A recent study (Bicer S and Reiser PJ. J Muscle Res Cell Motil 25: 623–633, 2004) suggested considerable variation in the apparent molecular mass (Ma), deduced from electrophoretic mobility, in fast-type myosin light chains (MLCF), especially MLC1F, among mammalian species. Furthermore, there was an indication that MLC1F Ma generally correlates with species body mass, over an ∼4,000-fold range in body mass. The results also suggested that Ma of other low-molecular-weight myofibrillar proteins is less variable and not as strongly correlated with body mass among the same species. The objective of this study was to test the hypotheses that the Ma of MLCs does, in fact, vary and correlate with species body mass. The electrophoretic mobilities of MLCF isoforms from 19 species, varying in size ∼500,000-fold, were quantitated. The results confirm that the Ma of MLC1F and MLC2F vary significantly among mammals, spanning a very broad range in body mass; the MLC1F Ma varies more than that of other low-molecular-weight myofibrillar proteins; and there is a significant correlation between species body mass and MLC1F Ma. Differences in MLC1F Ma among five species can be accounted for by differences in the reported amino acid sequence, especially the length of a common polyalanine region near the NH2-terminal actin-binding site. The possibility that the differences in MLC1F sequence among mammalian species, in and adjacent to the actin-binding region, are related to differences in modulation of cross-bridge kinetics in species with diverse locomotion kinetics is discussed.
- contractile proteins
- myofibrillar proteins
- body mass
nearly forty years ago, Barany (6) reported that differences in maximal shortening velocity (Vmax) among skeletal muscles of various species are strongly correlated with actomyosin ATPase activity. Fast-type and slow-type isoforms of myosin heavy chain (MHC), as well as isoforms of most of the other proteins of sarcomeric thick and thin filaments, are expressed in skeletal, cardiac, and smooth muscle. The results of a large number of studies, reviewed by Moss et al. (27) and by Schiaffino and Reggiani (35), have since convincingly demonstrated that the subunit of the hexameric myosin molecule that is primarily responsible for this correlation is MHC. That is, there is a general consensus that MHC subunits of myosin are the primary determinants of muscle shortening velocity and power output, because different isoforms of MHC are associated with variation in actomyosin cross-bridge kinetics and, therefore, in contraction kinetics.
Each myosin molecule is composed of two MHCs and four myosin light chains (MLCs). Two isoforms of regulatory MLCs are expressed in skeletal muscle, fast-type and slow-type MLC2 (MLC2F and MLC2S). Three isoforms of essential light chains are predominantly expressed in adult skeletal muscle, fast-type and slow-type MLC1 (MLC1F and MLC1S) and MLC3, the latter being a fast-type isoform. Although the role of the regulatory MLCs in skeletal muscle remains somewhat unclear, there is considerable evidence that the essential MLCs have a modulatory, or fine-tuning, role in the regulation of cross-bridge kinetics and Vmax (9, 13, 22, 38). Vmax in single fast fibers, expressing the same isoform of MHC or coexpressing the same set of MHC isoforms, increases as the MLC3/MLC1F ratio increases. We recently reported (8) that many individual slow fibers in limb muscles of several mammalian species coexpress MLC1F and MLC1S. We also reported from the same study that the electrophoretic mobility (hence, apparent molecular mass) of denatured MLC1F appears to vary among mammalian species and that the variation is generally correlated with species body mass (8). Specifically, results from the earlier study suggested that apparent molecular mass of MLC1F and species body mass are directly correlated, because electrophoretic mobility, under denaturing conditions, and molecular mass are inversely related. The latter was based on observations from six quadruped species (mouse, rat, rabbit, domestic cat, domestic pig, and domestic dog) and one biped species (cynomolgus monkey). MLC1F in small-size mammals was observed to have a higher (faster) electrophoretic mobility compared with that of larger-size mammals. Relatively small variations in electrophoretic mobility of other low-molecular-weight myofibrillar proteins, i.e., MLC1S, MLC2F, MLC2S, MLC3, fast-type troponin I (TnIF), and fast-type troponin C (TnCF), were observed in the same study, and these variations did not appear to be as strongly correlated with species body mass. The apparent molecular mass of MLC1S also varied considerably between the species studied, but there did not appear to be a strong correlation with species body mass. The variation in apparent mass of MLC2S was much less than that of MLC1S. Of all the proteins examined, MLC1F exhibited the greatest mobility difference and correlation with body size. The seven species in the earlier study spanned a broad range in body size, from mouse (species body mass ∼20 g) to domestic pig (species body mass ∼80 kg), with rat, rabbit, domestic cat, cynomolgus monkey, and domestic dog having intermediate body sizes. Our observation of an apparent correlation between electrophoretic mobility and species body size (8) was not expected and was not subjected to statistical analysis. We subsequently extended our examination of the differences in the electrophoretic mobility of the same myofibrillar proteins to 12 additional mammalian species, including 2 of the smallest North American species, both shrews, and the world's largest terrestrial mammal, the elephant. The results reveal that the mobilities of MLC1F and, to a lesser extent, MLC2F, are, in fact, correlated with species body mass. Differences in the reported MLC1F amino acid sequence of five species are discussed as possibly underlying the observed variation in electrophoretic mobility and potentially contributing to modulation of cross-bridge kinetics and muscle contractile properties across species that span a broad range in running speeds and limb kinematics.
