Mechanisms of Ageing and Development
122 (2001) 571– 594
www.elsevier.com/locate/mechagedev

Viewpoint

Why only time will tell
Siegfried Hekimi *, Claire Benard, Robyn Branicky,
´
Jason Burgess, Abdelmadjid K. Hihi, Shane Rea
Department of Biology, McGill Uni6ersity, 1205 A6enue Dr Penfield, Montreal,
Quebec, Canada H3A 1B1
Received 13 November 2000; accepted 15 November 2000

Abstract
The nematode Caenorhabditis elegans has become a model system for the study of the
genetic basis of aging. In particular, many mutations that extend life span have been
identified in this organism. When loss-of-function mutations in a gene lead to life span
extension, it is a necessary conclusion that the gene normally limits life span in the wild type.
The effect of a given mutation depends on a number of environmental and genetic
conditions. For example, the combination of two mutations can result in additive, synergistic, subtractive, or epistatic effects on life span. Valuable insight into the processes that
determine life span can be obtained from such genetic analyses, especially when interpreted
with caution, and when molecular information about the interacting genes is available. Thus,
genetic and molecular analyses have implicated several genes classes (daf, clk and eat) in life
span determination and have indicated that aging is affected by alteration of several
biological processes, namely dormancy, physiological rates, food intake, and reproduction.
© 2001 Elsevier Science Ireland Ltd. All rights reserved.

1. Introduction
Aging research is a vast field of scientific endeavour: a PubMed search with the
keyword ‘aging’ yields 125 000 entries. Recently, a small number of these entries
* Corresponding author. Tel.: +1-514-3986440; fax: +1-514-3981674.
E-mail address: shekim1@po-box.mcgill.ca (S. Hekimi).
0047-6374/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved.
PII: S 0 0 4 7 - 6 3 7 4 ( 0 1 ) 0 0 2 1 8 - 4

572

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

(218) can also be found with the keyword ‘elegans’ because the nematode
Caenorhabditis elegans, a classical model organism, is now being used more
frequently to investigate the biology of aging. This is part of a general impetus to
obtain insight into the molecular mechanisms of aging by harnessing the power of
genetic analysis in invertebrate and unicellular model systems, including the budding yeast Saccharomyces cere6isiae (148 entries) and the fruitfly Drosophila
melanogaster (818 entries).
Here, in an attempt to give an overview of what is known about genes and
processes that affect life span in worms, a series of independently standing
appendices are presented. We have tried to be concise rather than exhaustive and
we apologize to the people whose work was not included because of space
limitations. In contrast to the appendices, which summarize the current state of
knowledge, the main text is concerned with trying to discern the strengths and the
limits of the genetic approach when applied to the study of aging. We have drawn
all our examples from C. elegans, in keeping with the present authors biases, and
because the genetics of aging is well developed in this organism.
2. C. elegans as a model organism for aging research
A number of characteristics make the worm particularly suitable for the study of
aging. It is small ( 1 mm as adult) and thus large numbers of animals can be
grown and manipulated. As it is short-lived ( 15 days for the wild type), aging
experiments can be completed in a relatively short time. Also, worms can be frozen
indefinitely in liquid nitrogen such that a large number of genetically distinct strains
can be created and stored for future use at almost no cost. This also means that
unwanted genetic changes in a strain can be kept to a minimum as it is possible to
return to the original frozen strain if changes occur. C. elegans is an internally
self-fertilizing hermaphrodite. This is a particularly valuable characteristic as it
means that C. elegans tolerates almost total homozygosity. Thus, isogenic strains
can be easily obtained and the effect of mutations of interest can be studied with a
minimum of interference from other loci. C. elegans is also capable of cross-fertilization by males, which makes it possible to perform genetic crosses. Finally, the
worm is extremely hardy and resistant to genetic sickness. In fact, under the
conditions in which it is grown in the laboratory, the worm does not have to move
to feed or to reproduce. These characteristics are important because all mutations
known to date to affect aging, including those that prolong life span, have other
phenotypic effects, which can sometimes be severe.
3. Life span as a phenotype
It is difficult to define aging except by saying that it inevitably results in death.
Thus, the best measure of aging is the increase in the probability of death of the
organism with time. It is legitimate, therefore, to score life span as the main
characteristic when attempting to study genes that affect aging.

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

573

As a phenotype, life span is fickle because it is cumulative, and is therefore very
sensitive. This sensitivity makes it powerful as it allows one to score even very small
physiological differences between two strains as a difference in life span. But this
sensitivity is also a problem as worms are mostly cultured under relatively ill-controlled conditions. They are grown on agar plates on a lawn of E. coli. The agar
mix contains various nutrients for the bacteria, such as yeast extract, tryptone or
peptone, whose composition is not precisely defined and is likely to vary from batch
to batch. In addition to the agar, anything that affects the bacteria will indirectly
affect the worms, such as any genetic change in the standard bacterial strain, or the
age of the bacterial lawn at the time at which it is used as food for worms.
Moreover, although worm cultures are kept in temperature-controlled incubators,
one cannot ensure absolute temperature uniformity, especially over a lifetime of
weeks where the accumulated effect of even small differences in temperature are
likely to result in measurable differences of life span. These problems are partially
overcome by the inclusion of numerous control strains in every experiment and the
practice of carrying out repeat experiments. (Our experience is that at least three
independent experiments with a sample size of 50 for each are necessary to obtain
a reliable measure for life span differences of less than 50%.) The finding of a
relative difference in life span between strains scored at the same time is therefore
a much more meaningful measure than absolute differences in life span observed
among separate experiments, particularly when the experiments are conducted in
different laboratories.

3.1. Life span and genetic background
Another consequence of the special nature of life span as a phenotype is its
extreme sensitivity to genetic background. There are many reasons to suspect that
life span is a highly polygenic trait, and indeed most theories of aging support this
view. Although one of the strengths of C. elegans as a model system is the
possibility to obtain highly isogenic strains, potential differences in genetic background should still be considered and precautions should be taken to minimize
them. Indeed, in a recent study we carried out to investigate aging in mutants that
affect nervous system and muscle function (Lakowski and Hekimi, 1998), we have
found that a large proportion of strains harboured background mutations that were
not responsible for the main phenotype of the strain but that nonetheless affected
life span. Repeated backcrossing of the mutant strain to the canonical wild type
strain (N2) removed additional mutations in all cases but one.

3.2. Long li6es
The great power of C. elegans for aging research is that it is a genetic system in
which mutations can readily be identified. In fact, a number of mutations that
increase life span are being studied. It is an inescapable conclusion that the normal
activity of a gene must limit life span in the wild type when a loss-of-function
mutation of the gene results in an increased life span of the mutant.

