Oxygen convective uptakes in gas exchange cycles in early diapause pupae of Pieris brassicae

Breathing in many insects is characterised by discontinuous gas exchange cycles (DGCs), during which carbon dioxide is released periodically. To date, the focus of studies of DGCs has been on CO₂ release as measured by flow-through CO₂ respirometry, whereas O₂ consumption during the DGC has been little studied.

Classically, the DGC consists of three phases (Schneiderman, 1960; Lighton, 1996; Chown and Nicolson, 2004). During the open (O) phase, the spiracles are open and CO₂ is released in a burst. The O phase is followed by the facultative closed (C) phase, when the spiracles are closed and little or no gas exchange occurs. Within the C phase, sub-atmospheric pressure is created in the tracheae because of O₂ consumption by the tissues and CO₂ buffering by means of bicarbonates in the tissue and haemolymph (Wobschall and Hetz, 2004). After the C phase, the flutter (F) phase occurs, during which the spiracles open and close rapidly in succession (fluttering). When open, air is sucked through the spiracles into the tracheae by convection along the negative pressure gradient. This is known as passive suction ventilation (PSV); thus PSV occurs in the absence of muscular movement (see Miller, 1974; Miller, 1981). PSV during the F phase has been described as a mechanism for restricting water loss in insects (see Kestler, 1978; Kestler, 1982; Kestler, 1991). DGC is regarded as CFO cycles when CO₂ bursts are not actively ventilated by abdominal pumping, but as CFV (V, ventilation) cycles when the CO₂ bursts are associated with pumping movements (Kestler, 1985; Kestler, 2003).

Discontinuous O₂ uptakes and CO₂ release during the O phase has been demonstrated in some Carabidae beetles and in pupae of the Cecropia moth, Hyalophora cecropia, using heat conductivity detectors or diaferometers (Punt et al., 1957). Simultaneous measurements of O₂ uptake and CO₂ release by flow-through respirometry in the tok-tok beetle, Psammodes striatus, showed a peak of O₂ consumption at the beginning of the O phase, together with a burst of CO₂ release (Lighton, 1988). Oxygen uptake also peaked during the release of CO₂ in the giant burrowing cockroach, Macropanesthia rhinoceros (Woodman et al., 2007).

The F phase cannot be detected by flow-through O₂ respirometry without a significant diffusive component, because the inward bulk flow of air into the tracheal system is functionally equivalent to a minute and probably undetectable reduction in the flow rate of air through the respirometer chamber (Lighton, 1988; Lighton, 1994). Thus, single microopenings of the spiracles and air convective uptakes into the tracheae cannot be detected by flow-through respirometry.

Special techniques are required to record PSV during microopening of the spiracles in the F phase. Schneiderman used cannulated spiracles to measure partial pressure and thus described the rhythms of passive air uptake in silkworm pupae (H. cecropia) (Schneiderman, 1960). Sláma recorded a sawtooth pattern of abdominal retractions with contact transducers in lepidopteran pupae (including the large cabbage white Pieris brassicae) (Sláma, 1984; Sláma, 1988). This pattern was caused by the microopening of the spiracles and passive inspirations. A similar pattern of passive inspirations was recorded by Sláma and Neven in young pupae of the codling moth, Cydia pomonella (Sláma and Neven, 2001). Hetz et al. used miniaturized amperometric sensors to make direct O₂ measurements within the tracheal system of lepidopteran pupae (Hetz et al., 1994). Wobschall and Hetz recorded O₂ uptake directly in diapausing Atlas moth (Attacus atlas) pupae by simultaneous measurements of tracheal pressure and volume changes (plethysmometry) in the tracheal system, while combining CO₂ measurements by flow-through respirometry (Wobschall and Hetz, 2004). Coulometric (volumetric-manometric) respirometry has been used to directly record O₂ convective uptakes in diapausing 2–5 month old pupae of the cabbage moth, Mamestra brassicae (Jõgar et al., 2007), and P. brassicae (Jõgar et al., 2004; Jõgar et al., 2005; Jõgar et al., 2008). However, there is a lack of information about the gas exchange patterns, including O₂ convective uptake during the initiation phase of diapause (early diapause) (see Kostál, 2006; Belozerov, 2009).

Coulometric respirometry was combined with flow-through CO₂ respirometry. We suppose that the flutter events observed by coulometric O₂ measurements can usefully be directly compared with flow-through CO₂ respirometry.

