Presence in Human Skeletal Muscle of an AMP Deaminase-associated Protein That Reacts with an Antibody to Human Plasma Histidine-Proline-rich Glycoprotein

Abstract

Histidine-proline-rich glycoprotein (HPRG) is a protein that is synthesized by parenchimal liver cells. The protein has been implicated in a number of plasma-specific processes, including blood coagulation and fibrinolysis. We have recently reported the association of an HPRG-like protein with rabbit skeletal muscle AMP deaminase (AMPD). The results of the immunological analysis reported here demonstrate that an antibody against human plasma HPRG reacts with an AMPD preparation from human skeletal muscle. To probe the localization of the putative HPRG-like protein in human skeletal muscle, serial sections from frozen biopsy specimens were processed for immunohistochemical and histoenzymatic stains. A selective binding of the anti-HPRG antibody to Type IIB muscle fibers was detected, suggesting a preferential association of the novel protein to the AMPD isoenzyme contained in the fast-twitch glycolytic fibers.

Keywords

histidine-proline-rich glycoprotein AMP deaminase human skeletal muscle Western blot analysis immunohistochemical analysis histoenzymatic analysis

HISTIDINE-PROLINE-RICH GLYCOPROTEIN (HPRG) is a protein present at a relatively high concentration in the plasma of vertebrates. Its specific function remains unclear, although it has been implicated in several phenomena, including blood coagulation and fibrinolysis (Peterson et al. 1987). In a recent article we reported the isolation from purified rabbit skeletal muscle AMP deaminase (AMPD) of an approximately 75 kD novel peptide with an amino acid composition significantly different from that derived from the available AMPD cDNAs (Ranieri-Raggi et al. 1997). N-terminal sequence analysis of the fragments liberated by limited proteolysis revealed a striking similarity of this protein to rabbit plasma HPRG (Borza et al. 1996) although, in comparison with mature HPRG, the AMP deaminase-associated variant probably contains a unique N-terminal extension. Among the 60 amino acids sequenced up to now in the novel HPRG isoform, four substitutions were found with respect to the published rabbit HPRG sequence, all of them localized in the 472-477 region that also differs in five amino acid residues compared with the homologous region of the human protein (residues 461-466) (Koide et al. 1986). This divergence enabled us to raise a rabbit antibody against a synthetic peptide equivalent to residues 462-471 of human plasma HPRG. The similarity between the 461-466 region of human plasma HPRG (S-F-P-L-P-H) and the corresponding sequence of the HPRG-like molecule isolated from rabbit skeletal muscle (S-F-S-L-R-H) prompted us to utilize the antibody to probe the immunohistochemical localization of the putative HPRG-like protein in human skeletal muscle. The experimental results show that the anti-HPRG antibody reacts with an AMPD preparation from human skeletal muscle. Moreover, a clear positive reaction was detected at the level of Type IIB fibers, giving evidence of the presence of an HPRG-like peptide in human skeletal muscle. These observations suggest a correlation of this protein as well as AMP deaminase to the activity of muscle with a more developed anaerobic metabolism.

Materials and Methods

Collection of Samples

Eleven biopsy specimens from deltoid (n = 6), quadriceps femoris (n = 3), and tibialis anterior (n = 2) were selected among frozen muscle samples collected at the Division of Pathology, Department of Oncology, University of Pisa. Eight-serial sections were cut in a cryostat at −20C. Biopsy of human quadriceps femoris used for AMPD preparation was performed on a patient who underwent surgery for diagnostic purposes.

On examination, no histological or histochemical abnormalities were present in all the biopsies used. All the procedures followed were approved by the University Hospital Ethics Committee, Pisa, Italy.

Preparation of AMPD from Human Skeletal Muscle

AMPD was prepared from 2 g quadriceps femoris by the procedure previously described for the isolation of the enzyme from rabbit red muscle (Raggi and Ranieri-Raggi 1987). This method separates the AMPD isoenzymes by two successive elutions of a cellulose phosphate column with, respectively, 0.6 M KCl and 0.9 M KCl in 5 mM succinate/50 mM Tris (pH 6.5). Because the human enzyme fractions obtained by the two-step chromatography accounted for 60% and 5% of the original activity, respectively, only the first could be used for SDS-PAGE and Western blot analyses.