574

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

3.3. Short li6es
It is much more difficult to interpret mutations that shorten life span because one
cannot a priori distinguish between mutations that increase the rate of a life
span-limiting process and mutations that produce a pathology whose incidence does
not normally increase with age. We believe that this is not a semantic distinction.
One can easily imagine structures or processes that never fail within the life span of
wild-type animals, or do not fail more often with chronological age. In fact, the
existence of mutant strains whose average and maximum life span is several times
longer than that of the wild type implies the existence of highly resistant structures
which can support a life span longer than that of the wild type (Kenyon et al., 1993;
Larsen et al., 1995; Lakowski and Hekimi, 1996, 1998). Indeed, such long-lived
mutants are often anatomically very similar to the wild type. This means, for
example, that the worm’s collagenous cuticle, which serves as both a mechanical
support and a protection from the environment, is capable of resisting environmental assaults for much longer than ever needed in the wild type. A mutation that
would make the cuticle less resistant to the point of shortening the mutants’ life
span is unlikely to identify a function that identifies one of the causes of normal
aging.
The considerations above, however, do not mean that all genes whose loss-offunction phenotype results in a shorter life are irrelevant to aging. From examples
in C. elegans, we will see that with careful genetic analysis and prudent interpretation of epistatic relationships, some short-lived mutants can help to unravel the
molecular mechanisms of aging.

3.4. Big and small effects
The range of life-span increases brought about by single mutations is quite wide.
For example, daf-2 mutants can live more than two times longer than the wild type
(Kenyon et al., 1993), while clk-1, -2 and -3 mutants live only approximately 30%
longer than the wild type (Wong, et al., 1995; Lakowski and Hekimi, 1996).
Mutants that live very long are understandably easier to study. They can allow for
the unambiguous identification of epistatic or suppressor mutations (Kenyon et al.,
1993; Larsen et al., 1995; Taub et al., 1999). For example, mutations in either
daf-16 or ctl-1 dramatically shorten the life span of daf-2 mutants. In fact, the life
span of these double mutants is as short or even shorter than that of the wild type,
while mutations in daf-16 and ctl-1 shorten the life span of the wild type only
slightly.
On the other hand, that mutations in single genes lengthen life span only
modestly is not a clear indication of the importance of these genes for life span
determination. For example, mutations in clk-1, which increase life span relatively
little, can double the life span of daf-2 mutants under some environmental
circumstances (Lakowski and Hekimi, 1996). Similarly, double mutants of clk-1
with clk-2 or clk-3, live from two to three times longer than the wild type
(Lakowski and Hekimi, 1996). Another way to say this is that genes function only

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

575

in the context of other gene activities: a daf-2 mutant lives shorter than the wild
type in a daf-16 mutant background and a clk-1 mutant lives two times longer than
the wild type in a clk-2 mutant background. Thus, it appears that, although finding
mutants that increase life span is a prerequisite to identify processes that are
important for normal aging, the magnitude of the effect of a given gene mutation
cannot be directly interpreted.

3.5. En6ironmental effects
In the laboratory, C. elegans can be grown under a variety of conditions at
temperatures ranging from 15 to 25°C. Most commonly, for aging experiments,
worms are cultured on a dense lawn of E. coli on a solid substrate (agar). On dense
bacterial lawns, the pumping action by which the worm ingests bacteria is efficient.
Worms can also be grown under liquid conditions with either living bacteria or
bacteria killed by heat or irradiation, or in defined axenic medium, that is, entirely
without bacteria. Liquid culture is stressful for the worms as it makes it more
difficult for them to ingest food. This probably leads to caloric restriction (Appendix C), which would explain why most genotypes live longer when grown in
liquid culture (Braeckman et al., 1999). For some genotypes these conditions are
too stressful as many mutants cannot be grown at all in liquid culture or in axenic
medium. For example, clk mutants cannot grow in axenic medium and are infertile
in medium with killed bacteria (Braeckman et al., 1999).
The different conditions under which worms can be grown raises the question of
how to interpret differences of life span of a given genotype under different
environmental conditions. It is well known that temperature affects aging, and this
is true for wild-type as well as for mutant worms. Some genotypes live long only at
some temperatures. For example, clk-2 (qm37 ) is a temperature-sensitive embryonic
lethal mutation at 25°C (Hekimi et al., 1995), and daf-2 mutants arrest permanently
at a specific developmental stage at 25°C (Kenyon et al., 1993). On the other hand,
daf-2 mutants live a very long time when first grown at a permissive temperature
that allows them to complete development, and then transferred to 25°C for their
adult life. Clearly, daf-2 affects a process that limits life span, yet given that daf-2
is also required for this process to proceed normally, conditions (25°C) can be
found where the absence of daf-2 is clearly deleterious and leads to permanent
developmental arrest.
Similarly, when grown on bacterial plates, the life span of clk-1 is only marginally affected by daf-16 (Lakowski and Hekimi, 1996, 1998). In fact, it appears to
be shortened exactly by the amount by which daf-16 also shortens life span in a
wild-type background. However, in liquid culture daf-16 mutations limit much
more severely the life span of daf-16 ; clk-1 double mutants. We believe, however,
that this cannot be interpreted to mean that daf-16 is required for the long life of
clk-1. Indeed, we have argued above that it is very difficult to interpret the meaning
of short life spans because they can always be the result of a life-shortening
pathology that does not limit life span in the wild type. Therefore, when a
particular double mutant combination is short-lived only under particular environ-

576

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

mental conditions, this cannot be used to interpret the genetic meaning of the
interaction. The fact that an environmental condition can be found in which it is
deleterious to be simultaneously a clk-1 mutant and a daf-16 mutant, so that the
life span of the double mutant is shorter than that of clk-1, does not tell one much
about either gene, given that, under different conditions (bacterial plates), daf-16 ;
clk-1 double mutants live long. The only way in which the shorter life span of the
double mutant under the harsher conditions could be meaningful is if clk-1
increased life span in liquid culture by a different mechanism than by which it
increases life span on bacterial plates, only the former requiring daf-16.
4. Genetic interactions
As illustrated in the previous section, one way to gain insight into the function of
genes that affect life span is to study their interactions with other genes and with
each other by constructing and characterizing double mutants. This has been done
in a number of studies, some of which are described in detail in the appendices.
Here we examine the classes of possible outcomes when double mutants are
constructed (Fig. 1), and discuss how much one can or cannot conclude from such
studies.

4.1. Additi6ity and synergism
Some double mutants live longer than any of the two component single mutants.
The extension of life span in double mutants may arise from two types of genetic
interactions: additivity or synergism. When the life span of the double mutant is
approximated by adding the percent life span increases over wild type of the single
mutants, then the effect of the single mutations is additive. Any increase much
larger than an additive increase can be defined as synergistic. It is unclear at the
present time whether additive and synergistic interactions have different meanings.
An example of synergism is provided by the interactions of the three clk genes:
double mutants of any pair of clk genes live much longer than the single genes
(Lakowski and Hekimi, 1996). Such an outcome suggests that the two mutations
increase life span by mechanisms that are at least partially different from each
other. In other words, the gene products do not function in a linear cascade of
biochemical interactions. They might impinge on the same downstream process but
in manners that are sufficiently distinct that the process is altered more severely
when both mutations are present (see also Appendix A).
Fig. 1. The study of double mutants reveals different types of interactions between genes that affect life
span in the nematode C. elegans. In each of the five illustrated cases, the life span of a given strain is
indicated by the position of a schematic worm along the horizontal axis. The genotype of the strain is
given just below the worm. In every case, a black worm represents the wild type, a large red worm
represents the double mutant, and the small green or blue worms represent the constituent single
mutants. The relationships between life spans are schematized and not scaled. The functional significance
of these interactions is discussed in detail in the text (Section 4).

