Myosin Isoform Expression in Dog Rectus Muscles: Patterns in Global and Orbital Layers and among Single Fibers

Abstract

purpose. To quantitate the distribution of myosin heavy chain (MyHC) isoforms along the global and orbital layers of dog rectus muscles and determine MyHC and myosin light chain (MLC) isoform patterns among single fibers from both layers.

methods. Serial samples of both layers of rectus muscles were prepared for gel electrophoresis. Relative amounts of each MyHC isoform in each sample were determined with scanning densitometry. Single fibers were isolated from each layer for analyses of MyHC and MLC isoforms.

results. Nine MyHC isoforms were detected. Four prominent MyHC isoforms, and an additional MyHC isoform at very low levels, are expressed in the global layer. Evidence suggests that all nine MyHC isoforms are expressed in the orbital layer. There are marked gradients in the levels of some MyHC isoforms along the length of both layers. Complex patterns of coexpression of multiple MyHC isoforms exist in single fibers from both layers. Most fibers express conventional slow or fast MLC isoforms, in accordance with the type (slow or fast) of MyHC isoform(s) in a given fiber, with the exception that slow fibers in the orbital layer express the atrial/embryonic isoform of MLC1.

conclusions. MyHC isoform expression patterns differ markedly between and along global and orbital layers of dog rectus muscles, with greater complexity in the orbital layer. Heterogeneity in MyHC isoform expression in rectus muscles is much greater than in limb muscles and presumably is the basis for the broad spectrum of extraocular muscle (EOM) contractile properties in driving oculomotor functions.

Extraocular rectus muscles drive eye rotations that span a broad range of angular velocities, from movements during slow vergence to very rapid saccades. A pivotal role of myosin heavy chain (MyHC) isoforms in the regulation of contractile properties of muscle among multiple vertebrate species has been well established ^{1 2} and there is evidence in support of a similar role for MyHC isoforms in EOMs in determining the speed of eye rotations. ^{3 4 5} The expression of MyHC isoforms in EOMs is more complex, compared with that in limb skeletal muscle, with respect to the number and intramuscular compartmentalization of isoforms, ^{4 5 6 7 8 9 10 11 12 13 14 15 16 17 18} and this complexity presumably subserves the broad repertoire of eye rotations. 

Mammalian EOMs consist of two layers: a global layer that inserts directly onto the globe and rotates the eye when its fibers contract and an outer orbital layer. The two layers differ from each other with respect to fiber types ^{19 20} and protein expression patterns. ^{6 21 22 23} There is substantial evidence that the two layers differ with respect to their roles in driving eye rotations. For example, the EOM pulley hypothesis proposes that the orbital layer of each rectus muscle inserts onto a ring of connective tissue, ²⁴ rather than onto the globe, and that contractions in the orbital layer position a connective tissue pulley that constrains the path of the EOM, while the global layer inserts directly onto the sclera and rotates the globe. ^{25 26} Although this structural arrangement is distinct from that between the global fibers and the globe, onto which they directly attach, the functional significance remains controversial. ^{26 27 28}  

The effect of rectus muscle resection (for example, to correct strabismus) on postsurgical eye rotation kinetics could depend on the extent of tissue removal and possible nonuniformity in MyHC isoform distribution within the EOMs. The results of several previous studies have revealed differences in myosin isoform expression between the global and orbital layers of rectus muscles, and along the length of these layers in small mammals in which binocular vision is generally absent. ^{9 13 29 30} EOMs express the same set of MyHC isoforms as do limb muscles in the same species, plus additional isoforms that are not normally expressed in adult limb muscles. Because there are differences in the MyHC isoform composition of limb muscles between small-size (e.g., rodents, rabbits) and large-size mammalian species, it is likely that there are species-related differences in MyHC isoform expression in EOMs, as well. It is reasonable to expect that myosin isoform composition is a major determinant of contractile properties of EOMs, as is generally believed to be the case for limb skeletal muscles. Furthermore, given that rectus muscles drive eye movements during functions such as tracking and vergence, it was deemed important to conduct a study of the complexity of MyHC isoform expression in EOMs in a species in which vision is binocular, as it is in humans, which could provide advantages for studying EOMs in a binocular system. Therefore, a large mammal, the domestic dog, was selected as the animal model for this study. A recent study ³¹ in this laboratory revealed that variation in MLC isoform expression in dog thyroarytenoid muscle adds to the complexity of myosin isoform expression in this muscle and thus may contribute to modulation of contractile properties in individual muscle fibers beyond that imparted by variations in MyHC expression alone. The objectives of this study were to conduct a thorough, quantitative analysis of MyHC isoform expression along the entire length of both layers in dog rectus muscles and to determine the combinatorial patterns of MyHC and MLC isoforms in single muscle fibers, from both layers, in which multiple isoforms are expressed. 

Methods

Samples

The care and use of all the animals from which samples were obtained for this study were in accordance with protocols approved by The Ohio State University Institutional Animal Care and Use Committee and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research. All the animals were adults, either beagles or hounds, and of either sex. Some samples were obtained from animals after termination of acute cardiovascular studies for other projects. None of the treatments in these studies was deemed to have any impact on the parameters included in this report. 

