Introduction

The caddisfly, Micropterna lateralis (Stephens, 1837), is a species of Limnephilidae distributed across most of central, northern and eastern Europe outside of the Mediterranean Region (Fig. 1). Across its range, M. lateralis inhabits eucrenal, hypocrenal and epirithral zones of freshwaters and can be found up to 450 m a.s.l. in low mountain ranges (Graf et al., 2008). The species appears to be an indicator of small, organic streams that may be subject to periodic stream drying (Sommerhäuser, 1998). According to Sommerhäuser (1998), M. lateralis seems to be more abundant in intermittent than permanent habitats in the Lower Rhine area.

Fig. 1
figure 1

Distribution of M. lateralis across Europe based on data extracted from the data collection “Distribution Atlas of European Trichoptera” (Schmidt-Kloiber et al., 2015, 2017). The approximate location of our study sites is indicated as red star

Full size image

In order to overcome the waterless summer period in intermittent streams, many limnephilids evolved an imaginal diapause (Novak & Sehnal, 1963; Crichton, 1971; Denis, 1978), where adults emerge with underdeveloped ovaries and aestivate in dark and humid places (e.g., caves) until flow resumes. M. lateralis populations from Sweden and Britain show a short flight period in spring and summer (June/July) without diapause (Crichton, 1971; Svensson, 1972). Sommerhäuser (1998) reported adult emergence in May and June, while stating the existence of diapause to be not fully clarified. Longer flight periods are indicated in Tobias & Tobias (1981), lasting from May to October. In France, single imagines were regularly observed in subterranean environments (i.e., natural and artificial cavities; Bouvet, 1976), presenting a characteristic behaviour of diapausing species. Such geographic differences in adaptations have been observed in the congener Micropterna sequax McLachlan, 1875, where Swedish populations have no diapause (Svensson, 1972) but southern European populations do (Denis, 1976). Based on experimental evidence from Limnephilus rhombicus (Linnaeus, 1758), this may be explained by the differences in photoperiod length between higher and lower latitudes (Denis, 1978). Whether a similar geographic pattern is present in M. lateralis remains to be investigated.

M. lateralis larvae are mainly grazers and shredders, but also show predatory behaviour (Graf et al., 2008). They are expected to develop through five distinct instars within a year, yet relatively little is known about the size ranges of individual larval stages. The larvae are the main feeding stages and accumulate the necessary energy reserves for metamorphosis, as well as sustenance and reproduction of adults. The central energy storage in insects is the fat body, an organ unique to insects (Law & Wells, 1989). In the fat body, reserves are primarily stored as glycogen, triglycerides and proteins (Harrison et al., 2012). The amount of stored lipids is usually higher than that of carbohydrates across insects (reviewed in Arrese & Soulages, 2010), while the total amount of accumulated reserves varies between species. Proteins only present a minor metabolic energy storage component (e.g., in the caddisfly Sericostoma vittatum Rambur, 1842; Campos et al., 2016). The ability to store energy is an important ecophysiological trait fundamental in insect lives, yet autecological studies of growth-related energetics within natural habitats are rare.

Thus, our primary objective in this study was to provide an overview of life history characteristics that could affect the ability of M. lateralis to exploit intermittent habitats. Specifically, we aimed to clarify the existence of an imaginal diapause in M. lateralis by determining ovarian maturation of newly emerged adults. Notes on other caddisfly species present at the study streams will be given for comparative reasons. Further, we aimed to differentiate M. lateralis instars based on head capsule width and characterize growth, energy storage management and case material during larval development. In order to assess the adaptive potential of these traits for intermittent stream caddisflies, the characteristics are compared with those of two other intermittent stream caddisflies (M. sequax and Stenophylax permistus McLachlan, 1895).

Methods

Study site and field sampling

We studied the intermittent sections of the Auerbach (50°16′34.2″N, 9°25′58.7″E) and Klingbach (50°14′57″N, 9°26′23″E). Both are lower mountain streams (around 200–430 m a.s.l.) situated in the Rhine-Main-Observatory LTER-D site (Haase et al., 2016) in the Hessian Spessart (Germany) in mixed deciduous forest (dominant taxa: Fagus sylvatica L., Urtica dioica L. and Bryophyta). An approximate location of those sites is indicated in Fig. 1 showing the European distribution of M. lateralis based on occurrence data extracted from the “Distribution Atlas of European Trichoptera” (Schmidt-Kloiber et al., 2015, 2017). Sampling point coordinates were used to generate a map using R statistical software (R Core Team, 2016).

