The larvae of Parnassius apollo ssp. frankenbergeri Slabý originated from a semi-natural open-field breeding colony in the Pieniny National Park, set up for a recovery plan in the Pieniny Mts. The individuals for enzymatic analyses were obtained from additional eggs laid by Apollo females in the preceding year. In accordance with the permission from the Polish Ministry of Environment, the larvae of the fourth (L₄) and fifth (L₅) instars were chosen from the stock colony. Morphological and behavioural criteria allowed further separation of L₅ individuals into two subgroups: ‘young’ – larvae up to the third day after moulting, and ‘hyperphagic’ – larvae that most of the time spent on feeding. Hence, the gut size and the amount of digested food were the major differences between the subgroups.

The larvae were anaesthetised with CO₂ and kept at low temperature until dissection in 1.15% KCl solution at 4 °C. The midgut was cut out between a crop and a hindgut. The remains of trachea, fat body and Malpighian tubules were removed. The dissected midgut was cut open and rinsed in 0.1 M phosphate buffer (pH 7.4) containing 1 mM EDTA and 10% glycerol to obtain gut contents enclosed in peritrophic membrane. The cleared midgut wall was homogenised in the same buffer and then centrifuged for 10 min (10000g) at 4 °C. The midgut contents were centrifuged at 3000g for 10 min in order to remove the undigested leaf pieces. The resulting MT and MC supernatants were portioned out into Eppendorff tubes, and kept frozen at −66°C until the analyses of enzymes activities. Hence, from one individual, two samples were obtained: the midgut tissue together with glycocalyx (MT) and the liquid gut contents with peritrophic membrane (MC).

Appropriate pH optima for α-amylase, saccharase and α-glucosidase were assessed using 0.05 M citrate-phosphate (Mc Ilvaine's) (pH 2.6–7.0), 0.1 M phosphate (pH 5.8–7.8), 0.05 M Tris-HCl (pH 7.4–9.0) and 0.1 M bicarbonate (pH 9.2–10.7) buffers. Ionic strength of the buffers was adjusted to the values applied to other Lepidoptera (see Table 1 for references). Following the recommendation of the International Union of Biochemistry and Molecular Biology, all the enzymatic assays in P. apollo were done at 30 °C.

The procedures of the enzymatic assays as well as substrates used for analyses are summarized in Table 1. Routine procedures were applied with initial substrate concentration ensuring linear reaction for the incubation time. This was confirmed for the most active α-glucosidase. The reaction was linear at least four times longer than the incubation time (r=0.998) (Table 1). For this enzyme, temperature dependence (16–37 °C) of the reaction rate was measured as well. Enzymatic activities were assayed colorimetrically and expressed per milligram of protein content, determined according to Bradford [42]. Due to a limited number of available individuals for planned analyses, enzymatic assay for each larva was duplicated and the obtained mean taken as single result for further calculations.

The significance of differences was assessed with ANOVA/MANOVA. T Tukey's test for unequal sample size was applied for post hoc analysis. Test t for dependent variables allowed estimation of significant differences within MT and MC samples in the same larval instar (p<0.05). Pearson's linear correlation coefficients between activities of the enzymes were calculated at p<0.05.

The highest activity of α-amylase, measured in relation to various media, was stated at pH 9.0. For saccharase, it was at pH 5.6, and for α-glucosidase, in the 4.9–5.6 pH range in citrate–phosphate buffer for both enzymes (Fig. 1). For the latter, the increase of incubation temperature from 30 to 37 °C doubled the enzyme activity. In comparison with 16 °C, the rate of hydrolysis at 37 °C was 7.4-fold higher.

In general, the enzymes hydrolysing polysaccharidic linkages had higher activity in the samples of liquid gut contents (MC), contrary to the activities of disaccharidases and specific glycosidases that were higher in the midgut tissue with glycocalyx (MT). The ‘young’ and ‘hyperphagic’ L₅ larvae did not differ significantly in their efficiency toward polysaccharides and disaccharides digestion. Hence, these results were pooled for comparison with the corresponding activities measured in L₄ larvae (Table 2).

