Functional correlations of axial muscle fiber type proportions in the waterfall‐climbing Hawaiian stream fish Sicyopterus stimpsoni

Abstract

Assessing the factors that contribute to successful locomotor performance can provide critical insight into how animals survive in challenging habitats. Locomotion is powered by muscles, so that differences in the relative proportions of red (slow‐oxidative) vs. white (fast‐glycolytic) fibers can have significant implications for locomotor performance. We compared the relative proportions of axial red muscle fibers between groups of juveniles of the amphidromous gobiid fish, Sicyopterus stimpsoni, from the Hawaiian Islands. Juveniles of this species migrate from the ocean into freshwater streams, navigating through a gauntlet of predators that require rapid escape responses, before reaching waterfalls which must be climbed (using a slow, inching behavior) to reach adult breeding habitats. We found that fish from Kaua'i have a smaller proportion of red fibers in their tail muscles than fish from Hawai'i, matching expectations based on the longer pre‐waterfall stream reaches of Kaua'i that could increase exposure to predators, making reduction of red muscle and increases in white muscle advantageous. However, no difference in red muscle proportions was identified between fish that were either successful or unsuccessful in scaling model waterfalls during laboratory climbing trials, suggesting that proportions of red muscle are near a localized fitness peak among Hawaiian individuals.

Micrograph of cross‐section from posterior tail segment of a juvenile from the Hawaiian waterfall‐climbing gobiid fish Sicyopterus stimpsoni, stained for slow oxidative (red) muscle fibers. Comparisons showed no difference in axial red fiber proportions between successful and unsuccessful climbers, but did identify differences between fish from different islands that face different primary functional demands (predator escape on Kaua'i vs. sustained climbing on Hawai'i).

1. INTRODUCTION

Locomotor performance is critical to the survival of many animals (Blob and Higham, 2014; Biewener and Patek, 2018), determining the outcomes of tasks ranging from escaping predators and capturing prey, to finding mates and locating resources. In this context, assessing the factors that impact locomotor performance is crucial to understanding which individuals are most likely to survive and contribute offspring to the next generation, and which species are most likely to persist in changing and challenging habitats through time (Arnold, 1983; Biewener, 2002).

In vertebrates, muscles actuate movements of the body during locomotion, making muscle performance a key determinant of locomotor success (Gillis and Blob, 2001). One critical factor affecting the performance of anatomical muscle units is their proportions of different muscle fiber types. Different fiber types exhibit a gradient of contractile properties. As a result, muscles with different proportions of fibers perform with a range of capacities that can be suited for particular functional roles (Rome et al. 1988). These distinctions are particularly prominent in the axial muscles of fishes, which are composed predominantly of two primary types of fibers (Greer‐Walker and Pull, 1975; Zhang et al. 1996; Altringham and Ellerby, 1999). ‘White muscle’ comprises fast‐twitch, glycolytic fibers, suited for rapid contraction, whereas ‘red muscle’ comprises slow‐twitch, oxidative fibers suited to powering sustained activities (Bone, 1966; Jayne and Lauder, 1993, 1994). Fish that habitually use fast vs. slow locomotor behaviors often show differences in the relative proportions of fiber types that reflect these patterns. White, fast, glycolytic fibers typically constitute the majority (70–100%) of fish axial muscle (Greer‐Walker and Pull, 1975; Zhang et al. 1996; Altringham and Ellerby, 1999). However, species that engage in sustained cruising or extensive continuous locomotor activity will often show a greater proportion of red fibers than fish that are typically burst swimmers (Bone, 1966; Johnston and Moon, 1981; Jayne and Lauder, 1993, 1994; Gillis, 1998; Cediel et al. 2008). Differences in the proportions of red and white fibers can be readily quantified in fishes, because their axial red fibers are generally sequestered as bundles adjacent to the horizontal septum, and as a thin band along the periphery of the section (Cediel et al. 2008). This characteristic facilitates comparisons across groups of fishes that can test potential contributions of differences in axial muscle fiber composition to differences in locomotor behaviors or levels of performance.