Protocols for the care and use of all of the animals from which samples were obtained were approved by The Ohio State University Institutional Animal Care and Use Committee. Protocols for the acquisition of human samples were approved by The Ohio State University Institutional Review Board. Human myocardial samples were obtained through The Cooperative Human Tissue Network, which is funded by the National Cancer Institute. The species included in this study, in order of increasing species body mass, are as follows, with species body mass, taxonomic order, scientific name, and the number of animals of each species sampled (n) given in parentheses: masked shrew (0.004 kg, insectivore, Sorex cinereus, n = 1), mouse (0.015 kg, rodent, Mus musculus, n = 2), short-tail shrew (0.017 kg, insectivore, Blarina brevicauda, n = 2), bat (0.017 kg, chiropteran, Eptesicus fuscus, n = 2), meadow vole (0.036 kg, rodent, Microtus pennsylvanicus, n = 1), laboratory rat [0.143 kg, rodent, Rattus rattus (Sprague Dawley), n = 2], cotton rat (0.161 kg, rodent, Sigmodon hispidus, n = 1), guinea pig (0.393 kg, rodent, Cavia porcellus, n = 1), rabbit (1.52 kg; lagomorph, Oryctolagus cuniculus, n = 1), domestic cat (3.68 kg, carnivore, Felis silvestris catus, n = 2), cynomolgus monkey (4.3 kg, primate, Macaca fascicularis, n = 2), domestic dog (13.3 kg, carnivore, Canis lupus familiaris, n = 3), baboon (16.1 kg, primate, Papiocynocephalus anubis, n = 1), sheep (51.7 kg, artiodactyl, Ovis orientalis aries, n = 2), human (70.0 kg, primate, Homo sapiens, n = 1), pig (88.0 kg, artiodactyl, Sus scrofa, n = 3), cow (453 kg, artiodactyl, Bos taurus, n = 1), horse (614 kg, perissodactyl, Equus caballus, n = 2), and Asian elephant (2,128 kg, proboscidean, Elephas maximus, n = 1). Species body masses were calculated as the average value reported by Silva and Downing (36), except for guinea pig, sheep, cow, and horse; body masses for the latter four species were calculated as the mean of the values reported in three studies in which normal adult animals were reported to have been utilized (guinea pig, Refs. 1, 19, 42; sheep, Refs. 17, 18, 37; cow, Refs. 10, 18, 31; horse, Refs. 10, 18, 30). All of the animals from which samples were obtained for this study were adults, of either sex, with normal health status.
Limb muscles that are known to contain a high proportion of fast fibers in several species (3, 4) were removed following euthanasia and placed in cold relaxing solution (32). The tibialis anterior and/or extensor digitorum longus muscle was sampled in all species. Results from bat pectoralis fibers are illustrated. MLC1F from individual fibers had the same electrophoretic mobility within all species, regardless of the muscle from which fast fibers were isolated. The superficial portion of the tibialis anterior in those species in which the muscle is relatively thick was sampled to increase the chances of finding fast fibers. Bundles (∼1 mm thick) of fibers were cut and stored in glycerinating solution [relaxing solution with 50% (vol/vol) glycerol] for a minimum of 24 h. Single fibers were isolated from the bundles by dissection in cold relaxing solution. Typically, 10–15 fibers from muscles of each species were isolated and run on gels, for an initial screening, to ensure identification of conventional fast fibers (i.e., those expressing exclusively fast-type MLC isoforms). Conventional fast fibers, identified on the initial set of gels, from each of the species were then run together on a second set of gels so that the electrophoretic mobility of low-molecular-weight proteins could be quantitated using identical gel conditions. Slow fibers, identified as those fibers that expressed exclusively slow-type MLC isoforms, and fibers coexpressing fast-type and slow-type MLC isoforms were not included in the analysis.
