Respiratory dynamics of discontinuous gas exchange in the tracheal system of the desert locust, Schistocerca gregaria

Insects living in terrestrial environments face significant metabolic challenges. For example, they have to regulate internal O₂ levels necessary for aerobic metabolism and excrete metabolic by-products (e.g. CO₂), while minimizing water loss (e.g. Chown, 2002; Woods and Smith, 2010; Matthews and White, 2011a). Insects possess numerous morphological and physiological adaptations for coping with these simultaneous pressures. Among the most controversial of the proposed adaptations is the use of discontinuous gas exchange cycles (DGCs) associated with the possession of occludible spiracles (Lighton, 1994; Chown et al., 2006). Indeed, the underlying mechanistic and evolutionary reasons for variation in patterns of insect gas exchange are receiving renewed attention (White et al., 2007; Terblanche et al., 2008; Chown, 2011; Contreras and Bradley, 2011; Matthews and White, 2011b). These studies have all argued that a mechanistic understanding of insect respiratory dynamics and gas exchange patterns is central to understanding evolved variation in insect gas exchange patterns.

The respiratory organs of terrestrial insects consist of semi-rigid tracheal tubes that penetrate the body and tissues. These tracheae are air filled and transport gases between the atmosphere, spiracles and internal tissues or organs. The tracheae divide into tubes of decreasing diameter and eventually branch into microscopic tubes, or tracheoles. The tracheoles are the main sites of gas exchange with tissues (e.g. Wigglesworth, 1929) (for reviews, see Wigglesworth, 1965; Wigglesworth, 1983); they closely surround cells and might even penetrate them to lie close to mitochondria in some cases (Quinlan and Gibbs, 2006). In most insects the terminal part of the tracheole is filled with fluid (Chapman, 1998; Woods et al., 2009), and the final stages of gas transport between the tracheoles and haemolymph or mitochondria take place via diffusion through a liquid (Miller, 1964). Thus, the spiracles regulate the flow of gases and water vapour into and out of the tracheae (Hadley, 1994; Chapman, 1998). By closing the spiracles, tracheal P_O2 and pressure can be regulated, perhaps minimizing oxidative damage (Hetz and Bradley, 2005), while the loss of water from the internal environment of the insect to the external atmosphere can be limited (e.g. Loveridge, 1968).

In insects, efficient gas exchange is dependent on the morphological structure of the tracheal system, as well as on adequate ventilation. Passive diffusion or active muscular movements (i.e. convective gas exchange), and the selective opening of the spiracles can be used to modulate whole-animal ventilation patterns (Miller, 1964; Lighton, 1996; Chown et al., 2006). In small insects and in the inactive stages of some larger species, an adequate rate of O₂ uptake can be obtained predominantly by diffusion, and there is thus little need for respiratory movements (Kestler, 1985). However, in larger or more active insects diffusion has to be supplemented with convection to maintain an adequate O₂ supply (Chown and Nicolson, 2004). These larger insects probably rely on diffusion only for O₂ transport along the finer terminations of the tracheae, while mechanically ventilating the larger tracheal trunks (Wigglesworth, 1972). Although the significance of ventilation is widely acknowledged, the temporal dynamics of gas exchange and variation in pressure associated with such ventilatory movements have only been explored in a limited range of species (mostly Lepidoptera, some Orthoptera and one species of Hymenoptera; see Table 1). Most work to date, drawn primarily from the classic gas exchange model of diapausing moth pupae, suggests that the closed (C)-phase of the DGC is typically associated with negative intratracheal pressures, while during the open (O)-phase intratracheal pressure remains at or near atmospheric levels (Table 1). Moreover, only a handful of studies combine gas exchange, intratracheal pressure and body movement measurements simultaneously (Table 1). Given the diversity of gas exchange patterns among tracheated arthropods, and the fact that the DGC has evolved independently on at least five occasions within the Insecta (Marais et al., 2005), further investigation of patterns and mechanisms underlying gas exchange is required, perhaps encompassing a wider range of species.

