National Libra'Y Bibiiothrque nationale 1.1 of Canada du Canada Acquisitions and Direction des acquisitions et Bibliographie Services Branch des services bibliographiques 395 Wellington Street 395. rue Wellinglon Ottawa. Ontario Ottawa (Ontario) K1AON4 K1AON4
NOTICE AVIS
The quality of this microform is La qualité de cette microforme heavily dependent upon the dépend grandement de la qualité quality of the original thesis de la thèse soumise au submitted for microfilming. microfilmage. Nous avons tout Every effort has been made to fait pour assurer une qualité ensure the highest quality of supérieure de reproduction. reproduction possible.
If pages are missing, contact the S'il manque des pages, veuillez university which granted the communiquer avec l'université degree. qui a conféré le grade.
Sorne pages may have indistinct La qualité d'impression de print especially if the original certaines pages peut laisser à . pages were typed with a poor désirer, surtout si les pages typewriter ribbon or if the originales ont été university sent us an inferior dactylographiées à l'aide d'un photocopy. ruban usé ou si l'université nous a fait parvenir une photocopie de qualité inférieure.
Reproduction in full or in part of La reproduction, même partielle, this mic;roform is governed by de cette m!croforme est soumise the Cflnadian Copyright Act, à la Loi canadienne sur le droit R.S.C. 1970, c. C-30, and d'auteur, SRC 1970, c. C-30, et subsequent amendments. ses amandements subséquents.
Canada •
Mutations in the clk-l gene of Caenorhabditis elegans affect developmental and behavioural timing
AnneWong
Department of Biology
McGill University
Montreal, Canada
November 1994
A thesis submitted to the Faculty of Graduate Studies and Research
in partial fulfillment of the requirements for the degree of
Master of Science
© Anne Wong 1994 • National Library Bibliothèque nationale 1+1 of Canada du Canada Acquisitions and Direction des acquisitions et Bibliographie Services Branch des services bibliographiques 395 Wellinglon Street 395. rue Wellington Ottawa, Ontario Ottawa (Ontario) K1AON4 K1AON4 YOUf 1IIe Vol'er f~M'OflC(f
Our flle NoUIl Idfd/MCU
THE AUTHOR HAS GRANTED AN L'AUTEUR A ACCORDE UNE LICENCE IRREVOCABLE NON-EXCLUSIVE IRREVOCABLE ET NON EXCLUSIVE LICENCE ALLOWING THE NATIONAL PERMETTANT A LA BIBLIOTHEQUE LIBRARY OF CANADA TO NATIONALE DU CANADA DE REPRODUCE, LOAN, DlSTRIBUTE OR REPRODUIRE, PRETER, DISTRIBUER SELL COPIES OF mSIHER THESIS BY OU VENDRE DES COPIES DE SA ANY MEANS AND IN ANY FORM OR THESE DE QUELQUE MANIERE ET FORMAT, MAKING THIS THESIS SOUS QUELQUE FORME QUE CE SOIT AVAILABLE TO INTERESTED POUR METTRE DES EXEMPLAIRES DE PERSONS. CETTE THESE A LA DISPOSITION DES PERSONNE INTERESSEES.
THE AUTHOR RETAlNS OWNERSHIP L'AUTEUR CONSERVE LA PROPRIETE OF THE COPYRIGHT IN HISIHER DU DROIT D'AUTEUR QUI PROTEGE THESIS. NEITHER THE THESIS NOR SA THESE. NI LA THESE NI DES SUBSTANTIAL EXTRACTS FROM IT EXTRAITS SUBSTANTIELS DE CELLE MAY BE PRINTED OR OTHERWISE CI NE DOIVENT ETRE IMPRIMES OU REPRODUCED WITHOUT HISIHER AUTREMENT REPRODUITS SANS SON PERMISSION. AUTORISATION.
ISBN 0-612-05652-X
Canad~ • Abstract
Five allelie, matemal-effect mutations which affect developmental and
behavioral timing in Caenorhabditis e/egans have bcen identified. They result in a
mean lengthening of embryonic and post-embryonic development, the ccli cycle period,
and life span, as weil as the periods of the defecation, swimming, and pumping cycles.
These mutants also display a number of additional phenotypes I\::iated to timing. For
example, the variability in the length of embryonic development is several times larger
in the mutants than in the wild-type, resultir.g in the occasional production of mutant
embryos developing more rapidly than the most rapidly-developing wild-type embryos.
In addition, the duration of embryonic development and the length of the defecation
cycle of the mutants, but not of the wild-type, depends on the temperature at which
their parents were raised. Finally, individual variations in the severity of distinct
mutant phenotypes are correlated in a counter-intuitive way. For example, the animais
with the shortest embryonic development have the longest defecation cycle and those
with the longest embryonic development have the shortest defecation cycle. Most of
the features affected by these mutations are believed to be controlled by biological
docks, and we therefore cali the gene defined by these mutations clk-l, for "abnormal
function of biological docks".
• iii
• Résumé
Cinq mutations alléliques à effet maternel, affectant le déroulement temporel
du développement et du comportement chez le nématode CaellorlUlbditis elegmls, ont
été identifiées. Les durées moyennes du développement embryonmlire ct post
embryonnaire, du cycle cellulaire, ainsi que les périodes du cycle de défécation, de nage
et de pompage, ,;ont allongées par rapport à celles de animaux de type sauvage. La
durée de vie de ces animaux est aussi prolongée. De plus, ces mutants présentent
d'autres particularités liées au contrôle temporel. Par exemple, la durée de leur
développement embryonnaire est beaucoup plus variable qu'elle ne l'est chez les
animaux de type sauvage, de telle sorte que certains embryons mutants se développent
plus rapidement que les plus rapides des embryons de type sauvage. Aussi, la durée du
développement embryonnaire et de la période du cycle de défécation des mutants
dépende de la température à laquelle ceux-ci se se développent, ce qui n'est pas
observé chez les animaux de type sauvage. Enfin, les différents phénotypes montrent
une corrélation surprenante de leur sévérité lorsqu'ils sont mesurés chez un individu
patticulier. En effet, les animaux au développement embryonnaire le plus rapide,
montrent les cycles de défécation les plus longs; les animaux au développement le plus
long ont les cycles les plus courts. Le gène défini par ces mutations a été nommé dk-/,
pour suggérer un fonctionnnement anormal des horloges biologiques chez ces mutants
(du mot anglais dock pour horloge).
• iv
• Table of Contents
Abstraet ii
Résumé iii
Table of Contents iv
List of Figures vi
List ofTables vii
Aeknowledgements viii
Introduction 1
Materials and Methods 14
Materials 14
General methods and strains 14
Assays for developmental phenotypes 14
Assays for behavioral phenotypes 16
Effeets of temperature 17
Correlation assays 18
Results 19
A genetie sereen for matemal-effeet mutations affeeting development
and behaviour 19
clk- J mutations are pleiotropie 22
clk- J mutations ean be maternally reseued 34
Strict maternai effects on the length of embryogenesis 36
The length ofembryonic development is highly variable in clk-J mutants .42 • The cell cycle period is affected in clk- J mutants .47 v
Individual variations in the severity of clk-/ mutant phenotypes are • correlated in a non-intuitive way .48 Effects of temperature on the duration of embryonic development 51
Effects of temperature on the DMP 59
Discussion 63
References 68
Appendices 77
Appendix 1. Heritability of the rate of embryogenesis 77
Appendix 2. Egg laying rate 80
Appendix 3a. Self-brood size 82
Appendix 3b. Cross-brood size 82
Appendix 4. Defecation 84
Appendix 5. Duration of DMP 88 • Appendix 6. Pumping 90 Appendix 7. Swimming 92
Appendix 8. Temperature effects on defecation 94
Appendix 9a. Phenotypes offast embryos 97
Appendix 9b. Phenotypes of slow embryos 97
• vi
• List of Figures
Figure 1. Post-embryonic development 26
Figure 2. Life span 30
Figure 3. Defecation cycle 32
Figure 4. Duration of embryonic development 43
Figure 5. Quickly developing embryos 45
Figure 6. Effects of temperature on embryonic development 52
Figure 7. Effects of temperature shifts during development 57
• vii
• List of Tables
Table 1. Quantitative phenotypic analysis 23
Table 2. MaternaI, paternal and zygotic rescue of embryonic dcvcloplllcni 37
Table 3. Embryonic development of trans-heterozygotes 40
Table 4. Correlated variations in the severity of the clk-/ phenotypes 49
Table 5. Effects of temperature differences and shifts on embryogencsis 54
Table 6. Effects of temperature on defecation 60
• viii
• Acknowledgements
1 would first like to thank my supervisor, Siegfried Hekimi, for insight and
inspiration. My very sincere thanks to Tom Barnes and Jonathan Ewbank for their
expert and critical advice. Paula Boutis and Bernard Lakowski provided strains and
mapping data. 1 also thank Claire Bénard for the excellent translation of my abstract.
For my instruction in the ways of this world, 1 am grateful to Siegfried Hekimi
and Tom Barnes. 1 am indebted to Barbara Young and Walter Strapps for their
support and unconditional love.
• • Introduction Many aspects of Iife appear to be preciscly timed. The processes whieh
regulate timing in living organisms can be biological c1ocks. Most organisms appcar 10
have a circadian c1ock, an internalmechanism which allows themlo anticipalc. and
thus to keep in register with, the phenomena of the exlernal world, such as day and
night or the seasons. Numerous observations strongly supporl the existence or inlernal
circadian c1ocks. Among the most important is the exislence or free-running rhythms,
for example in animal behavioral activity patterns. Vertebrales and invertebrales show
peaks of locomotory activity at particular times of the day or night, that are still present
in the absence of external stimuli, for example when animais are kept in constanl
darkness or constant Iight.
Another insight into the nature of circadian c10ck mechanisms has come from
genetics (Dunlap 1993). In a number of organisms including fruit tlies (Konopka and
Benzer 1971), fungi (Feldman and Hoyle 1973), the hamster (Ralph and Menaker
1988) and the mouse (Vitaterna et al. 1994), mutations have been isolated which alter
the period of the free-running rhythms. In the fruit f1y Drosophi/a melal/ogaster, the
per/od (pel') locus is plays a fundamental role in the maintenance of circadian (-24
hours), as weil as ultradian, behavioral rhythms. Although the gene is expressed in a
varietyof tissues throughout development, only its function in the central nervous
system appears to be required for rhythmic behaviour. The locus encodes a protein of
1224 amino acids and distinct single amino acid substitutions lead to struclural changes
in the protein that can either shorten or lengthen the period of circadian behavioral
rhythms. For example, perS, an allele ofpel', shortens the period to 19 hours while the
perl mutation lengthens the period to 28 hours (Konopka and Benzer 1971). Also, • perS shortens the male f1y's courtship song cycle from 55 seconds to 40 seconds, while 2
• perllengthens it to 80 seconds. Flies that have the perO mutation have no detectable circadian rhythms and produce a truncated, apparently functionless protein (Baylies et
al. 1987, Yu et al. 1987b).
Gene dosage studies indicate that period length is determined by abundance of
the PER protein. The abundance of the per RNA and PER protein cycle with a
circadian periodicity (Hardin et al. 1990, Zerr et al. 1990) and there is negative
feedback of the PER protein on the abundance of the per RNA (Hardin et al. 1990).
Flies expressing lower levels ofper RNA produce circadian rhythms with
correspondingly longer periods (Baylies et al. 1987). As perl and pers mutants express
wild-type levels ofpel' RNA (Hardin et al. 1990), it has been suggested that the
mutants produce hypoactive and hyperactive PER proteins, respectively (Baylies et al.
1987). In addition, PER possesses a dimerization sequence motif of -270 amino acids
(termed PAS) which is also present in a number of basic-helix-Ioop-helix (bHLH)
transcription factors (Huang et al. 1993). This suggests that the activity of the PER
protein may regulate circadian gene transcription by interacting with one or more
bHLH-PAS proteins through a PAS-mediated protein-protein interaction. Such a
mechanism could underlie the feedback loop involving PER and the transcription of its
own messenger RNA (Huang et al. 1993).
Clock components have also been investigated in genetically and biochemically
tractable microbial systems such as Nellrospora crassa, a filamentous fungus in which
the dock regulates a clearly expressed overt rhythm in developmental potential and
morphology (Dunlap 1993). This rhythmicity is manifest through the cyclical
production of asexual spores called conidia, with peaks in spore production once every
24 hours. Genetic approaches have identified seven loci that comprise the Neurospora • clock and one of these,freqllency (frq), has been studied extensive1y. Thefrq gene is a 3
• likely candidate for encoding a central component of the oscillator as the ninc differenl frq alleles (jrqI-frq9) are altert:d differently in period Icngth and temperature
compensation. For example, the conidiation rhythm infr(/ is lengthened to 29 hours
andfrq9 is a recessive loss-of-function ailele that displays reduced conidialion and no
stable circadian rhythmicity (Loros and Feldman 1986). Ali nine jiYJ alleles are caused
by single amino acid changes, and it has been shown that these changes are rcsponsible
for the period alterations in the mutants (Feldman and Hoyle 1973, Gardner and
Feldman 1980). The amount of bothfrq transcripts and FRQ protein oscillates and it
has been shown that FRQ can affect its own transcript level (Gardner and Feldman
1980) and possibly regulate the expression of ether genes as weil. Although FRQ and
PER appear to be functionally homologous, it is important to note that FRQ is a
different protein l'rom PER. Indeed, there is no eVldence to suggest that they are
evolutionarily related at the structurallevel (McClung et al. 1989).
Circadian c10cks are responsible for the activation of behaviors at particular
times during the day-night cycle. Many behaviors are also periodic in a complclcly
different sense: as reiterated activation of a particular motor program. Walking,
breathing, gut peristalsis, heart rate and sleep patterns are only a few familial' cxamplcs
among the large number of such ultradian rhythms. Some of these rhythms arc vcry
regular, and the mechanisms underlying these stable oscillations are being investigatcd
at numerous levels (Glass and Mackey 1988, Friesen et al. 1993). For thc most
complex motor patterns, it is generally believed that neuronal c1ocks, also called ccntral
pattern generators, drive these rhythms (Cohen et ai. 1988). A central pattern
generator can be visualized as a neuronal network with an oscillatory output. The
neuronal bases of pattern generators have generally been investigated at the • physiologicalleveI. Insights at the genetic level have come l'rom studies on the cffects 4
• of per mutations on the Drosophila courtship song (Kyriacou and Hall 1980) and more rccently, from studies on the defecation cycle in the nematode Caenorhabditis elegans
(Thomas 1990, Liu and Thomas 1994).
C. elegans is a small (one millimetre long), free-living soil nematode. Il is
casily maintained in the laboratory, feeding on bacteria. There are two sexes,
hermaphrodites and males. Hermaphrodites produce both sperm and oocytes and
reproduce by internai self-fertilization. Males arise spontaneously at a low frequency
and they can mate and fertilize hermaphrodites. Hermaphrodites lay eggs which hatch
to give larval stage one (LI) worms. These LI worms moult through three further
larval stages (L2, L3, L4), and finally, into mature adult worms. C. elegans is a simple
organism: the adult hermaphrodite has only 959 somatic cells and the lineages of ail the
ceIls are known (Sulston et al. 1983). In addition, there is a large collection of mutants
that affect development and behaviour in the worm. The simplicity, transparency, ease • of cultivation, short life cycle (three and a half days), suitability for genetic analysis, small genome size and a wealth of pre-existing knowledge ail make C. elegans a useful
model system for the study of development and behaviour (Wood 1988).
One weil understood behaviour of C. elegans is its defecation cycle which
consists of the successive contraction of three distinct sets of muscles, repeated at
regular intervals and resulting in the expulsion of gut contents; this cycle is also
commonly referred to as the defecation motor program (DMP) (Thomas 1990). Each
cycle begins with the contraction of the posterior body muscles (referred to as pBoc,
for posterior body contraction) causing the contents of the intestinal lumen to be
squeezed anteriorly. Next, these muscles relax, causing the intestinal contents to
accumulate in the preanal region. One second later, the body muscles near the head • contract (referred to as aBoc, for anterior body contraction), driving the pharynx back 5
• into the anterior intestine, thereby pressurizing the intestinal contents. Meanwhile. the two intestinal muscles contract to further pressur'ze the intestinal contents. and the
single anal depressor muscle contracts to open the anus for expulsion (rcferred to ilS
Exp, for expulsion). This entire sequence of muscle contractions takcs about t'ive
seconds. In wild:type animaIs the DMP is activated just once for each dcfecation
cycle.