The cartilage of the temporomandibular arch was removed by blunt dissection to allow access to the origin of the rectus muscles, thereby assuring inclusion of the entire length of the rectus muscles when the eye was removed. Whole eyes were enucleated with all the rectus muscles attached and placed in cold relaxing solution (2.0 mM EGTA, 4.0 mM MgATP, 1.0 mM free Mg²⁺, 10.0 mM imidazole, and sufficient KCl to achieve an ionic strength of 180 mM [pH 7.0]). The rectus muscles were then carefully dissected along their entire length from the eye and placed in fresh, cold relaxing solution. The orbital and global layers of the ventral (inferior), dorsal (superior), lateral and medial rectus muscles from one eye of six dogs were separated from each other, and a sample of the entire cross-section of the belly of each layer was isolated. The inner transitional area between the two layers was not included when preparing the global and orbital layers, to avoid contaminating each sample with fibers from the other layer. This separation was easy to achieve in dog rectus muscles, as the distinction between the global layer fibers (relatively large diameters and very similar to limb muscle with less connective tissue) and orbital layer fibers (fibers with much smaller diameters surrounded by more connective tissue) was easy to discern with the use of a dissecting microscope (at ∼20× magnification). The orbital and global layers of the lateral rectus from one eye of three additional dogs were separated from each other and cut into 10 equally thick sections along their entire length. Sample mass was determined, and gel sample buffer ³² (30 μL/mg) was added. The orbital and global layers of the lateral rectus of one eye from two dogs were isolated along their entire length, and bundles of fibers from the insertion end, belly, and origin end were prepared by dissection and stored in glycerinating solution (same composition as relaxing solution but with 50% of the water substituted with glycerol). Segments (1–2 mm long) of single muscle fibers were isolated from the bundles by dissection and placed individually in 0.5-mL microcentrifuge tubes. The fibers were randomly selected. The goal of the portion of the study with the single fibers was to determine the MyHC isoform composition and the MLC isoform composition in each by analyzing separate aliquots with two gel formats. The MyHC isoform composition of two sets of 12 fibers from the insertion end of the orbital layer were analyzed because the fibers in the first set were too small for analysis of low molecular weight proteins, after the determination of their MyHC isoform composition. Therefore, 156 fibers were examined for MyHC isoform composition. The samples (cross-sectional samples of muscle layers and single fibers) were then prepared for analysis by gel electrophoresis, as described earlier. ²²  

Samples of the thyroarytenoid, right atrium, and right ventricle were prepared from one dog. The atrial and ventricular samples were prepared as described earlier for the cross-sectional samples of the global and orbital layers. The thyroarytenoid muscle was dissected as described elsewhere ³¹ and the sample prepared for this study was isolated from regions which together are known to express MyHC-I, -IIA, -IID and -IIB (designated as ML4, ML5, and ML6 in the earlier study ³¹ ). 

Gel Electrophoresis

The composition and staining of the gels, as well as the densitometric analysis to quantitate relative levels of MyHC isoforms in each sample (layer cross sections and single fibers) have been described. ²²  

Immunoblot Analysis

The proteins in the known MyHC region (based on previous gel runs) of one gel were transferred to nitrocellulose. An anti-MyHC antibody (MF 20, obtained from the Developmental Studies Hybridoma Bank [DSHB] at the University of Iowa and known to recognize all isoforms of sarcomeric MyHC in a broad range of animal species, spanning several phylogenetic classes) was used to identify MyHC bands on gels. An additional antibody, ALD-58 (also from the DSHB), was used to test whether dog rectus muscles express the tonic isoform of MyHC (MyHC-Ton) and to identify this isoform band. 

Mass Spectrometry

Samples were prepared for analysis by mass spectrometry (MS) for identification of specific MyHC isoforms on the gels), as described earlier. ³³ A Web search of protein databases was conducted on the MS results with a protein identification program (Mascot Daemon; Matrix Science, Boston, MA). Probability-based MOWSE scores greater than 56 are considered significant at P < 0.05. ³⁴ The identification of MLC isoforms on gels was based on the extracted myosin from single fibers which served as standards on subsequent gels, as described earlier. ²² The identification of the embryonic/atrial isoform of MLC1 (MLC1_E/A) in slow fibers of the orbital layer of dog rectus muscles, based on MS, was reported earlier. ³³  

Results

Identification of MyHC Isoforms

Seven protein bands were observed in the MyHC region of gels loaded with either EOM homogenates (Fig. 1)or single fibers from the same muscles. An immunoblot with monoclonal antibody MF 20 identified each band (Fig. 2)as MyHC. Four bands had the same mobility in samples from the global and orbital layers and comigrated with the four MyHC isoforms in adult dog thyroarytenoid muscle, which were identified using MS. ³¹ Nevertheless, all four global layer bands were identified in the present study with MS, and the identifications were identical with those in the previous study: MyHC-IIA, -IID, -IIB, and -I, in order of slowest to fastest gel mobility. The MS results are listed in Table 1 . The number of peptides matched for these four bands ranged from 147 to 410, and the MOWSE scores ranged from 3678 to 7866, far exceeding the cutoff score of 56 (see the Methods section), indicating a very high level of confidence in their correct identification. Others have reported that MyHC-extraocular (MyHC-EO) is expressed in both layers of some species (see the Discussion section). The ninth match for the band identified as MyHC-IID was with MyHC-EO. However, given the much lower MOWSE score with this match (5286, compared with the score for MyHC-IID of 7866), we assume that the global layer did not express MyHC-EO, but this possibility cannot be eliminated. The band identified as MyHC-IIB has never been observed in digastric (a jaw-opening muscle), diaphragm, or any limb muscle (tibialis cranialis, gastrocnemius, vastus lateralis, semitendinosus, and extensor digitorum, muscles which in rodents express high levels of MyHC-IIB) from more than 30 adult dogs in any study in this laboratory over a 10-year period. Furthermore, histochemically defined IIB fibers apparently do not exist in adult dog limb skeletal muscle. ^{35 36} The band identified as MyHC-I comigrated with the predominant MyHC (MyHC-β) in samples of adult dog ventricles (Fig. 1and illustrated for MyHC-I in dog thyroarytenoid in Bergrin et al. ³³ ). Others have reported that MyHC-I in limb skeletal muscle and MyHC-β in adult cardiac muscle are the same protein in several mammalian species. ^{37 38}  