At the Auerbach, the summer dry period starts around mid-June to mid-July and lasts for two to four months, but is interrupted by several, precipitation-driven, short flow events (years of observation: 2014–2016). Water temperature ranges between 1.1 (01/19/16) and 9.8°C (04/05/16), based on weekly measurements from October to April (2015 and 2016). Other environmental factors are stable over the flow period (pH 7.1 ± 0.1; electrical conductivity 158.4 ± 8.8 µS cm−1; oxygen concentration 11.3 ± 0.4 mg l−1; mean ± se October–April 2015/16, N = 17). Hydrological patterns are more variable at Klingbach. Drying starts between April and mid-July and lasts for five to eight months (years of observation: 2014–2016). During our study period, drying was not interrupted by intermediate flow events. Weekly measured water temperatures show little fluctuations ranging from 7.0 to 9.1°C, while other factors remained stable during the flow period (pH 5.9 ± 0.04; electrical conductivity 99.5 ± 0.9 µS cm−1; oxygen concentration 11.3 ± 0.1 mg l−1; mean ± se October–April 2015/16, N = 15).

At both streams, emergence traps were operated from April/May until September/October in 2014 to 2016. Specimens were preserved in ethanol and identified using Malicky (2004).

Larval specimens of M. lateralis were collected from mid-October 2015 until early April 2016 at Auerbach. Sampling was performed at a variable interval of one to three weeks from October to December, biweekly in February and weekly in March and April (five specimens per sampling occasion). In addition, four specimens of S. permistus and M. sequax were sampled in March (Klingbach) and April (Klingbach/Auerbach), respectively. In the field, animals were hand-picked from leaf packs, removed from their case and flash frozen in liquid nitrogen. In the laboratory, larvae were identified using DNA barcodes (mtCOI; Hebert et al., 2003).

Imaginal ovarian maturation and larval growth

In the laboratory, we dissected adult M. lateralis females caught in emergence traps (2014–2016; N = 49) to determine ovarian maturation, following the four-stage classification of Novak and Sehnal (1963): ‘A’: immature female; ‘B’: maturing female; ‘C’: mature female; and ‘D’: female after oviposition. Caddisfly species other than M. lateralis were also assessed for comparative purposes (N = 38).

To determine larval stages, heads of all collected larvae of M. lateralis (N = 83), M. sequax (N = 4) and S. permistus (N = 4) were cut off in order to measure head capsule width (HCW) as the maximum width of the head capsule using a binocular (Olympus SZX 16, Frankfurt, Germany). During head preparation, larval bodies were kept frozen and later freeze-dried and weighed to the nearest mg. Caddisfly cases were air-dried and assigned to three types of building material: organic, mineral and mixed organic/mineral.

Energy storage assay and analysis

Homogenized freeze-dried larval bodies were used to determine glycogen and triglyceride contents by enzymatic assays. Because of the minimal amount of tissue required for the enzymatic analysis, larvae were allotted to glycogen and triglyceride analysis based on their dry weight: individuals lighter than 1 mg were used in glycogen analysis and individuals of 1–2 mg in triglyceride analysis, whereas in individuals heavier than 2 mg both storage compounds were analysed. We used all collected specimens per sampling date in case of smaller instars (II, III) and three specimens each for instars IV and V (exception: N IV = 4 at 01/05/2016).

For analysis of glycogen concentrations, samples were extracted in perchloric acid (PCA). After neutralization with KHCO3 (2.0 M), glycogen was hydrolysed using amyloglucosidase (Sigma-Aldrich, Steinheim, Germany) during a 2-hour incubation (40°C) in acetate buffer (0.2 M; pH 4.8). The enzyme was thermally immobilised (5 min at 100°C), the samples were centrifuged (10 min at 18,600×g and 4°C, 200R, Hettich, Tuttlingen, Germany) and the supernatant was photometrically (Specord 200 Plus, Analytik Jena, Jena, Germany) analysed following Keppler & Decker (1984) in a modified enzymatic assay described by Becker et al. (2013).