A numbcr of mutations which can alter the period of this cycle havc now been
identified (D. Liu and J. Thomas, personal communication) and as for mutations in
circadian clock genes, they include mutations which lengthen, as wcll as shorten, thc
defecation cycle period. Many alleles of cha-/ (abnormal choline lIcetyltransferase)
and dec-/ (sa48) (defecation cycle defective) have a lengthened defecation cycle period
while aex-I (sa9) (aBoc and Exp defective) and aex-4 (sa22) have shortened periods.
Evidence suggests that the DMP rhythms are the product of an endogenous
pattern generator (Liu and Thomas 1994). In the presence of food, pumping and
defecation rates are very regular, but these rhythms are interrupted when worms are
away from food. When a worm leaves the food, it stops pharyngeal pumping after
several seconds and also does not activate the DMP. When they return to the bacterial
lawn, they resume feeding and defecating in phase with their previously established
rhythm (Liu and Thomas 1994). In addition, mutations that affect the motor program
do not alter the rhythm of the behaviour; the motor steps themselves are defectivc but
not the timing of their activation (Thomas 1990). Further evidence for an endogenous
clock is suggested by the ability of the DMP rhythm to be reset by mechanoscnsory
stimulation. When touched lightly with an eyelash, worms respond by moving away
from the source ofthe stimulation. At the same time, the DMP is reset to time zero • 6
• regardless of when the stimulation is applied with respect to the DMP activations (Thomas 1990, Liu and Thomas 1994).
Another rhythmic behaviour found in C. elegans is the pharyngeal pumping
responsible for food intake. The pharynx is a simple, self-contained neuromuscular
pump consisting of 80 cells (Albertson and Thomson 1976, Sulston et al. 1983) of
which 20 are neurons which regulate feeding (Avery and Horvitz 1987, 1939). The
pharynx is a bilobed structure made up of an anterior bulb connected to a terminal bulb
via the isthmus, a thin muscular tube. Bacteria pass through the anterior bulb and are
ground in the terminal bulb before entering the intestine thraugh the pharyngeal
intestinal valve. Fifty-two mutations that affect the pharynx have been found, defining
35 genes (Avery 1993a). Like defecation mutants, there are both mutations that slow
down and speed up the rate of the pharyngeal pumping. When ail 20 neurons of the
pharynx are ablated with a laser microbeam, the pharynx still pumps due to the inherent
myogenic activity ofthe pharyngeal muscles but the pumping is weak and irregular
(Avery and Horvitz 1989). The motor neuron M3 and the sensory neuron 15 influence
the timing of pharyngeal muscle motions: when M3 is ablated, the pump duration
increases and when 15 is ablated, the pump duration decreases (Avery 1993b). The
length of the pumping cycle appears therefore to be controlled by specifie neurons.
The ontogeny of multicellular organisms also requires mechanisms which
ensure that this extraordinarily complex sequence ofevents is correctly coordinated.
Indeed, much effort in the field of developmental biology is directed at determining the
spatial and temporal patterns of activation of genes defined by mutations affecting
development. In C. elegans, this has inc1uded the isolation of heterachronic mutations
which alter the succession of larval stage-specifie developmental events (Ambras • 1989). The proper timing of many postembryonic developmental events requires the 7
• normal activity of at least four heterochronic genes, the abnol111al celllilicage gencs, lin-4, lin-l4, lin-28 and lin-29 (Ambros and Horvitz 1984). Mutations in thesc gencs
result in either precocious development, where certain cclls exprcss fates n0I111ally
specific for cells later in the same Iineage, or in retarded developmcnt, wherc cells
reiterate fates normally specific for cells earlier in the Sllme lineage. The timing of a
particular stage-specific event in the development of the lateral hypodermis. the "Iarva
to-adult switch", has been studied with respect to these heterochronic gcnes. This
switch involves several coordinate changes in the behaviour of certain hypodcrmal cells
at the fourth (arval molt: cessation ofcell division by the seam cells. cells Ihat l'un along
the lateral hypodermis, formation of adult (instead of larval) cuticle, cell fusion of the
seam ceIls and cessation of the molling cycle. There appears 10 be a hierarchy of
regulatory interactions among the four lin genes to control the timing of this swilch
(Ambros 1989). The simplest model, as proposed by Ambros, is Ihat laIerai
hypodermal celIs have the potential to switch from larval to adult programs al
successive stages of development. The execution of that switch critically depends on
the activity of lin-29, the proposed trigger of the switch. lin-/4 and lin-28 inhibillin
29 at early stages of developmenl and thereby prevent premalure swilching. Latcr, lin
4 inhibits lin-l4 and lin-28, resulting in the activation of lin-29. In spile of the slriking
regularity in the timing of developmental events, including the total length of
development, no evidence for a general mechanism which would measure
developmental time has yet been presented (Gilbert 1991). In the absence of such a
developmental cIock, developmental timing would have to be seen as resulting enlirely
from the sheer precision of cascades of inductive interactions.
In muiticellular organisms, development is accompanied by, and largely based • on, an increase in the number of cells composing the organism. Thé length of Ihe ccII • cycle could therefore play a central role in determining the outcome of many developrnental events. Indeed, in sorne instances regulation of the cell cycle period
appears to be used as an effector of developmentaltiming. For example, in many
organisms a sudden lengthening of the early embryonic cell cycles is concomitant with
a number of developmental events at the so-called midblastula transition (Newport and
Kirschner 1982a, 1982b). The midblastula transition (MBT) in the Xenopus embryo is
a change in cell cleavage from rapid synchronous cleavages to slower asynchronous
divisions. At the MBT in Xenopus, the blastomeres become transcriptionally active for
the firsttime and the timing ofthis event depends on the embryo re!'.ching a critical
ratio of nucleus to cytoplasm (Newport and Kirschner 1982a, Kimelman et al. 1987).
The steadily increasing nuclear-to-cytoplasmic ratio during embryonic cleavage which
triggers the MBT is also seen in Drosophila development (Edgar et al. 1986). Strong
evidence suggests that the embryo knows that it has reached the critical nucleus-to
cytoplasm ratio by sensing the dep1etion of sorne component that is exhausted by
stoichiometric interaction with the increasing DNA content. In Xenopus, Newport and
Kirschner (1982a) have suggested that this factor is present in the unfertilized eggs and
is titrated by a criticallevel ofDNA. For Drosophila, Edgar et al. (1986) have
suggested thatthe hypothetical factors that are being titrated may be involved in the
formation of the nuclei and that when these factors become limiting, the cell cycle must
slow down and await the synthesis of new components. Regulation of the duration of
the cell cycle can clearly be used as a type of developmental clock.
Cell division plays a central role in alllife cycles, and the mechanisms driving
the cell cycle, (sometimes called the cell cycle clock) are being studied in great detail
(Norbury and Nurse 1992). The cell cycle can be considered to be composed offour • phases, the phase before DNA replication (01), the DNA synthetic phase (S), the 9
• phase after DNA replication (G2), and the mitotic phase (M) which culminatcs in ccli division (Darnell et al. 1990). Much attention has focused on the inteructions bctwecn
cellular kinases and phosphatases, especially the so-called "cyclins".
One protein kinase that has been investigated is the maturation-promoting
factor (MPF), a kinase necessary for transition from G2 to M phase (Newport and
Kirschner 1984). It is known that MPF activity is comprised of two protein spceies, a
protein kinase catalytic subunit p34Cdc2 and a B-cyclin protein (Gautier et al. 1988,
Gautier et al. 1990). However, p34Cdc2 was independently identified genetically in
yeast as having a function required for the onset of M phase (Nurse 1975) before it
was known to be a part of the MPF. The activity of p34Cdc2 is itself subjectto
regulation by phosphorylation (Gould and Nurse 1989, Krek and Nigg 1991) and the
predicted ATP-binding region ofthe protein is phosphorylated during interphase
resulting in inactivation of the pre-MPF complex until the onset of M phase (Gould and
Nurse 1989, Norbury et al. 1991). The biochemical role of the B-cyclin component
remains unclear but it is believed to be the regulatory subunit of the MPF (Norbury and
Nurse 1992).
At least sorne cyclins are c1early timing devices; they are unstable moleculcs
which slowly accumulate during various phases of the cell cycle, and their sudden
destruction corresponds to transitior.s from one phase to the next. For example, cyclin
B is first synthesized during S phase, accumulates in complexes with p34Cdc2 as cells
approach the G2 to M transition, and is abruptly degraded during mitosis (reviewed by
Sherr 1993). Cyclin A has a similar periodicity to cyclin B in its synthesis and
degradation, and both cyclins A and B interact with p34Cdc2 (Swenson et al. 1986,
Murray and Kirschner 1989), but cyclin A is synthesized and destroyed before cyclin B • during the cell cycle (Minshull et al. 1990, Pines and Hunter 1990). Most work with 10
• cyclins and associated molecules has focused on their involvement in the mechanisms which cnsure the proper succession of events of the cell cycle rather than the length of
the cycle as a whole (Hartwell and Weinert 1989). It has been shown that cyclins A
and B regulate the onset of Sand M phase (Pines and Hunter 1991, Fang and Newport
1991, Walker and Mailer 1991), but it is cyclins 0 and E that regulate GI phase
progression and S phase commitment (Ohtsubo and Roberts 1993, Ewen et al. 1993).
Recent work with the 0 cyclins, however, is addressing the role of cyclins as peripheral
sensors that integrate cycle clock (Sherr 1993). The final developmental events for most organisms are senescence and death. Research into the determinants of life span is uncovering environmental effects as weil as genetic components contributing to the duration of life. In C. elegans both avenues are being actively pursued (Klass 1977, Decuyper and Vanfleteren 1982, Johnson 1987, Friedman and Johnson 1987, Honda et al. 1993, Vanfleteren 1993). Environmental factors such as temperature and food concentration have been investigated specifically (Klass 1977). An increase in temperature reduces Iife span and a decrease in temperature was seen to increase life span during ail parts of the life cycle (Klass 1977). However, the only statistically significant changes in life span involved manipulating the temperatures during the worm's reproductive phase, suggesting that the reproductive phase is most sensitive to temperature with respect to Iife span. Dietary restriction leads to an increase in life span while overfeeding decreases life span (Klass 1977, Masoro 1985). Reduction of the food supply during any phase of the animal's life cycle led to an increase in life span with the greatest effect seen during the growth phase. • Il • The possibility of an internai mechanism determining life span is being investigated at the genetic level (Friedman and Johnson 1987, Johnson 1987, Kenyon et al. 1993). Klass (1983) reported the isolation of eight long-lived mutant strains of C. elegans and these strains are of interest because the mutation resulting in increased life span segregates as a single recessive gene called age-! (Friedman and Johnson 1987). age-} mutants have significant extensions in both mean and maximumlife span as compared to wild-type animais and this increase in life span is seen in a variety of environments (different media, different temperatures). Development is normal in age } animaIs, with no change in lengths of any larval stages or duration of embryonic development. In addition, there is a significant reduction in hermaphrodite self-fertility, although this decrease is not accompanied by a change in the length of thc reproduelive period of the animaIs, nor do age-} hermaphrodites have reduccd numbers ofspenn (Johnson and Friedman 1988). Lastly, there is no evidence to suggest thal age-! results l'rom mutations in a "dock" as is hypothesized to account for programmed aging (Johnson 1987). Exeept for the effect of lengthening life, lhere is no other evidenee that age-l has any effect on a "rate of aging"; for example, age-l does not play a role in determining the length of the fecund period of life. Il is believed that the lengthening of life results l'rom reduced hermaphrodite self-fertility or l'rom some other unkown metabolie or physiologie alteration and indeed, age-} mutants have been shown to have a redueed metabolie rate (Johnson and Friedman 1988). A different approaeh to studying the genetie basis underlying aging has come l'rom work with spe (defeetive spermatogenesis) genes (Van Voorhies 1992) and daf (dauerformation) genes (Kenyon et al. 1993). It is hypothesized that spermatogenesis deereases life span as mated wild-type males (males that have aetually maled with • hermaphrodites) have a shorter life span than unmated males (Van Voorhies 1992). 12 • The reduction of mated male life span seems to be caused by additional sperm production and not by the physical activity of mating. Furthermore, spe-26 hermaphrodites which are defective in spermatogenesis live longer than wild-type animais (Van Voorhies 1992). The dauer stage in C. elegans is a developmentally arrested larval form that is long-lived and is induced by crowding and starvation. The decision to enter the dauer stage, an alternative third larval stage (L3), must be made during the first larval stage (L 1). Dauer entry and exit are controlled by many genes, among them the dafgenes. daf-2 is a temperature-sensitive, dauer-constitutive mutant; at the non-permissive temperature, these worms become dauers even in uncrowded conditons with plenty of food (Vowels and Thomas 1992). When daf-2 larvae are raised at the permissive temperature until stages past LI, and then left to grow to adulthood, the adult hermaphrodites had a life span that was twice as long as wild-type hermaphrodites (Kenyon et al. 1993). There is a slight decrease in the brood size of daf-2 animaIs but it is not responsible for their longevity as both wild-type hermaphrodites with laser ablated precursors of the germ cells and somatic gonad, as weil asfem-3 (feminization of XX and XO animaIs) animaIs which do not produce sperm or self-progeny, have normallife spans (Kenyon et al. 1993). Il is not yet known how daf-2 mutations extend life span; however, daf- J6 is able to completely suppress the life span extension of daf-2 adult animaIs. daf- J6 is a dauer-defective mutant that functions downstream of daf-2 to promote dauer formation (Vowels and Thomas 1992). Il has been suggested that daf-J6 activity extends the life spans of daf-2 adults by triggering expression of a regulated life span extension mechanism that is normally coupled to dauer formation (Kenyon et al. 1993). This would imply two things: that the long life • • span of dauers is not simply a consequence of being developmentally arrested and that life span extension can be separated l'rom dauer formation. Above, a number of processl.:s II'hich appear to be, 01' cOllld be, llsing timing deviees have been olltlined. These regulatory meehanisms appear to possess very different eharaeteristies and no evidenee points to a cornmon mechanism. Here, however, 1 show that mutations in the gene clk-/ of C. eleg(lIls affect the timing of most of these features. 