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

Fig. 1.

577

578

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

4.2. Absence of additi6ity
Another possible outcome is that the double mutant lives no longer than any of
the component single mutants. For example, eat-2 (ad465 ); clk-1 (e2519 ) double
mutants live no longer than either eat-2 (ad465 ) or clk-1 (e2519 ) which produce a
similar increase in life span (Lakowski and Hekimi, 1998). Except for life span, the
phenotypes of these two mutants are very different from each other. clk-1 mutants
are slow developing and slow behaving, but display a fully wild-type appearance
and morphology (Wong et al., 1995). However, eat-2 mutants, which have reduced
food intake due to slow ingestion of bacteria, develop and behave at essentially
wild-type rates (except for food intake) but appear transparent and morphologically
stunted (Avery, 1993). The two mutations are fully additive except for life span, so
that the double mutants are both slow growing and slow behaving, and appear
transparent and stunted. These observations have been interpreted as meaning that
clk-1 and caloric restriction impinge on the same downstream process that affects
life span (Lakowski and Hekimi, 1998) (Appendix C). Another interpretation could
be that the double mutant might not succeed in living longer than it does because
of novel deleterious effects produced by the combined effects of the two underlying
mutations. This is an example of how the absence of positive effects on life span is
generally difficult to interpret.
The interpretation of an absence of additivity can be more straightforward in
some cases. For example, double mutants of daf-2 and age-1 do not live longer
than either of the single mutants (Dorman et al., 1995). However, the interpretation
that they impinge on the same process is much more clear than in the case of clk-1
and eat-2. Indeed, daf-2 and age-1 mutants have the same dauer constitutive
phenotype and are placed at the same level of the genetic pathway of dauer
formation (Appendix B). Furthermore, both genes have been cloned and shown to
encode gene products (an insulin-growth-factor like tyrosine kinase transmembrane
receptor and a PI3-kinase, respectively) that are known to interact in several
systems. Here, the absence of additivity, together with other data, indicates that the
life span increase in both mutants is due to altered regulation of the dauer
formation pathway, and not to some unknown consequences of mutations, unrelated to the normal functions of the gene products in the wild type.

4.3. Subtracti6ity
The occurrence of subtractivity is best illustrated by the interaction of clk-1 with
ctl-1 (Taub et al., 1999). ctl-1 mutants display decreased levels of catalase activity
and have a shortened life span compared to the wild type. However, the degree of
shortening of the life span of clk-1 by ctl-1 (the life span of the double is 73% of
the life span of clk-1 ) is very similar to the degree to which ctl-1 shortens life span
in a wild-type background (76%). Here it is the shortening of life span in one case,
and the lengthening in the other, that are additive, which suggests that the
mechanisms by which clk-1 and ctl-1 impinge on life span are substantially
different. However, because a shorter life span is observed in the double mutant, it

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

579

is not possible to conclude very firmly. Indeed, one cannot exclude the possibility
that ctl-1 is one of the genes that are necessary for the life span extending activity
of clk-1. Even so, it must not be the only target as the disruption of ctl-1 only
partially abrogates the life span extension conferred by clk-1 mutations. Furthermore, the clk-1 (e2519 ) mutation extends significantly the life span of ctl-1 ; daf-2,
which in a clk-1 ( + ) background lives no longer than ctl-1 mutants (Taub et al.,
1999) (Section 4.4, Appendix C).

4.4. Epistasis for life span
A mutation is epistatic to a second mutation when the phenotype of the double
mutant is indistinguishable from that conferred by the first mutation. One can look
at the epistatic relations between mutations by considering only one of the
phenotypes of the mutants. For example, when only life span is considered, then
ctl-1 appears to be fully epistatic to daf-2. Indeed, ctl-1 fully suppresses the
dramatically increased life span of daf-2, so that the double mutant lives as short
as ctl-1. However, ctl-1 suppresses none of the other phenotypes of daf-2 mutants,
in particular, their dauer constitutive phenotype. These results strongly suggest that
ctl-1 is required for the long life of daf-2 but do not indicate that the long life of
daf-2 is actually caused by an action of daf-2 on ctl-1 (+ ). In other words, because
a short life span is observed, it is possible that the conjunction of the mutations in
both genes is deleterious enough to produce a shortened life span without having
impinged on the process that daf-2 affects to prolong life span. It should be noted,
however, that in this particular case, a closer link between the actions of daf-2 and
ctl-1 is likely as various dauer constitutive mutants have been found to be stress
resistant and have elevated levels of enzymes that detoxify oxygen radicals (Larsen,
1993; Honda and Honda, 1999; Taub et al., 1999) (Appendix D). Furthermore,
ctl-1 mutants have lower catalase levels (Taub et al., 1999).

4.5. Full epistasis
The interaction between daf-16 and daf-2 is one of the cleanest cases of epistasis
(Kenyon et al., 1993). The phenotype of the double mutant is indistinguishable
from that of daf-16 mutants, for life span as well as for all other phenotypes
(Gottlieb and Ruvkun, 1994). For example, the double mutants are incapable of
forming dauer larvae as is daf-16. This suggests very strongly that the effect of
daf-2 on life span is mediated by daf-16. Consideration of the molecular identities
of the genes clarifies the situation even further: DAF-2 is a transmembrane receptor
in a classical signal transduction pathway and DAF-16 is a transcription factor (see
Appendix B).
As we have seen above, most interactions are not as clear cut as those that can
be observed between daf genes, whether daf-2 and age-1, or daf-2 and daf-16. One
possibility why this might be so is that the daf genes function in a signal
transduction pathway. Their biochemical activities might not directly impinge on
aging, as might repair or detoxifying enzymes, for example. Rather, their action

580

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

results in a downstream signal that likely has an effect on a large number of targets.
One might venture to predict that interactions between the targets of daf-16 will not
show such beautifully clear genetic interactions. In addition, the picture becomes
more complicated when other daf genes are taken into consideration. For example,
daf-12, which encodes a nuclear hormone receptor and acts in parallel to the
insulin-like signaling pathway, shows allele-specific interactions with daf-2 with
respect to both dauer formation and life span (Larsen et al., 1995; Gems et al.,
1998)

4.6. Allele-specific interactions
The study of allele-specific interactions can reveal different types of interactions
between genes. For example, two sets of clk-1 clk-2 and clk-3 clk-1 double mutants
were constructed: one set with the strong allele clk-1 (qm30 ), and another set with
the much weaker allele clk-1 (e2519 ). As described above, interactions between the
clk genes are all additive or synergistic. However, allele-specific interactions are very
different for clk-2 compared to clk-3. clk-3 (qm38 ); clk-1 (qm30 ) mutants are much
more severe than clk-3 (qm38 ); clk-1 (e2519 ) mutants, including for development,
life span and general health (Lakowski and Hekimi, 1996). This is the expected
result if clk-1 and clk-3 have partially overlapping functions: in the absence of
clk-3 ( + ) the phenotype becomes more severe as the activity of clk-1 is more
reduced in the more severe allele qm30. In contrast, the double mutants with
clk-2 (qm37 ) containing either e2519 or qm30 are identical for development as well
as for life span. This suggests that clk-2 functions either downstream or upstream
of clk-1 and that the residual activity of e2519, which makes it a weak allele of
clk-1, depends entirely on the presence of a wild-type copy of clk-2.