One fish system in which differences in axial fiber type proportions have been implicated as contributors to differences in locomotor performance are the waterfall‐climbing stream gobiids of Hawai'i (Cediel et al. 2008; Schoenfuss et al. 2013). The goby species of this assemblage are amphidromous, in which fry that hatch in streams are swept to the ocean, where they develop for several months before commencing a migration to adult habitats in freshwater streams (Smith and Smith, 1998; Schoenfuss and Blob, 2007). For three species, these adult habitats are located above substantial waterfalls that migrating juveniles must climb to establish breeding territories. These species exhibit two distinct styles of climbing (Schoenfuss et al. 1997; Nishimoto and Fitzsimons, 1999; Schoenfuss and Blob, 2003). Awaous stamineus and Lentipes concolor engage in powerburst climbing, in which a rapid adduction of the pectoral fins initiates a short burst of body axis undulation that quickly powers the fish up the waterfall, until it reattaches to the substrate using a ventral sucker formed from fused pelvic fins. In contrast, juvenile Sicyopterus stimpsoni use very little axial undulation and, instead, inch up surfaces by alternating the attachment of the pelvic sucker with a second, oral sucker, and advancing either the head or the body upward while the other segment adheres to the substrate. Inching movements are much slower than those of powerburst climbing, but are sustained for significantly longer periods of time (Schoenfuss and Blob, 2003; Blob et al. 2006). Correlating with these behavioral differences, S. stimpsoni show significantly greater proportions of red fibers in their body axis muscles than either species of Hawaiian powerburst climber (Cediel et al. 2008). The proportion of red fibers is also greater in posterior sections compared to anterior sections, to the point that S. stimpsoni have twice the fraction of red fibers in mid‐caudal sections as powerburst species (Cediel et al. 2008).

The correlations between locomotor performance and axial muscle fiber proportions that emerge from comparisons across goby species suggest that muscle fiber type proportions could also help to explain intra‐specific variation in locomotor performance. For the inching species S. stimpsoni in particular, climbing performance appears to vary across subpopulations from different islands across the Hawaiian archipelago. Selection experiments that challenged juvenile S. stimpsoni with a 2‐m climbing ramp showed a 68% success rate for fish from the Big Island of Hawai'i, but only a 49% success rate for fish from Kaua'i (Moody et al. 2017). These differences in climbing success may be correlated with a tradeoff between the predominant selection pressures to which juvenile S. stimpsoni are exposed on each island. Whereas streams on the younger island of Hawai'i typically contain near‐shore waterfalls that place a premium on climbing performance, streams on older Kaua'i have eroded further inland, forcing migrating gobies to endure longer periods of exposure to non‐climbing predators that could select for burst escape performance, at the expense of sustained climbing ability (Blob et al. 2010; Moody et al. 2015, 2017; Diamond et al. 2019). In this context, juvenile S. stimpsoni from Hawai'i might be expected to possess a greater proportion of axial red muscle fibers than juvenile S. stimpsoni from Kaua'i, enabling accentuated sustained climbing performance rather than accentuated escape performance. Beyond comparisons across subpopulations, individual fish within each climbing species also show diverse performance abilities, ranging from fish that fail to scale even modest barriers, to ‘super‐climbers’ that climb many times faster and further than other individuals (Blob et al. 2006, 2008, 2010; Kawano et al. 2013; Schoenfuss et al. 2013). In this context, for the inching species S. stimpsoni, individual fish with superior climbing performance might be predicted to have a greater proportion of axial red muscle fibers than unsuccessful individuals.

To test for potential contributions of axial muscle fiber type proportions to differences in goby locomotor performance, we collected migrating juvenile S. stimpsoni in stream estuaries to perform comparisons of muscle fiber type proportions in two contexts: between fish from the islands of Kaua'i and Hawai'i, and between successful and unsuccessful climbers. Through these comparisons, we sought to refine understanding of the factors contributing to locomotor performance in a system that places strong selective pressure on the successful execution of a demanding behavior (Blob et al. 2008, 2010; Kawano et al. 2013; Moody et al. 2017).

