Cuticular mechanics during larval development of the tobacco hornworm, Manduca sexta

During postembryonic development in insects, cuticular mechanical properties are altered to accommodate changing functional demands on the cuticle. For instance, the stiff inextensible cuticle of Rhodnius prolixus must be plasticized in order to accommodate expansion associated with a blood meal (Bennet-Clark, 1962; Reynolds, 1975a,b;,Nunez, 1963). During pupariation the soft larval cuticle in many flies becomes tanned and stiffened to support and protect the developing adult (Fraenkel & Rudall, 1940; Zdarek & Fraenkel, 1972). In female locusts the unusually extensible abdominal intersegmental membranes permit oviposition deep in the sand (Vincent, 1975). Furthermore, in connection with the characteristic phenomenon of discontinuous growth in arthropods, cuticular extensibility increases for a brief period during each moult (Neville, 1975).

During the fifth instar of the caterpillar Manduca sexta, cuticular growtH continues as the larva increases 10-fold in volume and four-fold in surface area (Wolfgang & Riddiford, 1981). On the fifth day growth ceases and the animal wanders off the plant, burrows into the ground, and forms a pupation cell. The aim of this study was two-fold. First, to measure larval cuticular mechanical properties throughout the fifth instar, in order to determine how they vary with respect to changing functional demands placed on the cuticle. Second, to reveal correlations between changes in mechanical properties and known changes in cuticular structure (Wolfgang & Riddiford, 1981) and protein composition (Wolfgang & Riddiford, 1986).

Caterpillars of Manduca sexta were reared under short-day conditions (12L: 12D) at 26±1°C and maintained on an artificial diet (Truman, 1972; Bell & Joachim, 1976). To ensure accurate staging, animals were selected that ecdysed to the fifth instar between 22.00 h AZT and 2.00 h AZT [Arbitrary Zeitgeber Time (Pittendrigh, 1965)].

To inhibit cuticle deposition the chitin synthesis inhibitor diflubenzuron (DFB) (Thompson-Hayward 6040) (Post & Vincent, 1973; Mitsui et al. 1980) was added to the diet at 10 p.p.m. Experimental animals were reared on a normal diet for 1 or 2 days and were then transferred to a diet containing DFB for the duration of the fifth instar.

In starvation experiments animals were allowed to feed for 36 h after ecdysis and all the food and frass were then removed from the rearing cup. Alternatively, the food was replaced with 2 % agar.

Haemolymph pressure was recorded from animals catheterized through the horn on a Gilson polygraph using a Statham P23Db pressure transducer. After each experiment the system was calibrated with a column of water and the data were subsequently converted into millimetres of mercury (1 mmHg = 133·3Pa).

The dorsal abdominal segments were cut from the larvae and the adhering fat body and muscle were removed in Manduca saline containing phenylthiourea to prevent blackening (Riddiford, Curtis & Kiguchi, 1979). Strips of cuticle 5 mm wide including the attached epidermis were cut with a pair of mounted razor blades. Mechanical properties of circumferential and axial strips were compared and found to be similar. This is consistent with the helicoidal distribution of microfibrils in the cuticle (Wolfgang & Riddiford, 1981). Since circumferential strips were easier to cut and showed less experimental variation than axial strips, all the data presented here are from circumferential strips.

Five strips were cut, one each from segments 3-7, such that no intersegmental membrane was included. Since no consistent variation was noticed between segments, the measurements were averaged to obtain a mean value for each animal. For each day of the instar, at least four, and usually 5-10, animals were tested and the values averaged. Throughout the dissection and during the testing, the integument was kept moist with Manduca saline.

Integumental strips (cuticle plus adhering epidermis) were tested on an Instron TM-S extensometer in the manner of Cleland (1967). Test pieces were extended at a constant rate of 3 mm min^-1 until a force of 1·0N was measured on the chart recorder.