5. Conclusions
The viewpoint we have tried to defend here is that the aging process can be
studied genetically in a model organism such as C. elegans, which allows one to
identify mutations that prolong life span and to quickly carry out genetic experiments. We believe that the identification and study of long-lived mutants will be
crucial to understanding aging as such mutations necessarily identify processes that
normally limit life span.
Beyond this, more elaborate questions can be asked and answered by performing
classical genetic analyses with these mutants. Yet, it is clear that life span is an
unusual phenotype, and that some of the interactions we can observe have to be
interpreted with great caution. This is because all genes serve functions that
contribute to the viability of the organism, and because life span studies measure
the occurrence of death. Altering the function of a gene may lengthen life span
under particular conditions, but will invariably contribute to make the organism
more susceptible to death under some other genetic or environmental conditions.

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

581

It is also clear from the examples we have given that knowledge of the molecular
identity of aging genes can greatly enhance our ability to interpret the outcome of
genetic interactions. One might think that when molecular functions are known
genetics becomes less important. To the contrary, we believe that molecular
identities are particularly precious when they contribute to a coherent picture of the
meaning of genetic interactions. Only the links between phenotype and molecular
function provided by genetics can identify the processes that are indeed causal to
aging.

Acknowledgements
This work was supported by a Medical Research Council of Canada Scientist
Award to S.H., a Hydro-Quebec McGill Major Fellowship to C.B., a McGill
´
Faculty of Graduate Studies Fellowship to R.B., and a Swiss National Foundation
Fellowship to A.K.H.

Appendix A. The clk genes
Several biological processes have been found to influence the life span of the
nematode C. elegans, including dormancy, food intake, and physiological rates1. A
group of at least four clk genes, namely clk-1, clk-2, clk-3, and gro-1 has been
proposed to affect life span by altering the latter2. Mutations in any one of the clk
genes produce a complex phenotype that includes the average slowing down of
development and rhythmic behaviors, as well as the lengthening of life span2,3 (Fig.
2A, B). For example, at 18°C, all Clk mutants have a mean life span that is at least
3 days longer than that of the wild type2. While the increase of the duration of
post-embryonic development of clk mutants contributes to their increased mean
and maximum life span, it is not sufficient to account for the increase in life span2.
clk double mutants exhibit a very pronounced increase in the duration of
development and life span (Fig. 2A, B). For instance, the clk-1 clk-2 double mutant
takes approximately 4 days longer than the wild type to develop and lives 14 days
longer at 18°C. The synergism of the effects of individual clk mutations on life span
indicates that they affect different molecular processes. However, they all appear to
extend life span by altering the same biological process, i.e. physiological rates.
Indeed, there is a strong positive correlation between the duration of development
and adult life span for both the wild type and strains carrying mutations in one or
two clk genes (Fig. 2C). It is reasonable to speculate that life span is determined, at
least in part, by the interplay between damage accumulation and repair. Thus,
slower physiological rates in clk mutants (illustrated by slower developmental rates)
should produce less damage, and the slow life of the clk mutants is probably causal
to their long life. Consistent with this idea is the observation that increasing clk-1
activity by overexpression of the wild-type gene accelerates physiological rates and
shortens life span4.

582

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

Fig. 2. clk mutations interact to determine life span in C. elegans: (A) Survival curves of clk-1 (e2519 )
and clk-2 (qm37 ) single and double mutants at 18°C. The graph shows the percentage of worms alive on
a given day after hatching. clk-1 clk-2 double mutants live longer than clk-1 and clk-2 single mutants;
(B) The length of development (from hatching to adult molt) and the mean life span at 18°C are given
for the genotypes presented in A; (C) The relationship between the length of post-embryonic development and mean adult life span at different temperatures. Circles represent the wild type and squares
represent different mutant strains containing clk-1, -2, -3 mutations, either singly or in double mutant
combinations. The positive correlation between post-embryonic development and adult life span
indicates that the effect of clk genes on development and life span is similar to the effect of lowering
temperature. This suggests that clk mutants live long because they live slowly. (After data in Lakowski
and Hekimi2).

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

583

Mutations in another group of genes, the dauer formation genes daf-2 and age-1,
also extend life span (Appendix B). The daf-2 and age-1 mutations alter life span
in a manner that is different from that of the clk mutations by at least two criteria.
First, while the life span extension of both daf-2 and age-1 depends on the activity
of another gene, daf-16, loss-of-function mutations in daf-16 cannot suppress the
long life of clk mutants. At most, a slight reduction of life span is observed in
daf-16 ; clk-1 /-3 double mutants, that corresponds to the life-shortening effect of
daf-16 on wild-type background2,5. Second, daf-2 and clk-1 mutations display a
synergistic effect on life span: daf-2 clk-1 double mutants live much longer than
either of the single mutants and can live over five times longer than the wild type.
The clk-1 gene has been cloned and encodes a small protein of unknown
biochemical function that is well conserved, both phylogenetically from proteobacteria to eukaryotes1, and functionally6,7. Indeed, both worm and rat sequences can
rescue the phenotype due to mutations in coq7, the yeast homologue of clk-1. coq7
mutants are defective in the synthesis of ubiquinone, an electron carrier that is
required for respiration, and therefore cannot grow on non-fermentable carbon
sources8. CLK-1 protein localizes to the mitochondria in all the cells of the worm4
and in yeast, where it has been shown to be part of the inner mitochondrial
membrane9. However, the level of respiration is only very mildly affected in clk-1
mutants compared to the wild type. It has been hypothesized that clk-1 could be
implicated in the cross-talk between the mitochondria and the nucleus to regulate
gene expression as a function of the level of energy metabolism in the cell4,10.
A.1. References (Appendix A)

1. Hekimi, S., 2000. Crossroads of aging in the nematode Caenorhabditis elegans.
In: Hekimi, S. (Ed.), The Molecular Genetics of Aging, Springer Verlag,
Berlin, pp. 81– 112.
2. Lakowski, B., Hekimi, S., 1996. Determination of life-span in Caenorhabditis
elegans by four clock genes. Science 272, 1010– 1013.
3. Wong, A., et al., 1995. Mutations in the clk-1 gene of Caenorhabditis elegans
affect developmental and behavioral timing. Genetics 139, 1247–1259
4. Felkai, S., et al., 1999. CLK-1 controls respiration, behavior and aging in the
nematode Caenorhabditis elegans. EMBO J. 18, 1783–1792.
5. Lakowski, B., Hekimi, S., 1998. The genetics of caloric restriction in
Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 95, 13 091–13 096
6. Ewbank, J.J., et al., 1997. Structural and functional conservation of the
Caenorhabditis elegans timing gene clk-1. Science 275, 980–983.
7. Jonassen, T., et al., 1996. Isolation and sequencing of the rat Coq7 gene and
the mapping of mouse Coq7 to chromosome 7. Arch. Biochem. & Biophys.
330, 285–289.
8. Marbois, B.N., Clarke, C.F., 1996. The COQ7 gene encodes a protein in
Saccharomyces cere6isiae necessary for ubiquinone biosynthesis. J. Biol. Chem.
271, 2995– 3004.