All seven bands in samples from the orbital layer were also submitted for identification by MS, with the number of peptide matches and MOWSE scores ranging from 62 to 264 and 1043 to 5469, respectively (Table 1) . The slowest-migrating band was identified as embryonic MyHC. This band has also never been observed in any dog limb muscle from more than 30 adult dogs in any study in this laboratory. The second and third slowest isoforms were identified as MyHC-IIA and -IID. These bands comigrated with the two identically identified MyHC isoforms in the global layer and with dog thyroarytenoid and limb muscles. ³¹ An additional match of the third slowest band with the second highest MOWSE score for this band predicted that this band was MyHC-extraocular (MyHC-EO). Therefore, although the highest probability is associated with this protein being MyHC-IID, it is possible that this gel band also contains MyHC-EO. Therefore, this band from the orbital layer is labeled MyHC-IID/EO. The fourth slowest band was identified by immunoblotting (Fig. 2) , because the MS results were ambiguous, possibly due to the relatively low amount of this protein. This band was recognized, with greater specificity than the other MyHC bands, by anti-MyHC-Ton ALD-58 antibody. It is noteworthy that this band, although present at a very low level, compared with other MyHC bands, stained the most intensely with ALD-58. This band was, therefore, identified as MyHC-Ton. The MS results and the patterns of comigration with other MyHC isoforms in single fibers (described below) suggest that two MyHC isoforms comigrate as the fifth slowest migrating band. This band comigrated with the band in the global layer that was identified by MS as MyHC-IIB. The band (orbital layer) from one dog was identified by MS as MyHC-IIB and from another dog as MyHC-α. The MS results from the second dog yielded five matches for this band with nearly equal probabilities that this band was either MyHC-β or -α. However, the migration of this band was distinctly slower than that of MyHC-β in cardiac ventricles. Human MyHC-β and -α share 96.2% homology at the deduced amino acid level. ³⁹ The high homology likely is responsible for the nearly equal MS probabilities of this band being MyHC-β or -α in the second dog. We, therefore, conclude that the fifth slowest migrating band in the orbital layer is either MyHC-IIB or MyHC-α. The isoform in this band was always coexpressed with other isoforms in individual fibers (described later). We assume that, when the isoform in this band was expressed in fibers with only fast-type isoforms (MyHC-IIA and -IID/EO), it was MyHC-IIB. We also assume that when this isoform was expressed in other fibers that expressed slow-type isoforms, it was MyHC-α. In support of these assumptions are the independent observations of MLC isoforms in the same fibers. Briefly, only fast-type MLC isoforms were observed in the former, and no fast-type MLC isoforms were observed in the latter. The sixth band was identified by MS as MyHC-I and from its comigration with MyHC-I in adult dog limb muscles ³¹ and with the predominant MyHC isoform (i.e., MyHC-β) in adult dog cardiac ventricles. The only MyHC isoform that was not identified was the fastest migrating band. MS results suggested that this protein was similar to MyHC-I. However, its migration was distinctly faster than MyHC-I. This protein was consistently expressed at higher levels in samples from the orbital layer, compared to the global layer (described later). Furthermore, this MyHC isoform was detected in most of the orbital fibers but was never exclusively expressed in single fibers. It is unlikely that this protein is a proteolytic fragment of MyHC-I, because it was observed at only very low levels in samples from the global layer, in which the level of MyHC-I is as high as in the orbital layer. Furthermore, this isoform was detected in fibers from the orbital layer that expressed only fast-type MyHC isoforms, as well as fibers expressing MyHC-I. It seems very unlikely that all the MyHC-I in the former fibers would have been proteolytically degraded. We, therefore, conclude that this is an additional MyHC isoform or a fragment of a different MyHC (from the MF 20 immunoblot) and that its identity remains to be determined. It is designated in this report as MyHC-Unidentified (MyHC-Un). 

Thus, the total number of MyHC isoforms detected in dog rectus muscles was nine, with apparently two pairs of MyHC isoforms comigrating: MyHC-IID with MyHC-EO, and MyHC -IIB with MyHC-α. 

MyHC Isoform Quantitation in Bellies of Global and Orbital Layers

The EOMs are much thicker in the belly region than in the tapering origin and insertion ends. Therefore, samples were initially prepared specifically from the belly region of the global and orbital layers of all four rectus muscles of one eye each from six dogs and the MyHC isoform composition of each sample was determined (Fig. 3)and quantitated for an initial assessment of isoform expression in the thickest portion of each rectus muscle. The results indicate that all four dog rectus muscles are quantitatively very similar with respect to the expression of specific MyHC isoforms and with respect to MyHC isoform differences between the two muscle layers (Fig. 4) . The belly region of the global layer of all four muscles uniformly expressed four MyHC isoforms: MyHC-IIA, -IID, -IIB, and -I, with MyHC-IIB consistently the predominant isoform (∼50% of total MyHC). MyHC-IID was the second most abundant isoform (25%–30% of total MyHC), whereas MyHC-I and -IIA were less abundant (both comprised 10% to 15% of total MyHC). Very low levels (∼1% of total MyHC) of MyHC-Un were also detected in the belly of the global layer of several rectus muscles. 

In the belly region of the orbital layer of all four rectus muscles, the MyHC-IID/EO band predominated (65%–75% of total MyHC), whereas MyHC-IIA, MyHC-IIB/α, and -I each comprised 5% to 15% of total MyHC in the same samples. MyHC-Emb and -Un were detected at very low levels (≤3%) and -Ton was not detected in the belly region. 