1 will diseuss how this pattern of phenotypes eould be produeed by mutations in a single gene, and speeulate that clk-/ might be involved in the function of a central biologieal cJoek whieh would enSllre the temporal coordination of ail features of the organism. • 14 • Materials and Methods Materials Nematode Growth Medium was prepared as described by Brenner (1974) using agar From Difco and chemicals From BDH, Sigma and VWR Scientific. Chemicals were analysis grade or better. Petri dishes were obtained From Fisher Scientific and Sarstedt. M9 buffer was prepared a~ described by Brenner (1974). The bacteria used was OPSO, a uracil-requiring mutant of Escherichia coli. General mcthods and strains C. elegans was cultured as described by Brenner (1974). AnimaIs were cultured at 200 C unless otherwise stated. Wild type animaIs were N2 Bristol strain. XX refers to hermaphrodites and XO refers to I1'.dles; if there is no designation • following a strain name, then the strain is hermaphrodite. Mutant animaIs used were CB4876 clk-/(e2S19). MQ28 clk-l(qml1), MQSO clk-/(qI1l30), MQ123 clk-l(qI1l47), MQI73 clk-l(qI1lS/), CB1978 dpy-17(e164) unc-32(e189), MQ27 dpy-J7(e/64) clk 1(e2S19), MQ236 unc-79(eJ030) clk-l(e2S19), MQ234 unc-79(eJ030) clk-l(e2S19) XX and XO, MQ233 unc-79(el030) clk-l(qml1), MQ237 unc-79(eJ030) clk-l(qml1) XX and XO and CB4S12 gro-l(e2400). Assays for developmental phenotypes Ail developmental phenotypes were scored at 20°C unless otherwise stated. Embryonic Development: Gravid adult hermaphrodites were placed in a drop • of M9 buffer and eut open with a razor blade. Two-celled embryos were chosen and 15 • individually placed onto fresh plates. The embryos werc monilorcd cvcry 30 minulcs until hatching. Heritability Study: The embryonic developmcntal timcs of 179 e2519 animais were scored, as above. They were then left to grow to adulthood. Onc lwo-ccllcd embryo from each adult animal was dissected and its embryunic dcvclopmcntaltinw scored. The developmental times fo:' each pair of animais ean be found in Appcndix 1. Post-emblyonic DeveloplIlent: Eggs were placcd on a frcsh platc and Icft for one hour to hatch. Larvae which had hatched during that perim\ wcrc choscn, plaecd individually on fresh plates, and monitored every three hours unlil malurity. Animais were scored as mature adults after they had undergone the final moll and a vlIlva eOlild be observed. Egg Production Rate: Twenty hermaphrodites which had reached adullhood 24 hours before were picked and placed on fresh plates for four hours, exccpl for qmll animaIs which were on the plates for five hours. The 20 hermaphrodites were then removed and the number of eggs laid during that four-hour or five-hour period was counted. The actual numbers can be found in Appendix 2. Self-brood Size: Larvae were placed individually on fresh plates and leflto develop until they had become adults and laid the first few eggs. The hermaphrodiles were then transferred onto fresh plates daily to prevent overcrowding until egg-Iaying ceased. The progeny were counted three days after removal of the parents. The actual number ofprogeny can be found in Appendix 3a. Cross-brood Size: Single fourth larval stage (U) hermaphrodites were placed on fresh plates with three young, adult N2 males. The hermaphrodites were transferred onto fresh plates daily to prevent overcrowding until egg-Iaying ceased and • the three males were replaced by three young, adult males every two days. The 16 • progeny were counted three days after removal of the parents. The actual number of progeny can be found in Appendix 3b. Life Span: Eggs were placed onto a fresh plate and left for one hour to hatch. Larvae which had hatched during that period were chosen, placed individually on fresh plates, and monitored once daily until death. The animais were transferred once daily while producing eggs to keep them separate From their progeny. Animais were scored as dead when they no longer responded with movement to Iight prodding on the head. Time Lapse Video Microscopy: Embryos at the pronuclear migration phase were mounted on a 2% agar pad in M9 buffer. They were visualized with a Zeiss Axiovert Microscope with a IOOx oil immersion objective. Recordings were made with a Hamamatsu CCD camera model C2400, and a Panasonic time-Iapse video cassette recorder model G720A. The temperature on the surface of the microscope slide was monitored and kept at 21°C. Ail time lapse experiments were done by Siegfried Hekimi. Assays for behavioral phenotypes Ail behavioral phenotypes were scored at 20°C on hermaphrodites which had reached adulthood approximately 24 hours before. Defecation: Individual hermaphrodites were placed onto fresh plates and maintained in a regulated incubator. They were brought out of the incubator and allowed to settle for five minutes before their defecation cycle length was measured directly under the dissecting microscope. Defecation cycle length was defined as the time between the first muscular contraction of one defecation and the first muscular contraction of the next defecation. For each animal, five defecation cycles were • measured with a stopwatch. The five inter-defecation periods for each animal can he 17 • found in Appendix 4. The temperature of the agar was taken with a probe thermometer before and after the scoring of defecation to ensure that the animais had remained at a constanttemperature; only if the temperature had not varicd by morc than 1°C were the data included in subsequent analyses. The duration of the three muscular contractions which constitute the DMP was scored in a separate experiment. The stopwatch was started atthe l'irst sign of the anterior muscular contraction and stopped after expulsion. Three defecations werc measured for each animal and the three values can be found in Appendix 5. Pumping: Pharyngeal pumping was visualizcd through the dissccting microscope under conditions as for the measurement of the defecation cycle. Hermaphrodites were placed individually onto fresh plates. The animais were scored for one minute and each animal was scored twice. The two values for each animal can be found in Appendix 6. Swimming: Animais were placed in a drop of M9 buffer on a glass plate and allowed to sink to the bottom of the drop before the simple rhythmic thrashing of the animais was counted. The animais were scored for one minute and each animal was scored once. The value for each animal can be found in Appendix 7. The temperature of the drop of liquid was taken with a probe thermometer after measuring the swimming rate to ensure that the temperature had remained constant. EtTects of temperature Embryonic development: Animais were cultured at three different temperatures, ISoC, 200 C and 25 0 C. Gravid hermaphrodites l'rom each temperaturc were eut open with a razor blade and two-celled embryos were chosen and returncd to • plates stored either at the same temperature at which the parents had been culturcd, or 18 • to 200 C for the remainder of embryogenesis. These embryos were monitored every 30 minutes until they hatched. Defecation: Animais were cultured at three different temperatures, 150 C, 200C and 250C. Hermaphrodites from each temperature were placed individually onto fresh plates stored at 200 C and allowed to settle for five minutes before their defecation cycle length was measured directly under the dissecting microscope as described above. The five inter-defecation periods for each animal can be found in Appendix 8. Correlation assays Embryonic development of two-celled embryos was monitored as above. "Quickly developing" and "slowly developing" embryos were kept for further analyses. Quickly developing embryos were defined as those which hatched before 13 hours for N2 and those which hatched before 12.5 hours for clk-l(e2519). Slowly developing embryos were defined as those which hatched after 13.5 hours for N2 and those which hatched after 21.75 hours for clk-l (e2519). Post-embryonic development and the adult defecation cycle period were sCJred on animaIs in each of the two classes. The embryonic development times, post-embryonic development times and inter-defecation periods for each of the animais in the two classes can be found in Appendix 9. • 19 • Results A genetie sereen for maternal-effect mutations affeeting development and behaviour In order to identify new genes affecting development, a genetic screcn wus performed to isolate mutants which present a morphological or behuvioural phenotype when derived from a homozygous mutant parent, but which are phenotypically rescued when derived from a parent carrying a wild-type allele of the affected gene. The rationale for this screen is that only gene products required early in development or in very small amounts could be deposited in the egg in sufficient amounts to be able to rescue the phenotype of adults, which are approximately 500 times more voluminous than the egg. Wild type (N2) animaIs were mutagenized with ethyl methane sulfonate (EMS) • and allowed to self-fertilize for two generations. Ali animaIs resulting from the second round of self-fertilization (F2) (approximately 10000 animais) are likely to be homozygous carriers of a number of new mutations with or without visible phenotypic effects. Such animais were examined and 50 of those which displayed wild-type growth rate as weil as overall wild-type morphology and behaviour were selected, picked singly cnte fresh plates and allowed to self-fertilize. Those animais which then produced an entirely mutant brood (F3) were analyzed further. Out of 30,000 F2 animais and their progeny analyzed in this way, five allelic mutations (e2519, qmll, qm30, qm47 and qm51), which confer slow developmelltal rates and display a slow defecation cycle, were identified. They define the clk-l gene. Very slowly growing strains have generally a very sick appearance, the animaIs being • thinner and more transparent than the wild-type. clk-l mutants, however, have an 20 • essentially wild-type appearance. The first clk-l mutation to be identified (e2519) was originally kept and studied because the mutants displayed this surprising association of apparent health and slow growth. It was later found that these mutants also displayed a slow defecation cycle. Only two other mutants (qm37 and qm38) were found, which are not allelic to clk-l, or to (~ach other, but have features in common with the clk-/ mutants, including slow growth, a wild-type general appearance, a slow defecation cycle and maternai rescue of the defects. These findings suggest that only a small number of loci can mutate toward a phenotype resembling that of clk-l mutants. After isolation, ail clk-l mutants were crossed to N2 males and the progeny of these animaIs examined. For each mutant, ail male progeny as weil as an approximately similar number of hermaphrodites were found to develop and defecate at wild-type rate, indicating that the presence of a wild-type copy of the gene in the zygote is sufficient to obtain a wild-type phenotype. The progeny (F2) of these rapidly growing (FI) hermaphrodites were also found to grow at wild-type rate. A number of such rapidly growing F2 animaIs were picked singly onto fresh plates. Ali these animais exhibited wild-type defecation cycle length but approximately one quarter of these animaIs produced only slow growing progeny ail of which as adults had slow defecation cycle lengths. These experiments indicate that ail the mutations are fully recessive, maternal-effect mutations for the post-embryonic growth rate and defecation-cycle length phenotypes. The possibility of a paternal effect was tested for clk-l (e2519) and 1 found that there was no paternal effec!. The results are described fully in a later section. Furthermore, the maternaI effect is not strict; zygotes from homozygous mutant mothers carrying one wild-type copy of the clk-l gene are wild type for post-embryonic growth rate and defecation cycle length. These results are • also described in a later section. 21 • Two types of complementation tests were performed. First, e2519 homozygous males were mated to homozygous mutant hermaphrodites of the other slow growing strains, and the male cross-progeny were tested for slow post-embryonic growth and slow defecation cycle. Four mutations, qmll, qm30, qm47 and qm51 failed to complement e2519 by this test. Second, self-progeny of trans-heterozygotcs for mutations which failed to complement by the first test were examincd. Il was found that ail self-progeny were slow growing. This indicates that these mutations failed to complement for the maternaI effect as weil; that is, trans-heterozygolls mothers failed to produce maternaI rescue. The second type ofcomplementation test described above was also a test for second-site non-complementation: trans-heterozygotes between two unlinked non complementing recessive mutations shollld produce 5/16 phenotypically wild-type progeny. In contrast, there should be no wild-type progeny if the two mutations are • closely linked (Avery 1993a). By this test, e2519, qmll, qm30, qm47 and qm51 wcre found to be closely linked. Ali complementation tests were performed by Paula Boutis. Three-point and two-point mapping were performed for clk-l (e2519). clk-l maps to chromosome III, approximately 0.2 map units to the right of dpy-17. This position suggested possible allelism to gro-l(e2400), which similarly confers overall slow growth, a mutation isolated from a wild strain by Jonathan Hodgkin (personal communication). An examination of the phenotype of gro-l(e2400), revealed a number of additional similarities to clk-l mutants, including maternai rescue and a slow defecation cycle. However, gro-l(e2400) complements ail five clk-l alleles. The mapping of e2519 was done by Siegfried Hekimi and Bernard Lakowski. The five clk-J mutations are probably loss-of-function alleles, as they arose at a • frequency 0/5000) expected for loss-of-function mutations when using the standard 22 • EMS mutagenesis protocol (Brenner 1974, Greenwald and Horvitz 1980, Avery 1993). Moreover, as shown below, the sarne set of features is affected in the five mutants, which can be arranged into an allelic series based on the length ofpost embryonic development or the period of the defecation cycle. In addition, the post embryonic development of the trans-heterozygote e2519/qm30 is of intermediate severity between e2519 and qm30. clk-l mutations are pleiotropic 1examined the development and behaviour of clk-l mutant and wild-type worms, and found that clk-l mutations have broadly pleiotropic effects (Table 1) but a detailed examination of the animaIs by differential interference contrast microscopy did not reveal any morphological abnormalities. Ali observed phenotypic effects are related to timing, and result in a mean lengthening of the duration of developmental stages and ofthe period of cyclic behaviors. This mean lengthening is sometimes, but not always, associated with an increase in the variability of the measured value, as indicated by an increase in the standard deviation of the sample. The same set of timing features is affected in ail five alleles of clk-l. The first allele to be isolated was e2519 and it remains the best-characterized of the five alleles. However, qml1 and qm30 have been analyzed in great detail as weIl. The genetic screen was an on going effort and qm47 and qm51 were the two alleles most recently identified; hence, only embryonic development and defecation cycle period length were quantified for them (see below). • 23 • Table 1. Quantitative phenotypic analysis Phenotypic comparison of N2 animais, clk-/ mutants and maternally-rescucd e25/9 mutants. Developmental and behavioural phenotypes were compared among N2 and mutants carrying either ofthree clk-/ alleles: e25/9, qm// and q1ll30. The numbers are given as mean ± standard deviation and the sample size is given in parentheses. In the header, m represents the maternai and z the zygotic contribution of clk-l activity respectively. Superscript '+' and '-' indicate the wild-type and mutant form of the gene respectively. Ali phenotypes were scored at 200 C. The maximum life span observed is also given in the life span row. The data used to obtain the values for embryonic development, post-embryonic development, life span and defecation arc shown in full form as Figures 4, 1,2 and 3, respectively. The raw data corresponding to Figure 3 is in Appendix 4. • • • Table 1. Quantitative phenotypic analysis Phenotypes Genotypes N2 qmll qm30 (matemally rescued) Embryonic development 13.3 ± 1.0 (n=500) 13.6 ±0.8 (n=50) 17.1 ± 3.9 (n=500) 16.1 ± 1.0 (n=IOO) 22.8 ± 5.0 (n=IOO) (hours) Post-embryonic 46.6 ± 3.0 (n=500) 50.6 ±3.