Jaurnal of
@
Insect Physiology
Pergamon
Journal of Insect Physiology 49 (2003)'17'l-784 www.elsevier.com/locate.,J
insphys
Selection for high diapause incidence in blow flies (Calliphora vicina) maintained under long days increases the maternal critical daylength: some consequences for the photoperiodic clock D.S. Saunders u.*, B. Cymborowski '
b
Institute oJ'Cell, Aninal and Population Biologt, university of Edinburgh, West blains Roael, Edinburgh EHg 3JT, Scotlarul, b Department of Invertebrute Physiologt, V[/arsav'(Jniversity, I lvtiec:nikowa Steet,02-096 llrarczawa, Pofuzrul
t]K
Received 6 February 2003; accepted 15 April 2003
Abstract Using a population of Culliphora uicirra from southern Scotland (55'N), showing a critical day length for maternal induction of diapause of about 14.5 h per day, strains of flies were selected for a high incidence of lan'al diapause under long day length (LD 16:8 h). Diapause incidence was raised from under 109'o to almost 100% within fir,e or six generations. The selected flies showed an increase in their critical day tength to over 16 h per day, and a high incidence of larval diapause under very long photophases. Selected flies, however, showed mean circadian periods for locomotor activity little different from the original stock, or from flies selected for high diapause under LD 12.12 h, and a Nanda-Hamner profile lacking peaks anrl troughs of diapause incidence at about 24 h inten'als. These results are interpreted to shorv that the photoperiodic regulation of diapause and the control of overt behavioural rhythmicity are 'separate' physiological systems. O 2003 Elsevier Science Ltd. Ail rights resened. Keyw'ords: Calliphora vicina; Diapause; Photoperiodism; Selection; Locomotor rhythms: Nanda-Hamner responses
1. Introduction
A
common strategy for ovenvintering arnong the
insects is a photoperiodically regulated diapause in the life cycle (Saunders, 2002). The blow fly, Calliphora vicina, is typical in this respect except that the diapause
occurs
in the
larvae but photoperiodic sensitivity is
rnatemal. Adult females exposed to autumnal short days
(or long nights) thus lay eggs giving rise ro larv-ae that enter a shallou, diapause at the end of the third instar after they burrow into the soil. On the other hand, flies experiencing the long days (short nights) of summer give rise to progeny that continue to the next generation without arrest- The maternal critical day length (CDL) for a population of C. vicina isolated in southern Scotland was about 14.5 h per day (Saunders, 1987). Much is norv knorvr About the formal properties of
-
Corresponding author. Tel.: +44-13l-650-555,I; fax: +44-l3l-
664-r 8 r9.
E-ruail address :
[email protected]
(D.S. Saunders).
0022-1910/03/$ - see front matler O 2003 Elscvicr Science Ltd. doi: I 0. I 0l 6/S0022-i9 I 0(03)001 I 4-8
All rights
photoperiodisrn, and of the endocrine regulation of the diapause it controls (l)enlinger, 1985; Saunders. 2000). However, the nub of the problem-photoperiodic tirne measurement (PPTM)-remains obscure and somewhat controversial (\reennan, 2001). On one hand there is a
substantial body
of
evidence
to
favour Biinning's
hypothesis, first proposed in 1936, that photoperiodic time measurement is a function of the insect's circadian system (Vaz Nunes and Saunders, 1999 Saunders. 2002). On the other hand there is a body ofevidence to suggest that night length measurement is accompiished by an hourglass-like timer quite distinct from the circadian system (Lees. 1973; Saunders.2002). Evidence both for and against involvement of circadian oscillators in insect PPTM conles, in paft, from the use of experimental protocols using light-dark cycles whose overall periods differ from 24 h. The most wideiy used of these are so-called Nanda-Hamner or 'resonance' experiments (Ilamner, 1960) in which insects are exposed to cycles containing a short light phase (say 10-12 h) coupled,.in different experimental sub-sets. to a variable dark period (say 8-72 h) to give light dark cycles spanning several rcserved.