To extract triglycerides, ice-cold hexane was added to the samples prior to 10-minute incubation on ice, followed by a centrifugation step (10 min at 19,900×g and 4°C) using a cooled micro-centrifuge (200R, Hettich, Tuttlingen, Germany). The supernatant was transferred into glass vials and placed under the fume hood to evaporate the solvent. Enzymatic analysis of triglyceride concentration was performed using a commercial kit (Triglyceride FS, DiaSys Diagnostic Systems GmbH, Holzheim, Germany) and a spectral photometer (Specord 200 Plus, Analytik Jena, Jena, Germany; for details see Winkelmann & Koop, 2007). Glycogen and triglyceride concentrations were expressed as μmol g−1 animal dry weight and % of animal dry weight, while energy content was expressed as kJ g−1 animal dry weight. Energy stored in glycogen and triglycerides was calculated using an energy content of 15.6 kJ g−1 glucose (180 g mol−1) and 37.3 kJ g−1 tripalmitin (807 g mol−1; Wieser, 1986), respectively. Tripalmitin is the dominant compound of storage lipids (Wieser, 1986). The total amount of stored energy is referred to as the sum of the energy stored as glycogen and triglycerides measured in the same individual and does not account for additional energy stored in the form of proteins. In instars III, glycogen and triglycerides could not be measured from the same individual (see above). Thus, the total stored energy is calculated from the average of energy stored in glycogen plus the average of energy stored in triglycerides (note that the standard errors were also summed).

Data analysis

To detect differences in dry mass or energy storage between instars, one-way ANOVAs with the factor ‘instar’ and the dependent variables dry weight, triglyceride concentration or glycogen concentration were performed. To achieve the normal distribution and the best approximation to variance homogeneity, data were transformed after analysing the optimal exponent for transformation via the ‘boxcox’ function in R (exponents 0.15 for dry weight and glycogen, no transformation for triglycerides). Differences in energy content were analysed using a t test, because sufficient data were available only for the fourth and fifth instars. We employed a piecewise linear regression analysis (Crawley, 2007) to detect potential break points in the seasonal dynamics of energy storage components. This analysis allows fitting separate regression slopes to characterize the relationship between two variables. The resulting model was compared to a simple exponential curve fitting. All statistical analyses were done in R statistical software (R Core Team, 2016).

Results

Emergence and ovarian maturation

Adult M. lateralis were caught in emergence traps from mid-May to August under wet and dry hydrological conditions in both studied streams. The species was found together with caddisflies that are typical of intermittent streams (Auerbach: M. sequax, Limnephilus cf. sparsus Curtis, 1834, Plectrocnemia conspersa Curtis, 1834, Rhyacophila praemorsa McLachlan, 1879, R. philopotamoides McLachlan, 1879, Beraea pullata Curtis, 1834, Synagapetus moselyi Ulmer, 1938, Wormaldia occipitalis Pictet, 1834; Klingbach: M. sequax, L. cf. sparsus, L. bipunctatus Curtis, 1834, P. conspersa, Sericostoma flavicorne Schneider, 1845, Stenophylax permistus, S. mitis McLachlan, 1875). A list of Plecoptera species is available in Supplementary Material Table S1.

Dissection of female M. lateralis revealed that specimens emerged with ovaries in stages B (N = 23), C (N = 9) or D (N = 17), but not in the immature stage A. However, freshly emerged females in ovarian development stage A were found in M. sequax, Limnephilus cf. sparsus, L. bipunctatus and S. mitis (Table S2).

Larval growth

Four M. lateralis instars were clearly distinguished within our samples, showing no overlap in head capsule width between subsequent instars (Table 1). The highest values matched those reported for final instars V in Waringer & Graf (2011) indicating the presence of instars II–V within our samples: instar II occurring in October (N = 4), instar III from October to December (N = 15), instar IV from November to January (N = 22) and instar V from November to April (N = 42). Mean weights of early instars II and III were similar, while a significant increase in weight occurred from instar III to V (Fig. 2a; Table 2). Most accrual of biomass happened in the long-lived instar V.

Table 1 Summary of growth and energy measurements expressed as mean ± SE for instars II–V of M. lateralis (number of samples given in brackets) and instar V of M. sequax (N = 4) and S. permistus (N = 4)
Full size table
Fig. 2
figure 2

Seasonal changes in the dry weight, concentrations of glycogen and triglycerides and the amount of stored energy of M. lateralis during larval development (mean ± SE)

Full size image
Table 2 Results of ANOVAs and post hoc tests of differences in dry weight and energy storage substrates between instars of M. lateralis (N as indicated in Table 1)
Full size table

The case building material in M. lateralis differed between instars and was characterized by a shift from purely mineral cases (instars II & III) to mixed mineral/organic cases (IV) to purely organic cases in the last instars (V). Final instars of S. permistus used organic case material, while M. sequax had mixed cases at the Auerbach and organic cases at the Klingbach.