\ (n= 100) 70.6 ±4.8 (n=500) 81.0 ±5.6 (n=loo) 99.2 ± 6.2 (n=IOO) development (hours) Egg production rate 6.0 ± 1.2 (n=200) 3.8 ±0.9 (n=25) 2.7 ± 1.0 (n=200) 2.1 ±0.5 (n=200) 0.9 ±O.l (n=200) (eggs 1hour) Self-brood size 302.4 ± 30.5 (n=20) 340.9 ±77.2 (n=20) 191.1 ±33.0 (n=20) 191.1 ± 17.6 (n=lO) 87.2 ± 37.2 (n=lO) (number ofprogeny) Cross-brood sire 659.1 ±137.7(n=15) Not determined 443.4 ±90.8 (n=15) 398.6 ±95.0 (n=lO) 180.0 ± 89.5 (n=3) (number ofprogeny) ±5.2 (n= 50) 19.9 ±5.3 (n=50) 26.0 ±9.0(n=50) 28.5 ± 10.3 (n= 50) 22.8 ±9.8(n=50) Life span (days) mean 18.6 maximum 27 33 45 46 46 Defecation (n~loo) (mean of five cycle periods, in 50.8 ± 5.6 (n=IOO) 54.7 ±8A (n=lI) 69.4 ±9.9 77.9 ±6.9 (n=25) 92.4 ± 15.0 (n=25) seconds) Duration ofDMP 5.0 ±0.2 (n=20) Not determined 4.9 ±OA (n=20) 5.5 ±0.5 (n=20) 4.4 ±0.8 (n=20) (seconds) Pumping 259.0 ± 23.7 (n=25) 232.3 ±26.7(n=lI) 156.0 ±29.0 (n=25) 169.2 ± 25.0 (n=25) 170.3 ±26.9 (n=25) (cycles 1minute) Swimming 120.7 ± 6.5 (n=25) 120.3 ±I3.2 (n=ll) 91.7 ±5.7 (n=25) 87.5 ±9.8 (n=25) 75.6 ±4.9(n=25) (cycles 1minute) 25 • El1lbryonic Developl1lent The duration of embryonic development was scored from the two-cell stage to hatching. Eggs were dissected from hermaphrodites and observed continuously (every 30 minutes) until hatching. The two-cell stage was chosen as the start point because of the ease with which it can be identified under the dissecting microscope. The mean duration of embryonic development is affected in ail five alleles up to a near two-fold mean lengthening in the strongest allele, qm30 (Table 1). However, even the slowest embryos of any allele (> 30 hours at 20°C) proceed to further development nOl'lnally after hatching. In ail five alleles this lengthening ofembryonic development is also reflected in a slowing down of the cell cycle (analyzed in more detail in a later section). The high variability of the developmental times of the mutants and the effects of temperature are described in more detail in later sections. Post-embryonic Development Post-embryonic development was scored in three of the five alleles by monitoring the developmental stage of the worms every three hours. 1observed a lengthening of the duration of ail four larval stages. The mean developmenlal lime is altered in the e2519, qm1 1 and qm30 alleles, with a more than two-fold lengthening in ql1l30 (Table 1). However, even the slowest animaIs (>100 hours) became fertile adults and had a generally wild-type appearance. The lengthening of post-embryonic development is the phenotype by which the mutations were originally isolated and which is followed in genetic crosses. For e2519, the weakest mutation for post embryonie development, the distribution ofdevelopmental times was non-overlapping with that of wild-type (Fig. 1). The distribution of developmental times of ail lhree • mutants examined is broader than that of the wild-type (Fig. 1 and Table 1). However, 26 • Figure 1. Post-embryonic development Duration of post-embryonic development of N2 and clk-J(e25J9) animais. Animais which hatched in a one-hour period were isolated and monitored every three hours until maturity. Animais were scored as mature adults when a vulva could be observed. Each bar represents the number of animais that underwent the last larval molt (L4 - adult) within a three-hour period ending at the indicated time. The animais were maintained at 2üoC throughout developmenl. • • • Post-embryonic development 240 200 ,. , 160 Number of 120 animais 80 40 n n ~ 1 • • lime (hours) o N2 n=500 • e2519 "=500 28 • the percentage increase in standard deviation in the mutants is not significantly greater than the percentage increase in mean developmental times. This suggests that whatever is responsible for the variability in post-embryonic development is the same in the mutants as in the wild-type. Egg Production Rate The egg production rate is one of the most strongly affected features in clk-l mutants, with a> 6-fold decrease in qm30. Self brood size is also strongly affected, with a> 3-fold decrease in qm30 (Table 1). The self brood size is a measure of the number of sperrn produced by the hermaphrodite (Ward and Carrel 1979). During gametogenesis, sperrnatogenesis occurs first, followed by a switch to oogenesis. e25/9, qml/ and qm30 can produce a much larger number ofprogeny when mated by males (Table 1). However, self brood size is only an indirect measure of sperm • number, and we cannot exc1ude the possibility that the mutants make sorne non functional sperm. In spite of the slow-down of gametogenesis, the adult gonad appears wild-type in the distribution of mitotic germ cells versus maturing oocytes. This suggests that mitotic germ line proliferation and gametogenesis are slowed down in a concerted manner. However, the reduction in usable sperm suggests that the slowing down of sperm production is not compensated by a corresponding lengthening of the duration of spermatogenesis. Life Span The mean and maximum life span were quantified in three ofthe five mutants, e2519, qmll and qm30, and these values can be found in Table 1. 1 have not • investigated the life span of qm47 and qm51. The increase in life span observed is not 29 • simply a result of delayed onset of mortality. Rather, the rate of dying of the population of mutants is slower than the wild-type rate lit ail chronologieal P'1CS, as iIlustrated by the survival curves for N2 and e2519 (Fig. 2). Gnly e2519 is depieted in Figure 2, but the survival curves of qll/II and qll/30 arc indistinguishable l'rom that of e2519. Defecation Defecation is effected by a series of three muscular contractions involving three distinct groups of muscles. The series of three contractions is repeated at regular intervals and often referred to as the DMP or defecation motor program (Thomas 1990). AnimaIs were scored for five inter-defecation periods and the mean and standard deviation of the five periods were calculated for each animal. Figure 3 shows the relationship between the mean and the corresponding standard deviation for eaeh of the 100 N2 and e25J9 animaIs. The values shown in Table 1 for N2 and three of the five c1k- 1 alleles are the mean and standard deviation of the means of ail individual animaIs scored. In addition to the three alleles in Table l, the values for qll/47 and qm51 are 75.1 ± 26.5 seconds (n=20) and 104.1 ± 24.1 seconds (n=20), respectively. The cycle period is strongly affected in ail alleles, with a two-fold lengthening in qm51. This lengthening of the cycle period, however, does not result in constipation unlike other defecation cycle mutants (Thomas 1990). In addition, the c1k-1 mutants appear to be not only slower than the wild-type but also more variable in two distinct ways. First, as can be seen in Figure 3 and as indicated by the larger standard deviation given in Table l, the distribution of the means is broader in the mutant: the mutant animais are more different from each other than are the wild-type animais. Second, individual • mutant animais have, on average, a more irregular cycle period (the abscissa of 30 • Figure 2. Life span Life span ofN2 and clk-/(e25/9) animaIs. AnimaIs which hatched in a one hour period were picked and monitored once daily until death. AnimaIs were scored as dcad when they no longer responded with movement to prodding on the head. The animaIs were maintained at2üOC throughout. • • • Lite Span 50 ••••-=- """"-~- 40 Number 30 of animais 20 10 o 2 4 6 8 101214161820222426283032343638404244 lime (days) --.- N2 n=50 -.- e2519 n = 50 32 • Figure 3. Defecation cycle Defecation cycle length ofN2 and clk-l(e2519) animaIs. Defecation cycle length of individual animaIs were scored at 2DoC for five consecutive defecations. Defecation is effected by a series of three muscular contractions. A defecation cycle was defined as the time between the first muscular contraction of one defecation and the first muscular contraction of the next defecation. Each point represents one animal. The ordinate corresponds to the mean length of five defecation cycles and the abscissa represents the standard deviation for that mean cycle length. • • • Defecation cycle 50 40 30 .& .& 0 ...... & .& Standard .& 00 6' ....& deviation 20 ...... & 0 ... .& ... 8 0.& 0.&. ~ o 0cJXJ*COÂ~, 10 Cb:> oQ) ~\ ..iii. \. o .& 0 30 45 60 75 90 105 Mean defecation cycle length (seconds) 0 N2 n = 100 • e2519 n = 100 34 • Figure 3). The average standard deviation for individual animaIs is 6.4 seconds for N2, while it is 10.4 seconds for e2519. Pumping ln ail three clk-/ alleles examined (e25/9, qm// and qm30), mutant pharyngeal pumping is slowed down approximately 1.4 fold. However, these mutants do not appear starved, in contrast to other mutants known to affect pumping frequency (Avery 1993b). For this phenotype, as for life span, e2519, qmll and qm30 are of similar severity. This is true both for the mean rate and the standard deviation of the sample. Pumping rate is the cyclic feature with the highest frequency that has been examined. Swimming On an agar surface C. e/egans moves forward by propagating rearward sinusoidal waves. For technical reasons, however, locomotory performance was measured by scoring swimming, a simple rhythmic thrashing, in a drop of M9 buffer. Ali three alleles examined, e2519, qm// and qm30, show a significant slowing of swimming frequency, with qm30 being the most severe mutation (Table 1). The swimming frequency ofqm47 and qm51 have not been quantified. clk·] mutations can be maternally rescued clk-l (e2519) homozygous mutants (e25/9/e25/9) descending from self fertilizing e2519/+ parents display a profound maternaI rescue of their development and behaviour (Table 1), which was the basis for selection in the mutant screen. To • obtain the scores shown in Table 1, whole broods of self-fertilizing e2519/+ mutants 35 • were scored for ail phenotypes and homozygous mutant animaIs identified later by lhcir entirely mutant brood. Most phenotypes are totally or partially rescued. The egg production rate is the least weil reseued phenotype, with a value intermediale bctwccn that of mutant and wild-type. In spite of the fact that adults arc at least 500 timcs larger than the egg, an almost complete phenotypic rescue extends to post-embryonic development and life span (shortened in rescued animais) as weil as to behaviours such as defecation, which are scored in adults. Furthermore, in maternally-rescucd e25/9 animais the self-brood size is excessive relative to the wild-type (Table 1). The increased mean brood size is also associated with a major increase in variability, and the maximal score recorded (580) is far greater than any seen in the wild-type (357). The degree of maternai rescue for embryonic development, defecation cycle period and self-brood size was also exarilined for our strongest ailele, q1ll30. Once again, whole broods of self-fertilizing qm30/+ mutants were scored for the abovc • mentioned phenotypes and homozygous mutant animaIs later identified by the:r entirely mutant brood. qm30 is not perfectly rescued for the duration of embryonic development (15.4 ± 1.I hours, n= 28) norforthe length of the defecation cycle period (67.8 ± 10.9 seconds, n=20) but the degree of rescue is remarkable considering thc severity of the allele. The self-brood size is rescued to wild-type numbers (304.2 ± 40.7, n=20), and like e2519, the maximal score recorded (401) is greater than thosc wc have observed in the wild-type. The profound maternai rescue suggests that either clk-1 is required carly in development or that the clk-1 product is required in extremely small amounts and that the perdurance of maternai product throughout development is sufficient for phenotypie rescue in adults. The incomplete maternai rescue of embryogenesis in • qm30, and the subsequent rescue of defecation to a large degree (a behaviour scored in 36 • the adult worm) suggests thatthe former hypothesis is correct. Ifperdurance of the product were responsible for the incomplete rescue of embryogenesis seen in qm30 animais, then it would imply that there was not enough of the product to rescue even the earliest developmental event. If there was not enough of the product to rescue embryogenesis, then defecation in the adult, a worm 500 times more voluminous than an embryo, should not be rescued at ail. In later sections 1describe experiments in which 1 manipulated the rate of development using temperature shifts and which further support this hypothesis. In C. elegans, the sperm contributes genetic material as weil as the centrosome (Ward anô Carrel 1979) so a paternal effect was possible. In contrast to the strong maternai effect, we observed only a very minor paternal effect which 1 investigated using e25/9 (Table 2). e25/9/e25/9 animais derived from e25/9/e25/9 mothers and e25/9/+ fathers appear entirely mutant except that embryonic development appears to • be slightly shortened relative to e25/9/e25/9 self-progeny «10%). That is to say, the wild-type copy (+) in the heterozygous male (e25/9/+) is not able to rescue the homozygous mutant progeny. Furthermore, 1 found that one zygotic copy of clk-/ (+) is sufficient for a wild type phenotype; that is, e25/9/+ animaIs derived from e25/9/e25/9 mothers appear almost entirely wild-type, except for a small «10%) delay in the embryonic development time compared to animaIs with a fully wild-type pedigree (Table 2). The slight lengthening ofembryogenesis seen in these e25/9/+ animaIs may be due to the delay in the start of transcription of the zygotic genome; until the zygotic genome is turned on, the reduced amount of clk-/ product supplied by the e25/9/e25/9mother is responsible for the development of the early zygote. • 37 • Table 2. Maternai, paternal and zygotic rescue of embryonic development Dunition of embryonic development of animais with difi'erent genotypes. +/+ represents animaIs that are self-fertilizing N2 and e25/9/e25/9 are se\f-fertilizing e2519/e2519 animaIs. e2519/+ animaIs are cross-progeny l'rom N2 males and homozygous e25/9/e2519 hermaphrodites; these animais were used to assess the presence ofzygotic rescue. Maternally-rescued e2519/e25/9 animais are the homozygous, mutant self-progeny of e25/9/+ hermaphrodites. The homozygous mutant animaIs were identified later by their entirely mutant brood. The homozygous, mutant cross-progeny of e2519/+ males and e2519/e25/9 hermaphrodites, (e2519/e2519), were used to assess the presence of paternal rescue. The hermaphrodites were also homozygous for dpy-17 (e /64) and therefore cross-progeny could be distinguished l'rom self-progeny by body shape. Self-progeny wcre dumpy while cross-progeny were wild-type for body shape. • • • Table 2. MaternaI, paternal