Here, we therefore simultaneously measured intratracheal pressure, V̇_CO2 and body movements (i.e. pumping body movements) of the desert locust Schistocerca gregaria. This study aimed to assess the role of ventilatory movements over the entire gas exchange cycle and the relative importance of convection vs diffusion in this species. Intratracheal pressure was determined for the different DGC phases, as well as for the start of each of the DGC phases. Specifically, we aimed to test whether the DGC is accompanied by a specific intratracheal pressure pattern. We hypothesized that the C-phase of the DGC is accompanied by a gradual decrease in intratracheal pressure, followed by rising pressure during the flutter (F)-phase, until the pressure reaches and stabilizes at or near atmospheric levels during the O-phase.

Adult S. gregaria (Forskål 1775) (Orthoptera: Acrididae) were supplied by Grillenzucht Hildner (Gerhardshofen, Germany). Individuals were kept separated from each other in plastic containers. Animals were fed fresh lettuce every second day and were kept at ambient temperature (18–22°C), relative humidity and air conditions. Both male and female adult grasshoppers were used. Animals were fasted for at least 8 h before respirometry commenced.

Twelve individuals were recorded for 8 h each at 15°C. Rate of CO₂ release (V̇_CO2), body movements (i.e. activity) and intratracheal pressure were recorded with 8 samples s^–1. Carbon dioxide release was measured in a custom-made flow-through respirometer, with a length of 60 mm and a volume of ∼10 ml. The temperature of the cuvette was controlled by a custom-made Peltier-cooling unit with a computer-controlled feedback (accuracy ±0.1°C). Ambient air (from outside the lab) was scrubbed of CO₂ and water vapour by passing it through two columns containing a 5 mol l^–1 NaOH solution. The scrubbed air was then driven through two 200 ml columns filled with distilled water, which were kept at a constant temperature of 8°C, to keep the water vapour pressure constant. The flow rate of the air was regulated at 250 ml min^–1 by a mass flow controller (MKS 1179, MKS Instruments, Wilmington, MA, USA). The air was then fed through the reference cell of a pressure-compensated differential infrared gas analyser (URAS 14, range 200 p.p.m., ABB Analytical, Frankfurt, Germany), whereafter it flowed through the respirometry cuvette, and then through the analysing cell of the gas analyser, to measure the CO₂ output of the locust. During a trial, the body movements of each individual were recorded with three different infrared sensors, located dorsally above the animal. These consisted of reflective interrupters (Osram SFH 9201, Osram, Germany) that sent infrared light (wavelength 850–950 nm) to the animal and measured the reflected changes in radiation due to body movements. Data were recorded to a computer hard disk using a customized recording program in TurboLab 4.03 (Bressner Technology, Gröbenzell, Germany).

Each individual was weighed before and after spiracle intubation, as well as after each trial. Body movements, V̇_CO2 and intratracheal pressure were recorded following methods outlined elsewhere (Moerbitz and Hetz, 2010; Wobschall and Hetz, 2004). To measure intratracheal pressure, the animal's mesothoracic spiracle was cannulated. Before intubation of the spiracle commenced, animals were restrained by binding the mesothoracic and metathoracic legs to the abdomen with adhesive tape (Fig. 1). The restraints did not confine the abdomen or physically impede abdominal ventilatory movements. These measures were taken to ensure that the tubing could not be removed from the cannulated spiracle by the animal's leg movements.

For the cannulation, a short piece of polyethylene tubing (PE10) was modified to include a thin steel tube (0.25 mm i.d.) at the end, which was inserted into the animal's spiracle. The surrounding of the spiracle was then sealed with dental wax. The tubing was connected to a precision micro-pressure transducer (SenSym SDXL010, SensorTechnics, Pucheim, Germany). The other port of the transducer opened into the respirometry chamber to measure pressure differences between the chamber and tracheal system. The pressure sensor electrically consisted of a Wheatstone bridge that was driven with a precision voltage reference of 6.000 V (REF02, Burr Brown, Texas Instruments, Dallas, TX, USA) buffered with a precision operational amplifier (OPA177, Burr Brown). Differential voltage output of the pressure sensor bridge was amplified with a differential amplifier (INA131, Burr Brown).