Distribution of MyHC Isoforms along the Length of Global and Orbital Layers

Four prominent MyHC isoforms, MyHC-I, -IIA, -IID, and -IIB, were expressed along the entire length of the global layer of dog rectus muscles (Figs. 5 6 ). Small amounts (1%–2% of total MyHC) of MyHC-Un were detected along the entire length of the global layer, as well. The relative amounts of MyHC-I and -IIA varied along the length of the global layer in a parallel manner, with greater relative levels of both isoforms at the origin and insertion regions. MyHC-IIB was the predominant isoform along the entire length of the global layer, except at the insertion and origin ends. Therefore, the relative level of MyHC-IIB varied along the length of the global layer in a manner that is opposite to that of MyHC-I and -IIA. The relative level of MyHC-IID was nearly constant along the length of the global layer. MyHC-Emb and -Ton were not detected in the global layer of dog rectus muscles. 

The pattern of MyHC isoform expression along the length of the orbital layer was more complex than along the global layer, due to the greater number of isoforms and marked differences in the amounts of the different isoforms. The relative levels of MyHC-I, -Emb, -IIA, and -Un varied similarly—greater at the insertion and origin regions (Fig. 7) . However, in contrast to MyHC-I, -IIA, and -Un, MyHC-Emb was present at extremely low levels in the mid (belly) region of the orbital layer. The relative levels of MyHC-IID/EO and the isoform(s) in the fifth band (MyHC-IIB/α) along the length of the orbital layer were greater in the mid (belly) region than in the insertion and origin. The relative level of MyHC-Ton also varied along the length of the orbital layer, but in contrast to the other MyHC isoforms, the pattern was skewed, with greater levels in the insertion region and gradually tapering to lower levels toward the origin. 

Distribution of MyHC Isoforms in Single Fibers from the Global and Orbital Layers

The MyHC isoform composition of single fibers from the insertion, belly, and origin of both layers was determined to evaluate what, if any, specific patterns of MyHC isoform expression and coexpression exist among individual fibers. Representative gels are shown in Figures 8 and 9 . Some general observations and comparisons between global and orbital layers are described first, followed by a description of layer-specific observations in the following two paragraphs. A majority of fibers in both layers (89% and 75% in the global and orbital layers, respectively) had one of only three MyHC isoform combinations. The three predominant combinations were different between the two layers. Two of the three combinations in both layers were of exclusively slow-type MyHC isoforms. Overall, 54% and 38% of the examined fibers in the global layer and orbital layer, respectively, expressed only fast MyHC isoforms. Only 3 of the 84 orbital fibers and none of the global fibers coexpressed fast and slow MyHC isoforms (designated as “hybrid” fibers). The number of MyHC isoform combinations (i.e., coexpressions) in single fibers was fairly constant between the insertion end, belly, and origin end within both layers (3, 4, and 5, respectively, in the global layer and 6, 8, and 6, respectively, in the orbital layer). 

Five patterns of MyHC isoform expression were observed in 72 fibers from the global layer (Table 2) , with the most predominant pattern being a fast-type, with coexpression of MyHC-IIA, -IID, and -IIB in each of 31 fibers. Two other types of fast fibers were detected in the global layer: All six fibers of one type coexpressed MyHC-IID and -IIB, in approximately equal relative amounts (as assessed visually), and two fibers of another fast type expressed MyHC-IIA and -IID. No single MyHC isoform consistently predominated among the fibers in these groups. The remainder of the global fibers was slow, with one type expressing exclusively MyHC-I and the other coexpressing MyHC-I and -Un. The fibers expressing only MyHC-I were the only fibers of any type from either layer that expressed a single MyHC isoform. 

All the orbital fibers coexpressed multiple MyHC isoforms, the majority (75%) expressing from three to six. Fifteen patterns of MyHC isoform expression were observed in 84 fibers from the orbital layer (Table 3)and, overall, approximately 60% of these fibers were slow. The most predominant type (27 fibers) was slow, with each fiber expressing five isoforms: MyHC-Emb, -Ton, -α, -I, and -Un. Four additional fiber types in the orbital layer were also slow, with all the fibers coexpressing MyHC-I and -α, with or without other MyHC isoforms. MyHC-I predominated in every slow fiber. Although MyHC-Ton was a relatively minor component of the total MyHC in homogenates of the orbital layer, it was detected in 40 of the examined orbital fibers, all being slow. All fibers that expressed MyHC-Emb also expressed MyHC-Un (the converse is not true), and always with either slow-type MyHC isoforms, fast-type MyHC isoforms, or both. MyHC-Emb and -Un were never expressed alone, either singly or together. Eight orbital fiber types were fast (32 fibers). All the fast fibers coexpressed two to five MyHC isoforms. 

Patterns of Expression of Individual MyHC Isoforms

Expression patterns of individual MyHC isoforms can also be gleaned from the information in Tables 2 and 3 . MyHC-Emb was always coexpressed with MyHC-Un, and, in some fibers, also with fast MyHC isoforms. MyHC-IIA was found in both layers and was consistently coexpressed, almost always with MyHC-IID (global layer) or with MyHC-IID/EO (orbital layer). MyHC-IID and -IID/EO was consistently coexpressed, almost always with MyHC-IIA. MyHC-Ton was found only in the orbital layer (in ∼50% of all orbital fibers) and was always coexpressed with MyHC-α (assuming that band 5 in these fibers was not MyHC-IIB) and -Un. MyHC-I was found in both layers, was always coexpressed in the orbital layer, and was expressed exclusively or with MyHC-Un in fibers in the global layer. MyHC-Un was found in fibers from both layers, was always coexpressed, was expressed in every fiber from the insertion and origin of the orbital layer, was expressed in half of the fibers in the belly of the orbital layer, and was expressed only in some global fibers that also expressed MyHC-I. MyHC-IIB was abundantly expressed in global layer fibers that also expressed MyHC-IIA and -IID and was never expressed alone. MyHC-α was expressed in every slow fibers of the orbital layer and appeared to be restricted to this layer. 