2. MATERIALS AND METHODS

2.1. Animal collection, climbing trials, and preservation

Juvenile S. stimpsoni were collected during the peak migratory season (March 2013) from Wailua stream (Island of Kaua'i) and Hakalau stream (Island of Hawai'i) using dipnets. Following protocols established in previous studies (Blob et al. 2008, 2010), fish were immediately placed in aerated buckets and transported (Kaua'i: expedited air freight; Hawai'i: car) to the Division of Aquatic Resources Fisheries Research Station in Hilo (Island of Hawai'i).

Fish from Hawai'i were subjected to climbing trials within 24 h of collection, according to established experimental protocols (Blob et al. 2008, 2010; Kawano et al. 2013). Briefly, groups of three to four fish were placed in a 40‐L basin below an artificial waterfall 2 m in height (~100 body lengths), constructed from a rain gutter chute coated with sand and angled at ~ 70° from horizontal. Stream water was gravity fed from a basin above the waterfall at a rate of ~200 mL min^‐1. Fish were allowed to climb for up to 1 h, and successful climbers were recaptured from the basin above the waterfall. Climbers and non‐climbers were placed (in separate groups) back in the climbing setup, and given an opportunity to climb a second time. Only fish that completed their climb in both attempts (‘winner’), or that attempted to climb but failed to scale the chute in both trials (‘loser’), were utilized for the collection of fiber type data.

Following arrival in Hilo from Kaua'i, or completion of climbing trials (in all instances within 48 h of capture), fish were killed using neutrally buffered MS‐222 and measured for total length. A total of 18 fish from Kaua'i (mean ± SD 28 ± 0.6 mm), 22 winners from Hawai'i (28 ± 0.7 mm) and 19 losers from Hawai'i (28 ± 0.9 mm) were processed. There was no significant difference in fish length; however, Kaua'i fish were significantly heavier than Hawai'i fish (anova, P < 0.05). Having been killed, fish were placed in cooled isopentane (2‐methylbutane) for 1 min, removed, and dissected to extract two specific transverse segments of the tail. Relative to the length of the tail (anal pore to caudal peduncle), an anterior segment was extracted to generate histological sections at a location one‐third of the distance posterior to the anal pore, and a second segment was extracted to generate histological sections at a location two‐thirds of the distance posterior to the anal pore. These locations differ from those of our previous study of muscle fiber types (Cediel et al. 2008), bracketing the most posterior section of that previous study. Each segment was ~5 mm in thickness, was immersed in optimum cutting temperature medium (OCT; Sakura Finetek) inside a tissue mold, and flash frozen using a suspension of dry ice in ethanol. Once frozen, tissue samples were tightly wrapped in aluminum foil and stored between two sheets of dry ice in a −20°C freezer until shipment on dry ice (within 1 week) to the St Cloud State University Aquatic Toxicology Laboratory. Samples were stored at −80°C upon arrival.

Permission for access to field sites and specimens was provided by the Division of Aquatic Resources, State of Hawai'i, and coordinated by Director Dr Robert Nishimoto. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (Institutional Animal Care and Use Committees of permits 1,207 and 0,113 (St Cloud State University) and 2011‐057 (Clemson University).

2.2. Histological processing

At St Cloud State University, tissue blocks from the extracted anatomical segments were shaved to expose the transverse sectional surface and sectioned at 7‐µm thickness using a Shandon cryostat (Thermo Fisher Scientific). Up to 12 sections from each tissue block were mounted on electrostatically charged glass slides (Newcomer Supplies, Middleton, WI, USA). To differentiate fast‐ and slow‐twitch muscle fibers, sections underwent a myofibrillar ATPase staining procedure to produce a brown precipitate in slow‐twitch red muscle, according to published protocols (Brooke and Kaiser, 1969; Cediel et al. 2008; Maie et al. 2011; Schoenfuss et al. 2013). Briefly, slides were preincubated in a Coplin jar for 2 min at 25°C in 50 mL 0.1 M buffer (pH = 4.3), 250 mg KCl, and final pH was adjusted to 4.6. Slides were then rinsed in deionized water and the ATPase reaction was initiated in an incubating solution (50 mL 0.1 M AMP buffer (pH = 9.3), 100 mg anhydrous CaCl₂, 60 mg ATP) for 30 min at 25°C, with final pH adjusted to 9.3. Slides were then transferred to a 2% cobalt chloride solution for 10 min before the reaction was stopped in deionized water. Slides were then placed in a 25% solution of ammonium sulfide [250 µL (NH₄)₂S in 50 mL deionized water] for 20 s. This final reaction was stopped in 50 mL deionized water. Slides were air‐dried and cover‐slipped using aqua mount solution. To avoid staining biases, we stained mixed batches of samples that contained an assortment of tissues from several specimens and both tail segments.