Synthesis of the HPRG Peptide and Preparation of a Polyclonal Antibody to Human Plasma HPRG

A synthetic peptide equivalent to residues 462-471 of human plasma HPRG was synthesized at the Krebs Synthesis & Sequencing Facility at the University of Sheffield, UK. Synthesis was performed on a Milligen 9050 Peptide Synthesizer using standard F-MOC active ester chemistry. The cleaved product was purified by HPLC using a Vydac C18 HPLC column (250 × 22 cm) (Phenomenex; Torrance, CA) and was validated by sequence determination and by mass spectrometry. Cysteine was incorporated at the C-terminus to allow conjugation of the peptide to porcine thyroglobulin using MBS (m-maleimidobenzoyl-N-hydroxysuccinimide ester) (Pierce; Rockford, IL).

A polyclonal antibody to human plasma HPRG was raised in a female New Zealand White rabbit by using the conjugated synthetic peptide. The antigen (200 μg) was emulsified in Freund's complete adjuvant (Sigma; St Louis, MO) and injected into multiple intradermal sites distributed over the back of the shaved rabbit. Six weeks later, a second injection of the antigen (50 μg) emulsified in Freund's incomplete adjuvant (Sigma) was carried out as described above, and after 4 weeks the animal was boosted with 50 μg of soluble antigen injected IV. Three months after the first injection, the animal was sacrified and exsanguinated and the serum was prepared, aliquotted, and stored at −20C. The IgG isolation from the rabbit antiserum was performed as described by McKinney and Parkinson (1987) and cross-reactivity to the antigen and to human plasma HPRG was tested by dot-blot and Western blot analyses.

Dot-blot, SDS-PAGE, and Western Blot Analyses

One μg, 100 ng, and 10 ng of the HPRG synthetic peptide and 5 μl of human plasma were spotted on a nitrocellulose membrane (Sigma) and air-dried. After a blocking step with 3% (w/v) BSA-0.5% (v/v) Tween-20 in PBS for 1 hr, the blot was washed once with PBS and then incubated with the rabbit anti-HPRG polyclonal antibody [1:10000 (v/v) in PBS] for 1 hr. A washing step with PBS was then followed by the incubation with a horseradish peroxidase-labeled goat anti-rabbit IgG (Sigma) for 1 hr [1:10000 (v/v) in PBS containing 0.01% (w/v) SDS]. The blot was extensively washed with 0.5% Tween-20 in PBS and the peroxidase reaction was developed using 3,3′-diaminobenzidine tetrahydrochloride (Sigma) as substrate. All the incubations were performed at room temperature (RT).

Electrophoresis in the presence of 0.1% (w/v) SDS was carried out on a 10% (w/v) polyacrylamide slab gel in 0.1 M Tris/0.1 M Bicine (pH 8.3). Prestained molecular weight standards (Bio-Rad; Richmond, CA) were used to calibrate the gel. Samples were run in duplicate for Coomassie Brilliant Blue staining and Western blot analysis.

After the SDS-PAGE, samples were electrotransferred to a nitrocellulose membrane. The blot was blocked with 3% BSA in PBS for 2 hr and then treated with the rabbit anti-HPRG polyclonal antibody diluted 1:10,000 (v/v) with PBS (overnight incubation at 4C). After washing with PBS, the blot was incubated with [¹²⁵I]-protein A (Amersham; Sunnyvale, CA) for 2.5 hr at RT and then extensively washed with 0.5% Tween-20 in PBS and air-dried before being processed for autoradiography.

Histochemistry

Serial sections from muscle biopsies were processed for the following histological and histochemical stains: hematoxylin-eosin, NADH tetrazolium reductase (NADH-TR), routine ATPase (pH 9.4), ATPase preincubated at pH 4.3 and 4.6 (Dubowitz and Brooke 1973) and AMPD activity (Fishbein et al. 1980).

Immunohistochemistry

Slides were fixed in cold acetone/methanol (1:1, v/v) for 10 min at −20C and then washed with PBS for 10 min at RT. After a 30-min blocking step for endogenous peroxidase activity with 3% (w/w) H₂O₂ in methanol at RT, slides were washed with PBS and incubated with the rabbit anti-HPRG primary antibody [1:100 (v/v) in PBS containing 0.5% BSA and 0.2% (w/v) gelatin] for 1 hr at 37C. As a second antibody, a horseradish peroxidase-labeled goat anti-rabbit IgG was used, diluted 1:250 (v/v) in PBS containing 0.5% BSA and 0.2% gelatin (1 hr at RT). The peroxidase reaction was developed using 3,3′-diaminobenzidine tetrahydrochloride as substrate. Controls included omission of primary antibody and substitution of immune serum by preimmune or nonimmune serum.