Gel electrophoresis and scanning.
All aspects of SDS-PAGE, including gel composition, running conditions, and staining, were identical to those we employed previously (8) to examine the low-molecular-mass protein composition of single muscle fibers. Three fast fibers from each of the species were run on three separate gels, which were scanned. Thin muscle strips were utilized for the short-tail shrew because the isolated single fibers were too fragile, presumably due to the tissue having been frozen before being stored in glycerinating solution. To quantitate the relative electrophoretic mobility of fast-type MLC subunits and of TnIF and TnCF in each species, we scanned the region of each gel loaded with a single fast fiber from each species using a gel densitometer (Hoefer model GS 300 densitometer, with GS365W software, version 3.01), from just above actin to just below MLC3. The migration distance from actin, for each protein of interest, was measured and is expressed in arbitrary scan units. Identical scanner settings (gain, sensitivity, resolution, and the zero, or baseline, limit) were utilized for all scans to ensure consistency in the determination of migration distance for all proteins of interest from all of the species. The electrophoretic mobilities that are reported and that were utilized to test for correlations with species body mass are the average values obtained from three gels.
Identification of myofibrillar proteins by SDS-PAGE.
To facilitate identification of MLC isoforms on gels, we extracted myosin from several fibers of each species and used the extracted proteins as visual standards to identify specific protein bands on subsequent gels (8). The objective of this portion of the study was to extract and identify fast MLC isoforms, TnCF, and TnIF, all of which migrate to the low-molecular-mass region of the gels used in this study. A single skinned fiber was isolated in a petri dish containing relaxing solution, using a dissecting microscope at ×10–40 magnification. The dimensions of the fiber were estimated by visual comparison to black nylon monofilaments of known thickness, to estimate diameter, and a submerged ruler, to measure length. The fiber volume was calculated, assuming a cylindrical shape. The fiber was then cut in half, lengthwise. One half, which contained all of the myofibrillar proteins in that fiber, was transferred to a microcentrifuge tube, and gel sample buffer was added (2 μl per nl of fiber volume). The other half was soaked in 5 μl of a myosin extraction buffer (in a microcentrifuge tube), consisting of relaxing solution with an additional 330 mM KCl (471 mM KCl total), a modification of the myosin extraction buffer of Margossian and Lowey (23). The fiber segment, containing the nonextracted proteins, was removed and transferred to another empty microcentrifuge tube to which sample buffer was added (2 μl per nl of fiber volume). The 5 μl of extraction buffer, containing the proteins extracted from the fiber, were mixed with 5 μl of gel sample buffer. All three tubes were kept at room temperature for 30 min, heated at 65°C for 2 min, immediately chilled on ice for 5 min, and stored at −40°C until run on gels. Aliquots (3 μl) from each tube for a given fiber were loaded in adjacent lanes on gels. This process, based on naive selection of fibers with respect to fiber type, was repeated until a conventional fast-type fiber (i.e., a fiber expressing exclusively fast-type isoforms of MLC1 and MLC2) from each species was identified. Thin strips of fast-twitch muscle were used for the extraction of myosin for both shrew species, because of the extremely small size of individual fibers and the fragility of the short-tail shrew samples (explained above). Peters et al. (29) and Savolainen and Vornanen (33, 34) did not find any evidence of slow-type MHC or MLC or of histochemically identified slow fibers in skeletal muscles of shrews. The same three fractions were prepared from a thin (∼200-μm diameter) ventricular strip from each species to identify ventricular MLC isoforms, which are identical to MLC1S (7, 20, 28) and MLC2S (11, 12, 21). The distinct migrations of ventricular MLC1 and of MLC1F verified the identification of MLC1F in fast fibers.
