, eds., Handbook of the Biology of Aging, 4th ed. New York: Academic Press. Orr, W.C., and R.S. Sohal 1994   Extension of life span by overexpression of superoxide dismutase and catalase in Drosophila melanogaster. Science 263:1128-1130. Partridge, L., and K. Fowler 1992   Direct and correlated responses to selection on age at reproduction in Drosophila melanogaster. Evolution 46:76-91. Puel, A., P.C. Groot, M.G. Lathrop, P. Demant, and D. Mouton 1995   Mapping of genes controlling quantitative antibody production in Biozzi mice. Journal of Immunology 154:5799-5805. Rose, M.R. 1984   Laboratory evolution of postponed senescence in Drosophila melanogaster. Evolution 38:1004-1010. 1991   Evolutionary Biology of Aging. New York: Oxford University Press. Sacher, G.A. 1978   Evolution of longevity and survival characteristics in mammals. Pp. 151-168 in E.L. Schneider, ed., The Genetics of Aging. New York: Plenum Press. Schachter, F., L. Faure-Delanef. F. Guénot, H. Rouger, P. Froguel, L. Lesueur-Ginot, and D. Cohen 1994   Genetic associations with human longevity at the APOE and ACE loci. Nature Genetics 6:29-32. Shepherd, J.C.W., U. Walldorf, P. Hug, and W. J. Gehring 1989   Fruit flies with additional expression of the elongation factor EF-1α live longer. Proceedings of the National Academy of Sciences U.S.A. 86:7520-7521. Shikama, N., R. Ackermann, and C. Brack 1994   Protein synthesis elongation factor EF-lα expression and longevity in Drosophila melanogaster. Proceedings of the National Academy of Sciences U.S.A. 91:4199-4203.

OCR for page 108
Shook, D.. A. Brooks, and T.E. Johnson 1996   Mapping quantitative trait loci specifying hermaphrodite survival or self fertility in the nematode Caenorhabditis elegans. Genetics 142:801-817. Sun, J.Y., S.P. Kale, A.M. Childress, C. Pinswasdi. and S.M. Jazwinski 1994   Divergent roles of RAS1 and RAS2 in yeast longevity. Journal of Biological Chemistry 269(28): 18638-18645. Takata, H., M. Susuki, T. Ishii, S. Sekiguchi, and H. Iri 1987 Influence of major histocompatability complex region genes on human longevity among Okinawan-Japanese centenarians and nonagenarians. Lancet 2:824-826. Tanksley, S.A. 1993 Mapping polygenes. Annual Review of Genetics 27:205-233. Van Broeckhoven, C. 1995   Presenilins and Alzheimer disease. Nature Genetics 11:230-232. VanVoorhies. W.A. 1992   Production of sperm reduces nematode life span. Nature 360:456-458. Varkey, J.P., P.J. Muhlrad, A.N. Minniti, B. Do, and S. Ward 1995   The Caenorhabditis elegans spe26 gene is necessary to form spermatids and encodes a protein similar to the actin-associated proteins kelsh and scruin. Genes and Development 9:1074-1086. Vaupel. J.W., K.G. Manton, and E. Stallard 1979   The impact of heterogeneity in individual frailty on the dynamics of mortality. Demography 16:439-454. Vaupel, J.W., T.E. Johnson, and G.J. Lithgow 1994   Rates of mortality in populations of Caenorhabditis elegans (Technical Comment). Science 266:826. Vijg. J.. and J. Papaconstantinou 1990   EURAGE workshop report: Aging and longevity genes: Strategies for identifying DNA sequences controlling life span. Journal of Gerontology 45:B179-B 182. Williams, B.D., B. Schrank, C. Huynh. R. Shownkeen, R.H. Waterston, and A.P. Wong 1992   A genetic mapping system in Caenorhabditis elegans based on polymorphic sequence tagged-sites. Genetics 131:609-624. Wong, A., P. Boutis, and S. Hekimi 1995   Mutations in the clk-1 gene of Caenorhabditis elegans affect developmental and behavioral timing. Genetics 139:1247-1259. Yu, C.-E., J. Oshima. Y.-H. Fu, E. M. Wijsman, F. Hisama, R. Alisch, S. Matthews, J. Nakura, T. Miki, S. Ouais. G. M. Martin, J. Mulligan, and G. D.Schellenberg 1996   Positional cloning of the Werner's Syndrome gene. Science 272:258-262. Yunis. E.J.. and M. Salazar 1993   Genetics of life span in mice. Genetica 91:211-223. Zeng, Z.-B. 1994   Precision mapping of quantitative trait loci. Genetics 136:1457-1468. Zwaan, B.. R. Bulsma, and R.F. Hoekstra 1995   Direct selection for life span in Drosophila melanogaster. Evolution 49:649-659.