and zygotic rescue ofembryonic development Parental genotypes Genotype Duration ofembryonic Nature of rescue development (hours) Wild-type (N2) +/+ 13.3 ± 1.0 (0=500) hermaphrodites e2519/e2519 hermaphrodites e2519/e2519 17.1 ± 3.9 (0=500) e2519/+ hermaphrodites e2519/e2519 13.6 ±0.8 (0=50) MaternaI +/+ fathers e2519/+ 14.3 ± 1.0 (0=104) Zygotic e2519/e2519 mothers e2519/+ fathers e2519/e2519 16.1 ± \.0 (0=106) Paternal e25191e2519 mothers 39 • Strict maternaI effects on the length ofembryogenesis 1 investigated the embryonie developmental time of e25 / 9/q1ll30 and qml/ /qm30 trans-heterozygotes when originating from 1110thers of different genotypes (Table 3). qm// and qm30 were ehosen because they are the weakest and strongesl alleles for embryonic development, respeetively, and e25/9 was used beeause il is the best-eharaeterized allele of the clk-/ gene. When the e25/9/qIll30 animais werc the eross-progeny of homozygous qm30/q1ll30 fathers and homozygous e25 /9/e25 /9 mothers, the duration of embryonie development was 17.6 ± 1.7 hours, not significantly different from e25/9/e25/9 self-progeny (17.1 ± 3.9 hours). In contrast, when the e2519/qm30 animais were the eross-progeny of hOl11ozygous e25/9/e25 /9 fathers and homozygous qm30/qm30 mothers, the duration of embryonic developl11ent was 20.6 ± 2.5 hours, more similar to qm30/qm30 self-progeny (22.8 ± 5.0 hours). It appears, therefore, that in the absence of a wild-type eopy of the gene in the zygote, the maternaI effeet can be strict: the duration of embryogenesis depends primarily on the maternaI contribution. Similar results were obtained for the qm/ I/qm30 trans heterozygotes (Table 3). • 40 • Table 3. Embryonic development of trans-heterozygotes Duration of embryonic developmenttime of trans-heterozygotes originating from mothers of different genotypes. Ali animais scored were the cross-progeny of the parents indicated in the left-most column. To distinguish self-progeny from cross progeny. double mutants of homozygous clk-l (e2519) unc-79 (eI030) and homozygous clk-/ (qmll) unc-79 (eI030) were used. Self-progeny would be uncoordinated and cross-progeny wou Id be wild-type for movement. The numbers given correspond to the means and standard deviations, in hours. The bottom portion of the table shows the duration ofembryogenesis for e2519, qmll and qm30 for ease of comparison (from Table 1). • • Table 3. Embryonic deve10pment of trans-heterozygotes Parental genotypes Genotype Duration of embryonic development (hours) e2519/e2519 mothers e2519/qm30 17.6± 1.7 (0=101) qm30/qm30 fathers qm30/qm30 mothers e2519/qm30 20.6 ± 2.5 (0=\02) e2519/e2519 fathers qmll/qmll mothers qmll/qm30 18.7 ± 1.3 (0=103) qm30/qm30 fathers qm30/qm30 mothers qml1/qm30 21.8 ± 2.4 (0=98) qmll/qmll fathers Se1f-progeny of e25191e2519 17.1 ± 3.9 (0=500) e2519/e2519 Se1f-progeny of qmll1qmll 16.1 ± \.0 (0=100) qmll1qmll Se1f-progeny of qm30/qm30 22.8 ± 5.0 (0=100) qm30/qm30 • 42 • The length ofembryonic development is highly variable in clk-l mutants 1scored the embryonic development of ail five clk-1 mutants and N2 animais; these results are shown as histograms in Figure 4 with the mean values given in Table 1 (except for qm47 and qm51, for which the corresponding values are 16.5 ± 0.7 (n=98) and 19.6 ± 2.9 (n=102), respectively). The distribution ofN2 animais centers around a single modal value. The distributions of qml1 and qm47, the clk- 1 alleles with the shortest mean embryonic development similarly center around a single value: the embryonic development of these animais is slow but as regular as the wild-type. In contrast, the distribution for e2519 and qm30 are extremely broad and do not seem to follow a single defined distribution. For e2519 a proportion of the values cluster around the same value as qml1 but the remainder are spread on either side. The phenotype of qm30 is even more severe in the sense that no clear clustering of values can be identified. The distribution of qm51 is not as broad as those of e2519 and qm30; it is, however, the second slowest of the five mutants. Figure 4 also shows that a number of e2519 embryos developed as rapidly as the most rapidly developing wild-type animais, and indeed, a small number appeared to develop faster than the fastest wild-type. To confirm this observation 1followed the development of an additional500 embryos of e2519 and N2 up to 12.5 hours after the two-cell stage. Figure 5 shows hatching times for the fastest of the total of 1000 embryos for each genotype. 3% of clk-1 embryos hatch earlier than 10 hours, while only 0.3% of wild-type embryos do so. A number of observations suggest that the variability in the embryonic developmental rates of e2519 is not due to genetic inhomogeneity. First, e2519 has been back-crossed more then 20 times to the wild-type, which suggests that • hypotheticai modifier mutations would have to he relatively tightly linked to clk-l. • Figure 4. Duration of embryonic developmenl Duration of embryonic development of N2 and e2519, qm51 animais. Two-celled embryos were dissected l'rom gravid hermaphrodites, placed singly and monitored every 30 minutes until they hatched. Each bar represenls the number of animaIs that hatched within a 3D-minute period. The embryos were maintained at 200 C throughout development. The sampie size was 500 for N2 and e25]9, and 100 for qm]], qm30, qm47 and qm51 . • 120 Ouration of embryonic development 100 E 80 M B • R 80 v Wild Type (N2) 0 S 40 20 0 24 20 18 E M B 12 R V 0 8 S qm11 4 0 40 30 20 E M B qm47 R 10 V 0 S 0 E M 40 B R e2519 v 20 0 S 0 E M Il B qm51 R 6 Y 0 3 S 0 E M Il B qm30 R 6 Y 0 3 • S 0 16 11 20 22 24 a 28 30 32 Tlme (hours) 45 • Figure 5. Quickly developing embryos Sorne clk-l(e2519) mutants develop l'aster during embryogcnesis lhan N2 animais. One thousand two-celled embryos of each genolype werc dissected l'rom gravid hermaphrodites, picked singly and monitored every 30 minutes until 12.5 hours. The remaining unhatched eggs were discarded. Each bar represents the number of animaIs that hatched within a 3D-minute period prior to 12.5 hours. Three percent of clk-l (e2519) embryos hatch earlier than 10 hours while only 0.3% of N2 embryos do so. • • • Quickly developing embryos 16 14 12 10 Number of 8 embryos 6 4 2 0 1 n 8 9 10 11 12 Time (hours) o N2 n =1000 • clk-1(e2519) n =1000 , 47 • Second, particularly slow or rapid development of clk- / mutant embryos is not heritable as indicated by the following experiment. 1scored the embryonic developmentaltimes of 179 e25/9 animaIs, which were then left to grow to adliithood. 1 then dissected one embryo l'rom each adult animal and scored its developmentaltime (Appendix 1). 1found no correlation between the two sets of limes (correlation coefficient 1'=0.026, as calculated by the method described by Zar (1984)). Together these observations suggest that clk- / mutations result in a dereglilation of the duration of embryonic development, and that the degree of deregulation is an intrinsic aspect of the severity of each allele. qm/l and qm47 are the alleles with the fastest and the most regular development and can therefore be considered to be the alleles with the weakest effect, at least on embryonic development. The cell cycle period is affected in clk-l mutants • Microscopy was used to observe embryos l'rom fertilization to hatching and it was found that the ccII cycles in alllineages are lengthened in mutant embryos compared to the wild-type. Nine early (pronuclear stage) N2 and e25/9 embryos were analyzed by time-Iapse video microscopy to deterrnine ifa slowing down of a particular phase of the ccli cycle is responsible for the delays observed. After fertilization, the pronuclei of the sperrn and oocyte fuse and the one-celied embryo is referred to as PO. It divides to give AB and Pl, the anterior and posterior daughter cells, respectively (Su1ston et al. 1983). The duration of cytokinesis and mitosis in e2519 animais is indistinguishable l'rom the wild-type. The only significant delay was observed for the interphases of AB and PI, which were lengthened \.6 fold compared to the wild-type. This lengthening is of similar magnitude to the effect of the mutation on the other • features scored (Table 1). In the wild-type, the interphases of the carly ccII cycles are 48 • entirely occupied by DNA synthesis with neither G 1 nor G2 observable (Edgar and McGee 1988). In the future, it will be informative to determine if the lengthening observed is due to a slow-down of DNA replication or to the appearance of aGI and/or G2 phase. In addition, not only the cell cycle but ail phases of embryonic development appear lengthened, as monitored by observation of the stage-specifie changes in the shape of the embryo. These delays, however, have not been precisely quantified. IndividuaI variations in the severity ofclk·] mutant phenotypes are correlated in a non·intuitive way As described in previous sections, a high variability among individual mutant animaIs was found in a variety of features. For example, the standard deviation of the sample of embryonie developmental times of e2519 mutants is almost four times larger than that of N2 (Table 1 and Fig. 4). Similarly, for the defecation cycle period, the variability among mutant animaIs is approximately twiee as large as that of the wild type (Table 1 and Fig. 3). The distribution of hatching times of the mutants appears not to follow a normal distribution, suggesting that the rapidly and slowly developing animais are actually different from each other in some unknown, non-heritable way. Do the animaIs with widely different hatching times remain recognizably different during later development and adulthood? To investigate this, 1selected wild-type and mutant animais with very fast or very slow embryonic development, and scored the duration of their post-embryonic development and defecation cycle (Table 4). For both these features, 1 observed a striking inverse correlation with embryonic development time. AnimaIs with fast embryonic development develop slowly during • post-embryonic development and have a slow defecation cycle, and animais with slow 49 Table 4. Correlated variations in the severity of the clk·l phenotypes Post-embryonic developmental time and defecation cycle period of ,mimais selected on the basis of embryonic developmental time. For each genotype a group of 29 embryos whose developmental times clustered closely around 12 hours (quickly developing embryos ofboth N2 and e2519 genotypes) or 14 hours (N2) and 22 hours (e2519) (slowly developing embryos) were selected. The post-embryonic developmental time, and later the adult defccation period of these animais was scored. Total developmental time corresponds to the total of embryonic developmental time plus post-embryonic developmental time. The numbers given correspond to the means and standard deviations, in hours for the developmental durations and in seconds for the defecation cycle length. • • Table 4. Correlated variations in the severity of the clk-l phenotypes Relative rate of Phenotype N2 clk-l(e2519) embryonic development Fast Embryonic development (hours) 12.1 ±0.7 (n=29) 11.7 ±O.7 (Il =29) Post-embryonic 49.5 ± 1.7 (n=25) 74.9 ±4.0 (n=25) development (hours) Total development 61.6 86.6 (hours~ - Defecation cycle period 52.7 ±5.0 (n=29) 71.6 ± 12.0 (n=29) (seconds) Embryonic development Slow 14.4 ±0.5 (n=29) 22.8 ±0.7 (n::29) (hours) Post-embryonic 47.2 ±2.0 (n=25) 64.3 ±4.1 (n=25) development (hours) Total development 61.6 87.1 (hours) Defecation cycle period 50.2 ± 3.8 (n=29) 57.3 ±6.5 (n=29) (seconds) • 51 • embryonic development develop rapidly during post-embryonic development and have a fast defecation cycle. Both slow and fast phenotypes l'ail within the normal clk- J mutant range. None of the other phenotypes which we have shown to be affectcd in clk-/ mutants appears to correlate in severity with the speed of embryonic development (data not shown). An additional observation is that the total time taken to develop (the sum of embryonic and post-embryonic development) is constant, regardless of the cohort. That is, fast embryogenesis is offset by slow larval growth and slow embryogenesis is offset by fast larval growth. These phenornena are apparent in the wild-type as weil as in the mutant, but the effects are much larger in the mutant, which has a IO-hour difference in post-embryonic development and a 14-second differenee in the defecation period between the cohorts. Overall, this suggests that animais with different embryonic developmental times remain different l'rom each other throughout • later development and as adults. Effects of temperature on the duration ofembryonic development C. elegans is capable of adapting its developmental rate to a relatively broad range oftemperatures, from 130 C to 25 0 C (Hedgecock and Russell 1975). 1tested the possibility that this ability was altered in clk- / mutants by examining the effcct of temperature on embryonic development. N2, three allelcs of clk-/ (e25/9, qm/ / and qm30), and matemally-rescued e2519 were included in the analysis. Two-cellcd embryos were dissected l'rom animais raised at one of three distinct temperatures (150 C, 200 C or 25 0 C), maintained at that temperature, and the duration of thcir embryonic development measured (Figure 6 and Table 5, top panel). For N2, qml1, e2519 and maternally-rescued e2519 animaIs, the duration of embryonic development • 52 • Figure 6. Effects of temperature on embryonic development Effects of temperature on the duration of embryonic development of N2, e2519, qmll and qm30 and maternally-rescued e2519 animais (e2519 resc. in the legend). Two-celled embryos were dissected from gravid hermaphrodites cultured at 150 C, 200 C and 250 C and allowed to develop atthe temperature at which their mothers had been raised. They were monitored every 30 minutes untilthey hatched. The means are plotted. See Table 5, top panel, for exact means, sample sizes and standard deviations. • • • Effects of temperature on embryonic development 50 45 40 Duration of 35 embryonic development 30 (hours) 25 20 15 10 5L...------...... ---1SCC 20°C 25°C Temperature during embryogenesis • N2 • e2519 )( e2519 resc. )1( qmIl • qm30 54 • Table 5. Effects of temperature differences and shifts on embryogenesis Embryonic developmental time ofN2, e2519, maternally-rescued e2519, qmll and qm30 under various incubation conditions. Two-celled embryos were dissected from gravid hermaphrodites cultured at 150 C, 200 C and 250 C, following which they were either allowed to develop at the temperature at which their mothers had been raised, or placed at 200 C for the remainder of embryogenesis. In ail cases they were monitored every 30 minutes until they hatched. The data is summarized graphically in Figures 6 and 7. • • • Table 5. Effects of temperature differences and shifts on embryogenesis Temperature Duration ofembryogenesis in hours when oogenesis and ernbryogenesis were at the indicated temperature N2 e25/9 111+ z· e25/9 qlll/ / qm30 (matemally rescued) 23.6 ±\.8 (n=160) 21.8 ± \.5 (n=161) 25.4 ±3.8 (n=98) 27.9 ± 1.3 (n=99) 46.1 ± 12.1 (n=94) 1 / , 13.3 ±J.O (n=5OO) 13.6 ±D.8 (n=50) 17.1 0<3.9 (n=5OO) .0.