The activities of α-amylase and cellulase in L₅ larvae were significantly higher in MC in comparison with MT (p=0.0002 and p=0.004, respectively). Similar, although not significant, results were obtained in the L₄ larvae. The activity of cellulase in a given compartment was much lower than that of α-amylase (Table 2, Fig. 2).

On the contrary, the activities of disaccharidases were higher in MT than MC samples. Assayed MT disaccharidases of the L₅ differed in specific activities among themselves (p<0.05, n=14). The highest total hydrolytic activity was stated for saccharase – it was twice as high as the activity of maltase (p=0.001, n=14). The activity of lactase was practically not detected in either of the fraction or the instars. The activity of disaccharidases in the L₄ did not differ significantly from those in the L₅, except for cellobiase activity in MT samples (p=0.048). The differences in the activity of enzymes in studied midgut compartments of the L₄ caterpillars were found only for maltase and trehalase (p=0.001 and p=0.0002, respectively), the activities of which were always higher in MT samples.

Assays of specific glycosidases activities provide another insight into the digestion of oligomeric compounds containing glycosidic linkages. Hence, MT samples were much more active in this respect than MC ones. In this compartment the total hydrolytic activity of glucosidases was higher than that of galactosidases, with the highest activity of α-glucosidase (Fig. 3). Among the assayed galactosidases, a higher activity was stated for β-galactosidase. Hydrolytic activities of both glucosidases and β-galactosidase in MT of the L₅ larvae were significantly higher than in the previous instar. Comparison of ‘young’ and ‘hyperphagic’ larvae revealed significantly higher activities of assayed glucosidases and galactosidases in MT samples of the latter (Fig. 4). Hence, activities of these enzymes in L₄ larvae were lower than in ‘young’ individuals of the last instar. No differences in the specific glycosidases activity were found between liquid midgut contents (MC) of both larval instars as well as between the midgut compartments in the L₄ (Table 2).

The activities of enzymes hydrolysing α-glucosidic linkages in starch (α-amylase, maltase/α-glucosidase) from the samples of the midgut tissue and the liquid gut contents of the L₅ larvae were highly and positively correlated with one another (Table 3). It is noteworthy that α-amylase activity was much higher than that of maltase in both gut compartments. This points out relative abundance of poly- and oligosaccharides (consisting of more than six saccharide units) in the leaf tissue. Significant positive correlations between amylase and α-glucosidase activities and body mass of L₅ larvae strongly indicate a nutritional role of these saccharides (Table 4).

Digestion of β-glucosidic linkages in plant β-glucosides presented a different picture. Negative correlation between activities of cellulase and cellobiase in the MT samples of the last instar was stated. There was no correlation between the enzymes in MC, as well as between the activities of cellulase and β-glucosidase in both compartments (Table 3). Cellulase activity in MC was negatively correlated with the body mass of that instar (Table 4). On the contrary, cellobiase activity very strongly correlated with β-glucosidase activities in MT samples of the L₅ (Table 3, Fig. 5). These findings indicate a negligible nutritional role of the cellulose, although shorter β-glucosides may have some importance.

Positive and significant correlations between various enzymes hydrolysing glycosidic linkages were stated for p-nitroanilide derivatives only in midgut tissue. They point out the possibility of simultaneous digestion of different saccharide-containing compounds (Table 3).

Alkaline pH optima for the assayed enzymes in P. apollo larvae were similar to the ones stated for the majority of the examined lepidopterans. The highest activity of α-amylase in Erinnyis ello (Sphingidae) and Ephestia cautella (Phyticidae) larvae was observed at pH 9.8 and 9.0, respectively [11,43]. Optimal pH for starch hydrolysis may have a broad range (e.g., from 5.5 to 9.0 in Earias vitella (Noctuidae)) [44], but this was not the case for Apollo larvae.

On the contrary, saccharase and α-glucosidase in the examined species had the highest activity in slightly acidic conditions. In other lepidopteran larvae, the optimal pH for disaccharidases and specific glycosidases ranges between 5.3 for trehalase in wax moth Galleria mellonella (Pyralidae) [45] and 8.0 for α-glucosidase in L₃ larvae of E. vitella [44].