Results

Antiserum Reactivity with Human Plasma HPRG

The reactivity of the IgG fraction isolated from rabbit antiserum was first checked by dot-blot analysis. A clear positive reaction was observed against human plasma and the HPRG synthetic peptide utilized as antigen for the immunization. No positive reaction was obtained when preimmune or nonimmune serum was used (not shown). Moreover, when human plasma was tested by Western blot analysis, a clear immunoreactive band at a molecular mass of approximately 50 kD was found. Reactivity of small amounts of more slowly migrating material (55-75 kD) was also observed (Figure 1, Lane 4).

SDS-PAGE and Western Blot Analysis of Human Skeletal Muscle AMPD

In a recent article we demonstrated the association of a HPRG-like protein to rabbit skeletal muscle AMPD (Ranieri-Raggi et al. 1997). To ascertain whether the presence of the HPRG-like molecule is also a feature of human enzyme preparation, the immunoblot analysis was performed with AMPD isolated from fresh quadriceps femoris as described in Materials and Methods. SDS-PAGE of this enzyme provided results similar to that described for rabbit skeletal muscle AMPD (Ranieri-Raggi and Raggi 1980), giving rise to a major protein band of an apparent MW of approximately 86 kD and a minor band of approximately 80 kD (Figure 1, Lane 2). Immunoblot analysis, confirming the data obtained by Coomassie Brilliant Blue staining, revealed the presence of an immunoreactive band at a molecular mass of approximately 86 kD and a minor band at approximately 80 kD (Figure 1, Lane 3). No other polypeptide was detected in the enzyme preparation.

Figure 1

Western blot analysis of human plasma and human skeletal muscle AMPD. Plasma and enzyme samples were run in duplicate on SDS-PAGE and processed as described in Materials and Methods. Lane 1, prestained molecular weight standards. Lane 2, Coomassie Brilliant Blue-stained gel of human skeletal muscle AMPD. Lane 3, autoradiograph of the immunoblot of human skeletal muscle AMPD. Lane 4, autoradiograph of the immunoblot of human plasma.

Histochemistry and Immunohistochemistry

Evaluation of muscle biopsies with NADH-TR histochemical staining enabled us to classify fibers as Type I (slow-twitch) and Type II (fast-twitch). Type II fibers varied from 30 to 65% among samples. Type I fibers appeared darkly stained, whereas Type II fibers were characterized by a pale gray cytoplasm with a diffuse blue granularity that, on the basis of its density, enabled us to distinguish two subtypes, light and intermediate, corresponding to IIB (fast-twitch glycolytic) and IIA (fast-twitch oxidative-glycolytic) fibers, respectively (Figure 2A, Case A). The fiber typing obtained with NADH-TR was confirmed by the ATPase staining after acid or alkaline preincubation (not shown).

Figure 2B shows a serial cross-section of Case A muscle biopsy immunostained for HPRG as described in Materials and Methods. All fibers showed a brown subsarcolemmal staining, but two fiber-types could be distinguished according to the sarcoplasmic shading (light and dark). The light fibers showed a granular positivity and a surrounding pale sarcoplasm, whereas dark fibers exhibited a similar granular pattern on a darker sarcoplasmic background. Comparison of this specimen with the serial section stained for NADH-TR (Figure 2A) showed a correspondence between the darker immunostained fibers and the IIB fiber type, whereas the lightly immunostained fibers corresponded to the I plus IIA fiber types.

The above observations indicate the presence of an HPRG-like protein in the glycolytic fibers, which are well known to contain the highest level of AMPD among muscle fibers (Raggi et al. 1969). To establish the enzyme localization in the different fiber types of our specimens, biopsies were analyzed using the AMPD histoenzymatic stain. In agreement with the results described by Fishbein et al. (1980), Type I fibers appeared lightly stained, showing granular blue densities on a clear gray cytoplasmatic background, whereas Type II fibers showed a reticular blue staining on a diffusely pink cytoplasm (Figure 2C). In the specimen of Figure 2C and in a further five cases from a total of 11, the histochemical assay revealed the presence of a third fiber type, characterized by granular blue densities on a blue-violet cytoplasm. Comparison of the serial sections gave evidence that these fibers are not immunoreactive against the anti-HPRG antibody (Figures 2B and 2C). Their correspondence to the NADH-TR intermediate fibers permitted us to ascribe them to the IIA type.