The gels shown in Fig. 1 illustrate the basis for the identification of actin, tropomyosin, TnC, TnI, and MLC isoforms in fast fibers in three of the species included in this study. Myosin was extracted from single skinned fibers and run on gels to identify the MLC isoform bands (see methods). The bands identified as actin, tropomyosin, TnC, and TnI were retained in the nonextracted fraction and identified on the basis of known molecular mass and stoichiometry (44, 45). Myofibrillar proteins were identified with the same extraction protocol for all species in this study. Extractions of myosin subunits from fast fibers of seven other species were illustrated in our previous study (8).
A gel loaded with a conventional fast fiber, expressing exclusively fast-type MLC isoforms, from each of the species is shown in Fig. 2. The three gels loaded with different single fast fibers from each species (strips for short-tail shrew, see methods) were scanned to quantitate the migration distances of the proteins of interest. The scan results are shown in Fig. 3. The fibers were loaded in order of decreasing electrophoretic mobility of MLC1F, from left to right, based on observations from previously run gels in which the order of MLC1F migration among all species was determined. Several observations can be made regarding the variation in MLC1F electrophoretic mobility across species. MLC1F stands out as the protein with the greatest variation in electrophoretic mobility (179 scan units) compared with MLC2F (121 scan units), MLC3 (134 scan units), TnCF (48 scan units), and TnIF (130 scan units). Generally, the mobility of MLC1F is greater in species with small body size, as suggested by our previous results (8), which were based on observations from only seven species. MLC1F in the insectivore species that were studied (shrews) and in the rodent species studied (mouse, rat, cotton rat, vole, and guinea pig) migrates faster than MLC1F in most of the large-size mammals that were studied, especially the members of the Artiodactyla (pig, cow, and sheep), Perissodactyla (horse), and Proboscidea (elephant) orders. MLC1F in the rodent species of smallest body studied (mouse) migrates faster than does MLC1F in the rodent species of largest body size studied (guinea pig). There is a significant inverse linear correlation (P < 0. 0001) between electrophoretic mobility and log of species body mass (Fig. 4). Bat MLC1F mobility was treated separately (i.e., was not included in the correlation), because the correlation between MLC1F mobility and log of body mass is very strong among small-size mammals and bat MLC1F deviates strikingly from this correlation. The slope of the relationship between MLC1F mobility and species body mass is −28.79 scan units/log of body mass (g). There is also a positive linear correlation (P < 0.0001) between the electrophoretic mobility of MLC2F and log of species body mass, but the slope of this relationship is much lower, 5.84 scan units/log of body mass, again excluding bat MLC2F because it deviates markedly from the relationship, especially among small-size species. There is not a significant correlation between MLC3 mobility or TnCF mobility and log of species body mass (P > 0.30 for both). A second-order polynomial fit between TnIF mobility and log of body mass is also significant (P < 0.02).
The MLC1F sequences (accession numbers in parentheses) in mouse (NP/20%067260), rat (P02600), rabbit (CAA37974), human (P05976), and dog (XP/20%536054) have been reported (Fig. 5). The molecular masses of MLC1F in each of these species, calculated from the reported sequences, are 20,575 kDa in mouse, 20,661 kDa in rat, 20,930 kDa in rabbit, 21,126 kDa in human, and 21,061 kDa in dog, in order of increasing electrophoretic mobility. The correlation coefficient for the inverse linear regression between the electrophoretic migration distance from actin and calculated molecular mass is 0.951. Regression of the electrophoretic mobility of MLC1F against the number of amino acid residues for mouse (188), rat (189), rabbit (192), human (194), and dog (194) indicates that 94.4% of the variation in mobility can be explained by the difference in the number of residues in MLC1F.
As mentioned above, migration of bat MLC1F and MLC2F compared with other species is noteworthy for two reasons. First, MLC1F in the species of bat studied (Eptesicus fuscus, big brown bat) migrated very similarly to rabbit MLC1F and human MLC1F, species with much greater body sizes. Second, bat MLC2F migrates more slowly than any other species. The overall variation in MLC2F migration among the 19 species studied (121 scan units) is much lower (62 scan units) if bat MLC2F is omitted from consideration. Therefore, the apparent molecular masses of bat MLC1F and MLC2F are markedly greater than predicted from the significant correlations between mobility and body mass for the other species.