The main aim of the present investigation was to describe the pattern of O₂ convective uptakes and associated body movements in young pupae of P. brassicae. To achieve this, coulometric O₂ respirometry was combined with flow-through CO₂ respirometry, and discrete O₂ discrete uptake was simultaneously recorded indirectly using an infrared (IR) actograph.

For laboratory experiments, eggs of Pieris brassicae (Linnaeus 1758) (second generation) were collected from cabbage fields near Tartu, Estonia (58°23′N, 26°41′E), during July and August 2009. They were reared in a laboratory under short-day conditions (12 h:12 h light:dark) at 21±1°C and ambient air humidity (55–65% relative humidity). The larvae that hatched from the eggs were fed on leaves cut from cabbage plants, which were replaced with a fresh supply daily. After pupation, each pupa was placed in an Eppendorff tube and kept in laboratory conditions.

For the experiments, twenty-five 2 week old (14±2 days) pupae were used. Each pupa was weighed to 0.1 mg with an analytical balance before experimentation (Explorer Balances, max. 62 g; Ohaus Corporation, Nänikon, Switzerland). Pupal body mass ranged from 0.383 to 0.411 g (Table 1). During respiratory measurements, temperature and humidity conditions were recorded using a digital HygroClip probe (HygroPalm, Rotronic Company, Basserdorf, Switzerland). All measurements were made at 21±1°C and ambient air humidity (50–55% relative humidity). The Eppendorff tube with pupa was used as the insect chamber in the respiratory systems; this avoided handling stress. The respiratory measurements of the first 30 min were discarded; recordings in both systems lasted at least 3 h. Tests with each individual were made twice. We confirmed, with preliminary experiments, that the switch from still air to flowing air (120 ml min^–1) and vice versa did not significantly change the frequency of the DGC. Body movements were visually observed under a stereomicroscope (SZ-ZTW, Olympus, Japan).

Coulometric respirometers usually work in an interrupted regime (on–off) of electrolysis (e.g. Heusner et al., 1982). By contrast, our coulometric respirometry (a volumetric manometric system) was characterised by a continuously (uninterrupted) O₂-compensating system (Kuusik, 1977; Kuusik et al., 1996; Tartes et al., 1999; Tartes et al., 2002). This setup has also been described by Lighton (Lighton, 2008). This respirometer ensures continuous and adequate replacement of consumed O₂ with electrolytically produced O₂. The insect itself plays an active role in this self-regulating system. The rates of O₂ production and O₂ consumption by the insect are indicated on graphs as V_O2 (ml h^–1). The system also records transient changes in the rate of release of CO₂. In our respirometer, we did not use the switching electrodes of electrolysis; instead, the electrolysis current was directly connected with a photoelement. High sensitivity of the respirometer to pressure changes in the respiration chamber was achieved by replacing the standard photodiode with the photosensitive element of a transistor (KT302A, Semitronics, Freeport, NY, USA), which has a very small photosensitive area (approximately 0.5 mm²). In this way, the smallest movement in the meniscus of ethanol inside the U-shaped capillary was reflected as a signal on the recording trace (Fig. 1). The electrolysis current depended on the intensity of the light falling on the phototransistor. The ethanol meniscus in the glass capillary served as a shutter to screen the photosensitive area from light. The electrochemical equivalent of O₂ generation has been reported as 209.5 μl O₂ mA^–1 h^–1 (Taylor, 1977). This value was used to convert the readings of the event recorder to O₂ consumption values (μl O₂ h^–1 or μmol O₂ h^–1).

The coulometric respirometer allowed simultaneous recording of O₂ consumption, sudden O₂ (air) uptake (known as PSV) by convection into the tracheae at microopenings of the spiracles, discrete CO₂ releases by bursts, abdominal pumping movements and heartbeat patterns (see Jõgar et al., 2004; Jõgar et al., 2007). Rapid changes in pressure (lasting seconds) in the insect chamber, caused by active body movements of the insect or other rapid events, were not compensated and led to corresponding rapid changes in the electrolysis current, reflected as spikes on recordings. Thus, our coulometric respirometer also served as an activity detector.

A rapid O₂ convective uptake resulted in adequate air volume decrease in the insect chamber and the ethanolic meniscus shifted down by a fraction of a millimetre. As a result, more light fell on the sensitive area of the transistor and an upward signal was recorded (Fig. 2). Downward signals indicated CO₂ release by bursts (Tartes et al., 1999). The volume of air uptake was estimated by extracting air from the insect chamber with a microsyringe (1 μl volume, Agilent Technologies, Espoo, Finland). The calibration of convective air uptake is shown in Fig. 2.