Figure 2

Serial sections from two normal deltoid muscle biopsies. Case A, specimens stained for NADH-TR (A) and AMPD (C) and immunostained for HPRG (B). Case B, specimens immunostained for HPRG (D) and stained for AMPD (E). Bars = 100 μm.

With the HPRG immunostaining, in one of the 11 cases examined (Case B, Figure 2D) three different fiber types were detectable (light, intermediate, and dark). The light and intermediate fibers, with a similar cytoplasmic reticular network, differed in intensity of the sarcoplasmic staining. The dark fibers, which exhibited a cytoplasmic shading that was similar to that of the intermediate fibers, could be identified by a pronounced network, related to a coarser stain thickening. When a serial cross-section of the same case was stained for AMPD (Figure 2E), only two fiber types were detectable, corresponding to Types I and II described by Fishbein et al. (1980). Comparison of the serial sections of Figures 2D and 2E revealed a clear correspondence of the dark plus intermediate HPRG-immunostained fibers with the Type IIB fibers. The fibers that stained weakly with the HPRG antibody could be divided into two groups on the basis of their response to the AMPD staining. The first group (Type IIA) showed a positive reaction for AMPD indistinguishable from that of the Type IIB fibers, whereas the other gave the weak response typical of Type I fibers.

Discussion

HPRG is a glycoprotein member of the cystatin superfamily that is synthesized by parenchymal liver cells (Koide and Odani 1987; Smith et al. 1988). It was originally isolated from human plasma (Heimburger et al. 1972) but its presence has been demonstrated also in platelets and milk (Leung et al. 1984; Hutchens et al. 1992). Its specific function remains unclear, although several biological roles have been suggested. Its interaction with heparin, fibrinogen and fibrin, plasminogen, and activated platelets has suggested that it could be a modulator of coagulation and fibrinolysis (Leung 1986; Peterson et al. 1987; Lerch et al. 1988). Moreover, HPRG has been shown to interact with components of the immune system and to inhibit the formation of insoluble immune complexes (Gorgani et al. 1997). This study provides the first evidence for a skeletal muscle protein that reacts with a plasma HPRG antibody, indicating that HPRG could play a new role in addition to those described for the plasma protein.

Immunoblot analysis demonstrated that the anti-HPRG antibody reacts with a plasma component with an apparent molecular mass of 50 kD. This value is in agreement with the molecular weight calculated for human plasma HPRG free of carbohydrate on the basis of its amino acid composition (Heimburger et al. 1972; Koide et al. 1986). The minor components that give rise to unresolved faint bands of higher molecular weight (55-75 kD) in our plasma immunoblot are probably related to the heterogeneity of HPRG forms with respect to carbohydrate attachment. Molecular weight values ranging from 60 to 81 kD have been reported for human plasma HPRG on the basis of its migration in SDS-PAGE (Rylett et al. 1981; Lijnen et al. 1983).

When the Western blot analysis was performed using freshly prepared human skeletal muscle AMPD, a clear immunoreactivity was detected, thereby extending to the human enzyme the association with a HPRG-like molecule, first observed for rabbit AMPD (Ranieri-Raggi et al. 1997). The anti-HPRG antibody recognizes in the human AMPD preparation a major 86-kD protein band and a minor 80-kD protein band similar to the sizes of the components revealed by Coomassie Brilliant Blue staining. This finding is in agreement with our previous observation that analysis by SDS-PAGE does not discriminate between the catalytic subunit and the HPRG-like component of rabbit skeletal muscle AMPD, both giving rise to a main band corresponding to their native status and a minor one corresponding to the product of a proteolytic process occurring during the enzyme purification (Ranieri-Raggi et al. 1997). The observation that the HPRG-like component of AMPD migrates in SDS-PAGE at a higher apparent molecular mass than that shown by the plasma protein could be explained by the presence of an N-terminal extension similar to that identified in the rabbit AMPD-associated HPRG variant (Ranieri-Raggi et al. 1997).

