Introduction

In recent years, it has become apparent that, if we want to understand the movement kinematics of a limb, we need to take into account limb size, together with active and passive properties of the muscles that drive its movements in addition to the motor output of the nervous system (Hooper et al. 2007a, 2009; Blümel et al. 2012a). One method to classify muscles apart from the physical measurements of muscular contractile properties is based on determining their metabolic or contractile protein profiles. Both are known to match physiological parameters, and can be measured by means of biochemical and histochemical techniques (Jahromi and Atwood 1969; Sherman and Atwood 1972; Silverman et al. 1987; Pette and Staron 1988; Müller et al. 1992; Pilehvarian 2015). One of these methods is the histochemical determination of myofibrillar adenosine triphosphatase (mATPase) activity, the enzyme that hydrolyzes ATP during the cross-bridge cycle of muscle contractions (Barany 1967; Barnard et al. 1971). The amount of hydrolyzed ATP in sections of muscle fibers over time gives a very good correlate for the contraction speed within the muscle fibers (Barany 1967; Müller et al. 1992; Maier et al. 1984, 1986), and, in invertebrates, may even reflect the innervation pattern by phasically or tonically active motor neurons (e.g. Rathmayer and Maier 1987; Müller et al. 1992; Günzel et al. 1993; Gruhn and Rathmayer 2002). The correlation of mATPase activity with contraction speed is based on the presence of mATPase isoforms, which show specific pH sensitivity and can be found at different expression levels in slowly to fast contracting muscle fibers (Padykula and Herman 1955; Staron and Pette 1986, 1987; Termin et al. 1989; Pette and Vrbová 1992; LaFramboise et al. 2000; Gruhn and Rathmayer 2002; Walklate et al. 2016). In invertebrates, different mATPase isoforms have been described in various crustaceans and insects (Stokes 1987; Peckham et al. 1990; Müller et al. 1992; LaFramboise et al. 2000; Hooper and Thuma 2005). This includes a study that compared mATPase activity in the extensor muscle of the middle leg of the two stick insect species Carausius morosus and Cuniculina impigra, and the locust Locusta migratoria (Bässler et al. 1996). However, neither the studies on the neural control of single legs nor biophysical or biochemical studies usually differentiate between the possible different functions of the appendages (Rathmayer and Maier 1987; Büschges and Wolf 1995; Cooper et al. 1995; Sasaki and Burrows 1998; Gabriel et al. 2003; Mentel et al. 2008; von Uckermann and Büschges 2009). Therefore, our knowledge about functional differences between the walking legs and the resulting implications for their neuro-muscular control is still rather sketchy, despite the fact that morphological and behavioral differences have been reported for different limbs (Mykles 1997; Grabowska et al. 2012; Theunissen et al. 2015; Pilehvarian 2015; Dallmann et al. 2016). For example, although all stick insect legs are similar in appearance, the three thoracic leg pairs differ in function (Dürr and Ebeling 2005; Gruhn et al. 2009b). The front legs have been reported to act as feelers during walking (Cruse 1976; Grabowska et al. 2012; Theunissen et al. 2015). The hind legs have the major role in creating propulsive forces as the center of mass is at or close to the metathoracic coxae (Cruse 1976; Dallmann et al. 2016). In addition, during each step cycle, swing movements are rather fast and stereotypical while stance durations and the forces produced in stance can vary considerably (Hooper et al. 2009; Rosenbaum et al. 2010; Schmitz and Dallmann 2014; Dallmann et al. 2016).

These observations imply either differences in the neural control of the legs, or differences in the muscular properties or both vary between segments. While we know that the neuronal control of different joints of different limbs can vary greatly during different behaviors (e.g. Büschges 2005; Berg et al. 2013; Gruhn et al. 2016), our knowledge on the muscular end is scarce and in invertebrates, the presence of different fiber types has been described but in single limb muscles of different insects and crustaceans (Müller et al. 1992; Günzel et al. 1993; Bässler et al. 1996; Stokes 1987). Complete contraction and force development profiles so far only exist for a single muscle, the extensor muscle of the middle leg femur tibia joint of the stick insect (Bässler et al. 1996; Guschlbauer et al. 2007; Blümel et al. 2012a). Therefore, the question as to whether fiber-type distribution of leg muscles can be generally correlated with specific functional roles of a leg or the segment of a leg in the locomotor system remains unanswered.

In an attempt to better understand the stick insect locomotor system at the muscular level, we therefore investigated as to whether previous findings on a single muscle of a single leg can be generalized for all legs, or if there are leg-specific variations that suggest functional adaptations. For this purpose, we histochemically characterized muscle fibers of the six major stick insect leg muscles extensor and flexor tibiae, levator and depressor trochanteris of all three thoracic segments, and protractor and retractor coxae of the meso- and metathorax, and compared the fiber-type proportions based on their mATPase activity.

Materials and methods

All experiments were carried out on adult female stick insects, Carausius morosus, from a colony maintained at the University of Cologne. The animals were kept under constant conditions at a temperature of 22–24 °C, a humidity of 60%, and a 12 h darkness/light rhythm. The animals were fed blackberry leaves, Rubus fruticosus, ad libitum.