The infrared gas analyser or flow-through CO₂ respirometer (Infralyt-4, Saxon Junkalor GmbH, Dessau, Germany) was used to confirm that the presumed CO₂ signals, i.e. the downward spikes on the recording trace of the electrolytic respirometer, were actually due to CO₂ bursts, and to measure them quantitatively. The respirometer was calibrated at different flow rates by means of calibration gases (Trägergase, Saxon Junkalor GmbH) and with gas injection. An air flow rate of 120 ml min^–1 was used. The insect chamber could be switched either to the flow-through CO₂ respirometer or to the coulometric respirometer without disturbing the insect (Fig. 1). During the measurements with coulometric respirometry, the empty respiration chamber served to determine the baseline of the measurements.

Both the coulometric (electrolytic) respirometer and the flow-through respirometer were combined with an IR insect cardiograph (opto-cardiography); we refer to this as the IR actograph, because it records not only heartbeats but also all other abdominal contractions, including muscular ventilation. An IR-emitting diode was placed on one side of the respirometer chamber near the ventral side of the abdomen, while an IR-sensitive diode (TSA6203, Mikrotechna, Prague, Czech Republic) was placed on the opposite side of the chamber (see Metspalu et al., 2001; Metspalu et al., 2002). The light from the IR diode (BP104, Mikrotechna) was modulated by contractions of the heart and skeletal muscles. The level of output voltage reflected the vigour of the muscular contractions of the insect (Hetz et al., 1999). Sudden extensions (lengthening) of abdominal segments (PSV) are recorded as relatively long upward spikes synchronous with microopenings of the spiracles (Figs 3, 4). Weak regular muscular contractions of the abdomen resulted in two-phased relatively short spikes on the recording traces we refer to as abdominal pulsations (Fig. 5). Regular, high-frequency, low-amplitude signals were interpreted as heartbeats (Fig. 3).

Computerised data acquisition and analysis were performed using DAS 1401 A/D hardware and TestPoint software (Keithley, Metrabyte, Cleveland, OH, USA) with a sampling rate of 10 Hz. Four bipolar channels allowed simultaneous recording of four events. Mean (±s.d.) standard metabolic rate was calculated automatically using STATISTICA (version 8, StatSoft, Tulsa, OK, USA). Statistical comparisons were made with one-way ANOVA (analysis of variance). Significant ANOVAs were followed with the Fisher's least significant difference (LSD) test. The significance level was set at P<0.05.

At the initiation phase of diapause (14±2 days old), P. brassicae pupae displayed DGCs lasting 40–70 min (Table 1), whereas the duration of CO₂ release by burst was 2–6 min (3.1±0.1). Recordings by flow-through respirometry showed a typical pattern. After the O-phase CO₂ emission had ceased, the C phase began, which was followed by the F phase with small bursts of CO₂ release (Fig. 3). No true C phase was found by flow-through respirometry. Shortly after the end of the O phase, coulometry revealed convective O₂ uptake. During this time, CO₂ release was suppressed by the inward-directed PSV. This convective O₂ uptake indicates an earlier beginning of the F phase than detectable with the flow-through system. Thus the F phase lengthened on account of the C phase. At each microopening of the spiracles and passive convective oxygen uptake event, signals of abdominal lengthening were recorded (Figs 3, 4). Recordings of coulometric respirometry showed clear, gradually shortened signals due to convective oxygen uptake (Fig. 4). Each of the two to three first microopenings and O₂ uptakes lasted less than 0.5 s. Oxygen convective uptakes during the interburst period increased metabolic rate by 5–6% compared with the metabolic rate when these uptakes were absent. Flow-through measurements also indicated abdominal lengthening concurrent with the small bursts of CO₂ during flutter (Fig. 5).

Between two large CO₂ bursts, a series of small bursts of CO₂ were recorded by coulometric respirometry (Fig. 6A). Each small burst started with a brief uptake of air into the tracheae, recorded by the IR actograph as a sudden extension of the abdomen, indirectly indicating air (O₂) uptake (Fig. 5). Between two consecutive small CO₂ bursts, a series of air uptakes (miniature inspirations) were recorded, which we considered as `miniflutter' (Fig. 6A,B; Table 1). These uptakes were irregular with respect to spike height and interval. During such miniflutters, no emissions of CO₂ were recorded.