According to van't Hoff's law (Q₁₀ law), the metabolic rate of an ectotherm doubles for every 10 °C increase in environmental temperature. Thus, increasing temperature results in internal changes that have an exponential effect on metabolism, hence various enzymes activities. Q₁₀ values for α-glucosidase activity in the L₅ P. apollo larvae were 2.4 for 16–24 °C and 2.5 for 30–37 °C range. Wide temperature range of this hydrolytic activity may reflect adaptation to high daily and seasonal temperature amplitudes (typical of mountain biotopes). Measurements done with P. apollo ssp. vinningensis proved that body temperature of dark brown caterpillar rearing in sunny places reached over 45 °C [46]. Such high temperatures are tolerated by Apollo larvae, although they seem to be above optimum for feeding and digestion, since the highest Q10 value (Q10=3.1) was stated for the 24–33 °C range. Tolerance of midgut glycolytic enzyme (trehalase) to high temperatures was stated also in tropical Orthaga exvinacea (Pyralidae) [47].

The activity of α-amylase in P. apollo was much higher than that of cellulase. Various authors use different basis for expression of the enzyme activities (per tissue fresh weight, per individual or per the whole organ). Hence, it is difficult to compare the obtained results with data provided by other authors [11,43]. The activity of α-amylase in P. apollo midgut detected mainly in the endoperitrophic space may result from a mechanism of enzyme production, storage and secretion by epithelium cells, similar to that stated in larvae of E. ello [43] and Spodoptera frugiperda (Noctuidae) [48]. These authors suggested that α-amylase is secreted into the gut lumen in small vesicles via microapocrine process. The enzyme is then released from the vesicle or, still bound with the vesicle membrane, is incorporated into the peritrophic membrane. Confirmation of this mechanism and identification of α-amylase distribution in P. apollo midgut needs further studies with immunocytochemical methods.

Cellulase is usually produced by the endosymbionts living in the insect alimentary tract [49]. In P. apollo larvae, this issue has not been studied so far. There is indirect evidence against their presence. Firstly, the cut leaf pieces pass quickly through the gut, giving little time for digestion by microbiotic organisms. Secondly, undigested fragments of orpine tissues, which can be easily observed during preparation of the gut, were not crushed and macerated. Thirdly, dissection of the gut revealed no structures, where the symbionts could survive. However, negative correlation between cellulase and cellobiase activity in MT samples and lack of correlation between cellulase and β-glucosidase activities (see Table 3) suggest that the cellulase activity detected in the samples was of exogenous origin. Since microbiota can reside in the gut also in the absence of any special structures (e.g., adhered to the peritrophic membrane or attached to the midgut surface), we cannot univocally say that they do not participate in the β-glucanase activity found in MC samples [50].

Disaccharidases and specific glycosidases activities detected in the midgut tissue of Apollo larvae indicate their presence in the glycocalyx or exoperitrophic space. They were also found in gut tissue of E. ello [43] and S. frugiperda [51]. Higher activities of these enzymes in Apollo MT samples support the view that final digestion in evolutionary advanced insects (e.g., Diptera, Lepidoptera) depends on the enzymes immobilised on the surface of midgut cells [52]. Our results suggest that in Apollo larvae saccharose is probably the most effectively hydrolysed disaccharide. This agrees with the finding that it is the main dicarbohydrate in host plant species S. telephium [36].

Trehalase activity in midgut tissue of P. apollo L₄ and L₅ larvae was 20- and 2-fold lower than in O. exvinacea and Lymantria dispar (Lymantridae), respectively [47,53]. Trehalose is an important energy source in insects and trehalase activity may serve as an indicator of energetic reserves resulting from the availability of carbohydrate compounds in their food [8]. Trehalose may also maintain nutrient absorption, osmotic regulation and other processes in midgut tissue [53]. Trehalose passes, along the concentration gradient, from the hemocel to the midgut lumen, hence low activity of trehalase in the Apollo midgut tissue may indicate low reserves of energy or a slight loss of this disaccharide resulting from its diffusion, and/or low intensity of the processes mentioned above. Possible inhibition of the enzyme activity by toxic β-glucosides and their aglycones from the Apollo food-plant remains to be justified [50]. Very low trehalase activity in the liquid midgut contents samples may also suggest the lack or negligible amount of trehalose in orpine leaves.