D.S. Saunders. B. Cymborowski /
Jourtal
circadian periods up to 84 h or more. In typical 'hourglassJike' responses, short night eflects (e.g. a low incidence ofdiapause) are observed until the duration ofthe dark period exceeds a critical night length; then long night responses (e.g. a high incidence of diapause) take over. Conversely, responses suggesting the iuvolvement
of the circadian system in
PPTM show
a similar
response to the shorter night lengths, but periodic peaks (and troughs) of diapause incidence, with a periodicity
close to 24 h, once night length exceeds the critical value. The controversies associated with the now voluninous literature on this subject are reviewed elservhere (Vaz Nunes and Saunders, 1999; Saunders,2002;
in preparation) and will not be detailed here' it to say, however, that one view ofthe apparent
Saunders,
Suffice
dichotomy between 'positive' and 'negative' responses might be that the photoperiodic oscillator is frequently heavily damped. In this view, extreme damping rvould equate with an hourglassJike timer. Experimental evidence for a darnping circadian oscillator in C. vicina's photoperiodic clock suggests that it occurs after about five cycles in continuous darkness (Vez Nrules et a1., 1990 a). In species with a large south-north distribution' geographical clines occur with respect to several diapause,
photoperiodic and circadian traits. In photoperiodism, critical day length increases to the north (Danilevskii, 1965; Lankinen, 1986; Saunders. 2001) but NandaHamner inter-peak intervals may shorten fv"az Nunes et al", 1990 b) or even disappear to form an hourglass-like response in more northerly populations (Thiele, 1977). Similarly, in circadian eclosion rhythms, the rhythm's period (z) frequently shortens to the north (Lankinen. 1986) or eclosion activiff becomes arrhythmic after a few days in constant darkness (Lankinen, 1986). An inverse correlation between circadian period and CDL rvith increasing latitude is what would be expected from circadian theory if a causal relationship exists between the trvo (see below). Horvever several authors. including Larnkinen ( 1986), have observ'ed a tight correlation betrveen CDL and latirude, and betrveen r and latitude, but a much looser correlation betw'een CDL and r. This rather undermines a clear causai relationship between PPTM and the circadian system-at least with that part of the circadian system controlling overt behavioural rhy,thms.
There are, in fact, a number of seemingly firndamental differences betrveen PPTM and overt rhythms that suggest that the tw'o systems are 'different'. In C. vicina. for example, r for the adult locomotor activity rhythm is clbse to 22.5 h (Kenny and Saunders, l99l; Cyrnbolorvski et ai..1993; I{ong anc{ Saunders,1998), rvhereas the inter-peak intervals in Nanda*Hamner expefiments are close to 24 h (Saunders, 1997). In
addition, the locomotor activity rhythm persists undamped into the fly's old age. whereas the oscillator
oJ'
Insect Physiolog, 49 (2003) 777-784
apparently involved in PPTM damps out within five cycles in darkness (Vaz Nunes et al., 1990a). The two aspects of time measurement, therefore, although both considered here to be products of the insect's circadian system and operating simultaneously in the adult fly,
in period and persistence have a different physiologistrongly suggesting that they show profound differences
cal origin. \fo'hat, however, does 'different' in this context actually mean? It could mean that the two processes ate carnpletely different, locomotor activity being rhythmic (circadian) rvhereas PPTM is non-rhythmic (hourglasslike). Or it could be that both are circadian based. but one (PPTM) involves a system of damping oscillators with a separate cellular location and even a different biochemical substrate. In this paper, these differences are addressed. in C. vicina, by the artificial seiection ofone trait, critical day length, by selectively breeding for a high incidence of diapause frorn flies maintained under long days. Strains with an increased CDL, rather like narurally occurring populations from a higher latitude, are then examined for changes in locomotor rhythm period, and for changes in the Nanda-Hamner response. Using this approach it was hoped that differences and silnilarities between the various aspects of time measurement could be further dissected, perhaps adding a neu' perspective to Biinning's hlpothesis.