Myosin Light Isoforms in Single Fibers

Five patterns of MLC isoform expression were detected among the single fibers in which MyHC isoform composition was also determined. Gels loaded with single fibers from the origin region of the global and orbital layers are shown as examples in Figures 8 and 9 , respectively. The fibers from the global layer expressed exclusively fast-type (MLC1F, -2F, and-3) or slow-type (MLC1S and -2S) isoforms, in accordance with their fast- or slow-type MyHC isoform composition. The pattern of MLC isoform expression among orbital fibers was more complex. All the fibers in the orbital layer that are indicated as being slow in Table 3(Fiber Type) expressed MLC1_E/A and slow-type MLC2, as reported earlier. ²² Some of these fibers also expressed slow-type MLC1. The fibers in the orbital layer that are designated as fast-type in Table 3expressed exclusively fast MLC isoforms. Hybrid fibers, found only in the orbital layer, expressed fast- and slow-type isoforms of MLC1 and -2. 

Discussion

The results of this study provide a detailed, quantitative analysis of the very complex pattern of MyHC and MLC isoform expression in the global and orbital layers of dog rectus muscles and among the single muscle fibers that comprise these muscles. Previous reports revealed unique features of myosin expression in rectus muscles of several mammalian species, including the presence of MyHC isoforms that are not normally expressed in adult limb muscles and striking regional compartmentalization of specific isoforms. For example, it has been reported that MyHC-EO is expressed in the rectus muscles of several mammalian species, including rat, ^{15 18 29} rabbit, ^{12 40} cow, ⁴ and domestic dog. ⁵ Others have reported that adult EOM fibers express embryonic and neonatal isoforms of MyHC. ^{12 18} MyHC-α and -Ton are expressed in human rectus muscles. ^{14 21} There is some species specificity in the expression of these MyHC isoforms, as not all are expressed in all examined species. The relatively low level of MyHC-Ton detected in dog rectus muscles in this study is consistent with a recent report of a very low level of this protein in human EOM. ⁴¹  

The results also show a general, reciprocal pattern in the distribution of MyHC isoforms that are typically associated with either slow or fast contractions in different regions of rectus muscles (Fig. 6) . Specifically, the relative levels of fast-type MyHC-IIB and -EO are greater in the longitudinal center, whereas the levels of MyHC-I and -IIA are greater at the insertion and origins ends. Similar variations in the distributions of fast (including MyHC-EO) and slow MyHC isoforms in rat and rabbit rectus muscles have been reported. ^{9 12 13 15 29 40} The functional implications of this pattern are not clear, but it may provide a mechanism for rapid, yet smooth accelerations during eye rotations, with the ends of the layers dampening sudden movements that may be driven by the central regions, at least under some circumstances such as smooth, yet rapid, pursuits. 

Only one type of fiber, identified on the basis of MyHC isoform composition, specifically the slow fibers in the global layer, expressed exclusively one MyHC isoform (MyHC-I). All other fibers in dog lateral rectus coexpressed two to six MyHC isoforms. Therefore, no MyHC isoform, other than MyHC-I, was expressed exclusively of other isoforms. Although some patterns of MyHC isoform coexpression were relatively common (e.g., MyHC-IIA, -IID, and -IIB coexpressed in 16 fibers of the global layer, other combinations were rarely or never detected (e.g., MyHC-I was not found to be coexpressed with the fast-type MyHC isoforms, MyHC-IIA, -IID, and -IIB. Furthermore, MyHC-α and -I were always expressed in fibers which also expressed MyHC-Ton. Therefore, although the number of MyHC isoforms detected in dog rectus muscles (up to nine) far exceeds the total number of MyHC isoforms in adult dog limb muscle (three: MyHC-I, -IIA, -IID), there appear to be strict controls on the regulation of expression of MyHC isoforms in adult rectus muscles. Only five patterns of MyHC isoform expression were observed in the global layer (Table 1) , and 15 patterns were observed in the orbital layer (Table 2) . It is possible, however, that additional MyHC isoform combinations in single fibers would have been detected if more fibers were analyzed. The theoretical maximum number of combinations with nine different MyHC isoforms in single muscle fibers is 511 (the number of binary combinations with each of nine isoforms being expressed or not expressed in a given fiber is 2⁹ = 512 − 1, because each muscle cell must express at least one MyHC isoform). This number can be conservatively reduced when considering only the number of MyHC protein bands (i.e., 127 possible combinations [2⁷ − 1]), because it is not absolutely certain whether two (or more) isoforms were present in the third and fifth slowest migrating bands. Therefore, although the number of possible MyHC isoform combinations actually observed in dog rectus muscles is much larger than the theoretical maximum number in limb muscles, which express only MyHC-I, -IIA, and -IID, (2³ − 1 = 7), the total observed is much lower than the theoretically maximum possible when considering the number of MyHC isoforms detected in dog EOMs. The mechanisms that regulate the expression of multiple MyHC isoforms within individual muscle fibers are not understood. It is reasonable to expect that the control mechanisms that govern the expression of MyHC isoforms in EOM muscles are elaborate, because only specific patterns of coexpression were observed, and many other potential MyHC combinations were never observed. 