Preparation and sectioning

Animals were pinned down dorsal site up, decapitated, and the intestine was carefully removed after a single cut to open the dorsal side of the animal along the midline. Pro-, meso- and metathorax (with protractor coxae and retractor coxae) in the meso- and metathorax, as well as coxae (with levator trochanteris and depressor trochanteris) and femora (with flexor tibiae and extensor tibiae) of all leg pairs were carefully collected from each animal including the cuticle to which the muscles are attached, taking care not to damage any muscles to be examined. The prothoracic pro- and retractor muscles are oriented dorso-ventrally, and were not further investigated due to the fact that they were not clearly distinguishable from one another and from other coxal muscles. The samples were all pinned out in a similar position on Sylgard (Dow Corning, USA) blocks, at a natural resting angle of all muscle fibers within the samples (Hooper et al. 2009). Then, a tin foil basket around each sample was formed, and the sample directly embedded in Tissue Tek (Sakura, Japan) without usage of saline. After that each sample was carefully frozen in liquid nitrogen (~−80 °C). Frozen samples were kept at −80 °C until used for sectioning.

For cryostat sectioning, frozen specimens were thawed from −80 to −22 °C for 1 h within the cryostat (Leica, CM3050S). 25 µm thin cross-sections of the meso- and metathoracic pro- and retractor coxae were obtained by sectioning transversally along the body axis from posterior to anterior. The coxae and the femora of all segments were cut transversally. Sections were mounted on three alternate gelatinized cover slips for later incubation with different staining protocols. Cryostat sections were dried either overnight at room temperature or for 1 h at room temperature and 1 h at 40 °C. The pinnate structure of the extensor and flexor tibiae muscles with a pinnation angle of 15.6° (Guschlbauer et al. 2007) creates a total increase of effective muscle cross-sectional area and also results in misestimating the total cross-sectional area of the muscles. Values of cross-sectional areas for these muscles must therefore only taken as approximation in a given section of the femur. For the single fiber length estimate caused by the pinnation angle, Guschlbauer et al. (2007) reported that the maximum error is less than 3.8% compared to a fictive arrangement of the fibers in parallel to the long axis of the leg. The equivalent error estimate in cross-sectional area for single fibers can be assumed to be similar and the same for all fiber types, and was thus neglected.

Myofibrillar ATPase staining

Myofibrillar ATPase (mATPase) activity was determined as previously described (Padykula and Herman 1955; Maier et al. 1984; Gruhn and Rathmayer 2002). For the identification of the presence of different isomyosines, alternate sections were pre-incubated at pH 10.1, 8.4, and 4.7 (Padykula and Herman 1955; Guth and Samaha 1969; Günzel et al. 1993). Starting pH values were taken from Bässler et al. 1996, and, incubation times, and the number of washing steps optimized to achieve the best possible staining results. NaN3 (1 M) was added to avoid any staining through mitochondrial ATPase. The staining protocol is shown schematically in Fig. 1a and the exact incubation times for each solution are shown in Fig 2. Stained sections were dehydrated in ascending alcohol concentrations (70/95/100% EtOH), cleared in xylol and mounted on slides in Entellan (Merck, Germany).

Fig. 1
figure 1

mATPase-based fiber typing. a Schematic of the staining protocol for serial cross-sections through the stick insect muscles investigated. The incubation times are given in Fig 2. b Example for the analysis steps following the staining procedure for a distal section of the metathoracic femur: after serial sections treated under alkaline (upper row) and acidic conditions (bottom row), fibers with a gray-scale value of more than 0.75 of the maximum were marked (asterisk). Corresponding fibers of the serial sections were compared and classified into slow contracting (squares, gray-scale value >0.75 under alkaline pre-incubation, <0.25 under acidic pre-incubation), and fast contracting muscle fibers (triangles, gray scale <0.25 under alkaline pre-incubation, >0.75 under acidic pre-incubation). Fibers that had under both pre-incubations a gray-scale value of less than 0.75 were classified as intermediate fibers. In single cases fibers with a gray-scale value of more than 0.75 under both conditions were found that were classified as dark–dark (dd) intermediate fibers (not present in this figure)

Full size image
Fig. 2
figure 2

Staining protocol with incubation times. Boxes show the solution listed (white part), and time of incubation in min (′) or s (″, grey boxes)

Full size image

Staining analysis and muscle fiber typing

Sections that showed freezing artifacts and irregular staining intensity were discarded. Sections of good quality were further analyzed with a microscope (Olympus BX61; imaging system: Olympus cellˆF) under bright field conditions. Pictures of the stained sections were taken at a total magnification of 100. Two alternate sections (after alkaline and acidic pre-incubation) were always compared with respect to staining intensities. Sections from the pro- and retractor coxae muscles of the meso- and metathorax were subdivided into a posterior (sections on cover slips 1–5) and an anterior part (sections on cover slips 6–10). One duplet of sections was chosen for fiber typing from each group and each thoracic segment. Due to their small size, one pair of sections only was taken from the middle of the coxal muscles levator and depressor coxae. The femur with the extensor and flexor tibiae muscles was subdivided into three parts a distal part, containing three cover slips with alternating sections, a medial part with three cover slips of alternate sections, and a proximal part with another four cover slips. One pair of sections from each femur part was compared with for fiber typing.