584

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

9. Jonassen, T., et al., 1998. Yeast Clk-1 homologue (Coq7/Cat5) is a mitochondrial protein in coenzyme Q synthesis. J. Biol. Chem. 273, 3351–3357.
10. Branicky, R., et al., 2000. clk-1, mitochondria, and physiological rates. Bioessays 22, 48– 56.

Appendix B. The insulin-like signaling pathway and aging
In C. elegans, a neurosecretory-signaling pathway regulates the decision to
continue growth or to enter an alternative, developmentally arrested larval stage
termed the dauer stage. Animals enter the dauer stage when environmental conditions do not favor growth and reproduction, such as when food resources are low
and population density and temperature are high. Dauer animals exhibit numerous
unique traits including increased resistance to environmental stress and a life span
that well exceeds that of adult animals1. Components of this neurosecretory
pathway include sensory neurons, hormone-like signaling molecules as well as
signaling cascades to transduce signals in the cells of downstream target tissues2.
Numerous mutations that enhance or suppress dauer formation have been isolated
and these have been ordered into a complex, branched genetic pathway based on
genetic epistasis. There appears to be at least three partially independent signaling
pathways that regulate dauer formation including a TGF-b like pathway, a cyclic
nucleotide-signaling pathway and an insulin-like pathway. Interestingly, the insulinlike signaling pathway appears to regulate longevity as well as dauer formation.
Mutations in the gene daf-2 3, result in constitutive larval arrest at the dauer stage
(at restrictive temperatures) and a 2-fold increase in life span (under more permissive conditions)4. daf-2, which encodes an insulin receptor-like tyrosine kinase,
transduces signals through age-1 5 a phosphatidylinositol-3-OH kinase family member, which generates the 3-phosphoinositide second messengers PtdIns-3,4-P2 and
PtdIns-3,4,5-P3 (Fig. 3). These in turn activate akt-1 and akt-2 6, which are believed
to directly antagonize the forkhead transcription factor daf-16 7,8. Mutations that
decrease signaling in this pathway, such as mutations in daf-2 and age-1, result in
upregulation of daf-16, leading to an increase in dauer arrest and to a dramatic
extension of life span. Mutations in daf-16 suppress these dauer formation and long
life span phenotypes.
There are at least twelve putative insulin-like signaling ligands in the C. elegans
genome9. It is not currently known which particular insulin genes are involved and
what processes regulate activation of the worm insulin-like pathway. Recent
evidence suggests that mutations that affect ciliated sensory neurons, which have
been previously shown to have a role in dauer formation, can regulate aging
through the insulin-like signaling pathway10. In particular, mutants with defects in
cilia structure, as well as animals in which sensory structures have been laser
ablated, exhibit increased life span. These mutations could not enhance the long life
span of daf-2 mutants, indicating that they function in the same pathway. In
support of this view, the long life span of the sensory neuron mutants is almost
fully suppressed by mutations in daf-16. This suggests that ciliated sensory neurons

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

585

Fig. 3. A model of the C. elegans insulin-like signaling pathway and the various processes that modulate
it. Under growth promoting conditions, an insulin-like signaling ligand activates DAF-2, which recruits
AGE-1. This results in the production of the second messengers PtdIns-3,4-P2 and PtdIns-3,4,5-P3,
which activate AKT-1 and AKT-2, which in turn antagonize the forkhead transcription factor DAF-16.
Absence of signaling through this pathway results in upregulation of DAF-16, increased dauer
formation, and long life. The ciliated sensory neurons control an insulin-like signal that stimulates
DAF-2 activity. The reproductory system produces mutually antagonistic signals. The gonad appears to
inhibit DAF-2 signaling while the germ line suppresses DAF-16 activity.

586

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

may control the release of an insulin-like signaling molecule that regulates aging. In
the absence of appropriate environmental cues or in the case of non-functional
sensory neurons, the level of this ligand falls, resulting in decreased activation of the
insulin-like pathway, and consequently in the upregulation of daf-16 and increased
life span (Fig. 3). In addition, two genes involved in Ca2 + -regulated secretion in
neurons, unc-64 and unc-31, have been demonstrated to enhance lifespan through
the insulin-like signaling pathway11. These two genes may be directly involved in
the release of an insulin-like signal. Together, these results suggest that neurons
modulate the release of an insulin-like signal that regulates life span through the
insulin-like signaling pathway.
A second process that appears to be involved in regulating the insulin-signaling
pathway is the reproductive system. Hsin and Kenyon1 propose a model in which
the germline produces a signal that suppresses long life by antagonizing daf-16
function (Fig. 3). Laser ablation of the germ line precursor cells (Z2 and Z3)
releases this suppression and results in increased lifespan. Conversely, the gonad
appears to produce a life span promoting signal that antagonizes daf-2 insulin-like
receptor function. In support of this view, whole gonad ablation in a daf-16 mutant
background, in which the germ line signal is blocked, results in a decrease in life
span. It appears that these two opposing signals are balanced as removal of the
whole gonad in wild-type animals has no effect on life span4.
How upregulation of daf-16, the terminal output of the insulin-like pathway,
increases life span is not known. However, it appears that upregulation of daf-16
might activate some aspects of a dauer specific program, including genes that
regulate metabolism as well as stress resistance (Appendices C and D).
B.1. References (Appendix B)

1. Klass, M., Hirsh, D., 1976. Non-ageing developmental variant of Caenorhabditis elegans. Nature 260, 523–525.
2. Riddle, D.L., Albert, P.S., 1997. Genetics and environmental regulation of
dauer larva development. In: Riddle D.L., (Ed.), C. elegans II., Cold Spring
Harbor Laboratory Press, New York pp. 393– 412.
3. Kimura, K.D., et al., 1997. daf-2, an insulin receptor-like gene that regulates
longevity and diapause in Caenorhabditis elegans. Science 277, 942–946
4. Kenyon, C., et al., 1993. A C. elegans mutant that lives twice as long as wild
type. Nature 366, 461– 464.
5. Morris, J.Z., et al., 1996. A phosphatidylinositol-3-OH kinase family member
regulating longevity and diapause in Caenorhabditis elegans. Nature 382,
536 – 539.
6. Paradis, S., Ruvkun, G., 1998. Caenorhabditis elegans Akt/PKB transduces
insulin receptor-like signals from AGE-1 P13 kinase to the DAF-16 transcription factor. Genes & Dev. 12, 2488–2498.
7. Lin, K., et al., 1997. daf-16 : An HNF-3/forknead family member that can
function to double the life-span of Caenorhabditis elegans. Science 278, 1319–
1322.