Metabolic energy storage

In M. lateralis, we observed a significant increase in glycogen concentration towards the fifth instar, while triglyceride concentrations did not change significantly (Tables 1, 2). Triglycerides were the main energy storage component in instars III and IV accounting for more than 59.5 ± 7.2% (mean ± standard error, N = 18) of total stored energy. In the last instar (V), however, triglycerides represented only 42.2 ± 2.8% (mean ± standard error, N = 32) of the stored energy and glycogen became the more important energy storage component. The weight-specific amount of total stored energy showed a significant difference between instars IV and V. Glycogen accounted for ~2% of total dry weight in instars II to IV and increased to ~10% in final instars. Triglyceride content slightly decreased with instar succession from 4.6 (III) to 2.8% (V) of total dry weight (Table 1).

The seasonal dynamics of metabolic energy storage in M. lateralis differed between glycogen and triglycerides. Glycogen concentrations increased over time and closely followed larval weight increase within instar V (Fig. 2b). This increase was steeper in spring compared to autumn and winter based on the segmented linear regression analysis (Fig. S1; better fit of linear than exponential curve). On the other hand, concentration of triglycerides showed no systematic seasonal change (Fig. 2c). The seasonal dynamics of total energy content resembled the pattern of triglycerides from November until early February (Fig. 2d). After that, the curve was more similar to the course of glycogen concentrations over time. The total amount of energy per individual followed a linear relationship with total dry weight, as did the amount of energy stored as glycogen and triglycerides per individual (Fig. 3). The increase is higher for glycogen than triglycerides, again showing that larvae tended to store more energy as glycogen compared to triglycerides.

Fig. 3
figure 3

Linear relationships of individual dry weight of final instars of M. lateralis against energy content per individual. Three regression lines are shown for the total amount of energy content as well as separate amount for glycogen and triglycerides. Coefficients for each model are as follows: Glycogen + Triglycerides: y = −0.023 + 0.00353 × x, R 2 = 0.816; Glycogen: y = −0.0179 + 0.00224 × x, R 2 = 0.789; Triglycerides: y = −0.0051 + 0.00129 × x, R 2 = 0.719

Full size image

In comparison, final instars of the species M. sequax (N = 4) had similar glycogen concentrations and energy stored as glycogen as M. lateralis, while the amount of energy stored in triglycerides was somewhat higher (Table 1). In S. permistus, the difference in glycogen and triglyceride concentrations, and consequently the amount of energy stored as glycogen or triglycerides, was even more pronounced (Table 1). In M. sequax and S. permistus, glycogen accounted for 53.3 ± 6.1 and 70.5 ± 4.3% (mean ± standard error, N = 4) of total stored energy, respectively.

Discussion

In this study, we characterize growth and energetics through four instars of M. lateralis in their natural habitat. Overall, developmental patterns are as expected showing a relative increase in head capsule width of approximately 1.5 following Dyar’s rule of a constant size ratio between moults (Dyar, 1890). A major finding is the accrual of glycogen reserves in the last instars, while lipid contents are more or less constant. More specifically, the species shifts from a period where energy is stored primarily as triglycerides to predominant glycogen storage before emergence. This is both surprising and highly unusual, as lipids typically are the major storage component throughout larval development as well as in adult stages of insects (reviewed by Arrese & Soulages, 2010). Among aquatic insects, this is, for instance, described in mayflies (Winkelmann & Koop, 2007) and mosquitoes (Timmermann & Briegel, 1999). In experimentally held caddisflies, two studies report a lower weight-specific glycogen than lipid content in final instars of L. rhombicus (Mondy et al., 2011) and S. vittatum (Campos et al., 2016). Most of the earlier work on caddisfly development primarily focuses on lipids rather than carbohydrates: lipids account for up to 14, 22 and 18% of individual dry weight in larval Potamophylax cingulatus (Stephens, 1837), P. nigricornis (Pictet, 1834) and Parachiona picicornis (Pictet, 1834), respectively (Sehnal, 1963; Otto, 1974). In M. lateralis, proportions of triglycerides are markedly lower (up to 8.5% of individual dry weight) and stabilize at roughly 3% in final instars, while glycogen represents around 10% of individual dry weight, indicating the high importance of glycogen as an energy storage compound in M. lateralis compared to the abovementioned limnephilids. This is also true for final stages of S. permistus and M. sequax. Independent of the specific energy-storing components, the total energetic content of M. lateralis individuals increases linearly with the accumulation of larval weight. A high amount of available energy at the end of the last instar is crucial for successful metamorphosis and to support life and reproduction as adults (Ziegler, 1991). Therefore, lipid reserves supply energy and fatty acids needed for the production of eggs (e.g., in mosquitoes; Van Handel, 1993) as well as energy for flight (reviewed in Beenakkers et al., 1985). Likewise, glycogen can be converted to trehalose as a substrate for insect flight as was shown in Anopheles gambiae Giles, 1902 (Kaufmann & Brown, 2008) or Locusta migratoria (Linnaeus, 1758) (Van der Horst et al., 1980). Glycogen is further used during pupation, e.g., in the form of glucose for the synthesis of chitin (e.g., in Phormia regina Meigen, 1826; Tate & Wimer, 1971). Other than that, the high glycogen content that we observe in M. lateralis and also M. sequax and S. permistus could be of adaptive significance in intermittent stream species. This hypothesis is based on the fact that glycogen can be converted to trehalose and sugar alcohols under drought stress (e.g. in Polypedilum vanderplanki Hinton, 1951; Watanabe et al., 2002) and binds bulk water as well, thus increasing the total pool of available water. Desiccation-resistant selection lines of Drosophila melanogaster Meigen, 1830 store high levels of carbohydrates that provide additional energy for dehydration acclimation and recovery (Chippindale et al., 1998; Djawdan et al., 1998). Similarly, L. rhombicus, a caddisfly species inhabiting small ponds that may be subject to drying, shows high carbohydrate contents (but not glycogen; Mondy et al., 2011). Further, a positive correlation between glycogen levels and desiccation resistance is reported for Aedes mosquitoes (Sawabe & Mogi, 1999). Thus, glycogen in final M. lateralis instars is potentially accumulated in large quantities to serve as nutrient store and desiccation protectant of larvae, pupae or egg masses.