2.3. Image analysis

The sections obtained from each tissue segment were assessed for section quality and completeness. Three satisfactory sections from each tissue block were designated and photographed using a ProgRes camera (model C12 Plus; Jenoptik) mounted on an Olympus SZH‐ILLD stereozoom microscope. Given the assumed bilateral symmetry of the tail musculature, one half of each section (excluding the vertebral column) was digitally traced on the photograph using Image J software (https://imagej.nih.gov/iij) to determine the total cross‐sectional area. The area composed of red muscle fibers was grouped with the small number of intermediate fibers (Cediel et al. 2008) and traced using the same procedure, allowing calculation of the proportion of red muscle fibers out of each total histological section. Visual bias was avoided in this procedure by labeling photographs using a simple numbering format that did not include the name of the treatment associated with each section. The mean of the three evaluations of the percentage of red fibers for each sectioning location was used for statistical analysis. A total of 369 tissue sections from 59 fish (18 Kaua'i, 22 Hawai'i winner, 19 Hawai'i loser) were evaluated (Figure 1).

Figure 1.

Representative histological images for muscle fiber distribution in axial musculature of Sicyopterus stimpsoni from Kaua'i and Hawai'i. (a) Juvenile S. stimpsoni with anal pore and caudal peduncle indicated by arrows. Location of section recovery from anterior and posterior tail segments indicated by dashed line. (b, c) anterior and posterior tail segment micrographs representative for juvenile gobies from Kaua'i. (d–g) Anterior and posterior tail segment of juvenile S. stimpsoni from Hawai'i, collected from ‘winners’ (d, e) and ‘losers’ (f, g) of climbing trials. Scale bars in (b) and (c) similar for all anterior or posterior segments, respectively. All images were processed for image clarity using the same functions in Adobe Photoshop

2.4. Statistical analyses

All values of red fiber percentage were arcsine transformed prior to analysis (Sokal and Rohlf, 1995). We ran all analyses using R version 3.6.1 (R Core Team, 2019). We explored the effects of section (S; anterior vs posterior), island (I; Hawai'i vs Kaua'i), and performance (P; winner vs loser) on the proportion of red muscle fibers among groups. The full model included three fixed effects (S, I, and P), and all interactions. To estimate effect sizes, terms (main effects and interaction effects) that added little information to the predictability of the response were excluded. All sub‐models of the fully factorial model were fit and ranked by ΔAIC (Burnham et al. 2011).

3. RESULTS

In all comparisons, sections taken from the anterior segment had markedly lower proportions of red muscle fibers than sections taken from the posterior segment, with all anterior sections averaging < 20% red fibers, and all posterior sections averaging >60% red fibers (Table 1, Figure 2). Anterior sections varied by <3% between Kaua'i and Hawai'i, and between ‘winners’ and ‘losers’ from Hawai'i. Sections taken from the posterior segments showed greater variation between comparison groups, with fish from Kaua'i showing a 5–6% smaller proportion of red muscle fibers than fish from Hawai'i; however, ‘winners’ and ‘losers’ from Hawai'i only differed in their proportion of red fibers by <1%. Comparisons of coefficients of variation (i.e. the amount of variation relative to the mean) showed similar values across the posterior sections of all groups; however, fish from Kaua'i showed greater variation for both anterior and posterior sections than either winner or loser fish from Hawai'i (Table S1). Variation across our samples was best explained by models that included the additive effects of section (anterior vs. posterior) and island (Kaua'i vs. Hawai'i), with the interaction between those two variables potentially also improving explanation of the variation (Table 2). Models including only the individual effects of section or island were less informative; however, section did explain more variation and could be driving the additive effect of the best model (Table S2). Climbing performance (‘winner’ vs. ‘loser’) contributed little explanatory power (Table 2).

