Abstract:
Circadian clocks enhance the chances of survival of organisms living under periodic
environment by enabling them to efficiently anticipate periodic events in their
environment, because precisely and appropriately timed behavioral and metabolic
processes are thought to confer greater adaptive advantage than randomly occurring
activities. Among clock regulated phenomena in insects adult emergence (eclosion)
rhythm is one of the most extensively studied and perhaps the best understood after
activity/rest cycle. Although each individual emerges as an adult only once in its life
cycle, gating of this event is under the control of an on-going oscillation present during
development (Saunders, 1992). Consequently, certain intervals of time in a day constitute
the “forbidden zone of eclosion”, whereas a brief period of time during which adults
emerge forms the “allowed zone” (also referred to as the “gate” of eclosion) (Pittendrigh,
1954). The gating is often so stringent that even if developing flies are mature enough to
emerge but fail to do so during the gate they remain within the puparium until the next
gate opens (Saunders, 1992). My thesis is motivated by the need for an unequivocal,
systematic and rigorous approach for studying the evolution of circadian waveform of
adult emergence in fruit flies Drosophila melanogaster. The main aim of my thesis is to
study the effect of selection on timing of adult emergence on circadian clocks, clockrelated
rhythm and life history traits. For this purpose I derived four genetically
independent, random mating, large populations each of early and late populations of D.
melanogaster by selecting for individuals that emerge during “lights-on” (morning hours)
and “lights-off” (evening hours) under 12:12 hr LD cycles. I assessed the direct response
to selection by comparing the number of flies that emerged out of the morning and
v
evening windows of selection in the selected and control populations at regular intervals
of 10-15 generations to trace the evolutionary trajectory of changes over 55 generations
of selection. I observed that D. melanogaster populations respond to selection on timing
of adult emergence, since the percentage of flies that emerge out of the morning window
after 55 generations of selection is about 60% in the early populations while it is reduced
to almost one third (~33%) in the late populations, and remain unchanged at ~45% in the
controls. The percentage of flies that emerged out of the evening window in the late
populations (~24%) is about thrice as much as in the early populations (~8%), while it
remain unchanged at ~16% in the controls. To investigate the consequence of selection
on circadian clocks we assayed eclosion and activity rhythms in the selected and control
populations under LD and constant dark (DD) conditions. Under 12:12 hr LD cycles, the
primary peaks of eclosion in the early populations are taller and occur earlier than the
controls, while those in the late populations are relatively flatter and occur later than the
controls. The early flies start and end activity earlier and are generally more active during
the morning hours, while the late flies start and end activity later and are more active
during the evening hours. In order to test the robustness of the circadian phenotypes of
the selected populations we assayed the eclosion and activity rhythms under short (8:16
hr), normal (12:12 hr) and long (16:8 hr) photoperiodic regimes at the 70th generation.
Although the overall eclosion and activity patterns of flies are influenced by the
photoperiodic conditions, the relative phase separations between the selected and control
populations are maintained; the time course and waveform of the early populations
remain phase advanced relative to the controls, while those of the late populations are
phase delayed relative to the controls. Consistent with the rhythmic expression of the
vi
selected populations under LD cycles, even the circadian periodicity of eclosion and
activity rhythms under DD condition is significantly longer in the late populations
compared to the early populations. Such alterations in circadian phenotypes, borne out of
heritable changes in genetic architecture in response to imposed selection pressure (and
not because of random genetic drift or some unknown environmental or non-genetic
effect) suggest that the time course and waveform of adult emergence and activity
rhythms in D. melanogaster evolve as correlated responses to selection on the timing of
adult emergence.
We further investigated the consequence of selection on life history traits such as
pre-adult developmental time and adult lifespan in unmated (virgin) flies from the
selected and control populations. The development time of the selected populations was
altered as compared to the controls; the early populations develop faster than the controls
under LD as well as DD conditions whereas the late flies take longer to develop under
both the regimes. This suggests that selection for early and late adult emergence causes
correlated change in the duration of pre-adult development. The lifespan of virgin flies
from the selected populations depends upon the timing of their emergence. For example,
the morning emerging early flies live longer than those that emerged during the evening,
while the evening emerging late flies live longer than those that emerged during the
morning. This, to the best of my knowledge is the first study of its kind demonstrating
that morning emergence is adaptive for early populations while evening emergence is
adaptive for the late populations.
In the last part of my thesis I have described studies on the molecular clocks of
the early, control and late populations. I have assayed the levels of transcripts of clock
vii
genes period and clock (per and clk) in the selected and control populations. Consistent
with the circadian phenotypes, the peak of per mRNA in the late populations is phase
delayed by about 4 hr compared to the early and control populations. Similarly, the peak
of the clk mRNA in the late populations occurs about 4 hr later compared to the early and
control populations. Besides the transcript levels even the ratio of spliced to unspliced
variants in the per 3’ UTR is significantly altered among the selected populations; the
early populations display an overall suppressed spliced form of per throughout the day
with a peak at ZT14 (2 hr after lights-off), whereas the late populations have generally
higher levels of spliced form of per with a peak at ~ZT20. In other words, the spliced
form of per peaks early in the early populations and later in the late flies. These findings
are consistent with earlier studies that implicated role of per splicing in the determination
of the phase of evening activity peak (Collins et al., 2004; Majercak et al., 2004).
Although the time course and waveform of the early and control populations
appear to have diverged from each other and from the late populations, they do not differ
from each other with regards to their free running period (τ), which raises a conundrum
as to how clocks with similar τ could result in different circadian phenotypes under LD
cycles. This is possible only if their clocks are differentially sensitive to light stimuli.
Indeed the results of my experiment reveal that, the early populations undergo greater
phase advance in their eclosion and activity rhythms in response to brief light stimuli
administered during the late subjective night (Circadian time 20, CT20) and a modest
delay during early part of the subjective night (CT14), whereas the late populations
display exactly opposite response at these two phases. Given that circadian clocks
underlying eclosion and locomotor activity in the selected and control populations
viii
respond differently to brief light stimuli presented at CT14 and CT20, we decided to
estimate the TIM levels at CT14 and CT20 to investigate the state of the molecular clock
of the selected populations following a brief exposure to light. The late flies show drastic
reduction in TIM levels at CT14 associated with large phase delays, while the early flies
show lesser reduction in TIM levels associated with smaller delays, whereas at CT20, the
light-dependent degradation is found to be greater in the early populations compared to
the controls. Interestingly, the late populations show similar reduction to those of the
early populations at this phase suggesting to a specific role for photoreceptor molecule
(CRY) in the late populations. However, the role of CRY is yet to be ascertained in our
selection lines.
The results of the experiments described in my thesis suggest that fruit flies D.
melanogaster evolve morning and evening preference for adult emergence and activity as
a result of periodic selection pressure imposed on the timing of adult emergence. As a
consequence the early populations develop faster than the late populations and morning
emergence becomes adaptive for the early populations and evening emergence for late
populations. These studies also underscore the significance of timing of rhythmic
phenomena for organisms living under periodic environments, and suggest a possible
mechanism by which circadian rhythms may have evolved and/or fine-tuned by periodic
biotic and abiotic factors of nature.