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

587

8. Ogg, S., et al., 1997. The fork head transcription factor DAF-16 transduces
insulin-like metabolic and longevity signals in C. elegans. Nature 389, 994– 999.
9. Gregoire, F.M. et al., 1998. Cloning and developmental regulation of a novel
member of the insulin-like gene family in Caenorhabditis elegans. Biochem. &
Biophys. Res. Comm. 249, 385–390.
10. Apfeld, J., Kenyon, C., 1998. Cell nonautonomy of C. elegans daf-2 function
in the regulation of diapause and life span. Cell 95, 199– 210.
11. Ailion, M., et al., 1999. Neurosecretory control of aging in Caenorhabditis
elegans. Proc.Natl.Acad.Sci. USA. 96, 7394–7397.
12. Hsin, H., Kenyon, C., 1999. Signals from the reproductive system regulate the
lifespan of C. elegans. Nature 399, 362–366.

Appendix C. Caloric restriction
Over 60 years ago, it was discovered that reducing the caloric intake of rats could
extend their mean and maximal lifespan. Caloric restriction (CR, also called dietary
restriction) has since been found to extend the life span of a wide range of animals
(reviewed in Weindruch and Walford1), and remains the only well-documented
experimental means of extending vertebrate life span. The effects of CR have been
best studied in rodents, where it has been shown that CR results in many
physiological changes, including reductions in body weight and body temperature,
blood glucose and insulin levels, and oxidative damage, and has also been found to
boost DNA repair capabilities (reviewed in Sohal and Weindruch2). The mechanism
by which CR increases life span is unknown, as it is unclear which of the many
changes brought about by CR is/are responsible for the life-span extension.
However, recent studies have shown that CR induces the production of growth
factors and stress proteins, and that production of these proteins increases the
resistance of the cells to stress and might help them cope with the adverse effects of
aging and age-related disease9,10.
Lowering food intake, in this case bacteria, has also been shown to increase the
life span of C. elegans. Klass3 showed that growing worms in liquid culture with
reduced levels of bacteria increases their life span. Lakowski and Hekimi4 showed
that mutations in eat genes, which affect the function of the pharynx, the organ that
pumps bacteria into worms, also increase worm life span. A number of observations suggest that these mutations do indeed prolong life span by restricting the
food intake of worms. In particular, eat mutants have a ‘starved’ appearance, that
is, they are more transparent and often smaller than the wild type5; mutations in
some eat genes, like eat-2, actually affect the rate at which the animal pumps in
food5,6, and for some mutations (eat-2 and eat-6 ) the life-span extension correlates
with the severity of the feeding defect4. Moreover, unc mutations, which affect
muscle or nervous system function, but not the pharynx, do not increase life span4
(Fig. 4).
In addition to caloric restriction, there are at least two other distinct genetic
mechanisms for extending life span in C. elegans. One mechanism involves genes

588

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

Fig. 4. Interactions between the different genetic mechanisms that extend life span in C. elegans. daf
genes, such as daf-2, and clk genes, such as clk-1, extend life span by distinct mechanisms. Caloric
restriction extends life span in a manner that is distinct from that of daf-2, but similar to that of clk-1.
CR and mutations in clk-1 might both extend life span by affecting the metabolic state of the animal.

that regulate dauer formation, such as daf-2 and daf-16 (Appendix B); the other
mechanism involves genes that affect physiological rates, the Clock genes, such as
clk-1 (Appendix A). Genetic interactions between eat-2 and daf-2 /daf-16 suggest
that dauer genes extend lifespan by a mechanism that is distinct from that of CR:
eat-2 ; daf-2 double mutants live substantially longer than either eat-2 or daf-2
mutants, and mutations in daf-16 do not suppress the long life of eat-2 mutants4.
In contrast, genetic interactions between eat-2 and clk-1 suggest that clk-1 mutations may extend lifespan by a mechanism similar to that underlying CR: eat-2 ;
clk-1 double mutants do not live significantly longer than animals that have only
eat-2 or clk-1 mutations4. However, given the very different phenotypes of eat and
clk mutants, it is unclear what process they both affect. One possibility is that CR
and mutations in clk-1 both affect the metabolic state of the animal. Although it is
not yet clear, in worms or in other organisms, whether CR results in decreased
metabolic rates, reduced caloric intake could result in a metabolic shift to lessen
energy consumption, a possibility that has also been raised to explain the phenotype
of clk-1 7,8.
C.1. References (Appendix C)

1. Weindruch, R.K., Walford, R.L., 1988. The retardation of aging and disease
by dietary restriction. (Springfield, I.L.: Thomas)
2. Sohal, R.S., Weindruch, R., 1996. Oxidative stress, caloric restriction, and
aging. Science 273, 59– 63.
3. Klass, M., 1977. Aging in the nematode Caenorhabditis elegans: major biological and environmental factors influencing life span. Mech. Aging and Dev. 6,
413 – 429.

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

589

4. Lakowski, B., Hekimi, S., 1998. The genetics of caloric restriction in
Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA. 95, 13 091–13 096.
5. Avery, L., 1993. The genetics of feeding in Caenorhabditis elegans. Genetics
133, 897– 917.
6. Raizen, D.M., et al., 1995. Interacting genes required for pharyngeal excitation
by motor neuron MC in Caenorhabditis elegans. Genetics 141, 1365–1382.
7. Braeckman, B.P., et al., 1999. Apparent uncoupling of energy production and
consumption in long-lived Clk mutants of Caenorhabditis elegans. Curr. Biol.
9, 493– 496.
8. Branicky, R., et al., 2000. clk-1, mitochondria, and physiological rates. Bioessays 22, 48– 56.
9. Duan, W., Mattson, M.P., 1999. Dietary restriction and 2-deoxyglucose administration improve behavioural outcome and reduce degeneration of dopaminergic neurons in models of Parkinson’s disease. J. Neurosci. Res. 57,
185 – 206.
10. Lee, J., et al., 2000. Dietary restriction increases survival of newly-generated
neural cells and induces BDNF expression in the dentate gyrus of rats. J. Mol.
Neurosci. 15, 105– 114.

Appendix D. Stress resistance and ROS
The life span of an organism likely depends on a balance between internal and
external damaging conditions and the capacity of the organism for resistance and
repair. This hypothesis is being pursued with the nematode C. elegans, whose
resistance to a number of external and intracellular stresses such as excessive
temperature and reactive oxygen species (ROS) can be tested (for review see
Lithgow1 and Ishii and Hartman2).
Mutants with an extended life span such as daf-2 and age-1 are resistant to
thermal stress3. daf-2 and age-1 function in the same signal transduction pathway
that regulates dauer larva formation (Appendix B). In particular, age-1 encodes a
homologue of phosphatidylinositol-3-OH kinase (PI-3K). Interestingly, a synthetic
inhibitor of PI-3K activity can increase thermotolerance in wild-type animals, as
well as their mean life span by 10%4. Also, induction of thermotolerance by a
transient temperature shift at the adult stage (3 h at 35°C) extends life span in
wild-type worms by 14%, but by 30% in age-1 mutants3.
More generally, if a common battery of genes is implicated in the response to
multiple stresses, one would expect that sensitization by a particular stress could
render the animal more resistant to another type of insult. Indeed, wild-type worms
that have experienced an oxidative stress display increased X-ray resistance5.
Molecular chaperones of the hsp family and superoxide dismutase (sod) genes could
be responsible for stress resistance, but the exact molecular mechanisms of this
adaptive response remain elusive.
The activites of two enzymes, superoxide dismutase (SOD) and catalase, protect
the cell from oxidative damage. SOD converts superoxide into hydrogen peroxide