The results of this study indicate that there is considerable variation in the apparent molecular mass of MLC1F among mammalian species, that this variation correlates with species body size, and that the other fast-type MLCs, as well as TnIF and TnCF, vary less with respect to apparent molecular mass. Therefore, MLC1F is unique among this set of myofibrillar proteins. Given the evidence that MLC isoforms differentially modulate cross-bridge kinetics, it is possible that the variability in apparent mass of MLC1F reflects structural differences that underlie variations in modulation of cross-bridge kinetics among different species. This may be one mechanism of fine-tuning contractile properties among animals with a broad range of kinetics of their predominant mode of locomotion: scurrying among small mammals, in contrast to slow roaming movements among large-size grazers.
The preliminary finding of a relationship between MLC mobility and body mass, based on results obtained from only seven species, was presented previously (8). The current study greatly expands the number of species and the species body size (mass) range covered in the previous study and provides quantitative data for the electrophoretic mobility of MLC1F, MLC2F, MLC3, TnIF, and TnCF among 19 mammalian species. This allowed us to test whether there are statistically significant relationships between the mobility of fast MLC isoforms, TnI, and TnC and species body mass. It should be noted that there was no a priori reason to expect what was initially observed (reported in Ref. 8) and formally pursued in the present study, that is, that there is a striking relationship between MLC1F and MLC2F mobilities and species body mass.
Several differences in the reported sequence of MLC1F between species could potentially explain the variation in MLC1F electrophoretic mobility. The most straightforward explanation would be a difference in the molecular mass of MLC1F. The high correlation between calculated molecular mass and electrophoretic mobility among the five species for which the entire MLC1F sequence has been reported suggests that the observed difference in MLC1F migration, at least across these species, is due to differences in the size of the molecule and not to differences in possible posttranslational modifications. The NH2-terminal residue of MLC1F in all studied species is trimethylated (16), as one instance of posttranslational modification. However, this modification is shared by MLC1F in all studied vertebrates and, therefore, cannot account for the variability in MLC1F electrophoretic mobility. Phosphorylation of ventricular MLC1 (i.e., slow-type MLC1) in vivo was reported by Arrell et al. (5), and it is well known that MLC2 in the myocardium can be phosphorylated, resulting in significant alteration of contractile properties (26).
All of the reported sequences of mammalian MLC1F are highly conserved, especially in the COOH-terminal two-thirds of the molecule. The NH2-terminal one-third, on the other hand, is more variable. Virtually all of the vertebrate MLC1 isoforms have a region at the very end of the NH2 terminus that is an actin-binding domain (41). The first four amino acids at this site are alanine-proline-lysine-lysine (some cardiac MLC1 sequences start with proline, instead of alanine). MLC1F differs from the other essential MLC in fast fibers (MLC3) by having an ∼40-amino acid NH2-terminal extension. Therefore, MLC3 does not have the actin-binding characteristics of MLC1F. There are two other characteristics of the 40-amino acid NH2-terminal extension of MLC1F that are shared in all of the reported sequences: a series of alanine-proline (A-P) repeats and a polyalanine region, with the latter located between the actin-binding domain and the series of A-P repeats. The potential significance of the A-P repeats (6 repeats in mouse and rat MLC1F, 7 repeats in rabbit, human, and dog MLC1F) is that this imparts a rigid structure to this region of the molecule, at the end of which is the actin-binding region. As discussed by Timson (40), this positions the actin-binding site at the tip of a rigid stalk. If this stalk is in fact longer in those species with a greater number of A-P repeats, then this could result in a greater probability of MLC1F binding to actin. Interactions between MLC1 and actin modulate cross-bridge kinetics in vitro and, apparently, in vivo (40). Therefore, it is possible that in those species in which MLC1F has a greater molecular mass and, hence, lower electrophoretic mobility, due in part to a greater number of A-P repeats, there is greater modulation of cross-bridge kinetics by MLC1F. Results from in vitro experiments indicate that there is a direct inverse correlation between the affinity of MLC1F for actin and the rate constant for actomyosin ATPase activity (41). This could translate into greater modulation of cross-bridge kinetics by MLC1F in those species with a greater number of A-P repeats, specifically slowing down cross-bridge kinetics in large animals and thereby accommodating slower limb kinematics during locomotion (14). Because muscle contractions in smaller mammals tend to be faster to accommodate faster limb kinematics and have greater energy costs (15), it is possible that the structural differences in MLC1F among mammals of varying size underlie differences in muscle economy, as well. Miller et al. (25) recently provided evidence that the NH2-terminal region of mouse cardiac MLC1 (i.e., slow-type MLC1) is important for tension production by demonstrating decreased Ca2+-activated isometric tension in skinned ventricular trabeculae from mice expressing MLC1S with an NH2-terminal deletion. This suggests that differences in the NH2-terminal region of MLC1 between species could be associated with differences in tension production.