Because the force extension curves for Manduca larval cuticle were not linear (see Fig. 3), the tangent modulus (Hepburn & Joffe, 1974) at 0·3 N force was used as a measure of intrinsic cuticular stiffness. As the name implies, the tangent modulus is the slope of the force extension curve at 0·3 N force and is calculated as follows:

An implicit assumption made when calculating the tangent modulus is that the structure of the material is homogeneous; this is not the case for Manduca cuticle (Wolfgang & Riddiford, 1981). Furthermore, the tangent modulus ignores the role of cuticular thickness in determining mechanical properties. Because the caterpillar must grow in a cuticle of a certain thickness that is structurally inhomogeneous, cuticular extensibility was also determined. Because this measure makes no assumptions about cuticle structure and takes into account cuticular thickness, it is a more meaningful measure of cuticular mechanical function than the tangent modulus (Hepburn & Joffe, 1976). Extensibility is the true strain that would be produced in a cuticular test strip if 1 N of force were applied to it. It was calculated by multiplying the inverse of the tangent modulus by 1/A (the inverse of the cross-sectional area of the test piece). The inverse of extensibility is comparable to the relative stiffness as defined by Hepburn & Joffe (1976).

It must be emphasized that the analytical methods used were developed for materials that respond to imposed stresses in a linear manner, which is not the case for Manduca cuticle. Thus, the data obtained are at best relative and should not be considered to represent absolute values. Since the test conditions were standardized, the data are meaningful for a comparison of the material properties of cuticles from larvae of different stages.

To permit relatively rapid processing of many specimens, cuticular thickness was determined indirectly. A standard curve of cuticular thickness versus the mass of a cleaned and desiccated (overnight at 100°C) disc of cuticle (4·75 mm diameter cut with a no. 2 cork borer) was linear (r=0·98). Therefore, cuticular thickness was determined from the mass of a dried standard disc of cuticle.

Because water loss is rapid from small isolated pieces of cuticle, the water content of the cuticle was measured in a manner similar to that described by Reynolds (1975b). Discs of cuticle and epidermis were extirpated and floated on a drop of saline. Immediately prior to weighing, the epidermis was removed with a cotton swab and the cuticular surface wiped dry. The loss of mass from the cuticle was recorded every 30 s for 3 min on a Mettler balance (Mettler M5). A plot of mass versus time was linear, allowing extrapolation back to zero time to determine the original wet mass. The disc was then dried overnight at 100°C and the dry mass was measured. The data were then expressed as percentage hydration [(wet mass — dry mass)/wet mass].

The effect of pH on cuticular hydration was tested by soaking cleaned cuticle discs for 1 h in low-salt (0·01 mol 1^-1) buffers: pH 4, sodium acetate; pH 6, sodium citrate; pH 8, Tris; pH 10, sodium carbonate; pH 12, sodium carbonate/sodium hydroxide.

During the first 4 days of the fifth instar (day 0 to day 3), the caterpillar feeds and grows 10-fold in volume and four-fold in surface area (Fig. 1) (Wolfgang & Riddiford, 1981). During this time, the cuticle increases in thickness from about 30 to 200 μm (Fig. 2). On the fifth day, the animal stops feeding and enters the wandering phase (w), burrows into the ground, and forms a pupation cell (Fig. 1). During the prepupal phase (w+1 to w+3), the animal becomes quiescent and shrinks to about 65% of its maximum length (Fig. 1). During this time, cuticular thickness increases from 200 to 300 μm (Fig. 2).

Throughout development, a basal haemolymph pressure of approximately 10 mmHg was maintained. During wandering, transient peaks in pressure, up to 100 mmHg, occurred in conjunction with the peristaltic waves associated with the locomotor behaviour.

Fig. 3A shows force-strain curves for growth- and wandering-stage larvae. They demonstrate the nonlinear relationship between force and strain characteristic of most biological materials. The general form of these curves indicates that the cuticle is not stabilized and is undergoing plastic deformation during the test (Hepburn & Joffe, 1976). If the cuticle is extended through a second cycle (Fig. 3B), the force-strain curve is shifted to the right, a further indication that permanent plastic deformation accompanies cuticular extension (Cleland, 1967).