A greater level of MyHC-Emb in the orbital, compared to the global, layer of mammalian rectus muscles appears to be a general pattern (rabbit ¹² and rat ^{9 15} ). McLoon et al. ¹³ also reported that a greater number of fibers in the orbital layer of rabbit rectus muscles express developmental (apparently embryonic) MyHC, compared to global fibers. Furthermore, the level of MyHC-Emb is greater in the proximal and distal portions of the orbital layer in both rats and rabbits, compared to the central end-plate zone, ^{12 13 15} near the muscle belly. Both of these observations are identical with the pattern of MyHC-Emb expression observed in dog rectus muscles in the present study. One difference between the results from dog EOM and those of Jacoby et al. ⁹ from rat EOM is that not all fibers in the central zone of the orbital layer in dog EOM express fast-type MyHC, whereas they do in rat EOM. Therefore, while some general patterns of MyHC isoform expression are shared among mammalian species, differences in other patterns exist. Rubinstein et al. ¹⁵ also reported variations in MyHC isoforms along the length of the orbital layer in rat EOMs and demonstrated a central localization of large NMJs in rat orbital layer and diffuse distribution of smaller NMJs throughout the layer. It is not clear to what extent length-dependent variations in MyHC isoform expression are mechanistically or just coincidentally related to the end-plate zone, as the results of a recent study demonstrate a diffuse distribution of neuromuscular junctions throughout rabbit and monkey EOMs. ⁴²  

Multiple reports demonstrate that mean muscle fiber length is much shorter than rectus muscle length in several mammalian species. ^{42 43 44} It is very likely, therefore, that different populations of muscle fibers were sampled at the origin, belly, and insertion in the present study. It was clear, from visual inspection (but not quantitated), that fiber length in the global layer of dog rectus muscles is much greater than fiber length in the orbital layer. This is consistent with less variation in MyHC isoform expression along the length of the global layer, compared with the orbital layer, as illustrated in Figure 2 . 

Our results also demonstrate that most dog EOM fibers coexpress multiple MyHC isoforms, especially in the orbital layer, where every fiber examined expressed multiple isoforms and 71 of 84 of these fibers expressed three or more isoforms. Coexpression of MyHC isoforms in single EOM fibers of several mammalian species has been reported by others (rat, ^{9 15 17 18 29 45} cat, ¹⁷ rabbit, ^{40 46} and human ^{21 47 48} ). EOMs are not unique in having a majority of single fibers coexpressing multiple MyHC isoforms, as this has also been reported for the dog intrinsic laryngeal muscles. ^{31 49 50} The functional consequences of the presence of multiple MyHC isoforms, all of which are assumed to be participating in crossbridge cycling, in individual fibers are not well understood. How this situation can exist in fibers, with some crossbridges cycling either rapidly or slowly or possibly generating different amounts of force, while maintaining structural stability of the fiber, is puzzling. All the single fibers in which MHC isoform patterns were examined were skinned (sarcolemma absent), so it is reasonable to assume that all of the observed isoforms were incorporated into thick filaments and rather than being soluble. Presumably, the extensive coexpression of MyHC isoforms among individual fibers in EOM muscles serves the broad range of eye movements during fixation, slow tracking, or saccades. Whether the different MyHC isoforms in a given EOM fiber are homogenously distributed among all sarcomeres or are differentially located among different myofibrils, as can occur in developing muscle, ⁵¹ is not known. Finally, it is possible that, if EOM fibers are polyneuronally innervated (for which some evidence exists ^{52 53 54 55 56 57 58} ) different patterns of activation by individual neurons in single fibers results in activation of multiple MyHC genes, a mechanism proposed by Hoh ⁵⁹ to explain coexpression of MyHC in single fibers of thyroarytenoid muscle in several mammalian species. 

Considerable evidence has accrued to suggest that MLC isoforms serve a modulatory role in the determination of contractile properties of skeletal muscle (reviewed in Refs. ^{1 2 60} ). The results of this study indicate that most fibers in dog EOM express the slow- or fast-type MLC isoforms expected on the basis of MyHC isoform composition, with the exception of slow fibers in the orbital layer expressing the MLC1 isoform characteristic of embryonic skeletal muscle and adult atrial muscle. Therefore, except for slow orbital fibers, variation in contractile properties among dog single EOM fibers is likely to be subserved by the extensive heterogeneity in the heavy chain subunit of myosin, with little, if any, contribution from light chains. 

Classification of EOM fibers frequently is made using a six-fiber-type scheme, which is based on location (global versus orbital layers) and type of innervation (single versus multiple innervation; reviewed in Porter et al. ⁶¹ ). The six types are orbital multiply innervated, orbital singly innervated, global multiply innervated, global red singly innervated, global intermediate singly innervated, and global pale singly innervated. It is not known whether or how dog EOM fibers fit this classification scheme. However, it appears that multiple subclasses of fiber types exist, distinguished on the basis of MyHC isoform composition, within at least some of the traditional six types, which are based on location and innervation, expressing different sets of MyHC isoforms. The six-type classification scheme is valuable with respect to providing a basis for categorizing fiber types on the basis of major differences, but it is somewhat limited in recognizing the full extent of muscle fiber heterogeneity in EOMs. Consideration of fiber types in EOMs, as well as other muscles (e.g., Bergrin et al. ³¹ ) on the basis of MyHC isoform composition provides another classification scheme, as suggested by McLoon et al. ¹³  