By combining the data of our immunohistochemical and histoenzymatic analyses, a clear indication is obtained of a preferential binding of the anti-HPRG antibody to Type IIB muscle fibers. This observation deserves consideration in the light of our recent finding (Ranieri-Raggi et al. 1997) of the association of a HPRG-like protein with rabbit skeletal muscle AMPD that can be assumed to represent white-muscle AMPD (Raggi and Ranieri-Raggi 1987).

As reported by Fishbein et al. (1980), the AMPD histoenzymatic stain discriminates between Type I and II fibers on the basis of color differences, the cytoplasm being pale gray or pink, respectively. Our results show that the response of Type II fibers could be further differentiated, because in six specimens from a total of 11, some fibers with a blue-violet cytoplasm were present in addition to the pink ones. The two subtypes corresponded to IIA and IIB fiber type, respectively, on the basis of the histochemical staining (ATPase and NADH-TR). This differentiation should be taken into account when the relationship between the AMPD histoenzymatic stain and the HPRG immunostaining is examined. Comparison of serial sections after the AMPD histoenzymatic stain and the HPRG immunostaining shows that the positive reaction obtained with the anti-HPRG antibody in Type IIB fibers is always associated with high levels of AMPD activity, whereas Type IIA and Type I fibers always gave a weak response to the antibody even in the specimens where Type IIA fibers showed an AMPD histoenzymatic staining similar to that of IIB fibers (Figures 2D and 2E). Together, these observations suggest that the different responses to the AMPD staining among the IIA fibers of different specimens can be ascribed to different levels of an AMPD isoform not related to the HPRG-like protein. At least two AMPD isoenzymes differing in regulatory properties and localization have been identified in various skeletal muscle fibers (Raggi et al. 1975; Van Kuppevelt et al. 1994) and a 4-10-fold higher level of AMPD in white muscle than in red muscle has been documented in different species (Raggi et al. 1969). By chromatography on cellulose phosphate of extracts of rabbit red muscles, two peaks of AMPD activity were obtained (forms A and B), the second having the same chromatographic properties as the single form present in white muscles (Raggi et al. 1975). Form B is inhibited by ATP, as is observed with white muscle AMPD, whereas form A is not affected by the nucleotide (Raggi and Ranieri-Raggi 1987). The chromatographic patterns of AMPD isoforms obtained with human muscle extracts do not fit with those observed in rabbit muscle. With human pectoralis maior, two peaks of AMPD activity were obtained, both in the elution region of isoenzyme B (Raggi et al. 1975), whereas the preparation from quadriceps femoris described here gave a major peak of activity in the elution region of isoform A. However, an examination of the kinetic properties showed that the enzyme was inhibited by ATP (unpublished results), indicating that the preparation used for the Western blot analysis reported here represents the white muscle isoform. This strengthens the hypothesis of a preferential association of the HPRG-like protein to the AMPD isoenzyme contained in fast-twitch glycolytic fibers.

Further studies are necessary to establish whether the novel protein participates in the structure of skeletal muscle AMPD or simply binds to the holoenzyme. However, attempts to define the physiological function of muscle HPRG should take into account the possible role of white muscle AMPD in regulating the relative concentrations of adenine nucleotides during sustained contractile activity (Ronca-Testoni et al. 1970; Ranieri-Raggi and Raggi 1990). In this context also, the observation that one of the 11 specimens examined showed two different responses to the HPRG immunostaining among Type IIB fibers deserves consideration because it suggests that the level of this protein could be related to different metabolic conditions of the fast-twitch glycolytic fibers.

Footnotes

Acknowledgements

Supported by a grant from the Italian MURST.

We thank Mr Stefano Mazzoni for assistance in rabbit immunization and Mr Piero Bertelli for skilled technical assistance.

References

Borza

Tatum

Morgan

(1996) Domain structure and conformation of histidine-proline-rich glycoprotein. Biochemistry 35:1925–1934.

Dubowitz

Brooke

(1973) Muscle Biopsy: A Modern Approach. London, Philadelphia, Toronto, WB Saunders.

Fishbein

Griffin

Armbrustmacher

(1980) Stain for skeletal muscle adenylate deaminase. Arch Pathol Lab Med 104:462–466.

Gorgani

Parish

Easterbrook Smith

Altin

(1997) Histidine-rich glycoprotein binds to human IgG and C1q and inhibits the formation of insoluble immune complexes. Biochemistry 36:6653–6662.