For muscle fiber typing, a custom written MATLAB script was used (Matlab, MathWorks) (script kindly provided by Dr. Till Bockemühl, University of Cologne). The procedure is visualized in Fig. 1b. The script allowed individual muscle fibers of each muscle in each section to be marked and analyzed. For each section an Excel-sheet (Microsoft Inc.) was created, containing a list of the marked fibers, their cross-sectional area and gray-scale values. With the help of the script, it was possible to follow each muscle fiber through several consecutive sections. Fibers were classified as dark colored when the gray-scale value was above 0.75. This procedure was repeated for corresponding sections stained after alkaline and acidic pre-incubation. Classification of the fibers after both pre-incubations was compared with each other. Fibers that did not stain or only stained lightly after alkaline treatment and darkly after acidic treatment were classified as fast contracting muscle fibers, while the reverse pattern marks slowly contracting muscle fibers (Brooke and Kaiser 1970; Bässler et al. 1996; Gruhn and Rathmayer 2002). Fibers stained faintly or darkly under both conditions were classified as intermediate [light–light (ll) intermediate; and very rarely dark–dark (dd) intermediate]. For figure preparation, digital images of the sections were only adjusted for lighting of the entire picture without changing relative grayscales using Corel Draw (Microsoft, X6) and Photoshop (Adobe, CS5, Vers.12.04). Statistical analysis was done using the paired t test. This is a parametric test that compares two paired groups that are derived from Gaussian populations. All statistical analysis was done with GraphPad software (GraphPad Software Inc, LaJolla, USA). All cross-sectional areas, percentages of fiber types, as well as the statistical values are given in Tables 1, 2, 3 and 4. If not otherwise stated, data for protractor/retractor from n = 6–7 muscles; for the levator/depressor, and for the flexor/extensor (except for the metathoracic proximal flexor, n = 2) from n = 4–6 muscles.

Table 1 Averaged muscle cross-sectional area in mm2
Full size table
Table 2 Averaged percentage of muscle fiber types in the individual muscles after mATPase-classification
Full size table
Table 3 Statistical results for intra-segmental comparison of the different muscle fiber types by paired-samples t test
Full size table
Table 4 Statistical results for comparison of the different muscle fiber types between analyzed muscle regions and between different thoracic segments using the paired-samples t test
Full size table

Results

We investigated the heterogeneity of muscle fibers in six major stick insect leg muscles of all three thoracic segments, based on the histochemical activity of the myofibrillar ATPase. We wanted to gain insight into the diversity and variability of contractile properties that may otherwise not be determined by direct biophysical measurements. First, we compared the fiber typing of the extensor and flexor tibiae muscles which control the femur–tibia joint with one another and between the three thoracic legs. We chose this muscle pair as starting point, because a previous study had presented evidence for a specific fiber-type distribution along the extensor tibiae muscle of orthopteran insects (Bässler et al. 1996).

Extensor and Flexor tibiae muscles

The stick insect femora contain the ventrally located flexor tibiae (short: flexor), and the dorsally located extensor tibiae (short: extensor). The flexor, fills, from distal to proximal, between two-thirds to approximately four fifths of the femur, and is a stance phase muscle that is activated shortly after touchdown and causes the tibia to bend around the femur–tibia joint, while the extensor is mostly involved in the extension of the tibia during swing. The femora differ slightly in their anatomy depending on the leg in question: while the meso- and metathoracic femora are quite similar, the prothoracic femur is longer and its diameter diminishes proximally, close to the coxa. We analyzed both pinnate muscles in the proximal, central and distal third of the femora of all three thoracic segments (Fig. 3). For the proximal prothoracic leg, this means that the analysis started distally of the region of small diameter.

Fig. 3
figure 3

mATPase fiber typing the extensor and flexor tibiae muscles: a Schematic drawing of a stick insect leg with its muscles extensor (gray) and flexor (black), and the location of the three sample areas in the distal (bi), medial (bii) and proximal (biii) femur. bi–biii Example sections of the distal (bi), medial (bii), and proximal (biii) part of the mesothoracic extensor and flexor tibiae muscles after acidic pre-incubation. The arrows mark the border between the two muscles. ci, cii Percentages of the fiber types normalized to the respective maximum cross-sectional area in the pro- (Pro), meso- (Ms), and metathoracic (Mt) flexor (ci) and extensor muscles (cii)

Full size image

Extensor

The cross-sectional area of the extensor always increases from proximal to distal, with the exception of the proximal extensor of the prothoracic femur, which has the largest cross-section among the extensors (Table 1).

Three fiber types are discernible in the extensor: fast fibers which stained darkly after acidic incubation, slow fibers that stained darkly after alkaline incubation, and intermediate fibers which stained faintly after pre-incubation at either pH. The percentage of fast fibers increases slightly, but not significantly from distal to proximal in all of the three legs (Table 2). The only significant change in fast fiber percentage was between the distal 33% and central 59% of the prothoracic extensor. The percentage of intermediate fibers was found to be variable but not significantly different along the extensor muscle of all legs, ranging between 37 and 52% (Table 2). The percentage of intermediate fibers was also not significantly different from that of the fast fibers. Slow fibers always made up less than a third of the total fiber percentage throughout all thoracic segments. Their highest concentration was in the distal femur and varied between 17% in the prothoracic, 23% in the mesothoracic, and 29% in the metathoracic distal extensor (Table 2), all not significantly different from each other. The percentage of slow fibers drops along the length of the muscle in all legs. These differences were significant (p < 0.01, Table 4), except for the difference between proximal and distal metathoracic slow fibers. Between the legs the relative percentages of all fiber types were not significantly different from one another apart from the percentage of slow fibers between the proximal pro- and mesothoracic extensor.