Fig. 3A shows that in wandering larvae the force—strain curve becomes much steeper when compared to the feeding stage, reflecting changes in cuticular extensibility and the tangent modulus. To determine if there was a consistent pattern of changes in cuticular mechanical properties during the instar, numerous force-strain curves were produced and analysed for each stage.

The tangent modulus was calculated for each animal’s cuticle to determine the intrinsic stiffness of the cuticle, independent of the size of the test piece (see Materials and Methods). The higher the tangent modulus, the greater the intrinsic stiffness of the cuticle.

During the growth phase the modulus decreased progressively from 3·0X10⁷ N m^-2 to 0·88X 10⁷Nm^-2 from day0 to day 3 (Fig. 4). Then the modulus was seen to increase significantly at wandering to a value of l·4x10⁷Nm⁻². The modulus then declined on the next day and remained low during the prepupal period.

Although a clear pattern of changes in the tangent modulus is observed during larval development, this measurement is of limited relevance to the mechanical functioning of the cuticle because it assumes structural homogeneity and ignores cuticular thickness (see Materials and Methods). In contrast, cuticular extensibility (as defined in Materials and Methods) has neither of these problems and therefore provides a more realistic view of cuticular mechanical function. Extensibility is the true strain produced in a given test strip when a force of 1 N is applied. From day 0 to day 3 extensibility declined gradually from 0·183 to 0·135 (Fig. 5). At wandering, a further dramatic decline to 0·071 occurred. The extensibility then increased slightly on the day after wandering (w+1) and remained relatively constant during the prepupal period. Thus, cuticular extensibility was relatively high during the feeding stage, decreased sharply to its minimum at wandering, and then remained low throughout the prepupal stage.

Cuticular water content was measured throughout the fifth instar. Fig. 6 shows that during larval growth (day 0 to day 3) the cuticle became progressively hydrated [55% ±0·9% water to 68% ±0·6% (mean ± S.E.)]. At wandering, a transient dehydration to 63 % water was observed. Then the cuticle rehydrated to about 72% water during the prepupal phase. The tangent modulus appears to be inversely related (not necessarily linearly) to the degree of hydration (Fig. 7). A similar relationship was observed when the modulus at 0·05 N force was plotted against cuticular hydration (data not shown).

Reynolds (1975b) reported that cuticular hydration in Rhodnius was strongly affected by pH such that pieces of cuticle soaked in pH 4 buffers were about 16-fold more hydrated than those in pH 7 buffers. In Manduca, soaking pieces of cuticle in different pH buffers had no such dramatic effects on cuticular hydration (Table 1).

The flexural stiffness of the cuticle was calculated using the modulus at 0-3 N force. Fig. 8 shows that the cuticle remained relatively flexible during larval growth despite increasing cuticular thickness (Fig. 2), because of the decline in the tangent modulus (Fig. 4). At wandering, the flexural stiffness increased dramatically, primarily because of a transient increase in the modulus since cuticular thickness increased little at this time. The further increase in flexural stiffness the day after wandering was apparently due to the large increase in cuticular thickness associated with axial shortening of the caterpillar, since the modulus decreased at this time. Thus, during growth the cuticle remained relatively flexible, but showed a marked increase in flexural stiffness once growth was completed.

To determine whether a decrease in cuticular extensibility and an increase in the tangent modulus is a general response to the cessation of growth, the effects of starvation were examined. Animals that were starved for 96 h (Table 2) showed a significant (P< 0·005, Student’s i-test) increase in the modulus and an almost two-fold decrease (significant to P< 0·005 level, Student’s t-test) in cuticular extensibility for their mass class, which is similar to that of the day 1 control. These changes were apparent by 19 h (extensibility= 0·101, the tangent modulus = 2·4X10⁷ N m^-2; significance compared to day 1 control P< 0·005, P< 0·05, respectively), by which time the rate of cuticular deposition was greatly reduced (Table 2 and unpublished data). In animals that were starved for 22 h and then fed for another 26 h, the modulus decreased significantly (P< 0·005) from starvation levels and was about the same as in control, unstarved larvae. Interestingly, cuticular extensibility remained low and showed no significant increase (0·3<P<0·4) after 24h of refeeding. Feeding for 48 h after 96 h of starvation caused an even more pronounced decline in the modulus, and the extensibility recovered to nearly normal levels (Table 2). No growth was observed in animals fed a diet of 2% agar, but their cuticles had the normal mechanical properties for their mass class (Table 2). Thus, bithough starvation stiffened the cuticle and reduced extensibility, the response of the agar-fed animals suggested that these effects may have been due to dehydration rather than to lack of cuticle deposition caused by starvation.