In this study, we focus on larval stages for energy storage analyses and do not include pupal or adult specimens as well as eggs. Thus, the exact fate of energy reserves and the allocation of resources remains unknown. Generally, resource allocation at metamorphosis depends on the life history strategy and the future requirement of the species (Boggs, 1981). For example, the adult life span determines the amount of energy resources used for somatic investment compared to reproduction (Karlsson & Wickman, 1989). In a study on adult limnephilid caddisflies, Stevens (2000) tested this idea by investigating male thorax (somatic investment) and abdomen (reproductive investment) dry mass in species with different life histories. He found variation in the change of dry mass among the studied species, which was, however, more related to mating system (i.e., monandry or polyandry) than flight period length (Stevens, 2000).

The studied population of M. lateralis shows a flight period length of over three months (mid-May to August) without an imaginal diapause. Thus, our results support the classification of M. lateralis into a group of species characterized by “a shorter fight period, without a diapause, in spring and summer, and sometimes extending into autumn” following Crichton’s (1971) assessment of populations in Great Britain. Consequently, there is no evidence for a geographic difference in adult adaptation between Swedish (Svensson, 1972), British (Crichton, 1971) and German (this study) populations. Regarding other caddisfly species present at the study streams (Table S2), both an extended flight period into autumn and an imaginal diapause were confirmed in M. sequax (Denis, 1976), L. cf. sparsus (Novak & Sehnal, 1963; Svensson, 1972) and S. mitis (Bouvet, 1978). In line with lacking a diapause, adults of M. lateralis must lay their eggs well before stream flow resumes. As shown in other temporary-water limnephilids, we expect eggs to be laid in dry substrates or damp vegetation, where first instars develop within gelatinous masses. Likely, the larvae quiesce for several months (Stevens, 2000) and actively leave the gelatin following long-lasting rainfall or total submersion in water (e.g. Novak & Sehnal, 1963). Once they enter the water, they are expected to show fast larval development as we observed in the remaining four instars.

In nature, both larvae and pupae experience and resist drought, as stream drying does not terminate emergence of imagines in M. lateralis and other caddisflies (also see Bohle, 2000). Another life history characteristic of potential importance is the observed shift from mineral to organic case material in final instars. While this behaviour is not unique, organic cases can hold water more efficiently than mineral cases (Zamora-Muñoz & Svensson, 1996). Hence, the shift in building material could increase the species’ ability to exploit intermittent habitats and present an advantage for both final instars and pupae. This also applies to the final instars of S. permistus and M. sequax that use organic or mixed mineral/organic cases.

In conclusion, our results show that the studied limnephilid caddisfly species M. lateralis as well as M. sequax and S. permistus present an exceptional case of resource accumulation that differs from the commonly observed pattern in both terrestrial and aquatic insects. This result highlights the importance of species-specific investigations of energy storage management during insect development in order to capture the full diversity of energetic strategies. Moreover, our findings represent a promising baseline for future comparative studies in order to understand the link between energy storage and drought adaptation. We further emphasize the need for continued studies of the complex life histories of insects inhabiting highly dynamic systems such as intermittent streams.