Table 1.

Proportions of red muscle fibers from histological sections of the Hawaiian climbing goby, Sicyopterus stimpsoni

	Kaua’i (anterior, n = 18)	Kaua’i (posterior, n = 18)	Hawai'i 'winner' (anterior, n = 22)	Hawai'i 'winner' (posterior, n = 22)	Hawai'i 'loser' (anterior, n = 18)	Hawai'i' loser' (posterior, n = 19)
Red muscle fiber proportion, %	17.49 ± 3.17	60.36 ± 10.25	17.97 ± 1.80	66.63 ± 10.62	19.78 ± 1.65	65.84 ± 9.61

Figure 2.

Proportion of red (and intermediate) muscle fibers in tail sections of Sicyopterus stimpsoni. Box represents 25–75th percentile of distribution, horizontal line represents the median, and whiskers indicate total data range. Sample size (n) for each tail segment is in parentheses below each plot. Posterior tail segments are indicated by grey boxes. The coefficient of variation for each treatment is indicated above each box‐and‐whisker plot. ant, anterior; post, posterior

Table 2.

Best models (ΔAkaike's information criterion < 2.0) to explain variation in red muscle fiber proportions of the Hawaiian climbing goby, Sicyopterus stimpsoni

Model	ΔAIC	Adjusted R 2	P
S + I + S:I	0	0.909	0.144
S + I	0.2	0.908	0.01
S + P + S:I	1.8	0.909	0.146

Abbreviations: AIC, Akaike's information criterion; I, island (Kaua'i vs. Hawai'i); P, performance ('winner' vs. 'loser'); S, section (anterior vs. posterior).

4. DISCUSSION

Differences in the proportion of axial red muscle fibers were evident across our comparative samples from S. stimpsoni, but the scales at which they contribute to differences in locomotor performance in this species did not entirely match our predictions. By far the greatest source of variation across our samples related to the antero‐posterior location from which sections were extracted, with the proportion of red fibers increasing posteriorly to the point that posterior sections contained a majority of red fibers (Figures 1 and 2). This pattern parallels that found by Cediel et al. (2008) in their analysis of more anterior segments from S. stimpsoni, and further emphasizes the important role of slow, sustained muscular activity in this species. Compared to the results of Cediel et al. (2008), our more posterior sections did not show red fibers along the periphery of the body wall near the horizontal septum; instead, these fibers were concentrated in internalized locations (a location also noted in that study). Given the strength of staining in other locations, and the presence of near‐peripheral red muscle fibers in the superior and inferior aspects of other sections (as well as lateral aspects of anterior sections), the lack of lateral peripheral red fibers in posterior sections seems unlikely to be artifactual. On the contrary, these patterns appear to represent a shift in red muscle fiber distributions for the most posterior portions of the tail, from peripheral to internal. The functional significance of the reduction in peripheral red muscle fibers in this region is unclear; however, it is possible that sustained contraction of peripheral muscle to stiffen the body wall, which could reduce drag in high‐velocity currents (see below), becomes less critical as the tail tapers toward the caudal peduncle.

Our sample groups showed little difference in fiber type proportions across their anterior segments. Instead, the divergence that was identified was concentrated in their posterior segments, with S. stimpsoni from Kaua'i showing a smaller proportion of red fibers than fish from Hawai'i. This disparity in fiber type proportions between fish from these two islands could reflect differences in the primary functional demands faced by fish on these islands. With longer stream reaches before waterfalls on Kaua'i, gobies on this island may need to engage more frequently in fast‐start escape responses from predators than fish from Hawai'i (Blob et al. 2010; Maie et al. 2014; Diamond et al. 2016, 2019), potentially making a relatively greater proportion of white, fast‐twitch muscle fibers in the tail advantageous (Jayne and Lauder, 1993). However, it is noteworthy that even fish from Kaua'i maintain a majority proportion of red fibers in their distal tail. It is possible that a sufficient scope of sustained activity is still important for S. stimpsoni from Kaua'i, such that a smaller proportion of red fibers would be disadvantageous. Sustained activity may be associated with the high stream velocities common in the adult habitat of this species (Schoenfuss and Blob, 2007). Although adults use their pelvic suction disk to maintain station in high‐velocity currents, the axial musculature may remain contracted to stiffen the body during station holding, thereby limiting flapping of the posterior body as a result of pressure differences imposed by the passing water (Schultz & Webb 2002; Tandler et al. 2019).