590

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

and water, and catalase in turn produces oxygen and water from hydrogen
peroxide. The activity of SOD certainly contributes to the dauer stress resistance
traits6, where it is increased 5-fold compared to 2 day old adults7. A positive role
of SOD in lengthening life span has been proposed in age-1 7, and daf-2 worms,
where the expression of sod-3 is upregulated and correlated with the DAF-2
phenotype8. It is interesting to note that overexpression of SOD can be sufficient to
increase life span. For example, the overexpression of sod-1 in motorneurons in
Drosophila prolongs life span by 40%9. The role of catalase has also been explored.
A mutant locus that reduces catalase activity (ctl-1 ) suppresses the long life of both
daf-2 and age-1 10 (Appendix B). An increase in catalase activity and expression of
a cytosolic catalase with age has been reported for the age-1 mutant7,10 but is has
not been described for daf-2 mutants. However, contradictory evidence exists for a
role of catalase in dauer larvae resistance traits. Indeed, as compared to young
wild-type adults, daf-2 dauer larvae exhibit an increase in catalase activity10, while
wild-type and age-1 mutant dauer larvae do not7.
Small molecules (salen– manganese complexes) acting as catalytic scavengers of
ROS, have also been used to try to detoxify ROS11. Remarkably, the life span of
wild-type worms is extended upon treatment with such molecules12. However, due
to the high concentrations used in these experiments and the absence of dose-dependency, it is difficult to conclude that they act only by detoxifying ROS. Also, the
relationship between life span and the products of oxygen-dependent metabolism
has been further studied by the analysis of me6-1 mutants13. Mutations in the gene
me6-1, which encodes a mitochondrial enzyme (succinate dehydrogenase cytochrome b540), affect electron transport from succinate to ubiquinone. The mean
life span of me6-1 (kn1 ) mutants is shortened to 8.5 days at 25°C in the presence of
increased oxygen levels (60%), as compared to 13.0 days for the wild-type.
Consistent with this observation, me6-1 mutants are also hypersensitive to
paraquat, a compound that generates oxygen radicals13. These phenotypes are
probably caused by an oxygen concentration-dependent intracellular increase in
ROS due to the mutation.
Finally, signal transduction proteins have also been implicated in stress response,
as illustrated by the discovery of methuselah in Drosophila, a gene encoding a
G-protein-coupled receptor14, and by studies on tkr-1 in worms15. In Drosophila, a
partial loss-of-function mutation in methuselah causes a 35% increase in life span as
well as resistance to starvation, high temperature and ROS. In worms, the putative
receptor tyrosine kinase gene tkr-1 lengthens life span by 65% when overexpressed,
and induces resistance to heat and UV treatment15. The cascades of events triggered
by these receptors remain to be determined.
D.1. References (Appendix D)

1. Lithgow, G.J., 2000. Stress response and aging in C. elegans. In: Hekimi, S.,
(Ed.), The Molecular Genetics of Aging, Springer Verlag, Berlin, pp. 81–112.

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

591

2. Ishii, N., Hartman, P.S., 2000. Oxidative stress and aging in Caenorhabditis
elegans. In: Hekimi, S., (Ed.), The Molecular Genetics of Aging, Springer
Verlag, Berlin, pp. 149– 164.
3. Lithgow, G.J., et al., 1995. Thermotolerance and extended life-span conferred
by single-gene mutations and induced by thermal stress. Proc. Natl. Acad. Sci.
USA 92, 7540– 7544.
4. Babar, P., et al., 1999. P13-kinase inhibition induces dauer formation, thermotolerance and longevity in C. elegans. Neurobiol. Aging 20, 513– 519.
5. Yanase, S., et al., 1999. Oxidative stress pretreatment increases the X-radiation
resistance of the nematode Caenorhabditis elegans. Mutation Research 426,
31 – 39.
6. Klass, M., Hirsh, D., 1976. Non-ageing developmental variant of Caenorhabditis elegans. Nature 260, 523–525.
7. Larsen, P.L., 1993. Aging and resistance to oxidative damage in Caenorhabditis
elegans. Proc. Natl. Acad. Sci. USA. 90, 8905–8909.
8. Honda, Y., Honda, S., 1999. The daf-2 gene network for longevity regulates
oxidative stress resistance and Mn-superoxide dismutase gene expression in
Caenorhabditis elegans. FASEB J. 13, 1385–1393.
9. Parkes, T.L., et al., 1998. Extension of Drosophila lifespan by overexpression of
human SOD1 in motorneurons. Nature Genetics 19, 171–174.
10. Taub, J., et al., 1999. A cytosolic catalase is needed to extend adult lifespan in
C. elegans daf-C and clk-1 mutants. Nature 399, 162–166.
11. Melov, S., 2000. Mitochondrial oxidative stress. Physiologic consequences and
potential for a role in aging. Annals of the New York Academy of Sciences
908, 219– 225.
12. Melov, S., et al., 2000. Extension of life-span with superoxide dismutase/catalas mimetics. Science 289, 1567–1569.
13. Ishii, N., et al., 1998. A mutation in succinate dehydrogenase cytochrome b
causes oxidative stress and ageing in nematodes. Nature 394, 694–697.
14. Lin, Y.J., et al., 1998. Extended life-span and stress resistance in the Drosophila
mutant methuselah. Science 282, 943–946.
15. Murakami, S., Johnson, T.E., 1998. Life extension and stress resistance in
Caenorhabditis elegans modulated by the tkr-1 gene. Curr. Biol. 8, 1091–1094.

Appendix E. Aging and metabolism
The dauer larval stage of C. elegans is a developmentally arrested, alternative
third larval stage that is entered under unfavorable environmental conditions.
Dauer larvae live much longer than adults, and mutations in the genetic pathway
that regulates dauer larva formation affect the aging process (Appendix B). These
observations have prompted investigations into possible physiological and biochemical particularities of dauer larvae and dauer formation mutants.
The transition from egg to adult is normally accompanied by a switch in
metabolism from the glyoxylate cycle to the tricarboxylic acid (TCA) cycle1. This