It is reasonable to presume that the extraordinarily complex patterns of MyHC isoform expression in EOMs subserve the broad repertoire of eye rotations during normal oculomotor events. Li et al. ³ reported that cross-bridge kinetics in single fibers from rabbit EOM span a broader range than those in limb slow and fast muscle fibers. More recently, Toniolo et al. ^{4 5} demonstrated that fibers in dog and cow EOM that express different MyHC isoforms have distinct sets of contractile properties with respect to force-generating ability and rates of contraction. The quantitative data in the present study, when considered in the context of these findings, should allow for more precise predictions regarding the role of the muscle fibers in a specific layer or in a region within a layer in driving eye movements or the potential consequences of muscle resections for correction of strabismus. Concerning the latter, resection of the distal global layer may be expected to compromise slow oculomotor functions, at least initially, as MyHC-I and -IIA, the latter being associated with the slowest contractions driven by the fast-type MyHC isoforms, comprise a greater amount of total MyHC in this region (Fig. 6) . The number of muscle fibers expressing the neonatal isoform of MyHC increased after resection in rabbit EOM in one study, ⁶² and after injection of botulinum toxin A, frequently used to pharmacologically treat strabismus, in another study. ⁶³ Other studies have also demonstrated malleability of MyHC isoform expression in EOMs. In vivo injections of growth factors (bone morphogenic protein-4, transforming growth factor-β1, sonic hedgehog, and Wnt3A, a member of the Wingless transcription factor family) into rabbit EOMs induced changes in MyHC isoform expression. ⁶⁴ Generally, the numbers of fibers expressing MyHC isoforms in the fast-type family (adult fast MyHC, neonatal MyHC, and developmental [presumably embryonic] MyHC) were lower after injection of these growth factors. Injection of insulin-like growth factor II caused a large decrease in the number of fibers expressing MyHC-neo in another study. ⁶⁵  

The fibers of the global layer insert directly onto the sclera and thereby act on the globe, whereas evidence suggests that the orbital layer inserts onto a ring of connective tissue and, according to the EOM pulley hypothesis, does not directly drive eye rotations. ^{24 25 26} The functional significance of a different structural relationship between the global and orbital fibers and the globe remains controversial. ^{27 28} In the context of a putative functional distinction between the two layers, it is noteworthy that only the three MyHC isoforms found in dog limb muscles (MyHC-I, -IIA, and -IID) and MyHC-IIB were detected in the global layer, whereas at least seven MyHC isoforms were detected in the orbital layer. Furthermore, the number of combinations of MyHC isoforms among single fibers was much greater in the orbital layer (i.e., 15) compared with the global layer (i.e., only 5). Thus, the fibers acting directly on the eye are more like limb muscle fibers with respect to MyHC isoform expression, with the striking exception being the additional expression of MyHC-IIB, whereas the orbital fibers are very different. The possibility that additional isoforms are expressed in the global layer of dog rectus muscles, which were below the level of detection in this study (silver stain), cannot be excluded. The greater complexity of MyHC isoform expression in the orbital layer suggests that this layer does, in fact, have a very important role during oculomotor tasks, but apparently do not directly drive eye rotations. 

Although significant progress has been made during the past approximate two decades in understanding MyHC and MLC isoform expression patterns in mammalian EOMs, some limitations and uncertainties remain. Two limitations are antibodies that recognize more than one MyHC isoform, leading to false-positive results, and the lack of antibody recognition of specific MyHC isoforms in some species, leading to false-negative results. Electrophoretic comigration of multiple MyHC isoforms, despite efforts to separate all isoforms, is also a significant problem. Despite these persistent concerns, a large body of knowledge of myosin isoform expression in mammalian EOMs, from multiple laboratories and based on an array of experimental approaches, has evolved. The results of this study provide new quantitative data on normal patterns of myosin isoform expression in a large animal model with binocular vision. The results reveal elaborate patterns of myosin isoform expression between and along the global and orbital layers and among individual fibers in these layers and can be used to quantitatively assess differences in expression patterns during either pathophysiological states or after experimentally induced perturbations. 

 Supported by Grant IOB 0133613 from the National Science Foundation. The MF 20 and ALD-58 antibodies, from Donald A. Fischman, Cornell University, were obtained from the Developmental Studies Hybridoma Bank, developed under the auspices of the National Institute of Child Health and Human Development of the National Institutes of Health and maintained by The University of Iowa, Department of Biological Science, Iowa City, IA.

 Submitted for publication June 10, 2008; revised July 23, 2008; accepted October 13, 2008.

 Disclosure: S. Bicer, None; P.J. Reiser, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

MyHC region of a silver-stained gel loaded with homogenates of right ventricle (RV), right atrium (RA), orbital and global layers of the lateral rectus (LRO and LRG) muscle, and thyroarytenoid (TYA) muscle. RV, RA, and TYA samples were used on this gel as sources of standards for MHC-I/β (RV); -α (RA); and -IIA, -IID, and -IIB (TYA) in dog skeletal muscle, which were identified in a previous study. ³¹

Figure 2.

 

Lateral rectus orbital (LRO) layer analysis. One lane from an SDS gel (left) loaded with a LRO homogenate, two lanes of an immunoblot (middle) of the LRO probed with MF20 antibody, and one lane of an immunoblot (right) of the LRO probed with ALD-58 antibody.

Table 1.

 

Identification of Myosin Heavy Chain Isoforms by Mass Spectrometry

Band No.*	Protein Identification	Accession No.	No. of Peptides Matched	PBMS, †
Global Layer
2	MyHC-IIA	Q076A7	330	6281
3	MyHC-IID	Q076A6	410	7866
5	MyHC-IIB	XP_861854	147	3678
6	MyHC-I/β	P49824	391	6555
Orbital Layer
1	MyHC-Emb	NP_001107184	104	2466
2	MyHC-IIA	NP_001070263	62	3155
3	MyHC-IID	XP_862232	85	1465
	MyHC-EO	NP_003793	68	1043
4	MyHC-tonic, ‡
5	MyHC-IIB	XP_861854	120	2845
	MyHC-α	P02563	264	5144
6	MyHC-I/β	BAB20630	221	5469
7	MyHC-I/β	XP_848707	93	3251

*  Electrophoretic migration order (1, slowest; 7, fastest).