To examine the role of increased flexural stiffness during the prepupal phase, larval cuticle deposition was inhibited by DFB in a manner similar to that used by Mitsui et al. (1980). Table 3 shows that when larvae were fed a diet containing lOp.p.m. DFB from day 1 to day3, the flexural stiffness of the cuticle on the day after wandering (w+1) was reduced 16-fold as compared to untreated control larvae. Exposure to DFB for only 2 days caused a four-fold decrease.

This treatment had no effect on larval growth. Nor did it affect pupal cuticle deposition or tanning, presumably because the DFB was metabolized in the 3 days between the end of larval cuticle deposition and the beginning of pupal cuticle deposition. However, the shapes of the pupae were aberrant. The body was often bent and the cuticle was buckled. Furthermore, the body was always flattened with an oblate rather than round cross-sectional profile. This distortion was apparent in the larval cuticle of the prepupa. Thus, it appears that low flexural stiffness of the larval cuticle during the prepupal stage resulted in deformed pupae.

During the fifth instar, Manduca larvae pass through three distinct developmental phases: growth, wandering and prepupal. Pronounced changes in the mechanical properties of the cuticle accompany the transition from one stage to the next.

During the growth and feeding phase (day 0 to day 3), extensibility remained relatively high and flexural stiffness remained low, despite increasing cuticular thickness. It is possible that the progressive hydration of the cuticle during this phase permitted maintenance of an extensible and flexible condition. Alternatively, mechanical properties of the cuticle may be determined by a small region of the cuticle (for instance, the inner five lamellae) while the bulk of the cuticle is mechanically non-functional in terms of the properties that were measured. This could occur because continual cuticle deposition during larval growth results if greater strain in the outer than in the inner lamellae. When we consider that the load-strain curves are non-linear, it seems likely that lamellae that have undergone varying degrees of strain may have different mechanical properties.

During the wandering stage, a dramatic decline in cuticular extensibility was associated with the end of growth and excavation of an underground pupation cell (Figs 3, 5). Appositional deposition of cuticle also ceases as indicated by the lack of further incorporation of chitin precursors (Wolfgang & Riddiford, 1981), although a few high molecular weight proteins are still deposited by intussusception (Wolfgang & Riddiford, 1986). No further distension of the cuticle occurs, despite the elevated haemolymph pressures associated with crawling and burrowing. Thus, at a time when no further feeding occurs, a reduction of cuticular extensibility would help to prevent further undesirable cuticular growth. Starvation, which also halts growth and stimulates locomotion, appears to cause a similar decrease in cuticular extensibility, which returns to near normal values upon refeeding. However, the lack of such changes in the cuticular properties of agar-fed animals suggests that changes associated with starvation may be a secondary effect of dehydration and are not directly attributable to starvation.

In the prepupa, extensibility remained low, while flexural stiffness increased and reached a plateau at its highest levels on the day after wandering (w+1). This increase in flexural stiffness results from the greatly increased thickness of the cuticle associated with axial shrinkage of the larvae. Such an increase in stiffness converts the cuticle from a hydrostatic to a self-supporting skeleton as can be readily demonstrated by the retention of cuticular shape in an eviscerated post-wandering larva. The importance of increased flexural stiffness of the larval cuticle for the development of a normal pupa was shown by the use of the chitin deposition inhibitor DFB. Animals given a dose that had no effect on growth but prevented much of the normal increase in larval cuticular thickness and flexural stiffness had abnormal pupal shapes. These results indicate that the tremendous increase in cuticular thickness during larval growth serves initially to support the expanding larva and then later as a rigid mould for the formation of the pupa.