After intubation, animals were allowed 20–30 min to recover from the handling stress of the intubation procedure, before recording of their CO₂ release commenced. After this period animals no longer struggled and appeared to ventilate normally.

All 12 individuals exhibited DGCs when recorded at 15°C (mean ± s.d. body mass 1.761±0.325 g). There was no significant relationship between body mass and any DGC phase duration, V̇_CO2 during the O-phase and interburst period, or volume of CO₂ released during the O-phase (P>0.39 in all cases).

The mean duration of the interburst period increased with mean DGC duration (r=0.922, N=12, P<0.0001). Mean interburst V̇_CO2 was positively correlated with mean interburst duration (r=0.608, N=12, P=0.0358). Mean O-phase volume was positively related to mean interburst duration (r=0.734, N=12, P=0.0066). Moreover, the interburst period of a DGC commenced at any intratracheal pressure and was not restricted to sub-atmospheric pressures (Fig. 2, Table 2).

Mean O-phase V̇_CO2 was negatively, but not significantly, related to mean O-phase duration (r=–0.524, N=12, P=0.0806). Mean O-phase volume was positively correlated with mean O-phase duration (r=0.726, N=12, P=0.0075).

To examine whether any relationship exists between the duration of the different phases of a DGC and the intratracheal pressure pattern of an individual, correlations between these variables were assessed. No significant relationship existed between mean O-phase duration and mean pressure during an O-phase within those same DGCs (r=–0.509, N=12, P=0.0914), or between mean interburst duration and mean pressure during that same interburst period (r=–0.098, N=12, P=0.7615).

There were no characteristic patterns for intratracheal pressure changes that were common among all individuals, but rather intratracheal pressure traces showed high among-individual variability. In S. gregaria the O-phase could include periods of negative intratracheal pressure, while the interburst period sometimes showed high positive pressure peaks (Table 3).

The mean intratracheal pressure of individual S. gregaria during the interburst period ranged from –112.1 to 224.6 Pa, with four individuals having a negative mean intratracheal pressure during the interburst period and eight individuals having a positive mean intratracheal pressure. Mean V̇_CO2 during the interburst period ranged from 1.49 to 14.67 nmol g^–1 min^–1. In eight individuals, V̇_CO2 during the interburst period was statistically greater than 0, regardless of whether the mean intratracheal pressure was negative or positive. Four of the 12 individuals recorded had a negative mean intratracheal pressure during the interburst period, and in three out of these four individuals V̇_CO2 during the interburst period was statistically greater than the infrared gas analyser baseline (see e.g. Fig. 2, individual 2).

Examples of DGCs, together with intratracheal pressure and body movements, are presented in Fig. 2. These traces clearly show that S. gregaria employs two of the four different ventilatory movements described by Miller for resting locusts (Miller, 1960). Dorso-ventral muscles are contracted at a high rate, while longitudinal muscles are used rarely, but coincide with higher positive pressure pulses (Fig. 2, individual 1 and individual 2). Also, strong pumping actions do not necessarily lead to CO₂ release, and large outputs of CO₂ are not always associated with abdominal pulsation body movements (Fig. 2, individual 3).

Several authors have demonstrated that a comprehensive understanding of gas exchange dynamics in insects should include measurements of intratracheal pressure, as these measurements can provide novel information about both the kinetics of gas exchange and insect respiration in general (e.g. Buck, 1958; Levy and Schneiderman, 1966; Terblanche et al., 2008). This was certainly the case here, where we found for the first time that for S. gregaria the O-phase can include periods of negative intratracheal pressure, as well as pressures well above atmospheric levels (Table 3), and the interburst period can start at any pressure amplitude (Table 2) and can include periods of high intratracheal pressure (Table 3; Fig. 2). Rather than the patterns found here being unusual forms of the DGC, several aspects of the DGCs recorded conformed well with established patterns for Orthoptera and other arthropod taxa. For example, the positive relationship observed between O-phase duration and DGC duration is similar to that reported previously for a variety of dung beetles (Davis et al., 1999) and for Garypus californicus (Lighton and Joos, 2002) (for review, see Chown and Nicolson, 2004).