Simultaneous recording with the IR actograph during flow-through CO₂ respirometry indicated that pupae differed in the type of body movement associated with the respiratory patterns of CO₂ release. In some pupae, CO₂ bursts were always concurrent with abdominal ventilating movements (CFV cycles) (always group, N=9) (Fig. 7A), whereas in others (occasionally group, N=10), only some CO₂ bursts were concurrent with abdominal ventilating movements (Fig. 7B). In a few pupae (never group, N=6), CO₂ bursts occurred without active ventilation (CFO cycles) (Fig. 7C). Ventilating movements (amplitude 1–2 V) associated with CO₂ bursts were visible externally as twisting abdominal movements.

Active ventilation during the bursts of CO₂ showed individual variation in the vigour of contractions and their number (from one to 15). In pupae with only one to five muscular (active) ventilating movements accompanying the burst, as well as in those lacking active ventilation, a relatively low level of CO₂ release was observed. In contrast, pupae with vigorously ventilated bursts showed a significantly higher level of CO₂ release (Fig. 7A). Each burst lasted 3–6 min in the always group of pupae, but 2–2.5 min in the occasionally and never groups. Statistical comparison of CO₂ release frequency (ANOVA, F_24.2=41.8, P>0.05) did not show a significant difference. The energy cost of muscular ventilation during a burst was not studied.

Abdominal two-phase regular contractions (5–7 min^–1) of low amplitude (0.2–0.3 V) (referred to as abdominal pulsations) (Fig. 5, Fig. 6B) occurred periodically; these were not visible externally.

In some pupae showing no active ventilation during CO₂ release by bursts, very regular low amplitude (0.1–0.2 V) pulsations (57–70 min^–1) were recorded; these we interpreted as heartbeats (Fig. 3).

Our results showed that, in the initiation phase of diapause, P. brassicae pupae display relatively short DGCs (40–70 min), with CO₂ bursts lasting 2–6 min. This contrasts with earlier studies, using P. brassicae pupae more than 2 months old, which displayed longer DGCs (8–23 h) (Harak et al., 1999; Jõgar et al., 2004; Jõgar et al., 2005) with CO₂ bursts lasting 13–18 min (Harak et al., 1999; Kuusik et al., 1980; Tartes et al., 1999). The young pupae we used with their short DGCs were convenient for studying flutter events. They had a relatively high metabolic rate; in 2–3 month old pupae metabolic rate is at least two times lower (12–28 ml O₂ g^–1 h^–1) (Kuusik 1977; Jõgar et al., 2004; Jõgar et al., 2005).

Commonly, after the O phase, a period with no CO₂ release occurs (C phase); later, the CO₂ level was marginally elevated (F phase) (Chown et al., 2006). The present study revealed that in young P. brassicae pupae, the DGC measured with flow-through respirometry was characterised by a C phase, at the end of which a series of O₂ convective uptakes was found. Thus, the C phase in those pupae was not as closed as previously thought. Between two large CO₂ bursts, almost regular small CO₂ bursts were recorded. Each small burst started with sudden uptake of air into the trachea (PSV). The main finding in the present study was a series of irregular microopenings of the spiracle(s) with convective O₂ uptakes (mini-flutter) found between small bursts. During the mini-flutter, no recordable CO₂ emission occurred.

There are several examples where the interburst period consists of discrete small CO₂ bursts. Such bursts were described by Lighton (Lighton, 1988) in the tok-tok beetle, P. striatus; in this beetle, each burst was accompanied by active abdominal movement. Discrete CO₂ emissions during the F phase have also been reported by Duncan et al. in the tenebrionid beetle, Pimelia grandis (Duncan et al., 2002), and by Kovac et al. in resting honeybees Apis mellifera (Kovac et al., 2007). However, spiracle openings within the F phase were commonly observed to be irregular with respect to frequency and amplitude, if inferred from the CO₂ release pattern (e.g. Wobshall and Hetz, 2004).