†  Probability-based MOWSE Score.

‡  Identified by immunoblot analysis.

Figure 3.

 

MyHC region of a silver-stained gel loaded with homogenates of the belly region of the orbital (O) and global (G) layers of the lateral, medial, dorsal, and ventral rectus muscles (LR, MR, DR, and VR, respectively). One MyHC isoform, MyHC-Ton, was detected at very low levels in the belly region of the orbital layer of rectus muscles and is not evident on this gel.

Figure 4.

 

Relative amounts (percent of total) of MyHC isoforms in homogenates of the belly region of the global (G) and orbital (O) layers of the lateral, medial, dorsal, and ventral rectus muscles (LR, MR, DR, and VR, respectively). Each bar represents the mean ± SEM of the values obtained from six dogs.

Figure 5.

 

Top: Images of the global and orbital layers of the lateral rectus muscle. Vertical lines demarcate the 10 sections prepared from each layer. Bottom: two gels onto which homogenates of the 10 sections prepared from the global and orbital layers of lateral rectus were loaded.

Figure 6.

 

Relative amounts (percentage of total) of MyHC isoforms in homogenates of 10 sections prepared from the global layer of the lateral rectus. Each bar represents the mean ± SEM of the values obtained from three dogs.

Figure 7.

 

Relative amounts (percent of total) of myosin heavy chain (MyHC) isoforms in homogenates of the 10 sections prepared from the orbital layer of the lateral rectus. Each bar represents the mean ± SEM of the values obtained from three dogs.

Figure 8.

 

SDS gels onto which single slow-type (S), and fast-type (F) fibers from the global layer of lateral rectus (LRG) were loaded. Two gels were used to separate myosin heavy chain (MyHC) isoforms (top) and low-molecular-weight proteins (bottom) in the same fibers. A homogenate of the LRG was loaded in the center lane of both gels. Tm-α and Tm-β are isoforms of tropomyosin.

Figure 9.

 

SDS gel onto which single hybrid (H), slow-type (S) and fast-type (F) fibers from the orbital layer of lateral rectus (LRO) were loaded. Two gels were used to separate myosin heavy chain (MyHC) isoforms (top) and low molecular weight proteins (bottom) in the same fibers. A homogenate of the LRO was loaded in the center lane of both gels. Tm-α and -β, isoforms of tropomyosin; E/A, embryonic/atrial.

Table 2.

 

Combinations of MyHC Isoforms Observed among 72 Fibers from the Global Layer, Listed Separately for the Insertion End, Belly, and Origin End of Lateral Rectus Muscle from Two Dogs

MyHC Isoform Combinations	Global Layer of Lateral Rectus
		Insertion	Belly	Origin	Total	Fiber Type
MyHC-I	0	14	4	18	Slow
MyHC-I, Un	7	2	6	15	Slow
MyHC-IID, IIB	0	5	1	6	Fast
MyHC-IIA, IID, IIB	16	3	12	31	Fast
MyHC-IIA, IID	1	0	1	2	Fast

 The underscored isoform in MyHC-IID, IIB is the predominant one. The last column indicates the general fiber type (fast, slow, or hybrid), based on MyHC isoforms expressed in fast (MyHC-IIA and MyHC-IID) and slow (MyHC-I) fibers of limb muscles.

Table 3.

 

Combinations of MyHC Isoforms Observed among 84 Fibers from the Orbital Layer, Listed Separately for the Insertion End, Belly, and Origin End of the Lateral Rectus Muscles of Two Dogs

MyHC Isoform Combinations	Orbital Layer of Lateral Rectus
		Insertion	Belly	Origin	Total	Fiber Type
MyHC-Emb, Ton, α, I, and Un	20	0	7	27	Slow
MyHC-Ton, α, I, and Un	0	2	9	11	Slow
MyHC-α, I, and Un	3	0	0	3	Slow
MyHC-Ton, α, and I	0	2	0	2	Slow
MyHC-α, and I	0	6	0	6	Slow
MyHC-IID/EO, and IIB	0	2	0	2	Fast
MyHC-IIA, IID/EO, and IIB	0	2	0	2	Fast
MyHC-IIA and IID/EO	0	2	0	2	Fast
MyHC-Emb, IIA, IID/EO, and Un	6	3	1	10	Fast
MyHC-Emb, IIA, IID/EO, IIB, and Un	1	0	0	1	Fast
MyHC-IIA, IID/EO, and Un	0	5	2	7	Fast
MyHC-IIA, IID/EO, IIB, and Un	5	0	0	5	Fast
MyHC-IIA and Un	0	0	3	3	Fast
MyHC-Emb, IIA, IID/EO, α/IIB, I, and Un	0	0	2	2	Hybrid
MyHC-Emb, IID/EO, α/IIB, I, and Un	1	0	0	1	Hybrid

 The predominant isoform, if any, in each combination is underscored. The last column indicates the general fiber type (fast, slow, or hybrid), based on MyHC isoforms that are known to be expressed in slow (MyHC-I) or fast (MyHC-IIA and MyHC-IID) fibers of dog limb muscles. Hybrid fibers are those with combinations of fast and slow MyHC isoforms. Results suggest that MyHC-α and MyHC-IIB comigrated on the gels. Designation of one or the other of these two isoforms in this table is based on coexpression in individual fibers of other MyHC isoforms and on myosin light chain isoform composition.