In summary, the three fiber types of the extensor appear to be unevenly distributed within all three thoracic legs: while slow fibers can make up to 30% of the distal extensor their percentage can drop to as little as 1% proximally. Conversely, fast and intermediate fibers make up approximately half of the remaining percentage each and fast fibers show a tendency for a decrease in percentage from the proximal to the distal extensor. The respective fiber-type proportions between the three thoracic legs are very similar.

Flexor

Among the three thoracic segments the prothoracic flexor has the largest cross-sectional area throughout the femur just distal of the narrow proximal segment, while the mesothoracic and metathoracic flexor have their largest cross-sectional areas in the center section (Table 1). Again, three fiber types are clearly distinguishable: fast, intermediate and slow. In only one case did we find a single fiber that stained darkly after acidic and alkaline pre-incubation.

Figure 3bi–biii show example sections from the distal, central, and proximal part of the muscle in the mesothoracic femur. It is clearly visible that the amount of slow fibers decreases from distal to proximal while the opposite is true for the amount of fast fibers (Table 2). For example, the percentage of fast fibers increases in the mesothorax from 11% distally, through 21% in the center of the respective femora up to 33% in the proximal femur. The opposite is true for slow fibers which decreased in the middle leg from 26% distally, through 3%, down to single fibers with a mere 1% of the cross-sectional area in the most proximal part of the femur. Fast and slow fiber percentages were relatively similar in the distal flexor, but become progressively different towards the proximal flexor in all legs. The intermediate fibers always form the largest proportion of fibers throughout the flexor muscles. Their proportion remains relatively constant or increases slightly from distal to proximal (Table 2). For example, mesothoracic intermediate fibers make up 64% of the distal, 76% of the central, and 66% of the proximal flexor. The respective proportions of all fiber types are in no case significantly different between the three legs. The relative proportions of the different fiber types normalized to the total cross-sectional area for the flexors of the pro-, meso- and metathorax are shown in Fig. 3ci, and again all values are given in Tables 1, 2, 3 and 4.

Apart from the observed change in fiber-type percentage along the muscle, we also observed specific cross-sectional patterns of the three fiber types within the flexors muscles of all three legs. In the distal part of the femur, the fast fibers are concentrated in the center of the flexor muscle, and dorsally of the flexor tendon, while the slow fibers are located towards the cuticle of the femur (Fig. 3bi). With decreasing distance to the coxa, fast fibers are also found ventrally of the flexor tendon, and the band in which they are located widens. At the same time, slow fibers become concentrated in a stretch of fibers that spans from the tendon towards the anterior and posterior cuticle (Fig. 3bii). Finally, in the proximal part of the femur, fast fibers form the majority of all fibers and slow fibers are only interspersed along the midline of the flexor at the level of the tendon (Fig. 3biii). Throughout the muscle, the intermediate fibers are usually found at the boundary between the fast and slow fibers without forming distinct clusters. The above described changes in the fiber-type location are also recognizable in the respective sections of the femur in all other thoracic segments and the relative proportions are not statistically different between the corresponding sections of the femora of the three body segments.

In summary, in analogy to our findings in the extensor muscle, the muscle fiber-type distribution in the femoral flexor is characterized by the gradual decrease of fast fibers from proximal to distal with concomitant increase in the percentage of slow fibers and a relatively constant high percentage of approximately two-thirds of intermediate fibers independent of thoracic leg.

Other thoracic leg muscles

We then investigated as to whether the inter-segmental similarity in fiber-type distribution and the similarity of fiber-type distribution between antagonists found for tibial muscles are general characteristics in the stick insect leg muscle control system. For this purpose we analyzed the antagonistic muscles of the next two proximal leg segments, the levator and depressor of the trochanter, and the protractor and retractor of the coxa.

Levator and depressor trochanteris muscles

First, we investigated the muscles in the coxa, which move the next proximal segment, i.e. the trochanter. The coxae of the different segments all have roughly the same size and structure. Within the coxa, the levator trochanteris (short: levator) is located dorsally, arises in the coxa and attaches to the dorsal edge of the trochanter. It is a swing phase muscle, responsible for lifting the leg. Its antagonist, the depressor trochanteris (short: depressor) is located ventrally, but arises partly in the coxa and partly in the thorax (Marquardt 1940; Schmitz 1986). As the depressor has its largest muscle division in the coxa (Schmitz 1986), and only two smaller divisions located inside the body cavity whose functions are still not clarified (Marquardt 1940; Schmitz 1986; Cruse et al. 1993; Brunn 1998; Brunn and Heuer 1998), we focused in our study on the coxal part of the depressor muscle. Functionally, it is a stance phase muscle, responsible for lowering the leg and providing lift to the body. Both muscles are easily distinguishable based on their muscle fiber arrangement. Since both muscles are short, they were not subdivided into different regions of interest in our analysis. Figure 4a shows a schematic with the location of the two muscles. A cross-section of the coxa after histochemical staining following acidic pre-incubation is shown in Fig. 4b. A white line marks the boundary between the two muscles, and all values are given in Tables 1, 2, 3 and 4.