> ±\.O (n=loo) 22.8 ±s.o (,=100) 10.3 ±D.8 (n=98) 11.1 ±D.8 (n=26) 13.1 ±\.5 (n=IOO) 12.6 ±D.? (n=90) 12.7 ±\.l (n=65) Temperature Duration of embryogenesis in hours when oogenesis was at the indicated temperature and embryogenesis at 20°C N2 e25/9 111+ z· e25/9 qllll/ qm30 (matemally rescued) 13.6 ±DA (n=107) 14.8 ±J.9 (n=45) 20.7 ±2.5 (n=I().l) 18.5 ±2.0 (n=I().l) 32.4 ±9.5 (n=6l) 20°C 13.3 ±l.O (n=5OO) 13.6 ±D.8 (n=50) 17.1 ±3.9 (n=5OO) 16.1 ±J.O (n=loo) 22.8 ±s.o (n=loo) 13.1 ±D.8 (n=loo) 14.5 ±O.6 (n=42) 13.7 ±\.O (n=101) 13.7 ±l.O (n=97) 14.2 ±1.3 (n=79) 56 • is dependent on temperature, and to a similar degree. In contrast, for qm30, the amount of change between any two temperatures is much greater than for N2 or the other alleles. At 250 C, however, the three clk-l alleles are indistinguishable, producing animais that develop at the same rate, albeit still significantly slower than the wild-type or maternally-rescued e2519. These observations suggest an altered ability of the qm30 mutants to respond to various temperatures. l have shown above that clk-l mutants display a profound maternaI rescue when derived from mothers carrying a wild-type copy of the gene. One possible interpretation of this phenomenon is that the clk-l-sensitive processes are, or can be, set carly du ring development. As embryonic development in the wild-type and clk-l mutants is sensitive to temperature, l attempted to manipulate the rate ofembryonic devclopment with a tcmperaturc-shift paradigm. For wild-typc and mutant animais, two-celled embryos were dissected from gravid hermaphrodites raised at 150 C, 200 C • and 25 0 C, but were then transferred to plates at 200 C, and the duration of embryogenesis measured. For N2 and maternally-rescued e2519, shift to 200 C resulted in a length of embryogenesis typical of200 C regardless of the temperature of the oocyte and early embryo up until the two-celled stage (Fig.? and Table 5, bottom panel). In striking contrast, the rate of embryonic development of clk-l mutant oocytes and early embryos does depend on the temperature experienced up until the two-cell stage. When originating from mothers raised at 150 C the rate of embryonic development at 200 C was less than that for embryos derived from mothers raised at 200 C. Conversely, when embryos originated from mothers raised at 250 C, the rate of embryogenesis at 200 C was greater than that ofembryos originating from mothers raised at 200 C. Furthermore, for each allele, the effects produced by early heat (250 C) • 57 • Figure 7. Effects of temperature shifts during development Effects of temperature shifts on the duration of embryonic developmcnt of N2. clk-l(e2519. qmll, qm30) and matemally-rescued e25/9 animaIs (e25/9 rcsc. in the Iegend). Two-celled embyros were dissected l'rom gravid hermaphrodites that had been cultured at 150 e. 200 e and 25 0 e. These embryos were picked and plnccd al 200 e for the remainder of embryogenesis. They were monitored cvery 30 minutcs untilthey hatched. The ploued points correspond to the means. Sce Table 5, boltom panel. for exact means. sample sizes and standard deviations. • • Effects of temperature shifts during development 35 30 Duration of 25 embryogenesis (hours) 20 at 20°C 15 10 L...------o~-----_- 1SoC 20 C 25°C Temperature during oogenesis -.- N2 -e- e2519 ·-X- mat. resc. e2519 -.- qm11 _ .•- qm30 S9 • or by early cold (IS0 C) are of very similar magnitude (resulting in straight lines connecting data points in Fig. 7). The results of incubation at constant temperature (Fig. 6) might be used to infcr that qlll30 shows a graded cold-sensitivity. However. two observations suggest a more complex phenomenon. First. early cold (1S 0 C) can slow down subsequent development at 200 C. Second. shifting the embryos l'rom 2S 0 C to 200 C makes the subsequent development of clk-1 animais more similar to that of wild-typc trcated in the same way. than does continuous development at 2S 0 C. Effects of temperature on the DMP Just as C. elegans is capable of altering its developmental rate in response to a range of temperatures. so it also can alter its behaviour (Hedgecock and Russell 1975). 1 tested the possibility that this ability was altered in clk-1 mutants by examining the • effect that the temperature at which development had taken place had on the defecation cycle at a fixed temperature. N2 and e2519 were used in the analysis. Animais were raised at 1S°C. 20°C and 2S°C and their defecation was scored at 20°C. The results of the mean and standard deviation of l'ive inter-defecation periods are shown in Table 6. Regardless of the temperature at which they were raised, when scored at 20°C, N2 worms are able to defecate at a rate typical of 20°C. In contrast, the rate of defecation of e2519 animaIs at a fixed temperature is somewhat sensitive to the temperalUre at which they experienced development. For example, when e2519 animais are raised at 2S°C, the defecation cycle period is 67.0 seconds while it is 84.4 seconds for animais raised at 1S°C. Similar to emblyonic development, the temperature at which e2519 animaIs were raised seems to matter. Hence. it appears that wild-type animais can • 60 • Table 6. Effects of temperature on defecation Defecation cycle periods of N2 and e2519 animais raised at various temperatures were measured at 200C. The animaIs were cultured at 150C, 200C or 25°C and hermaphrodites which had matured in the past 24 hours were scored at 20°C, as described in Materials and Methods. Five defecation cycles were scored for each animal and the numbers given correspond to the means and standard deviations, in seconds. • • Table 6. Effects of temperature on defecation Growth Genotype Defecation Cycle Temperature Period at 20°C (s) 15°C N2 53.0 ± 3.4 (n=25) e2519 84.4 ± 10.7 (n=25) 20°C N2 50.8 ±5.6 (n=IOO) e2519 69.4 ±9.9 (n=IOO) 25°C N2 54.5 ±2.7 (n=25) e2519 67.0 ±4.7 (n=25) • 62 • sense and/or adapttheir behaviour to their surroundings but clk-l (e2519) animaIs cannot sense and/or adaptto the same degree. • • 63 • Discussion We have isolated l'ive ailelie mutations which detine the gene clk-/. The phenotype resulting from these mutations indicates that the normal function or clk-/ is necessary for the correct timing of a variety of features of the organism. These features include the cell cycle, embryonic and post-embryonic development, gametogenesis, death rate, and a number of high and low frequency behaviollral cycles. ln clk-/ mutants, these features are, on average, slowed down or decreased to an allele-speeific degree. In addition, the timing of these features appears to be deregulated, as indicated by a substantial increase in the variability among animais. A partieularly striking example of this phenomenon is provided by clk-/ embryonic developmenl, for which 1 have found that about 3% of the animais develop raster than even the fastest wild-type. These observations suggest that mutations in clk-/ lead to a partial uncoupling of these functions l'rom a mechanism which normally regllimes their rate. Furthermore, temperature shift experiments sllggest that embryos actively sense and adapt to temperature, and that clk-/mutant embryos appear to be partially uncoupled from this regulatory input (Fig. 7 and Table 6). This suggests that clk-/ is involved in a process which is required for the regulation of developmental rate by temperature. How can a single function or a single mechanism rcgulate events of very different nature and on very different time scales? 1 have shown that the process affected by clk-J is required for the normal activity of a number of specifie biological clocks devoted to particular features. Behavioral and developmental clocks are known to exist in C. elegans. For example, observations on wild-type and mutants which alter • the defecation cycle suggest that this eycle is regulated by a biological clock (Thomas 64 • 1990, Liu and Thomas 1994), and, as in other organisms, the regulation of cell cycle length ha~ been shown to have c1ock-like properties (Schierenberg 1984, Schierenberg and Wood 1985, Schnabel and Schnabel 1990). clk-/ could be involved in a function which, by regulating the level of activity I)f several biological c1ocks, would ensme the proper temporal coordination of ail features of the organism. Such a function might be required to obtain a coordinated response to fluctuations in the environment. Temperature is a good example of an environmental parameter to which one expects organisms to react actively. Indeed, 1 have shown that the response to changes in temperature is profoundly altered in clk-/ mutants. In the future, we plan to test if clk-/ could also be required for the sensing of or the reaction to changes in other environmental parameters, for example, nutrient availability. The biological c10cks which are altered in clk-/ mutants are very distinct from each other, expressed in very different tissues at different times and with very different periods. For example, the cell cycle and the defecation cycle have nothing obvious in common. What kind of mechanisms couId affect such disparate processes? Variation in intracellular Ca2+ concentration is a cellular mechanism with regulatory propcrties which is sufficiently universal that its alteration could have highly pleiotropic effects. A vast Iiterature describes the intracellular calcium oscillations which are observed in numerous cell types (for reviews see Berridge 1993, Berridge 1990). Variations in Ca2+ concentration and other second messengers inter::.cting with Ca2+ have been Iinked to almost ail aspects of cell function. Furthermore, theoretical consideration suggests that regulatory controlloops involving second messenger system could forrn the basis of high frequency biological rhythms (e.g. Rapp and Berridge 1977). An alteration in the availability of ATP is also a condition which would impinge • on ail functions of the organism, including biological c1ocks. It is possible that clk-l 65 • mutations alter sorne part of normal cellular energy procurement. Indeed, glycolytic oscillations, primarily controlled by phosphofructokinase, have been demonstrated in several systems (Hess and Plesser 1979, Tornheim 1979, Andres el lIl. 1990). Given the ubiquitous distribution of glycolytic enzymes, it has often been proposed that metabolic oscillations could drive other cyclic physiological processes, including rhythmic smooth muscle contraction (Connor 1979), electric activity of nellrons (Meech 1979), of heart muscle cells (O'Rourke el li/. 1994), and hormone secrction (Corkey el li/. 1988). It would be illuminating if future investigations would show a link between the clk-l gene and metabolic oscillations, as it would provide a rationale to understand the pleiotropic effects of clk- J mutations. A different question is to ask whether the clk-l mutant phenotype could be the result of a simple decrease of ATP production or availability. The following arguments suggest that this is unlikely to be the case. 1) clk-l mutants do not appear sick or anatomically abnormal, while metabolically restricted worms, for example because of insufficient food intake, have a typical "starved" appearance (Avery 1993). 2) The maternai effect on the clk-l phenotype is profound and extends to adult phenotypes. This makes it unlikely that clk-l couId encode a major structural component (e.g. a glycolytic enzyme) of the pathway which extracts energy from food. 3) The results of our genetic screen indicate that probably only few genes exist which can readily mutate towards a clk-l-like phenotype. If the clk-l phenotype were a secondary consequence of a non-specifie reduction in metabolism, one would imagine that a larger number of genes could mutate to bring about such an effeet. 4) Finally, of ail the features ofthe clk-l mutant phenotype, only the decreased mean growth rate would have been predicted for mutations which decrease metabolism. In fact, clk- J mutations uncover processp-s sueh as, for example, the eomplex temperature- • .' . 66 • dependence of developmental rates (Table 5) and the complex correlation between developmental and behavioural rates (Table 4), which could not have been extrapolated from the idea of a metabolic impairment. How does the observation that clk-f mutants can be rescued by maternai product fit into a hypothesis? The model we envisage is that continuous presence of clk-f(+) activitYis not required to regulate the feature-specific clocks. The LI larva which hatches from the egg is already very similar to the adult in the behaviours it displays as weil as in most details of its body. It is possible therefore, that many feature-specific clocks are already being established during embryogenesis. The clk-f gene product could be only required to set, or modify, the setting of feature-specific clocks, and not being continuously required to maintain a particular setting. Once set, the feature-specific clocks would keep ticking at the wild-type speed even after clk-f activity had been diluted by post-embryonic growth. It is readily conceivable that the clk-f{+) product could have been deposited in the egg to set specific clocks at a wild type level, or as close to a wild-type level as possible. In this context, it is notable that the features which are not weil rescued are those involving the germ line (Table 1). The germ line is remarkable for undergoing a major expansion during post-embryonic development from a two-cell primordium. Furthermore, because the germ line gives rise to a new generation it is likely that many epigenetic features present in the parent animal, such as the setting of biological clocks, would be erased in the germ cells. 1observed a complex correlation between developmental and behavioural features (Table 4). It indicates that animaIs with different embryonic developmental times remain phenotypically different during post-embryonic development and as adults. This suggests that part of the variability among animaIs which we observe for • most phenotypes, could be due to an initial variability in the setting of the putative 67 • clk- / -sensitive central clock. This is consistent with our hypothesis that the clk-/ sensitive clock is set early in development. We hypothesize further that in the absencc of wild-type clk-/ activity, the clock might still be set early but at nmdom activity levels. Il would be this random setting which makes the developmental times or clk-/ embryos so variable. Furthermore, this early setting could influence aspeets of later development and behaviour as we have observed for the apparent setting produeed by matemally derived wild-type activity which results in maternaI rescue. Our current hypothesis, however, does not explicitly account for the puzzling observation that animais with fast embryonic development show slow post-embryonic developmcnt and an irregular and slow defecation cycle period, and vice versa. In the future, molecular analysis of the clk- / gene and analysis of other genes involved in the timing phenomena revealed by the mutations in clk- / will allow us to c test further the hypothesis that clk- / regulates biological clocks. • 68 • References Albertson, D.G. and Thomson, J.N. (1976). The pharynx of Caenorhabditis elegans. Philos. Trans. R. Soc. Land. Sel'. B 275: 299-325. Ambros, V. and Horvitz, H.R. (1984). Heterochronic mutants of the nematode Caenorhabditis elegans. Science 226: 409-416. Ambros, V. (1989). A hierarchy of regulatory genes controls a larva-to-adult developmental switch in C. elegans. Cell 57: 49-57. Andres, V., Schultz, V. and Tornheim, K. (1990). Oscillatory synthesis of glucose 1,6-bisphosphate and frequency modulation of glycolytic oscillations in skeletal muscle extracts. J. Biol. Chem. 265: 21441-21447. Aronson, 8.0., Johnson, K.A., Loros, J.J. and Dunlap, J.C. (1994). Negative feedback defining a circadian c1ock: autoregulation of the c10ck gene frequency. Science 263: 1578-1584. Avery, L. (1993a). The genetics offeeding in Caenorhabditis elegans. Genetics 133: 897-917. Avery, L. (1993b). Motor neuron M3 controls pharyngeal muscle relaxation timing in Caenorhabditis elegallS. J. Exp. Biol. 175: 283-297. Avery, L. and Horvitz, H.R. (1987). A cell that dies during wild-type C. elegans development can function as a neuron in a ced-3 mutant. Cell 51: 1071-1078. Avery, L. and Horvitz, H.R. (1989). Pharyngeal pumping continues after laser killing of the pharyngeal nervous system of C. elegans. Neuron 3: 473-485. 