Heimburger

Haupt

Kranz

Baudner

(1972) Humanserumproteine mit hoher Affinitaet zu Carboxymethylcellulose, II. Physikalisch-chemische und immunologische Charakterisierung eines histidinreichen 3,8S-α₂-Glykoproteins (CM-protein I). Hoppe Seylers Z Physiol Chem 353:1133–1140.

Hutchens

Yip

Morgan

(1992) Identification of histidine-rich glycoprotein in human colostrum and milk. Pediatr Res 31:239–246.

Koide

Foster

Yoshitake

Davie

(1986) Amino acid sequence of human histidine-rich glycoprotein derived from the nucleotide sequence of its cDNA. Biochemistry 25:2220–2225.

Koide

Odani

(1987) Histidine-rich glycoprotein is evolutionarily related to the cystatin superfamily. Presence of two cystatin domains in the N-terminal region. FEBS Lett 216:17–21.

Lerch

Nydegger

Kuyas

Haeberli

(1988) Histidine-rich glycoprotein binding to activated human platelets. Br J Haematol 70:219–224.

10.

Leung

LLK

(1986) Interaction of histidine-rich glycoprotein with fibrinogen and fibrin. J Clin Invest 77:1305–1311.

11.

Leung

LLK

Nachman

Harpel

(1984) Complex formation of platelet thrombospondin with histidine-rich glycoprotein. J Clin Invest 73:5–12.

12.

Lijnen

Rylett

Collen

(1983) Physicochemical, immunochemical and functional comparison of human histidine-rich glycoprotein and autorosette inhibition factor. Biochim Biophys Acta 742:109–115.

13.

McKinney

Parkinson

(1987) A simple, non-chromatographic procedure to purify immunoglobulins from serum and ascites fluid. J Immunol Methods 96:271–278.

14.

Peterson

Morgan

Blackburn

(1987) Histidine-rich glycoprotein modulation of the anticoagulant activity of heparin. Evidence for a mechanism involving competition with both antithrombin and thrombin for heparin binding. J Biol Chem 262: 7567–7574.

15.

Raggi

Bergamini

Ronca

(1975) Isozymes of AMP deaminase in red and white skeletal muscles. FEBS Lett 58:19–23.

16.

Raggi

Ranieri-Raggi

(1987) Regulatory properties of AMP deaminase isoenzymes from rabbit red muscle. Biochem J 242:875–879.

17.

Raggi

Ronca-Testoni

Ronca

(1969) Muscle AMP aminohydrolase. II. Distribution of AMP aminohydrolase, myokinase and creatine kinase activities in skeletal muscle. Biochim Biophys Acta 178:619–622.

18.

Ranieri-Raggi

Montali

Ronca

Sabbatini

Brown

Moir

AJG

Raggi

(1997) Association of purified skeletal-muscle AMP deaminase with a histidine-proline-rich glycoproteinlike molecule. Biochem J 326:641–648.

19.

Ranieri-Raggi

Raggi

(1980) Effects of storage on activity and subunit structure of rabbit skeletal-muscle AMP deaminase. Biochem J 189:367–368.

20.

Ranieri-Raggi

Raggi

(1990) Regulation of skeletal-muscle AMP deaminase. Evidence for a highly pH-dependent inhibition by ATP of the homogeneous derivative of the rabbit enzyme yielded by limited proteolysis. Biochem J 272:755–759.

21.

Ronca-Testoni

Raggi

Ronca

(1970) Muscle aminohydrolase III. A comparative study on the regulatory properties of skeletal muscle enzyme from various species. Biochim Biophys Acta 198:101–112.

22.

Rylett

Sia

Mundy

Parish

(1981) Autorosette inhibition factor: isolation and properties of the human plasma protein. Eur J Biochem 119:641–646.

23.

Smith

Alam

Tatum

Morgan

(1988) Histidine-rich glycoprotein is synthesized by the liver and is a negative acute-phase response protein. J Cell Biol 107:584a.

24.

Van Kuppevelt

Veerkamp

Fishbein

Ogasawara

Sabina

(1994) Immunolocalization of AMP-deaminase isozymes in human skeletal muscle and cultured muscle cells: concentration of isoforms M at the neuromuscular junction. J Histochem Cytochem 42:861–868.