Fig. 4
figure 4

mATPase fiber typing the levator and depressor trochanteris muscles: a schematic drawing of a stick insect leg with its muscles levator (gray) and depressor (black). b Example section of the mesothoracic levator and depressor trochanteris muscles after acidic pre-incubation. The arrows mark the border between the two muscles. ci, cii Percentages of the fiber types normalized to the respective maximum cross-sectional area in the pro- (Pro), meso- (Ms), and metathoracic (Mt) depressor (ci) and levator muscles (cii)

Full size image

Levator

The cross-sectional area of the metathoracic levator is the largest of the three (0.36 mm2) and decreases to meso- and prothorax which have similar cross-sectional areas with 0.26 mm2 in the meso- and 0.22 mm2 in the prothorax, respectively (Table 1). Within the levator, fast fibers are mostly located in the ventral part of the muscle, towards the center of the coxa, while intermediate and slow fibers are found interspersed at the dorsal side of the muscle, towards the cuticle. In two levator muscles, a single fiber each was found that stained darkly after acidic and after alkaline pre-incubation. The distribution of the different fiber types in the levator is shown in Fig. 4cii.

In all segments, fast muscle fibers make up the largest part of the total cross-sectional area (between 68% in the pro- and 73% in the metathorax) (Table 2). Their relative proportion is not significantly different between segments. Intermediate and slow fibers are also present in similar quantities between segments (IF between 13 and 27%, SF between 6 and 13%) (Table 2). The percentage of slow or intermediate fibers is significantly different from that of the fast fibers, with the exception of the difference between intermediate and fast fibers in the prothorax (p = 0.051).

In summary, the levators of the pro- and mesothorax are slightly smaller in cross-sectional area than that of the metathorax. The majority of fibers in all three segments is made up of the fast type.

Depressor

As among the depressor muscles, the metathoracic depressor has the largest cross-sectional area with 0.48 mm2, compared to 0.41 and 0.38 mm2, respectively, in the pro- and mesothorax (Table 1). Figure 4ci shows the normalized distribution of fiber types in the depressor muscles of the three thoracic segments. Again, the three fiber types can be distinguished: fast fibers are mostly located in the dorsal part of the muscle, and in a small medial band that cuts through the muscle to the ventral side of the coxa. In addition, a single layer of fast fibers surrounds two distinct groups of slow fibers located anteriorly and posteriorly in the ventral coxa. Located between the slow and fast fibers is an additional layer of intermediate fibers.

The relative proportions of fast, slow and intermediate fibers are similar between the three thoracic segments. The majority of the muscle is made up by similar percentages of fast fibers covering 48, 56 and 49% of the depressor cross-sectional area in the pro-, meso- and metathorax, respectively (Table 2). The second largest area is covered by intermediate fibers which cover between 28 and 1% in all segments. More interestingly, slow fibers, which make up the smallest percentage of the cross-sectional area, show a significant increase in percentage from 11% in the prothorax to 18% in the metathorax (Table 2).

In summary, fast fibers make up the largest proportion of the depressor muscles, followed by the intermediate fiber type and the slow fibers, which show the smallest percentage. The proportion of fast fibers in the depressor is smaller than in the levator, conversely that of slow fibers larger. The percentage of the latter increases significantly towards the metathorax.

Protractor and retractor coxae muscles

The protractor and retractor coxae muscles are major leg muscles located in the cavity of the thorax and not within each leg (see schematic in Fig. 5a). By moving the coxa, they are responsible for the forward and backward movement of the leg along the long axis of the body. During forward walking, the protractor acts as swing, while the retractor acts as a stance phase muscle. Size and location of both muscles differ slightly between the three thoracic segments (Marquardt 1940). The mesothorax is the largest segment compared to the somewhat smaller meta- and very short prothorax. In the meso- and metathorax the protractor coxae (hence protractor) is positioned ventro-laterally below the dorso-laterally located retractor coxae (hence retractor). In both segments, both muscles attach at an apodeme anterior to the coxa (Igelmund 1980) and continue anteriorly to attach to the side of the thorax at approximately 200–700 µm anterior of the coxa. In the very short prothorax both muscles run directly dorsally from their internal attachment site in the ventral part of the thorax and in very close proximity to one another. Since it was not possible to distinguish between protractor and retractor in the prothorax, no quantitative analysis was done for their muscle fibers in this segment. Figure 5a shows a schematic with the location of the two muscles in the mesothorax. Figure 5bi shows a cross-section of the anterior and Fig. 5bii of the posterior part of the dorsally located retractor and the ventrally located protractor after histochemical staining following acidic pre-incubation. A white line marks the border between the two muscles, and all values are given in Tables 1, 2, 3, and 4.