8aylies, M.K., Bargiello, T.A., Jackson, F.R. and Young, M.W. (1987). Changes in abundance or structure of the pel' gene product can affect periodicity of the • Drosophila dock. Nature 326: 390-392. 69 • Berridge, M.J. (1990). Calcium oscillations. J. ofBiol. C"em. 265: 9583-9586. Berridge, M.J. (1993). lnositol trisphosphate and calcium signalling. Nall/re 361: 315-325. Brenner, S. (1974). The genetics of Caellor"abditis elegal/s. Gelletics 77: 71-94. Cohen, A.H., Rossignol, S. and Grillner, S. (1988). Neural comml (~r r"yt"mic movemellts in vertebrates. Wiley-lnterscience, New York. Connor, J.A. (1979). On exploring the basis for slow potential oscillations in the mammalian stomach and intestine. J. Exp. Biol. 81: 153-173. Corkey, B., Tornheim, K., Deeney, J.T., Glennon, M.C., Parker, J.C., Matschinsky, F.M., Ruderman, N.B. and Prentki, M. (1988). Linked oscillations offree Ca2+ and the ATP/ADP ratio in permeabilized RlNm5F insulinoma cells supplemented with a glycolyzing cell-free muscle extraet. J. Biol. Chem. 263: 4254-4258. • Darnell, J., Lodish, H. and Baltimore, D. (1990). Cancer. In: Molecular Cell Biology, 2nd ed., pp.959-960, W.H. Freeman and Company, New York, N.Y. Decuyper, C. and Vanfleteren, J.R. (1982). Nutritional alteration of life span in the nematode C. elegans. Age 5: 42-45. Dunlap, J.C. (1993). Genetic analysis of circadian c1ocks. Ail/lU. Rev. Physiol. 55: 683-728. Edgar, B.A., Kiehle, C.P. and Schubiger, G. (1986). Cell cycle control by the nucleo-cytoplasmic ratio in carly Drosophila development. Cell 44: 365-372. Edgar, L.G. and McGhee, J.D. (1988). DNA synthesis and the control of embryonic gene expression in C. elegans. Ce1l53: 589-599. • 70 • Ewen, M.E., Sluss, H.K., Sherr, C.J., Matsushime, H., Kato, J.-y. and Livingston, D.M. (1993). Functional interactions of the retinoblastoma protein with mammalian D-type cyclins. Ce1l73: 487-497. Fang, F. and Newport, J.W. (1991). Evidence that the GI-S and G2-M transitions are controlled by different cdc2 proteins in higher eukaryotes. Cell 66: 731 742. Feldman, J.F. and Hoyle, M.N. (1973). Isolation of circadian clock mutants of Neurospora crassa. Genetics 75: 605-613. Friedman, D.B. and Johnson, T.E. (1987). A mutation in the age-l gene in Caenorhabditis e/egans lengthens life and reduces hermaphrodite fertility. Genetics 118: 75-86. Friesen, W.O., B1ock, G.D. and Hocker, c.G. (1993). Formai approaches to understanding biological oscillators. A/mu. Rev. Physiol. 55: 661-681. Gardner, G.F. and Feldman, J.F. (1980). The frq locus in Neurospora crassa: a key element in circadian clock organization. Genetics 96: 877-886. Gautier, J., Minshull, J., Lohka, M., Glotzer, M., Hunt, T. and Mailer, J.L. (1990). Cyclin is a component of maturation-promoting factor from Xenopus. Ce1l60: 487-494. Gautier, J., Norbury, C., Lohka, M., Nurse, P. and Mailer, J. (1988). Purified maturation-promoting factor contains the product of a Xenopus homolog of the fission yeast cell cycle control gene cdc2+. Ce1l54: 433-439. Gilbert, S.F. (1991). Deve/opmenta/ Bi%gy. Sinauer Associates, Inc., Sunderland. Glass, L. and Mackey, M.C. (1988). From docks 10 chaos. the rhythms of/ife. Princeton Paperbacks, Princeton. • 71 • Gould, K.L. and Nurse, P. (1989). Tyrosine phosphorylation of the fission yeast ede2+ protein kinase regulates entry into mitosis. Nutllre 342: 39-45. Greenwald, I.S. and Horvitz, H.R. (1980). wle-93(eJ500): A behavioral mutant of Caenorhabditis eleguns that defines a gene with a wild-type null phenotype. Geneties 96: 147-164. Hardin, P.E., Hall, J.C. and Rosbash, M. (1990). Feedback of the Drosophila period gene produet on circadian cye!ing of its messenger RNA levels. Nutllre 343: 536-540. Hartwell, L.H. and Weinert, T.A. (1989). Cheekpoints: contr~ls that ensure lhe order ofcell cycle events. Seience 246: 629-634. Hedgecock, E.M. and Russell, R.L. (1975). Normal and mutant thermotaxis in the nematode Caenorhabditis eleguns. Proe. Natl. Aead. Sei. USA. 72: 4061 ~\ 4065. Hess, B. and Plesser, T. (1979). Temporal and spatial order in biochemieal syste!îls. Ann. N. Y. Aead. Sei. 316: 203-213. Honda, S., Ishii, N., Suzuki, K. and Matsuo, M. (1993). Oxygen-dependenl perturbation of !ife span and aging rate in the nematode. J. Gemlllai. 48: 857 61. Huang, Z.J., Edery, 1. and Rosbash, M. (1993). PAS is a dimerizalion domain common to Drosophila period and severaI transcription faclors. Nature 364: 259-262. Johnson, T.E. (1987). Aging can be genetically dissecled into component proccsses using long-lived !ines of Caenorhabditis elegans. Proe. Natl. Aead. Sei. USA. 84: 3777-3781. • 72 • Kenyon, C., Chang, J., Gensch, E., Rudner, A. and Tabtiang, R. (1993). A C. elegans mutantthatlives twice as long as wild type. Nature 366: 461-464. Kimelman, D., Kirschner, M. and Scherson, T. (1987). The events of the midbla~tula transition in Xenopus are regulated by changes in the cell cycle. Ce1/48: 399-407. Klass, M.R. (1977). Aging in the nematode Caenorhabditis elegans: major biological and environmental factors influencing life span. Mechanisms Ageing Dev. 6: 413-429. Konopka, R.J. and Benzer, S. (1971). Clock mutants of Drosophila melanogaster. Proc. Natl. Acad. Sci. USA. 68: 2112-2116. Krek, W. and Nigg, E.A. (1991). DifferentiaI phosphorylation of vertebrate p34Cdc2 kinase atthe G Ils and G2/M transitions of the cell cycle: identification of majorphosphorylation sites. Embo. J. 10: 305-316. Kyriacou, c.P. and Hall, J.C. (1980). Circadian rhythm mutations in Drosophila melanogaster affect short-terrn fluctuations in the male's courtship song. Proc. Natl. ACl/d. Sci. USA. 77: 6729-6733. Liu, D. and Thomas, J.H. (1994). Regulation of a periodic motor program in C.elegallS. J. of Neurosci. 14(4): 1953-1962. Loros, J.J. and Feldman, J.F. (1986). Loss of temperature compensation of circadian period length in the frq-9 mutant of Neurospora crassa. J. Biol. Rhythms 1: 187-198. Masoro, E. (1985). State of knowledge on action of food restriction and aging. In: Genetic Effects on Aging, pp.5-39, Woodhead, A.D., Blackett, A.D. and Hollaender, A., eds. Alan R. Liss, New York, N.Y. • 73 • McClung, C.R., Fox, B.A. and Dunlap, J.C. (\989). The Neurospora elock gene frequency shares a sequence element with the Drosophila eloek gene paiod. Nature 339: 558-562. Meech, R.W. (1979). Membrane potential oscillations in molluscan "burster" neurones. J. Exp. Biol. 81: 93-112. Minshull, J., Golsteyn, R., Hill, C.S. and Hunt, T. (1990). The A- and B-type cyclin associated cdc2 kinases in Xenopus tum on and off at different times in the cell cycle. Embo. J. 9: 2865-2875. Murray, A.W. and Kirschner, M.W. (1989). Cyclin synthesis drives the early embryonic cell eyele. Nature 339: 280-286. Newport, J. and Kirschner, M. (1982a). A major developmentaltransition in early Xenopus embryos: II. Control of the onset of transcription. Cel! 30: 687-696. Newport, J. and Kirschner, M. (l982b). A major developmentaltransition in early Xenopus embryos: I. charaeterization and timing of cellular changes at the midblastula stage. Cel! 30: 675-686. Norbury, C., Blow, J. and Nurse, P. (1991). Regulatory phosphorylation of the p34Cdc2 protein kinase in vertebrates. Embo. J. 10: 3321-3329. Norbury, C. and Nurse, P. (1992). Animal cell cyeles and their control. All1lU. Rev. Biochem. 61: 441-470. Nurse, P. (1975). Genetic control of cell size at cell division in yeasl. Nature 256: 547-551. Ohtsubo, M. and Roberts, J.M. (1993). Cyelin-dependent regulation of G 1 in mammalian fibroblasts. Science 259: 1908-1912. • 74 • O'Rourke, B., Ramza, B. and Marban, E. (1994). Oscillations of membrane current and excitability driven by metabolic oscillations in heart cells. Science 265: 962-966. Pines, J. and Hunter, T. (1990). Human cyclin A is adenovirus ElA-associated protein p60 and behaves differently from cyclin B. Nature 346: 760-763. Ralph, M.R. and Menaker, M. (1988). A mutation of the circadian system in golden hamsters. Science 241: 1225-1227. Rapp, P.E. and Berridge, M.J. (1977). Oscillations in calcium-cyclic AMP control loops form the basis of pacemaker activity and other high frequency biological rhythms. J. Theor. Biol. 66: 497-525. Schierenberg, E. (1984). A1tered cell-division rates after 1aser-induced cell fusion in nematode embryos. Dey. Biol. 101: 240-245. Schierenberg, E. and Wood, W.B. (1985). Control of cell-cycle timing in early embryos of C. elegans. Dey. Biol. 107: 337-350. Schnabel, R. and Schnabel, H. (1990). Early determination in the C.elegans embryo: a gene. cib-f, required to specify a set of stem-cell-like blastomeres. Deyelopment lOS: 107-119. Sherr, C.J. (1993). Mammalian G 1 cyclins. Ce1l73: 1059-1065. Sulston, J. and Hodgkin, .J .. (1988). Methods. In: The Nematode Caenorhabditis elegans, pp.587-606, Wood, W.B., ed. Cold Spring Harbor Laboratory, Cold Spring Harbor. Sulston, J.E., Schierenberg, E., White, J.G. and Thomson, J.N. (1983). The embryonic celllineage of the nematode Caenorhabditis elegans. Dey. Biol. 100: 64-119. • 75 • Swenson, K.I., Farrell, K.M. and Rudermau, J.V. (1986). The clam elllbryo protein cyclin A induces entry into M phase and the resulllption of Illeiosis in Xenopus oocytes. Cel! 47: 861-870. Thomas, J.H. (1990). Genetic analysis of defecation in Caellorlwbditis elegalls. Genetics 124: 855-872. Tornheim, K. (1980). Co-ordinate control of phosphofructokinase and pyruvate kinase by fructose diphosphate: a mechanism for amplification and step changes inthe regulation of glycolysis in liver. J. 7'lleor. Biol. 85: 199-222. VanVoorhies, W.A. (1992). Production of sperm reduces nematode life span. Nature 360: 456-458. Vanfleteren, J.R. (1993). Oxidative stress and ageing in Caellorlwbditis elegalls. Biochem. 1. 292: 605-608. Vitaterna, M.H., King, D.P., Chang, A.M., Kornhauser, J.M., Lowery, P.L., McDonald, J.D., Dove, W.F., Pinto, L.H., Turek, F.W. and Takahashi, J.S. (1994). Mutagenesis and mapping of a mouse gene, Clock, essential for circadian behaviour. Science 264: 719-725. Vowels, J.J. and Thomas, J.H. (1992). Genetic analysis of chemosensory control of • dauer formation in Caenorhabditis elegans. Genetics 130:) 05-123. Wood, W.B. (1988). Introduction to C. elegans biology.ln: The Nematode Caenorhabditis elegans, pp.I-16, Wood, W.B., ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Walker, D.H. and Mailer, J.L. (1991). Role forcyclin A in the dependence of mitosis on completion of DNA replication. Nature 354: 314-317. Ward, S. and Carrel, J.S. (1979). Fertilization and sperm competition in the • nematode Caenorhabditis elegans. Dev. Biol. 73: 304-321. 76 • Yu, Q., Colot, H.V., Kyriacou, c.P., Hall, J.c. and Rosbash, M. (1987a). Behaviour modification by in vitro mutagenesis of a variable region of within the period gene of Drosophila. Nature 326: 765-769. Yu, Q., Jacquier, A.C., Citri, Y., Hamblen, M., Hall, J.C. and Rosbash, M. (1987b). Molecular mapping of point mutations in the period gene that stop or speed up biological docks in Drosopizila melanogaster. Proe. Nat/. Aead. Sei. USA 84: 784-788. Zar, J.H. (1984). The correlation coefficient. In: Biostatistical Analysis, Second Edition, pp.306-309. Prentice Hall, Englewood Cliffs, NJ. • Zerr, D.M., Hall, J.C., Rosbash, M. and Siwicki, K.K. (1990). Circadian fluctuations of period protein immunoreactivity in the CNS and the visual system of Drosophila. J. Neurosei. 10: 2749-2762. •,. 77 • Appendices Appendix 1. Heritability of the rate of embryogenesis Comparison of the embryonic developmenttimes of e25J9 hermaphrodites and a single offspring of that hermaphrodite. 179 pairs of numbers are presented. The upper number of e.1ch pair is the length of embryonic development for a given worm in 'hours:minutes' . The lower numbers correspond to the length of embryonic development of one of the former worm's progeny. • • 78 • Table Al. Heritability of the rate of embryogenesis 1 14:30 25 18:45 49 12:40 73 10:20 97 15:30 16:10 17:10 17:50 16:10 21:35 2 13:50 26 17:40 50 20:30 74 16:25 98 24:20 17:30 18:05 20:00 15:30 17:45 3 15:10 27 18:20 51 19:25 75 22:10 99 Il:45 17:05 17:30 16:10 18°35 14:50 4 14:50 28 19:30 52 24:30 76 17:30 100 18:20 18:00 17:50 19:00 10:40 15:50 5 15:30 29 13:00 53 14:25 77 19:20 101 13:35 12:15 17:10 18:15 17:40 17:50 6 12:35 30 15:10 54 22:20 78 13:45 102 22:30 17:50 16:30 17:20 19:30 17:00 7 13:20 31 24:00 55 17:35 79 22:15 103 24:15 16:50 15:50 17:15 17:10 18: 15 8 14:30 32 12:30 56 15:45 80 15:35 104 18:15 13:30 17:00 17:00 24:10 23:00 9 16:10 33 17:15 57 16:20 81 16:50 105 15:45 16:40 14:50 17:40 16:25 17:15 10 17:50 34 14:30 58 12:45 82 17:30 106 25:30 22:00 18:00 16:30 18:35 18:10 Il 19:00 35 l6:05 59 12:40 83 15:20 107 13:40 15:10 19:20 18:50 16:15 19:30 12 13:15 36 25:30 60 24:00 84 22:50 108 15:35 17:10 17:40 13:50 17:30 24:10 • 13 18:35 37 18:55 61 18:35 85 25:50 109 13:10 18:30 17:30 17:15 12:20 19: 15 14 17:40 38 19:45 62 16:30 86 21:40 110 25:10 Il:55 12:40 14:10 17:45 20:00 15 15:30 39 18:05 63 13:55 87 16:40 111 11:50 19:00 21:50 11:15 14:20 17:30 16 19:20 40 Il:00 64 17:25 88 13:10 112 21:00 14:30 13:10 15:40 17:10 18:40 17 22:15 41 17:00 65 26:10 89 17:20 113 16:00 19:30 17:30 17:00 13:50 13:50 18 16:30 42 15:40 66 21:15 90 17:20 114 14:25 18:25 15:30 16:35 17:40 14:30 19 24:30 43 24:25 67 17:30 91 11:30 115 22:45 17:10 16:50 17:00 18:30 17:10 20 14: 10 44 13:30 6~ 16:10 92 24:00 116 24:15 17:40 17:05 23:30 20:05 17:50 21 21:30 45 21:10 69 24:25 93 19:00 117 10; 1.î 16:50 18:10 17:30 15:45 16:45 22 Il:00 46 16:40 70 14:50 94 18:00 118 17: 15 15:30 14:10 18:40 11:40 15:15 23 17:10 47 24:20 71 19:10 95 14:30 119 26:30 16:20 17:50 14:20 °18:30 17:15 24 19: 15 48 18:45 72 18:30 96 :.'0:10 120 12:20 • 18:25 22:15 16:50 18:30 16:25 79 • 121 17:05 144 26:30 168 14:05 18:30 17:40 17:00 122 18:30 145 13:55 169 21:00 Il:00 17:50 23:05 123 13:30 146 21:30 170 15: 10 17:40 18:00 17:30 124 21:40 147 16:25 171 24:15 18:15 14:30 14:10 125 19:10 148 19:30 172 12:30 15:10 18:00 1~:40 126 Il:10 149 17: 15 173 22:00 19:00 17:15 16:50 127 20:30 150 18:30 174 26:50 16:20 15:10 18:00 128 16:15 151 14:20 175 13:50 14:15 18:30 17:10 129 17:10 152 16:00 176 17:50 17:00 12:15 12:30 130 24:55 153 12:00 177 20:00 17:50 15:50 15:50 131 15:35 154 17:50 178 18:50 18:00 13:20 17:30 132 17:20 155 14:30 179 19:10 • 24:00 17:10 17:00 133 14:20 156 22:25 16:45 14:00 134 17:20 157 24:15 17:00 17:35 135 18:30 158 18:00 18: 15 17:35 136 24:35 159 17:30 17:30 17:40 137 19:10 160 19:40 18:05 15:00 138 24:05 162 13:20 17:30 16:30 139 17: 15 163 13:25 17:10 17:10 140 16:45 164 25:30 17:00 16:15 141 12:10 165 21:50 22:00 18:30 142 18:35 166 20:30 17:15 18:00 143 15:05 167 25:05 • 11:15 16:10 80 • Appendix 2. Egg laying rate Comparison of the egg laying rate of N2, e2519, qm1 1, qm30 and maternally rescued e2519 animaIs. For N2, e2519. qmll and qm30, 20 hermaphrodites were placed on a plate and removed four hours later, (five hours in the case of qm1 1), at which time, the number of eggs on the plate was counted. This was repeated 10 separate times with different hermaphrodites and the results of each trial are shown in the second column. For maternally-rescued e2519 animaIs, a single worm was placed on a plate for four hours and upon removal, the number of eggs was counted. This was repeated 25 times and the number of eggs from each trial is shown in the fourth column of the table. • 81 Table A2. Egg laying rate • Genotype Number of Eggs Genotype Number o!" Eggs ,f N2 388504416496300 Matcrnnlly-rcscucù 181210169 646410554516572 e25/9 2523 15 1810 1520232123 e25/9 146162254124144 1917221721 122296312248344 1418201819 qlll30 70 76 87 75 89 62 79 82 59 68 qmll 230183179231294 179312251 162141 • 82 • Appendix 3a. Self-brood size Comparison of the number of progcny from N2, e2519, maternally-rescued e2519, qmll, qm30 and maternally-rcscued qm30 animais. L4 hermaphrodites were individually placed on plates and transferred daily to prevent overcrowding until the cessation of egg-Iaying. The progeny on the plates were scored three days following transfer of the parent. The number of progeny is shown in the second column of the table, each number corresponding to the number of animaIs from a single parent of the indicated genotype. Appendix 3b. Cross-brood size Comparison of the number ofprogeny from N2, e2519, qmlland qm30 animais when fertilized by N2 males. L4 hermaphrodites were placed individually on plates with three N2 males and the hermaphrodite and males were transferred daily to new plate. The males were replaced with new N2 males every two days until the cessation ofegg-Iaying. Progeny were scored three days following the transfer of the hermaphrodite. Each number in the second column represents the total number of cross-progeny from a single hermaphrodite of the indicated genotype. • .' H3 • Table A3I. Self·brood size Genotype Number of Progeny Genotype Number of Progeny N2 270 334 332 345 281 qmll 192192202217156 297 286 297 308 341 203204184196165 261357285292315 331 235 274 312 294 e2519 163171 193120174 qm30 14010510913438 155 156 165226184 12659 55 51 55 259 195 226 203 208 222154217227204 Matemally- 304305 550 297 372 Maternally- 29031431230835H rescued 312 269 580 304 351 rescued qm30 401 284278333308 e2519 257 305 364 332 314 337 261 249371 296 330319326279307 324 267 276 283 234 338384 Table A3II. Cross-brood size Genotype Number of Progeny Genotype Number of Progeny N2 485 725 785 750645 qmll 183 489 276445 495 871 801 554 602 508 473406452392375 734 723 596 763 345 e2519 347442485 322542 qm30 26022555 255466530496447 311431482569526 • 84 • Appendix 4. Defecation Comparison of the inter-defecation periods of N2, e2519, qml1, qm30, matemally-rescued e251 9 and maternally-rescued qm30 animais. The inter-defecation period is the length of time from the first muscular contraction of one defecation until the first muscular contraction of the next defecation. The first column indicates the number assigned the scored animais. The second column shows length of the first five inter-defecation periods, in seconds, for each worm. • N6 Genotype Fivc intcr-del'cciltion pcrim.ls Gcnotvpe Fivc inlcr-dt:ll.'