The changes in cuticular structure and mechanical properties that prepare Manduca larvae for pupal ecdysis have a number of parallels in the pupariation of flies. In both, upon completion of larval growth, the cuticle shrinks axially to assume a pupal shape and to form a rigid mould in which pupation can take place (for fly studies see Fraenkel & Rudall, 1940; Zdarek & Fraenkel, 1972). The principal difference is that in Manduca the larval procuticle is resorbed, and the epicuticle is shed once the thick pupal cuticle is deposited and tanned. In most flies the tanned puparium is not shed and the pupal cuticle is thin and untanned (Fristrom, Doctor & Fristrom, 1986).

Throughout Manduca development, the tangent modulus calculated at a force of 0·05 N was always less (approx. 3-to 6-fold) than that measured at 0·3 N. The consequence of the non-linear response of the cuticle to applied tension (Fig. 3) is that as stress increases the cuticle becomes stiffer (tangent modulus increases). Thus, under the low stresses applied during larval feeding and growth, the cuticle is’ relatively more extensible than under the high stresses associated with burrowing.

In contrast with Manduca’s. long period of growth in the fifth instar (4 days), postecdysial morphogenetic events in many insects are relatively rapid and elevated cuticular extensibility is transient. Cottrell (1962) demonstrated that a transient increase in sclerite extensibility permits cuticular expansion after imaginal ecdysis in the blowfly Calliphora erythrocephala. In locusts, cuticular softening occurs just prior to emergence of the adult to allow expansion of adult structures after ecdysis (Elliot, 1981). In adult Manduca, post-eclosion wing expansion is also associated with increased cuticular extensibility (Reynolds, 1977). Rhodnius also shows a transient increase in cuticular extensibility associated with feeding and abdominal expansion (Bennet-Clark, 1962; Reynolds, 1975a).

The means by which the mechanical properties of the cuticle are controlled are under considerable debate. Vincent & Hillerton (1979) and Hillerton & Vincent (1979) believe that dehydration is the proximal cause of cuticular stiffening whereas Andersen (1979) and Hackman & Goldberg (1971) invoke covalent cross-linking of cuticular proteins. In Manduca larvae the procuticle (everything but the epicuticle) is completely resorbed prior to the pupal moult, probably indicating that little or no covalent cross-linking of proteins has occurred. Furthermore, the inverse relationship between the modulus and cuticular hydration suggests that the degree of hydration may determine the material properties of the cuticle; unlike that of Rhodnius (Reynolds, 19756), the cuticular hydration of Manduca is independent of pH. The decreased extensibility from day 3 to wandering is also associated with the synthesis and deposition of at least three new cuticular proteins, while 11 other cuticular proteins cease to be synthesized (Wolfgang & Riddiford, 1986). This change in cuticular protein composition is correlated with the formation of cuticular lamellae that are ten times thinner than those previously deposited. In addition, at wandering, synthesis of lamellae ceases (Wolfgang & Riddiford, 1981) and several new proteins are deposited by intussusception (Wolfgang & Riddiford, 1986). Thus, altered cuticular protein composition, degree of hydration and morphology are temporally correlated with changes in cuticular mechanical properties. Although experimental confirmation is not yet available, we suggest that in Manduca larvae changes in cuticular composition and structure that are independent of sclerotization play a role in stabilizing the cuticle once growth is completed. Consequently, the extensible hydrostatic skeleton characteristic of growing larvae can be converted into a rigid structure in which pupation may occur.

We wish to thank Dr Robert Cleland for providing instruction in the use of his Instron mechanical testing device and for a critical reading of the manuscript. We also wish to thank Dr David S. King for his helpful discussions. This study was supported by NSF PCM 80-11152 to LMR and NIH training grant IT32HD07183 to WJW.