Fig. 5
figure 5

mATPase fiber typing the protractor and retractor coxae muscles: a schematic drawing of a stick insect leg with its muscles protractor (black) and retractor (gray), and the location of the two sample areas in the anterior (bi) and posterior (bii) part of the muscles. bi, bii Example sections of the anterior (bi) and posterior (bii) part of the mesothoracic protractor and retractor coxae muscles after acidic pre-incubation. The arrows mark the border between the two muscles. ci, cii Percentages of the fiber types normalized to the respective maximum cross-sectional area in the meso- (Ms) and metathoracic (Mt) protractor (ci) and retractor muscles (cii)

Full size image

Protractor

We analyzed muscle fiber types and distribution of the meso- and the metathorax in the anterior and posterior parts of the muscle. The posterior parts of the protractor have similar total cross-sectional areas with means between 0.36 and of 0.4 mm2 (Table 1). They taper down to between 0.19 and 0.23 mm2 in their anterior section. Neither posterior nor anterior cross-sectional areas were significantly different between meso- and metathorax. All three types of fibers are present in the protractor: fast fibers make up most of the part of the muscle towards the inside of the thoracic cavity and a single layer towards the cuticle; slow fibers are located exclusively in a ribbon-like arrangement centrally from this single layer of fast fibers; and intermediate fibers, located mostly next to and around the slow fibers.

In the mesothoracic protractor, fast fibers were most abundant, covering up to 78% of the muscle fibers. In contrast, in the metathorax, intermediate and fast fibers were similarly abundant (Table 2). Slow fibers always make up the smallest portion of the muscle, typically with 10% or fewer of the fibers. The relative proportions of the different fiber types normalized to the total cross-sectional area for the protractors of the meta- and mesothorax are shown in Fig. 5ci.

In summary, the majority of the protractor is made up of fast and intermediate fibers, while slow fibers make up only a small fraction of the muscle, and only in its larger posterior part.

Retractor

The retractor muscles of both, the meso- and metathorax, the antagonist of the protractor muscle, have a larger cross-sectional area in the posterior than the anterior parts. In contrast to the protractors, the posterior mesothoracic retractor has a significantly larger cross-sectional area than its metathoracic analog. However, while the mesothoracic retractor tapers down to approximately two-thirds of the posterior cross-sectional area, the metathoracic cross-sectional at its anterior attachment site is just slightly smaller than the posterior one (Table 1).

The histochemical staining of the retractor muscle is shown in the exemplary Fig. 5bi, bii. Slow fibers are exclusively found in a bundle towards the lateral and dorsal side of the muscle. They are surrounded by intermediate fibers, while fast fibers make up most of the central part of the muscle, and form an additional layer around the slow fibers towards the cuticle.

In contrast to the protractor, retractor-fiber-type proportions are more constant throughout the muscle and vary less between anterior and posterior end. Fast fibers occur in much smaller quantities in the retractor compared to the protractor. Their percentage only amounts to up to 43% in the meso-, and between 32% in the metathorax, respectively (Table 2). Slow fibers occur in much larger percentages in the retractor than the protractor and cover up to 17% of the meso-, and up to 33% of the metathoracic retractor (Table 2). Only in the mesothorax is the slow fiber percentage smaller than that of fast fibers (p < 0.05 and p > 0.01). For both, the anterior and the posterior parts of the retractor, the increase in the proportion of slow fibers from the meso- to the metathorax was significant (p < 0.01). Intermediate fiber cross-sectional area is around 40% along the muscle in both segments (Table 2). The relative proportions of the different fiber types normalized to the total cross-sectional area for the retractors of the meta- and mesothorax are shown in Fig. 5cii.

In summary, the retractors of the meso- and the metathorax contain fewer fast, and more slow fibers than the protractors while the amount of intermediate fibers is similar. There is a significant increase in slow fiber cross-sectional area towards the metathoracic retractor where slow fibers amount to one-third of the muscle.

Discussion

Little data currently exist that demonstrate that muscular properties and specific behavioral limb functions correlate. Some examples demonstrate that muscle fiber-type differences are connected to morphological specializations of appendages such as the crusher and pincer claws of lobsters (Ogonowski et al. 1980; Govind and Kent 1982; Mykles 1997). However, it is unclear, whether morphologically similarly structured limbs can express muscle fiber diversity related to specific functional roles of a limb.

One effective method to investigate muscle properties is to determine the activity of myofibrillar ATPase (mATPase) in cross-sections of muscle fibers. Conversion of ATP to ADP and phosphate by mATPase is a crucial step in the muscular cross-bridge cycle determining contraction speed during muscle contractions throughout the animal kingdom (for a review see Walklate et al. 2016). The enzyme activity for this process resides in myosin heavy chain (Barany 1967; Hooper and Thuma 2005), and is dependent on the pH at which cross-sections of the muscle tissue are pre-incubated (Guth and Samaha 1969, 1970; Brooke and Kaiser 1970; Hämalainen and Pette 1993; Bässler et al. 1996; Gruhn and Rathmayer 2002). Studies comparing single identified locust and crustacean muscle fiber properties have shown that fibers with slow contraction velocities express acid labile/alkaline stable mATPase isoforms. A high contraction velocity, on the other hand, correlates with high mATPase activity after acidic pre-incubation (Brooke and Kaiser 1970; Guth and Samaha 1970; Maier et al. 1984; Rathmayer and Maier 1987; Müller et al. 1992).