\..'iuion pcripd:- e2519 ('2519 • ~8 1 766869 73 51 72 88 6! 73 65 2 6461698080 52 86908298 140 3 8373626771 53 72 6N 69 62 63 4 72 66 67 72 69 54 626260 (1159 5 656170115260 55 6563657361 6 605960122129 56 65 69 65 65 70 7 64 66 64 7562 57 72 76 65 77 67 8 6962606964 58 7771717067 9 6257535952 59 6566766265 10 5657565554 60 68 70 76 67 69 Il 6766686766 61 62 5N 56 54 62 12 6667656769 62 6764 63 68 64 13 6963626370 63 61 67575897 14 6263657064 64 5961595961 15 7569646467 65 6268696765 16 70 64 68 70 68 66 62 71 66 67 108 17 7676758765 67 677261 666N 18 70 66 68 65 103 68 56585861 61 19 827764114130 69 8263676565 20 6961 59 161 103 70 6363506191 21 6571 72 68 59 71 6862595763 22 6971686766 72 6061636454 23 64 66 66 66 67 73 6262625061 24 6361646060 74 64 76 64 63 65 25 6057515151 75 71 7470 65 83 26 6667656462 76 6362716162 27 6366627065 77 6965666662 28 7962676768 78 75 82 69 64 70 29 6861 667573 79 77 62 63 64 62 30 72 74 70 68 72 80 6264666964 31 7174667271 81 6356576353 32 6365606372 82 7686868076 33 5664716268 83 6163655657 34 5964 60 5361 84 6660666365 35 59 f;2 58 5259 85 596761 7059 36 6771686760 86 6863686364 37 6072 65 6269 87 6461637371 38 6475636670 88 6670696759 39 6958586063 89 66698873137 40 6857555757 90 59697170139 41 77 73 689866 91 7480787976 42 7379798077 92 6174666971 43 8373857676 93 767677 78 70 44 7380737075 94 86686874210 45 74908684 146 95 7166616969 46 60707170136 96 6262664462 47 64 63 75 68 70 97 6269696567 48 5962655958 98 6774756467 49 6461647271 99 7071716969 • 50 696964 64 67 100 7072687171 87 Genotype Pive intcr-dcfccution pcriods Genotype Five inlcN!rrccation pcriods lfmll 911130 1 8586868388 1 78 97 III 78 87 • 7077777674 2 7563989395 2 3 69656265142 3 r9 809485 87 4 7672 69 70155 4 14172 64 86 89 5 7158697271 5 12676818588 6 9571 767471 6 977077 94 97 7 7269677165 7 76 138 90 76 &3 8 8080747677 8 1137899102100 9 8270787165 9 897812285102 10 7365656570 10 12810490104100 Il 7976798573 Il 98 103 96 99 102 12 89788371 102 12 88989391100 13 87799790120 13 9396908282 14 8388967982 14 6063686983 15 8081 888583 15 99256115 176108 16 8789688371 16 93112135 103 115 17 7880817986 17 8375757872 18 7072 92 72 68 18 72 74 75 99 87 19 705972 73 75 19 74817576104 20 75 86 81 5074 20 9790868078 21 817679 86 ~7 21 83 72 98 88 108 22 75 84 87 81 88 22 75868294175 23 7069737179 23 77 89 82 95 85 24 6969696872 24 8183877080 25 78 ','9 82 70 73 25 9088917784 GenotYpe Five intcr-dcfccation pcriods Genotype Fivc intcr-dcfecation periods Matcmully-rcscucd e25/9 Matemally-rescued qm30 5745747665 1 6262656661 2 5050515156 2 6868696462 3 6768597477 3 5866746363 4 6162636559 4 7070737669 5 6262637374 5 7968636466 6 5454526447 6 62707176100 7 99818851119 7 999494109153 8 657163101 119 8 5858606161 9 5152444862 9 6363585962 10 4952484546 10 5457535560 Il 5051414847 Il 6673747175 12 44 44 47 49 53 12 7374756871 13 5052524854 13 6565667658 14 7349505455 14 6263636469 15 5151524849 15 64 64 64 68 63 16 6768686577 16 5254535357 17 6162625759 17 6667687172 18 4849495045 18 6465656669 19 6161698089 19 6969696668 • 20 6965606168 • Appendix 5. Duration of DMP Comparison of the duration of DMP for N2, e25/9, qm/ / and qm30 animais. The DMP is made up of three well-characterized muscular contractions and the length of time from the beginning of the first muscular contraction (aBoc) untiJ the expulsion of waste (Exp) was timed in three consecutive defecations for each animal. The duration of each of these DMPs is shawn in the second column. Twenty worms of each genotype were scored for the duration of the DMP. , • 89 Table AS. Duration ofDMP • Gcn 1 4.404.524.91 1 5.27 5.35 5.30 1 5.57 5.866.09 1 5.175.284.93 2 4.504.874.89 2 5.104.834.70 2 5.045.405.31 2 5.32 5.204.88 3 5.23 5.15 5.23 3 4.78 4.99 4.97 3 6.296.896.53 3 3.924.303.79 4 4.864.45 5.16 4 5.21 4.594.99 4 5.845.65 5.58 4 5.394.965.18 5 4.374.875.11 5 5.124.905.00 5 4.765.305.01 5 3.873.57 3.93 6 5.07 5.32 5.08 6 4.78 4.49 4.40 6 5.68 6.38 5.50 6 3.48 3.76 3.65 7 4.805.194.94 7 ...804.524.63 7 5.44 5.67 5.54 7 3.203.91 3.02 8 4.884.745.G3 8 4.31 4.524.08 8 5.41 5.205.34 8 3.203.27 3.44 9 5.095.365.41 9 4.984.904.75 9 5.05 5.31 5.41 9 3.563.543.41 10 4.864.925.38 10 4.86 4.27 4.57 10 5.87 6.13 5.65 10 3.243.10,.24 Il 5.21 5.474.96 Il 4.75 5.06 4.23 11 5.55 5.84 5.37 11 4.98 4.98 4.86 12 5.085.145.08 12 4.025.004.58 12 5.21 4.934.86 12 5.505.21 5.37 13 4.965.044.66 13 4.72 4.62 4.81 13 4.494.674.78 13 4.404.184.31 14 5.225.345.20 14 4.57 5.51 4.87 14 5.365.765.48 14 5.205.41 5.50 15 4.664.864.73 15 4.375.084.54 15 5.896.116.31 15 3.23 3.68 3.71 16 5.005.154.97 16 5.345.305.05 16 5.77 5.415.36 16 4.214.744.83 17 5.275.075.25 17 5.205.57 5.52 17 5.265.465.81 17 5.41 5.635.58 18 5.074.754.98 18 4.174.914.60 18 5.17 5.624.97 18 5.865.65 5.76 19 5.154.824.93 19 5.074.544.75 19 4.985.184.86 19 4.373.984.15 • 20 4.894.905.21 20 5.265.07 5.44 20 5.434.745.36 20 3.884.183.91 • 90 • Appeodix 6. Pumping Comparison of the of number of pumping cycles pel' minute between N2. e2519, qmll, qm30 and maternally-rescued e25/9 animaIs. The number of pumps pel' minute was counted by scoring every rive pumps because of the very high frequency of pumping, hence, 5 Cycles/minute. Each animal was scored twice, and these two numbers are shown in the second column. • • 91 Tllble A6. Purnping • Genntype 5 Cycles! Genotype 5 Cycles! Genotype 5 Cycles! Genotype 5 Cycles! Genotype N2 Minute e25/9 Minute qm!! Minute qm30 Minute Maternally-rescue, .25!9 1 4850 1 4036 1 3430 1 3834 1 2 4750 2 4238 2 3038 2 4345 2 3 5150 3 2831 3 3034 3 2326 3 4 5356 4 2830 4 3533 4 3534 4 5 5348 5 41 36 5 3833 5 3638 5 6 64 56 6 3235 6 3127 6 2829 6 7 5143 7 3033 7 3028 7 3742 7 8 4851 8 3230 8 3428 8 3027 8 9 4947 9 3941 9 4039 9 3236 9 10 5155 10 2624 10 4948 10 3636 10 Il 3740 Il 3028 Il 2528 Il 3437 II 12 5654 12 2326 12 3331 12 3332 13 5754 13 3840 13 3529 13 3741 14 5453 14 3330 14 3433 14 3643 15 4952 15 2823 15 3631 15 3431 16 6362 16 3834 16 4133 16 3940 17 5854 17 4538 17 3233 17 3335 18 5653 18 3227 18 2933 18 2928 19 5250 19 3027 19 4341 19 3035 20 5350 20 2725 20 41 43 20 3635 • 21 5555 21 2425 21 31 29 21 4143 22 5046 22 3031 22 2926 22 2427 23 5553 23 31 25 23 3428 23 3335 24 5450 24 2428 24 3436 24 2322 25 4846 25 3431 25 3235 25 3735 • 92 • Appendix 7. Swimming Comparison of the rythmic swimming mo:ions of N2, e2519, malcrna\ly rescued e2519, qmll and qm30 animaIs. The swimming motions of the worms werc scored while the worm was in a drop of M9 bul'fer. Eaeh worm was scored once for number of swimming cycles pel' minute. • • 93 Table A7. Swimming • Genotype CycieslMinute Genotype CycieslMinute N2 120120135125131 qmll 84928174101 122125123118119 89 79 86 87 103 121 122124114109 84938192110 117111120105120 9687929486 131 126120121 118 6994887967 "2519 95999297 100 qm30 697481 7882 94899186100 7768788169 77 85 949183 7376837073 9395879293 767581 6576 91 87859998 7182807478 Maternally- 118102108121111 rcscucd 104128130121 119 "2519 105140106137118 141 113102121 • • 94 • Appendix 8. Temperature effects on defecation Comparison of the inter-defecation periods of N2 and e25/9 animaIs raiscd at three different temperatures, and scored for length of inter-defecation period at 20°C. N2 and e25/9 worms were raised at 15°C, 2QoC and 25°C. These animais werc then scored for defecation at 2QoC, as described previously. Twenty-five worms wcrc scored for each genotype at each of the three indicated temperatures and five inter defecation periods are shown in column two. For the inter-defecation periods of worms raised at 2QoC, see Appendix 4. • • 95 Table AS. Temperature effects on defecation • Genolype Five inter-defeealion Genotype Fi ve inter-defeeation Raised al 15"C periods at Raiscd at 15'C pcriods al N2 20'C e25/9 20'C 1 5754565355 1 6773737476 2 55565457 107 2 77 78 78 80 87 3 5656624855 3 8183848580 4 5857565761 4 87986882112 5 5455575959 5 91988384100 6 5753545551 6 7375767880 7 5750524848 7 8086798291 8 4547494650 8 9494949098 9 4646474749 9 8793848586 10 48 55 67 6046 10 747471 7387 Il 4852505049 11 767677 77 80 12 50515147103 12 7373726681 13 4848505145 13 838672 7779 14 5055535764 14 8990509393 15 5354504766 15 129119131 114114 16 4446535960 16 8082848896 17 5151515564 17 8893100103109 18 5152524950 18 858972 79 147 • 19 485051 5449 19 7175757677 20 5052495453 20 818277 79 86 21 4550475049 21 90909987 100 22 5054555955 22 8876807678 23 5858565748 23 7475757575 24 4959545551 24 9194967888 25 50 50 55 60 72 25 7169747464 • % Genotype Five intcr-dclccutÎon Genotype Five intcr-dclccation • Raised at 25"C periods at Raiscd at 25°C pCl'iods al N2 20"C e2519 20"C 1 5051 525249 1 5858596366 2 5454544353 2 716364 66 68 3 5454505155 3 7575757677 4 5555554957 4 6565636661 5 4752545558 5 6162656758 6 5759616670 6 5455606364 7 5959595958 7 6061626467 8 4852535458 8 5071717274 9 6062575859 9 6061666773 10 51 52535354 10 6970707073 Il 31 53575860 Il 6768686962 12 4949515546 12 71 7364 65 66 13 5555525261 13 6168595965 14 5456535758 14 6262576764 15 5656565253 15 7171737668 16 5256575858 16 6568656971 17 5053565859 17 6268655970 18 5053545454 18 9468697182 19 5354555759 19 6266676859 • 20 5354555860 20 7071746468 21 5051 535655 21 7072 74 65 69 22 5052555658 22 5566636971 23 5254576061 23 64 65 69 69 87 24 5555535457 24 747677 86 89 25 5454535350 25 6366676769 • 97 • Appendix 9a. Phenotypes of fast embryos N2 and e25/9 embryos that were selected on the basis of embryonic developmental time. A group of 29 embryos whose developmental times c1ustered c10sely around 12 hours were selected. The post-embryonic developmental time, and later the adult defecation period of these animais was scored. The numbers given correspond to the duration of embryogenesis and post-embryonic development in hours, and five inter-defecalion periods, in seconds. The post-embryonic development of four of the 29 embryos was not measured. • Appendix 9b. Phenotypes of slow embryos N2 and e25/9 embryos that were selected on the basis of embryonic developmental time. A group of 29 embryos whose developmental limes c1ustered c1oselyaround 14 hours for N2 and 22 hours for e25/9 were selected. The post embryonic developmental time, and later the adult defecation period of these animaIs was scored. The numbers given correspond to the duration of embryogenesis and post-embryonic development in hours, and five inter-defecation periods, in seconds. The post-embryonic development of four of the 29 embryos was not measured. • • T8ble A9I. Phenotypes of fast embryos N2 e2519 Duralion of Duralion of Fivc intcr- DuratÎol1 of Duratinn or l'ive intcr- cmbryogcncsis post- dcfccation cmbryogcncsis post- dcl'ccatioll cmbryonic pcriods clnhryonic pcrillds dcvclopmcnl dcvclopmcnt 1 12 49 4951494953 13 72.5 7ll65'71797ll 2 13 46.5 4596484154 13 77.5 71776974128 3 12 51 5151444746 12 73 67 70 62 69 1:16 4 13 48.5 5144474649 13 79 75757273 Ill9 5 11.5 48.17 5454515053 11.5 73 75767679 115 6 Il 48.5 53 51 57 50 50 Il.42 76.33 74 75 74 8073 7 12.75 46.75 5252494953 11.42 78.33 6772 78 7167 8 1\ 51.17 5251 534849 11.5 77.83 75 75 7668 7ll 9 12.58 50.67 53485651 54 Il 80 737493121 124 10 12.75 49.75 4748495050 Il.42 67 6161626364 Il 12.67 49.25 48484851 52 10.58 68.67 585564 46 61 12 12.75 50.33 5052545455 Il.42 72.67 71 71 71 6974 13 12 49.17 9950515247 11.33 7ll.75 59 60 62 62 64 14 11.33 48.67 495051 5859 11.33 71.25 57585961 62 • 15 11.25 48.5 5053555699 11.5 77.5 64 65 6669 Isll 16 11.5 52 64 57 57 54 56 12.33 73 66697072245 17 1\ 47 4952545848 11.67 75.5 6ll 6ll 61 6366 18 12.25 47.5 5757545649 12.33 72 6ll 60 62 63 58 19 12.5 49.5 64 64 58 58 50 12.5 72 60 6ll 54 58 64 20 12.25 49 5353525866 10.67 80.5 63 63 60 64 fJ7 21 12.5 51 4747505145 10.75 78 6262616459 22 12 52.5 3637373839 1\ 79.5 64 64 6ll 62 65 23 1\.75 48.5 4547485051 12.5 70 5959 61161 67 24 11.5 53 5252555658 12.25 81 68786579 % 25 12.25 51 6163525542 1l.25 76.5 5758595462 26 12 6064 68 55 58 12 7374758384 27 12 5353545449 1\.25 69 69 62 78 8ll 28 12.5 51 52525660 12 6669717563 29 1l.75 4749495062 11.5 806272 75 68 • 99 • Table A9I1. Phenotypes of slow embryos N2 e2519 Ouralion of Ouralion of Fivc intcr- Ouralion of !Juralion of Five intcr- cmbryogcncsis post- defeealion embryogenesis posl- defeealion embryonie periods embryonie periods developmenl development 1 14 48 455351 4549 24 61.5 6259606159 2 15 45.5 4947484946 24 69 5864576060 3 15 49.5 4344454347 23 67.5 67595763 125 4 14 48 4848494445 23.5 63.5 5764665763 5 14.75 45.17 4950515254 22.5 63.75 5354555657 6 14 46 4949515157 21.75 64.5 6163646566 7 13.67 45.33 4647444646 22.17 70 5656595953 8 14.5 44.5 5050535351 22 69 6162636459 9 14.83 45.5 4747484351 23 69.75 5858596566 10 15 48.83 4343456389 22.67 65.5 5354555658 Il 14.75 48 4346494037 22.92 61.33 5859595961 12 14.5 50.75 5050504652 23.17 60.5 5356505765 13 14 50.75 5455565748 23.83 59.67 5859606163 14 14 51.l7 4747474445 23 54.5 5254565555 • 15 13.75 47 4951 535354 23 65.5 61 55565657 16 14.33 46.5 5051515151 23 70.5 5161575860 17 14.75 46 4951525342 23 68 6262636469 18 15 45 4949515354 22 67.5 5555555256 19 14.5 47.5 50 52 52 48 113 22 60.5 5759525455 20 15 47 4848505162 22 63 505051 5052 21 14.75 44.5 4749494551 22.33 60 4747484950 22 14.25 48 5656575052 22.5 62.5 6049494550 23 14 48.5 51 53554549 22.5 68.5 4748494950 24 13.5 45 4546495156 24 59.5 4546475051 25 14.33 49 4544535445 23.75 63 5051515237 26 14 5151505246 22.67 5253534448 27 13.75 4849535357 22.25 51 52525249 28 14.33 5052524356 23.5 6264 64 58 73 29 14.5 44 46 56 49 52 22 45474961108 •