The pumping movements of an insect's abdomen lead to volume changes of the tracheal system, which in turn lead to pressure fluctuations within the internal body (detected as intratracheal or haemolymph pressure changes) of the insect (Buck, 1962). In locusts, the pumping movement of the abdomen is generally thought to ventilate the tracheae and larger tracheal trunks (Wigglesworth, 1972), and both inspiration and expiration are active processes (Miller, 1964). In an actively ventilating orthopteran, the first four pairs of spiracles open during the inspiratory phase of abdominal ventilation, while the remaining pairs open during the expiratory phase (Weis-Fogh, 1967; Harrison, 1997), leading to an anterior to posterior (unidirectional) flow of air through the insect (Henderson et al., 1998). However, the path of this unidirectional flow of air is subject to variation, and occasionally any of the spiracles might serve for both inspiration and expiration (Steen, 1971). If there is a unidirectional flow of air through the insect, it could be possible for the insect to have a negative pressure in one part of its tracheal system, while a positive pressure is present in another part, depending on coordinated spiracle and ventilation movements. This could explain the periods of negative intratracheal pressure during the O-phase (Table 3).

The pressure patterns we recorded in S. gregaria are similar to those reported for other grasshoppers [e.g. Melanoplus differentialis (Krolikowski and Harrison, 1996)] in which tracheal pressures varied substantially among individuals, although this is unlike the situation for lepidopteran pupae (e.g. Levy and Schneiderman, 1966; Wobschall and Hetz, 2004; Terblanche et al., 2008). Moreover, the negative intratracheal pressure during the O-phase is unexpected as the spiracles typically remain open, indicating that the intratracheal pressure should be close to atmospheric levels. This result also differs from findings in lepidopteran pupae in which intratracheal pressure during the O-phase was maintained at levels similar to atmospheric pressure, and intratracheal pressure never rose much above atmospheric levels (e.g. Wobschall and Hetz, 2004; Terblanche et al., 2008). By contrast, in M. differentialis grasshoppers, intratracheal pressure measured in the thoracic trachea never reached atmospheric levels when these individuals were at rest (Krolikowski and Harrison, 1996). In lepidopteran pupae, intratracheal pressure generally falls during the C-phase, rises during the F-phase and remains constant near atmospheric levels during the O-phase (Schneiderman and Schechter, 1966; Hetz et al., 1993). However, for S. gregaria, the C-phase can start at any intratracheal pressure (above or below atmospheric pressure) (Table 2) and can include periods of high intratracheal pressure (Table 3; Fig. 2). Therefore, in S. gregaria the respiratory cycle is not accompanied by a clearly distinguishable sub-atmospheric intratracheal pressure cycle, as is the case with lepidopteran pupae.

The body movements accompanying gas exchange at rest are similar to those reported previously for this species. Indeed, it was established more than 50 years ago through visual observation and mechanical recordings that dorso-ventral contractions and longitudinal telescoping movements of the abdomen occur in grasshoppers at rest (e.g. McCutcheon, 1940; Weis-Fogh, 1967; Lewis et al., 1973). More recent work that included gas exchange recordings showed that gas exchange may also take place when no abdominal pumping occurs, suggesting that gas exchange is primarily diffusive (although some micro-ventilation pumping actions can be employed, are barely visible to the naked eye and may aid convection) (e.g. Harrison, 1997; Greenlee and Harrison, 1998). However, our study is unique because we recorded body movement patterns, gas exchange and intratracheal pressure simultaneously, which has never previously been done for any Orthoptera (Table 1).