In this study we asked as to whether the fiber-type composition of the main pairs of leg muscles in the three leg pairs of the stick insect locomotor system is characterized by a common fiber-type heterogeneity, or whether the distribution of fiber types and their ratios relate to the functional role of a muscle in a leg segment and specific thoracic leg in a walking insect.

Differences in muscle fiber-type ratios within and between flexor and extensor tibiae muscles

Based on mATPase histochemistry, we found that the six major leg muscles of all thoracic segments are heterogeneously composed of at least three different muscle fiber types. This is not an unexpected finding, as heterogeneous mATPase distributions within muscles are regularly found in both vertebrates and invertebrates (Brooke and Kaiser 1970; Stokes 1987; Fowler and Neil 1992; Rivero et al. 1996; Bässler et al. 1996; Gruhn and Rathmayer 2002). A heterogeneous fiber-type distribution was also previously reported for the extensor tibiae of the two stick insect species Carausius morosus and Cuniculina impigra, and the locust Locusta migratoria (Bässler et al. 1996). We confirmed their findings on a specific distribution of fast and slow fibers along the extensor muscle with a larger relative percentage of fast fibers in the proximal compared to the distal portion, and the opposite for slow fibers. In their study the maximal percentage of fast fibers in the proximal part increased to almost 100%, compared to 60% in our study. This difference is most likely due to the fact that Bässler et al. used only the staining intensity after alkaline incubation for their classification, while we compared alkaline pre-incubated and acidic pre-incubated sections for further fiber-type differentiation. The additional pre-incubation pH increased our number of intermediate fibers, as we classified our sections automatically by the grayscale values measured (see “Materials and methods”). Therefore, our data are in good accordance with the fiber-type distribution shown by Bässler et al. (1996).

The fiber-type distribution in the extensor is mirror imaged in the much larger antagonist, the flexor tibiae muscle. The percentage of slow fibers increased from proximal to distal with a concomitant decrease of fast fibers, and a relatively constant percentage of approximately two-thirds of intermediate fibers. For the flexor muscle of the stick insect Eurycantha calcarata L., Pilehvarian recently described the mATPase histochemistry as a heterogeneous distribution of predominantly fast fibers (Pilehvarian 2015). The difference may be the result of our use of pre-incubation at different acidic and alkaline pH, which allows for a more differentiated picture, but could also be the result of species specific differences such as shown for the extensor muscles of the two stick insect species C. morosus and C. impigra by Bässler et al. (1996).

The fiber-type distributions in extensor and flexor tibiae appear to support the role of the FTi-joint in locomotion. Both muscles serve posture control while at rest in the standing animal (Bässler 1967, 1974, 1993), as well as movement of the tibia during leg stance in the step cycle (Graham 1985; Büschges et al. 1994). Previous studies have shown that both muscles are co-activated in these cases (Graham 1985; Büschges et al. 1994). A large number of fast fibers in the proximal portion of the extensor is needed for generating fast tibia extension movements during swing in the step cycle (Bässler et al. 1996; Bässler and Stein 1996). Activation of extensor muscle fibers arises from innervation by two excitatory motoneurons only, the SETi and the FETi, as well as a single inhibitory motoneuron, the common inhibitor 1 (CI1). The innervation pattern of the extensor tibiae muscle largely resembles the distribution of slow, intermediate and fast muscle fibers, suggesting a correlation between the motoneuronal innervation and the muscle fiber type, at least for this muscle (Bässler et al. 1996). The flexor is innervated by as many as 25 excitatory motoneurons and 2 inhibitory motoneurons in a not yet fully characterized pattern (Goldammer et al. 2012). The similar distribution of fiber types in both muscles suggests a similar functional division for both muscles in postural duties involving co-contraction, and propulsive duties. This, however, remains to be tested. It is interesting to note that we did not find any statistical difference between the fiber-type distributions of the three thoracic leg pairs, suggesting that the role of the two muscles moving the tibia might be similar between legs. This is an unexpected finding as Dallmann et al. (2016) found differences in the torques of femora of different thoracic segments.

Differences between swing and stance muscles, and between thoracic segments

The lack of differences between the extensor tibiae and the flexor tibiae could suggest that there exists a general similarity in fiber-type distribution in leg muscles across leg segments and between legs. However, contrary to this assumption, we found that other muscles serving the generation of leg swing, i.e. levator trochanteris and protractor coxae both contain significantly fewer slow fibers than their respective antagonists, i.e. the depressor trochanteris and retractor coxae, which generate force, when the legs are on the ground. While the percentage of fast fibers in levator and protractor amounted to between 68 and 78%, the number of slow fibers was found to be about 10% in both muscles. In contrast, the number of fast fibers in depressor and retractor muscles ranged between 26 and 56%, while the percentage of slow fibers in these muscles ranged from 11 to 33%.

In addition, we also detected a difference between the proportions of slow fibers between the stance muscles of the three thoracic segments. Their percentage increases significantly between the pro- and metathoracic depressor, and also significantly between the meso- and metathoracic retractor.