Comparative data on both climbing and fast‐start performance between S. stimpsoni juveniles from Hawai'i and Kaua'i would be informative for evaluating the extent to which observed differences in muscle fiber type proportions might translate into differing functional capacities between these subpopulations. In this context, the much higher coefficient of variation for muscle fiber proportions in the anterior segment of juveniles from Kaua'i (17.2%) when compared to both winners and losers from Hawai'i (10.1% and 8.5%, respectively) is noteworthy, as it implies a broad range of axial muscle fiber proportions in migrating juveniles entering streams on Kaua'i. Previous selection experiments (Blob et al. 2010) suggest that predation exerts selective pressures on migrating juvenile S. stimpsoni. The greater range of muscle fiber proportions observed in the anterior tail segment of migrating juveniles from Kaua'i, prior to predation, may represent a potential range of functional performance for these juveniles that is greater than for Hawaiian juveniles. In contrast, the lesser variation in muscle fiber proportions of anterior segments in Hawaiian juveniles may indicate that localized fitness peaks for climbing traits have reduced variation through stabilizing selection (Moody et al. 2017).

We observed no significant differences in axial muscle fiber type proportions between successful (‘winner’) and unsuccessful (‘loser’) climbers (Table 1, Figure 2). This result was surprising, considering the expected associations between locomotor performance and muscle fiber type proportions identified among fish from different islands (Table 1, Figure 2). This result also stands in contrast to expectations based on comparisons between species, in which the derived, inching mode of climbing (Cullen et al. 2013; Blob et al. 2019) is accompanied by an elevated proportion of red muscle fibers compared to multiple species of powerburst climbers (Cediel et al. 2008). In addition, it also contrasts with in‐stream observations of S. stimpsoni, in which fish from farther upstream reaches possess greater proportions of red fibers in climbing muscles than conspecific adults that reside in habitats in lower stream reaches (Schoenfuss et al. 2013). It is possible that the range of variation in this feature is limited in S. stimpsoni, and that subpopulations under high selective pressure for climbing performance are essentially ‘pushing the boundary’ with limited opportunity for further increases in red muscle proportions, as appears to be the case for several features of external morphology (Moody et al. 2017). Comparable analyses of axial muscle fiber type proportions from fish with differing fast‐start performance would provide useful perspective for evaluating whether performance variation in this alternative functional trait could be correlated with the relative abundance of red and white muscle.

Overall, however, our comparisons continue to highlight the extent to which functional performance depends on the integration of hierarchical levels of organization (Schoenfuss et al. 2013). As in the case of adhesive function of the sucker for this species (Maie et al. 2007; Maie et al. 2012; Schoenfuss et al. 2013), locomotor performance of S. stimpsoni likely depends on ‘internal’ features, such as muscle physiology, as well as external morphological traits. In contrast to comparisons of the muscles controlling the sucker, which vary within populations from a single stream exhibiting different levels of performance (Schoenfuss et al. 2013), results from this study suggest that correlations of axial muscle fibers with performance may only be evident at a broader scale, such as across subpopulations from different islands exposed to differing primary selection pressures (Blob et al. 2008; Blob et al. 2010; Moody et al. 2017), or even across species with different locomotor behaviors (Cediel et al. 2008). As these comparisons indicate, continued tests of how variation across multiple structural levels is associated with performance can provide a rich foundation for understanding the diversity of animal function.

CONFLICTS OF INTEREST

None declared.

Supporting information

 

 

Blob RW, Baumann T, Diamond KM, Young VKH, Schoenfuss HL. Functional correlations of axial muscle fiber type proportions in the waterfall‐climbing Hawaiian stream fish Sicyopterus stimpsoni . J. Anat. 2020;236:1160–1166. 10.1111/taja.13169