2. Ilaterials and methods 2.1. Flies and uilture metlrcds The blow flies (Calliphora vicina R.-D.) used in this investigation rvere drawn from a laboratory population originating fiom flies captured near Edinburgh, Scotland (55 "N) in the late 1980s. The adult flies were maintained in small (26 x 21 x 19 cm) garlze covered cages held in a constant temperature room at 25 oC under continuous illumination. The flies rvere provided with water and sugar ad libihmr, and with a daily suppiy of fiesh meat (beef muscle) from about day 4 post emergence. Eggs laid on the meat were used to establish larval cultures using a further piece of meat supplemented with a lan'al medium made from dried milk, yeast and agar. Mafure larvae were allowed to disperse into dry sawdust to form puparia. Flies emerging from these puparia were then used to establish the next generation of flies. Experimental cultures of flies were kept in a similar
fashion,'but the cages were enclosed
in
lighrtight
rvooden boxes (37 x 36 x 31 crn inside diameter) held oC. Each in a rvalk-in constant temperature room at 16 box rvas provided with a 4 W fluorescent strip iight. con-
trolled b-v a timer to give a light cycle of LD 16:8 h. Eggs deposited on days lG-12 were allowed to hatch at
D.S. Saunders, B. Cymborowski / Journal of Insect Physiolog' 49
16 "C and the hatchlings used to establish larval cultures oC. (see above) in constant darkness at ll-12 this lower
temperature being necessary for the full expression of lanal diapause (Vinogradova and Zinor.jeva, 1,972; Yaz Nunes and Saunders, 1989) should it occur. Newly forrned puparia were collected daily from these cultures to establish cumulative pupariation records. Further generations of ffies were established by selecting those individuals showing delayed pupariation (see below). Diapause destined and diapausing larvae of C. vicina show no overt signs of their physiological state other than delayed pupariation; it is only on this criterion that they may be distinguished frorn their nondiapausing siblings. In this and earlier papers, therefore, diapause was considered to have begun in larvae failing to form puparia by day 30 (at 1l*12 "C). Support for this arbitrary age is afforded by observations that haemolymph ecdysteroid titres fall to very low levels, and the larval ring gland becomes unresponsive to exogenous PTTH, by this time (Richard md Saunders, 1987). 2.
2. Selection procedttre
Selection for a high incidence of larval diapause u,as initiated using 'parental' populations of flies maintained at 16 "C under long day length (LD 16:8 h), conditions that normally induce less than about 10% of diapause (Saunders, 1987). In each generation thereafter, the last larv'ae to form puparia, or those that remained lan'ifonn at the end of their cohort. were separated from the rest and used for onward breeding. Such selection was continued for six generations in three independent popuiations. From generation three, however, one ofthe lines was split, giving rise to four lines of high-diapause flies from this generation onwards.
(200i) 777*784
779
by computer and assembled into the conventional 'double-plotted' actogram format (Kenny and Saunders. 1991; Cymborouski et al., 1993). Activity records were usually continued for about 20-25 days. Free-running circadian periods (r) were calculated for each fly, using periodogram analysis (based on Enright, 1965, and Williams and Naylor, 1978) or other specifically written software (Saunders et al., 1994), as appropriate. Mean values ofc were calculated for each selected line offlies for different sections of their respective actograrns (i.e. for the initial, middle or later stages of their activity record).
2.4. Nanda-I lanmer etperiments Cages of adult flies were exposed to a series of lightdark cycles each containing a 13 h photophase coupled, in different experimental sub-sets. to dark periods ranging from 5 to as many as 67 h; this provided a series of photocycles (Q ranging from 18 to 80 h, thereby covering three circadian cycles. Lawae hatching from eggs laid on days 10-12 were then used to establish cultures at ll*12 "C under continuous darkness to determine the incidence of larval diapause. Diapause incidence was ascertained on day 30 ofeach larval culture, the arbitrary age at which diapause is considered to begin (Richard and Saunders, I 987), and again on day 40. Such NandaHamner experiments w'ere performed here using the strain of C. vicina selected for a high incidence of diapause under otherwise long day (LD 16:8 h) conditions. These results were then compared with NandaHamner profiles for the 'parental' population published earlier (Vaz Nunes et a1., 1990 a; Kenny and Saunders. l99l; Saunders, I997).