Hamilton (Hamilton, 1964) found that in adult Schistocerca bursts of CO₂ coincided with ventilation, and postulated this as being a means to conserve respiratory water. Kestler also explained these coordinated ventilatory movements (involving both diffusive and convective exchanges) as being a means to reduce respiratory water loss (Kestler, 1978). By compressing the abdomen and closing the spiracles, pressure inside the abdomen can be increased, thereby also compressing the tracheal system and gases within the tracheae. But what could the advantage of such high pressures achieved during the interburst period be? One potential explanation is that this might slow down evaporation of water. The pressure cycle theory suggests that changes in intratracheal pressure could help with the regulation of the influx and efflux of water vapour, by making it possible for the animal to regulate the hydrostatic pressure within the trachea (Corbet, 1988). Within the tracheal system, respiratory gases occur at a vapour–liquid interface between the air-filled trachea and insect body fluids. Water can move from the body fluids to the trachea and vice versa, and by regulating intratracheal pressure (and thereby also hydrostatic pressure) it should thus be possible to reduce the rate or reverse the direction of water movement. Water evaporates into subsaturated air, but if the vapour pressure within the air exceeds the supersaturation ratio, water will condense. An increase in intratracheal pressure will lead to increased hydrostatic pressure within the tracheal system, and an increase above the supersaturation ratio will cause condensation. The trachea are lined with cuticle that is impermeable to water, but the tracheoles are liquid filled and have permeable walls, making it possible for water and respiratory gases to move through these walls (Krogh, 1941; Wigglesworth, 1965; Kerkut and Gilbert, 1985; Mill, 1985). Therefore, increasing the hydrostatic pressure in the tracheal system above the supersaturation ratio will lead to the condensation of water and the movement of water from the trachea into the tracheoles. The remarkably high intratracheal pressure pulses observed during the interburst period in S. gregaria could be high enough to meet, or perhaps even exceed, the supersaturation ratio for water vapour, thus reversing the direction of water movement. According to Henry's law, at a given temperature the partial pressures of gases, like water vapour, within a gas mixture are elevated when the total pressure is elevated. In contrast, evaporation from a surface (like the tracheoles) occurs faster if the pressure is lower. The osmolality of the fluid within the tracheoles should set the equilibrium water partial pressure within the tracheal system.

If the osmolality of the tracheole fluid is assumed to be 300 mosmol l^–1 – which is equivalent to a water chemical activity (a_w) of 0.995, or 99.5% relative humidity, and approximates values for insect haemolymph (see Edney, 1977) – at this high relative humidity tracheal pressure can be directly interpreted as water vapour pressure, assuming the system is in equilibrium. Therefore, the highest pressure pulses within an individual of around 1971 Pa are equivalent to water vapour pressure of 14.78 mmHg. At 15°C the saturation partial pressure of distilled water is 12.76 mmHg. This partial pressure exceeds the saturation pressure of water vapour and air in the tracheal system is therefore likely to be oversaturated during the interburst phase of the DGC. Another possibility is that these high pressure pulses could be used to ensure effective mixing of gases in the closed tracheal system of the insect, which could perhaps allow the insect to increase the duration of the interburst period as a result of more efficient distribution of O₂.

Terrestrial insects display at least three different patterns of gas exchange when at rest (for reviews, see Chown et al., 2006; Quinlan and Gibbs, 2006; Bradley, 2007). A DGC consists of three phases, which are typically described in terms of spiracular behaviour; namely a C, F and O spiracle phase (Schneiderman, 1960). In some of the recorded individuals, although a true C-phase of the DGC could be identified by means of subatmospheric intratracheal pressure recordings, some CO₂ continued to be released (e.g. Fig. 2, individual 2). This particularly novel finding has important implications for the identification of patterns solely on the basis of V̇_CO2 traces, although to date discussion has mostly focused on the use of appropriate flow rates or temperatures for pattern identification (e.g. Gray and Bradley, 2006; Terblanche and Chown, 2010; Contreras and Bradley, 2011).

In conclusion, this study highlights the importance of using multiple parameters when examining the mechanisms of gas exchange and suggests that much is still to be learned about the fundamentals of insect respiration.

We thank Leigh Boardman and two anonymous referees for constructive comments that helped to improve an earlier version of the manuscript. B.G. thanks members of the Hetz Lab for assistance and support during experiments.