592

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

occurs relatively rapidly during the L2 larval stage and the major fuel source
consequently changes from fat reserves to ingested matter. In dauer larvae there
is a marked (10-fold) reduction in flux through the TCA cycle relative to young
adults2. A concomitant reduction in oxygen consumption3 (per dry weight), ATP
levels1, and mitochondrial oxidative capacity (3-fold) relative to young adults2
has also been observed. The contribution of the glyoxylate cycle to dauer
metabolism remains debated. Two reports have suggested that the key glycolytic
enzymes malate synthase and isocitrate lyase are decreased1,4, while another has
observed a 3-fold increase of both components5. Intuitively one imagines that the
dauer must subsist off its stored energy reserves, but whether catabolic routes
other than the glyoxylate cycle are utilized to do this remains unclear.
Many genes that are normally ubiquitously expressed during other larval
stages (such as actin, histones and ubiquitin), are heavily or completely shut
down in the dauer larvae6. Conversely, key survival genes such as RNA polymerase (Rpo) I and III, which transcribe rRNAs and tRNAs, as well as genes
that encode stress-resistance proteins such as hsp-90, sod-3 and ctl-1, are maintained at high levels6. An intriguing finding has been the observation that in
dauers, Rpo II activity, which normally transcribes mRNA, is depressed despite
the presence of normal Rpo II protein levels. Heat shock rapidly overrides this
block to induce a dramatic increase in stress-response mRNAs for genes such as
hsp-16 and hsp-70 6.
daf-2 and age-1 mutants are defective in their ability to properly regulate
dauer development and live a long time (Appendix B). A logical prediction
might be that the increased life span observed in mutant adults may result from
the expression of dauer-like metabolic properties. At least one study has observed elevated levels of malate synthase and isocitrate lyase in age-1 and daf-2
animals5. Conversely, levels of the TCA enzyme isocitrate dehydrogenase are not
decreased in these animals and the levels of this enzyme did not decrease with
age as it does in a control population5. Other studies have also found no
significant decreases in the O2 consumption levels of age-1, or indeed for any of
the clk mutants7,8 (Appendix A). Additional studies employing CO2 production
as a measure of metabolism corroborate the finding for age-1 9. In the latter
study, however, CO2 production rates were found to be significantly reduced
(4-fold) in both a daf-2 (e1370 ) mutant and a daf-2 (e1370 ) clk-1 (e2519 ) double
mutant. This is surprising because studies have placed age-1 immediately downstream of daf-2 (Appendix B), and the effect of these genes on life span or
dauer formation are not additive (see main text). Thus, if the rate of oxidative
metabolism of daf-2 mutants is decreased, this is not causal to their long life.
An intriguing observation has suggested that both a clk-1 (e2519 ) and, even
more so, a clk-1 (e2519 ) daf-2 (e1370 ) double mutant have markedly elevated
levels of total ATP relative to age-matched controls7. Given the hypometabolic
phenotype of clk mutants (slow growth and slow reproductive rates), one possibility is that ATP production is intact in these animals but that ATP is used at
a lower rate.

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

593

E.1. References (Appendix E)

1. Wadsworth, W.G., Riddle, D.L., 1989. Developmental regulation of energy
metabolism in Caenorhabditis elegans. Dev. Biol. 132, 167–173.
2. O’Riordan, V.B., Burnell, A.M., 1989. Intermediary metabolism in the dauer
larva of the nematode Caenorhabditis elegans-I. Glycolysis, gluconeogenesis,
oxidative phosphorylation and tricarboxylic acid cycle. Comp. Biochem. &
Physiol. 92B, 233– 238.
3. Anderson, G.L., 1982. Superoxide dismutase activity in dauer larvae of
Caenorhabditis elegans (Nematoda: Rhabditidae). Can. J. Zool. 60, 288–291.
4. O’Riordan, V.B., Burnell, A.M., 1990. Intermediary metabolism in the dauer
larva of the nematode Caenorhabditis elegans-II. The glyoxalate cycle and
fatty-acid oxidation. Comp. Biochem. & Physiol. 95B, 125–130.
5. Vanfleteren, J.R., De Vreese, A., 1995. The gerontogenes age-1 and daf-2
determine metabolic rate potential in aging Caenorhabditis elegans. FASEB J. 9,
1355 – 1361.
6. Dalley, B.K., Golomb, M., 1992. Gene expression in the Caenorhabditis elegans
dauer larva: developmental regulation of Hsp90 and other genes. Dev. Biol. 151,
80 – 90.
7. Braeckman, B.P., et al., 1999. Apparent uncoupling of energy production and
consumption in long-lived Clk mutants of Caenorhabditis elegans. Curr. Biol. 9,
493 – 496.
8. Vanfleteren, J.R., De Vreese, A., 1996. Rate of aerobic metabolism and superoxide production rate potential in the nematode Caenorhabditis elegans. J. Exp.
Zool. 274, 93– 100.
9. Van Voorhies, W.A., Ward, S., 1999. Genetic and environmental conditions
that increase longevity in Caenorhabditis elegans decrease metabolic rate. Proc.
Natl. Acad. Sci. USA 96, 11 399–11 403.

References
Avery, L., 1993. The genetics of feeding in Caenorhabditis elegans. Genetics 133, 897 – 917.
Braeckman, B.P., et al., 1999. Apparent uncoupling of energy production and consumption in long-lived
Clk mutants of Caenorhabditis elegans. Curr. Biol. 9, 493 – 496.
Dorman, J.B., et al., 1995. The age-1 and daf-2 genes function in a common pathway to control the
lifespan of Caenorhabditis elegans. Genetics 141, 1399 – 1406.
Gems, D., et al., 1998. Two pleiotropic classes of daf-2 mutation affect larval arrest, adult behavior,
reproduction and longevity in Caenorhabditis elegans. Genetics 150, 129 – 155.
Gottlieb, S., Ruvkun, G., 1994. daf-2, daf-16 and daf-23 : genetically interacting genes controlling Dauer
formation in Caenorhabditis elegans. Genetics 137, 107 – 120.
Hekimi, S., et al., 1995. Viable maternal-effect mutations that affect the development of the nematode
Caenorhabditis elegans. Genetics 141, 1351 – 1364.
Honda, Y., Honda, S., 1999. The daf-2 gene network for longevity regulates oxidative stress resistance
and Mn-superoxide dismutase gene expression in Caenorhabditis elegans. FASEB J. 13, 1385 – 1393.
Kenyon, C., et al., 1993. A C. elegans mutant that lives twice as long as wild type. Nature 366, 461 – 464.

594

S. Hekimi et al. / Mechanisms of Ageing and De6elopment 122 (2001) 571–594

Lakowski, B., Hekimi, S., 1996. Determination of life-span in Caenorhabditis elegans by four clock
genes. Science 272, 1010 –1013.
Lakowski, B., Hekimi, S., 1998. The genetics of caloric restriction in Caenorhabditis elegans. Proc. Natl.
Acad. Sci. USA 95, 13 091 –13 096.
Larsen, P.L., 1993. Aging and resistance to oxidative damage in Caenorhabditis elegans. Proc. Natl.
Acad. Sci. USA 90, 8905 –8909.
Larsen, P.L., et al., 1995. Genes that regulate both development and longevity in Caenorhabditis elegans.
Genetics 139, 1567 –1583.
Taub, J., et al., 1999. A cytosolic catalase is needed to extend adult lifespan in C. elegans daf-C and
clk-1 mutants. Nature 399, 162 –166.
Wong, A., Boutis, P., Hekimi, S., 1995. Mutations in the clk-1 gene of Caenorhabditis elegans affect
developmental and behavioral timing. Genetics 139, 1247 – 1259.

.