2.3. Recording locomotor activit.y* rhytltms
3. Results
Activity monitors were constructed by gluing a 5 cm diameter Petri dish to the centre of a larger, 9 cm, Petri dish to form a circular track. The inner dish contained cotton wool soaked in sugar solution, u,ith some of the cotton drawn through lateral holes to provide a source of sugar and water for the fly. A single female of C. vicina vvas then placed within the hack. Two such Petri dishes were mounted on a rvooden platform which also provided support for the infra-red emitters (Radio Spares, type396-077) and detectors (RS, type 306-083). The infra-red light beam u,as arranged vertically across
3.1. Selectiott for a high incidence of lanal diapause under long da1,s
the side of the circular track containing the fly. The vvhole assembly was then 'enclosed in a lighrtight rvooden box held in a rvailc-in constant temperature room at 20 oC in constant darkness.
Activity events were registered as the number of times the moving fly broke the infra-red light beam within successive l0 min intervals. These numbers were recorded
Starting with
a 'parental' population of C.
vicina
maintained at 16 oC under long days (LD 16:8 h), several
parallel lines were selected for a high incidence of diapause by breeding from those larvae that pupariated last or were still larviform when the rest of their siblings had pupated. In the latter case, diapausing lawae were subjected to an upshift in temperature from 16 to 25 "C to initiate pupariation and to synchronise the group for
Fig. I shorvs that the proportion entering diapause rose steadily through six generations of selection, Iiom about 10ohto ahnost I00%. Selection rvas temporarily suspended betteen generations t*'o ancl three; this resulted in no change in diapause incidence at this point. The success of this selection procedure onrvard breeding.
should be compared to that described eisewhere
D.S. Saunders, B. C-vmborowski / Jountal
780
o-f
Insect Physiolog, 49 (2003) 777-781
selected for a high incidence ofdiapause under LD 12:12 (Saunders, 2001) also showed a CDL of about 14.5 h, indicating that selection had strengthened the diapause response under short days, but not altered the critical
s0
Eto o $so
value. The result for selection under long days (LD 16:8), however, was quite different: the incidence of
]co 3ro
diapause was raised under all photoperiods tested, and the CDL was lengthened from about 14.5 to over 16 h per day. Roughly half of the larvae also entered diapause under the very long photoperiods of LD 19:5 and LD 20:4, suggesting that selection under long days had produced a strain showing a strong tendency toward univoltinism.
o
$so 6 .zo
'**"irn 'r
5
6
l. Selection for a high incidence oflarval diapause in the blow ffy, Calliphora vicina,kept as adults under long day length (LD 16:8 h) at 16 "C. AII larvae'*ere maintained in continuous darkness @D) Fig.
at I l-1 2 oC. No selection occurred between generations two and three. After generation three, one of the lines was split, giving rise to four
3.2. Rhythrns of locomotor activity in selected and stock strains
selected populations.
Fig. 3 shows the mean circadian periods (r values) of the locomotor activity rhythm in constant darkness for
(Saunders, 2001) in which both high and low diapause lines rvere selected from a diapause-weakened labora-
diapause under
tory population.
Using flies from generation five, a photoperiodic response curve (PPRC) was determined for the selected population of flies by measuring diapause incidence in groups offlies exposed to a range ofexperimental photoperiods. This PPRC is compared in Fig. 2 with that for the original stock population (Saunders, 1987) and with that for the high diapause line selected earlier from the *'eakened laboratory strain (Saunders. 2001). The stock PPRC showed a sharp critical day length of about 14.5
h per day (Saunders, 1987). That for the line of flies
lines
of C. vicina selected for a high incidence of
LD l2:I2 h (A{) and LD 16:8 h (D, E). Values of r for the initial free-runs (A and D) w'ere 23.05 + A.252 and 22.75 + 0.110 h, respectively (Table l). These values should be compared rvith those obtained earlier for the original stock population: 22.53 t 0.78 h (n = 78) (Cymbororvski et al., 1993) and 22.60 + 0.60 h (n = 200i) (Hong and Saunderc, 1998). For the strain selected under LD 12:12 h, the mean free-running period for older flies (Fig. 3C) was significantly longer than that for young flies (Fig. 3A) (r = 2.596, df = 34, p < 0.05); likewise that for oider flies selected under LD 16:8 h (Fig. 3E) was significantly longer than that for young flies (Fig. 3D) (t = 6.233, df = 46, p <