The ability to hydrolyse cellobiose may result either from a distinct enzyme, which ensures the complex cellulose digestion, together with β-glucosidase, or from the activity of A class β-glycosidase (former class 1 [52]), the active centre of which can bind a wide spectrum of substrates [50]. The earlier investigations revealed that S. frugiperda larvae have only one hydrolytic enzyme breaking down β-glycosidic bonds. The activity of this enzyme towards various substrates was the same in all cytosol fractions of the midgut tissue. Apart from this fact, the larvae have a well-developed detoxifying system for the neutralisation of toxic aglycones, released during digestion of plant glycosides. According to Ferreira and co-workers, this feature is characteristic for the more evolved insects [13,14]. To utilize plant glycosides as an energy source and avoid poisoning, less specialised insects selectively decrease the activity of β-glucosidase enzymes, responsible for the release of toxic aglycones [12]. Then the range of various β-glycosidic enzymes synthesized in the gut of particular species may reflect this coevolutionary trade-off [14]. Very high, in comparison to other Lepidoptera, β-glucosidase activity found in Apollo larvae suggest that its feeding strategy allows utilisation of at least some compounds containing β-glycosidic bonds. High activities of detoxifying enzymes in P. apollo midgut tissues [33] support this suggestion.

Yu [12] found that β-glucosidase activity increases in subsequent noctuid larval instars. Cellobiase/β-glucosidase, β-galactosidase and also α-glucosidase activities were higher in L₅ than in L₄ also in P. apollo larvae (Table 2). Such an increase in glycosidic enzymes activities may result from the greater food intake by older instars. Higher average glycosidic activity (Fig. 4) in MT samples from hyperphagic larvae fits well with this suggestion. Fast passage of chyme along an alimentary tract forces an insect to enhance the production of digestive enzymes to a level ensuring sufficient amounts of nutrients for absorption. Similar differences in cellobiase and β-glucosidase activities in hyperphagic and younger larvae, as well as strong positive correlation between these enzymes (Fig. 5, Table 3), suggest involvement of the same enzyme in digestion of glucose-containing plant compounds.

The lowest activity of α-galactosidase stated in P. apollo larvae may result from its low affinity for the substrate used – NP-α-Gal (p-nitrophenyl α-galactopyranoside). This was described in S. frugiperda (high Km for NP-α-Gal), hence the enzyme activity towards this substrate resulted only from high Vmax [51]. It does not seem possible that α-galactosidase activity in Apollo caterpillars is of endosymbiotic origin.

Yu [12] and Ferreira and co-workers [14] demonstrated that there was no correlation between the glycosidic activity and extent of insect polyphagy. However, the studied monophagous P. apollo subspecies [31] should develop a strategy of efficient utilisation of nutritional components, also secondary compounds, present in the host plant. One of the reasons is the very high water content of the leaves (⩾90%) [27].

It is concluded then that Parnassius apollo ssp. frankenbergeri caterpillars are equipped with a wide range of glycolytic enzymes. High α-amylolytic, but also other glycosidic activities, indicate that apart from starch also other compounds, including plant allochemicals, are utilized as energy sources. The set of digestive enzymes present in the gut of Apollo larvae probably reflects strict adaptation to its host plant phenology. This is supported by our earlier studies on larval development on other Sedum plants [31], as well as by other researchers and Apollo caterpillars breeders [6,54–56]. Incompatibility in caterpillars and host plant development (caused by, e.g., weather anomalies [24,26,28]) or experimental changes in food quality (e.g., caused by heavy metals pollution [25,30,56]) affected larval development into adult and their survival [31].