Our flow-through CO₂ measurements showed no CO₂ release for a short time after the O phase. The coulometric respirometry and IR actographic recordings showed rapid and clear uptakes of air shortly after the O phase, indicating the beginning of the F phase. An earlier study by Tartes et al. revealed that air convective uptakes began immediately after the O phase (Tartes et al., 2002). Air convective uptakes, shortly after large CO₂ bursts, also occurred in old diapausing M. brassicae pupae (Jõgar et al., 2007). We suggest that, at the beginning of the flutter, air uptakes were convective but later were diffusive-convective. These results concur with the plethysmometry flow-through measurements of Wobschall and Hetz (Wobschall and Hetz, 2004), revealing that the convective uptakes of O₂ dominate at the beginning of the flutter phase but that, in the later F phase, diffusion takes over from convection as the chief mechanism of O₂ uptake. Wobschall and Hetz showed that, in diapausing moth pupae (A. atlas), uptake of air into the tracheal system at the beginning of the F phase along the negative hydrostatic pressure gradient may initially inhibit CO₂ release from the tracheae (Wobschall and Hetz, 2004). We suppose that, at the beginning of the F phase of P. brassicae pupae, CO₂ emission was also inhibited. One may suppose that water is conserved only at the beginning of the F phase when clear convective O₂ uptakes (PSV) occur, but not later when diffusion is the dominating mechanism of the fluttering period.

The duration of the F phase may be underestimated, as the F phase may start before the CO₂ measurements can detect it (Hetz et al., 1994; Wobschall and Hetz, 2004). In pupae of P. brassicae, the duration of the F phase may also be longer than estimated by the flow-through system, as far as CO₂ release was prevented at the beginning of the F phase by convective air uptakes. Wobschall and Hetz showed that small volume and pressure decreases occurred between the microopenings in the F phase (Wobschall and Hetz, 2004). This confirmed a small but significant contribution of suction ventilation during each microopening (see also Kestler, 1985). In our measurements in P. brassicae pupae, each miniature inspiration was synchronised with rapid extension of the abdomen, confirming that a convective component was always present in O₂ uptakes. We observed in P. brassicae a relatively longer flutter phase compared with that in some other insects, such as the carabid Pterostichus niger (Kivimägi et al., 2011) and the bumblebee Bombus terrestris (Karise et al., 2010).

Gas exchange patterns are known to vary between and within individuals (see Chown, 2001; Chown et al., 2002; Marais and Chown, 2003). In the present study, variation was found between individuals in the duration of CO₂ release and metabolic rates, but not in DGC frequency. Individuals displayed different gas exchange cycles. Most showed DGCs with all CO₂ bursts actively ventilated, a few with no bursts actively ventilated, and others with only some bursts actively ventilated.

Beside active ventilation, some other types of body movements, differing in frequency and amplitude, were observed in the present study. The regular but periodically occurring abdominal pulsations in P. brassicae pupae correspond, in our opinion, to the extracardiac hemocoelic pulsations described in lepidopteran pupae (Sláma, 1984; Sláma, 1999). These pulsations and other abdominal movements play an important role in the regulation of pupal respiration and haemolymph circulation (Sláma and Neven, 2001). In P. brassicae pupae, we interpreted high-frequency but low-amplitude signals, accompanied by CO₂ release in bursts, as heartbeats. In a previous study using thermographic measurements, we demonstrated heartbeat reversal, correlated with gas exchange cycles and twisting abdominal movements in diapausing P. brassicae pupae (Jõgar et al., 2005). Heartbeat reversal correlated with gas exchange cycles has also been reported in saturnid moth pupae (Wasserthal, 1996; Hetz et al., 1999; Sláma, 2003).

Manometric O₂ respirometry methods have been criticized and their readings mistrusted because these methods usually do not allow separation of active and resting metabolism (see Van Voorhies et al., 2008). Nevertheless, some volumetric manometric methods, including our coulometric respirometry, are regarded as useful (Klok and Chown, 2005). We are convinced that in gas exchange studies of insects, coulometric respirometry supplemented by the flow-through method has clear advantages. Lighton pointed out that coulometric continuously recording respirometry deserves to be more widely used (Lighton, 2008).

In summary, in this study we have shown that the pattern of gas exchange in P. brassicae pupae may be effectively investigated by the combined use of coulometric respirometry and flow-through CO₂ systems by switching the same respiration chamber from one system to the other without disturbing the insect. By combining IR actography in parallel with both types of respirometry, it was possible to record rapid air uptakes. Thus, on our recording traces, the patterns of microopenings of the spiracles were clearly indicated.

We thank the editor and two anonymous reviewers, who provided helpful and much-appreciated comments on earlier drafts of the manuscript.