What can be the functional significance of such specificities in fiber-type distribution, namely the relatively high quantities of fast fibers in the protractor and levator on the one hand, and the comparatively large percentage of slow fibers in the retractor and depressor on the other? Slow fibers are optimized for generating tonic force, however, with a reduced maximum power output, while fast fibers are generally adapted to producing fast, phasic movements with large power, under high ATP expenditure (Jahromi and Atwood 1969; Costello and Govind 1984; Galler et al. 1997; Hooper et al. 2007b; Walklate et al. 2016). For vertebrates, it has been shown that isometric tension does not change much between fibers that express fast and slow MHC isoforms (Bottinelli and Reggiani 2000), but during isometric contractions, mATPase activity differs between isoforms up to threefold, resulting in a threefold higher ATP consumption in fast type muscle fibers (Walklate et al. 2016). If this were true for invertebrates as well, this could explain the quantities of fast fibers in the protractor and levator muscles. In the case of the levator this muscle is exclusively, and in the case of the protractor the muscle is predominantly responsible for the relatively fast swing movement during the stepping cycle (Gruhn et al. 2009a; Rosenbaum et al. 2010). At the same time, a higher percentage of slow fibers within muscles such as the depressor, the metathoracic retractor or the distal flexors and extensors, would increase the suitability of these muscles or a section thereof for isometric contractions and hence a role in stiffening of a joint. This may also be functionally related to the role of the depressor as the muscle that is not only responsible for lowering the leg but also for supporting the body weight at rest during posture control. One might argue that stick insects generally display slower movements than other terrestrial insects, reducing the comparability of the findings with other invertebrates. However, the time course of single twitch contractions of the stick insect leg muscles is similar to those of faster moving locusts and cockroaches (Guschlbauer et al. 2007).

The increase in slow fiber proportion towards the metathoracic legs is interesting in light of the fact that the center of mass of the stick insect is at or slightly posterior to the coxae of the hind legs (Cruse 1976). This also matches the role of the metathoracic legs as the ones that appear to be more involved in producing propulsive force than the mesothoracic or even prothoracic legs (Cruse 1976; Dallmann et al. 2016). Dallmann et al. (2016) recently presented force measurement data from stick insects that showed that the depressor is largely responsible for producing propulsive force during level walking, supporting our interpretation of the muscle fiber-type distribution. Interestingly, however, in this study the retractor was found to produce far less torque than previously assumed based on kinematics. It could be that changes in the behavioral paradigm from level walking to the frequently occurring climbing may reveal such a stronger role for the retractor as well.

One has to note, that we only analyzed the coxal parts of the depressor trochanteris, and not its thoracic muscle units. While the function of the coxal part of the depressor has been intensively studied (Pearson and Iles 1970; Iles and Pearson 1971; Graham and Bässler 1981; Epstein and Graham 1983; Stokes 1987; Full and Ahn 1995; Rosenbaum et al. 2010), the function of its thoracic parts is still largely uncertain. A functional comparison between insect species must be undertaken with great care as the homolog muscles may have different roles depending on leg geometry. The few stick insects studies on these depressor parts showed that they are often activated during the leg swing phase, but can also be active during stance, similar to the coxal depressor, and could thereby be responsible for shaping the leg’s swing phase (Cruse et al. 1993; Brunn 1998; Brunn and Heuer 1998). Further studies, including electrophysiological, histological and microCT approaches will reveal more insights into this topic. Interestingly, the results by Dallmann et al. (2016) also point towards functional differences in the use of the flexor/extensor tibiae muscles in the different legs, which do not seem to exhibit fiber-type differences. This emphasizes the fact that muscle activity is not only dependent on fiber types but also innervation and neuronal control, which may vary greatly between the legs.

Given the fact that the acquisition of data on the biomechanical properties of stick insect leg muscles is challenging (Guschlbauer et al. 2007; Blümel et al. 2012a), our study helps to compensate for this problem and draw some conclusions about parameters such as force–velocity relationships. This can now be attempted, because experiments on single skinned muscle fibers in vertebrates and invertebrates alike have shown that both force–velocity relationship and maximum shortening velocity, are positively correlated with the mATPase activity and corresponding MyHC isoforms (Bottinelli et al. 1991, 1994; Galler and Rathmayer 1992). Furthermore, there is a positive correlation between fiber type and the dynamics of stretch-activation, where fast fiber types also show faster force transients (Galler et al. 1994). For the extensor tibiae muscle there now exists an extensive data set on both biomechanical properties and histochemical profile, including this study (Bässler et al. 1996; Bässler and Stein 1996; Guschlbauer et al. 2007; Hooper et al. 2007a, b; Blümel et al. 2012a, b, c). Theoretical studies implementing fast and slow fiber characteristics into a neuromuscular model for the pro- and retractor muscles of the walking stick insect created much more realistic outputs (Toth et al. 2013a, b).

We now have evidence for a specific heterogeneous fiber-type distribution between six major leg muscles of the three pairs of legs in a wingless orthopteran insect. These fiber-type distributions closely match the functions of the respective muscles: the distribution within flexor and extensor muscles in the femur support their dual role for generating fast tibial movements during leg swing, and co-contractions in the generation of leg stance and posture control (Bässler et al. 1996; Bässler and Stein 1996; Gabriel and Büschges 2007). The prominent presence of fast fibers in protractor and levator muscles parallels their function as swing muscles in most behavioral settings (Hooper et al. 2009; Rosenbaum et al. 2010). Finally, the relatively higher proportion of slow fibers in the retractor and depressor muscles underscores their role in stance phase and posture control (Cruse 1976; Dallmann et al. 2016).