The other structural difference among the reported MLC1F sequences is the length of the common polyalanine region in the NH2-terminal region. Alanine is the amino acid with the greatest propensity to form an α-helix (24). Those species, for which the MLC1F sequence has been reported, with slower migrating MLC1F have a longer polyalanine region. The number of alanines in tandem in this region is three in mouse, four in rat, five in rabbit, six in human, and seven in dog. Those species with a greater number of alanines in tandem might have a greater helical content in the NH2-terminal region. Williams et al. (43) provided evidence for lysine residues as providing stabilization for helices with high alanine content. All of the reported sequences of MLC1F include two lysine residues at positions 7 and 8, followed by proline. The polyalanine region begins at position 10 for four of the species (a valine is interspersed between lysine-9 and alanine-10 in human MLC1F). Therefore, it is possible that the lysine-7 and lysine-8 residues provide stability of the helicity associated with the polyalanine region that is in close proximity. Given the position of the polyalanine region with its putative helical content, relative to the actin-binding region at the very NH2 terminus and the A-P repeat region, one can envision a rigid stalk of variable length, followed by an α-helix of variable length, terminated by the actin-binding site as one approaches the NH2 terminus of MLC1F. A variable-length α-helix could provide a variable amount of flexibility in this region of the molecule, which might contribute to differences in modulation of cross-bridge kinetics by MLC1F among species. Furthermore, several alanine-rich proteins, such as myristoylated alanine-rich C kinase substrates, display anomalous migration (i.e., slower than expected from their molecular masses) on denaturing gels (2, 39). It is possible that the increased number of amino acids and the fact that the additional amino acids are exclusively or primarily alanine synergistically decrease the electrophoretic mobility of MLC1F among different mammalian species.
The identification of the gel bands that contain MLC1F and TnIF was determined from the extraction and electrophoresis of fast-type myosin and from the retention of TnIF in the nonextracted fraction for each species. It is clear, therefore, that in some species MLC1F and TnIF comigrate, at least with the gel protocol employed in this study. The importance of this observation is that for studies in which the amounts of MLC1F and/or TnIF are to be quantitated, possible comigration of these two proteins must be considered and avoided. Furthermore, identification of low-molecular-weight proteins on SDS gels for a given species, based on the migration order of the same proteins in a different species, could lead to errors, given that the order of these proteins is not constant across all species. For example, TnIF migrates slower than MLC1F in four of the rodent species, but this order is reversed among the species with larger body mass, beginning with cat.
In conclusion, the apparent molecular mass of MLC1F varies among mammalian species and does so to a greater extent than does the mass of the other fast-type MLCs, TnIF, and TnCF. The variation in MLC1F apparent mass is strongly correlated with species body mass and with differences in limb kinematics during locomotion. Modulation of cross-bridge kinetics by MLC1F may provide a mechanism to fine-tune muscle contraction kinetics at rates that are appropriate for varying limb kinematics among diverse mammalian species. Whether the variation in MLC1F apparent mass reflects a structural change that underlies differences in cross-bridge kinetics and/or the economy of muscle contraction remains to be determined.
This project was supported by National Science Foundation Grant IOB 0133613.
We are grateful to Dr. Michael Barrie (Director of Animal Health, Columbus Zoo and Aquarium, Columbus, OH) and Dr. Denise Schwahn, both of whom provided very valuable assistance in procuring samples of elephant muscle. We are also very grateful to Joy Pietkiewicz and Dr. John Harder for providing shrews and to Dr. W. M. Masters for providing the bats for this study. We also are grateful to Drs. Jack Rall and Jonathan Davis for providing valuable suggestions from an early draft of the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society