Identification of a putative collagen-binding protein from chicken skeletal muscle as glycogen phosphorylase

225

Biochimica et Biophysica Acta, 1i 22 (1992) 225-233 © 1992 Elsevier Science Publishers B.V. All rights reserved 0167-4838/92/$05.00

BBAPRO 3~66

Identification of a putative collagen-binding protein from chic,,,-:n skeletal muscle as glycogen phosphorylase Douglas J. Law and James G. Tidbal] Laboratories for Muscle Cell Biology, Department of Physiological Science, Unirersity of California, Los Angeles, CA (USA) (Received 5 December 1991)

Key words: Skeletal muscle; Muscle; Glycogen phosphorylase; Gelatin affinity; lmmunolocalization

We have purified and generated antisera to a 95 kDa skeletal muscle protein that constitutes the l~irgest mass fraction of gelatin-agarose binding proteins in skeletal muscle. Preliminary results indicated that this 95 kDa chicken skeletal muscle protein bound strongly to gelatin-agarose and type IV collagen-agarosc, suggesting a possible function in muscle cell adhesion to collagen. However, N-terminal sequencing of proteolytic fiagments of the 95 kDa protein indicates that it is the chicken skeletal muscle form of glycogen pbosphorylase, the binding of which to gelatin-agarose is unlikely to be biologically relevant. Further characteriz~tion showed that the skeletal muscle form of glycogen phosphorylase is immunologically distinct from the liver and brain form:, in the chicken, and suggests that, unlike mammalian skeletal muscle, chicken skeletal muscle may have two phosphoryta~,e isoforms. Furthermore, immunolocalization data and solubility characteristics of glycogen phosphorylase in muscle extraction c~periments suggest the enzyme may interact strongly with an unidentified component of the muscle eytoskeleton. Thus, this st .~dyyields a novel purification technique for skeletal muscle glycogen phosphorylase, provides new information on the distribur, on and isoforms of glycogen phosphorylase, and provides a caveat for using gelatin affinity chromatography as a primary step in purifying collagen-binding proteins from skeletal muscle.

Introduction The attachment of skeletal muscle cells to the extracellular matrix is mediated in part by known structural proteins such as fibronectin, laminin, and the integrin superfamily [1-3], but may also involve muscle-specific structural proteins unrelated to the above. The present study originated as a search for such a muscle adhesion protein, using gelatin affinity chromatography as an initial step in the purification of proteins from chicken skeletal muscle. Incubation of extracts from non-muscle tissues with agarose bound to gelatin or native collagens has been a widely-used technique in the enrichment and purification of collagen-binding proteins such as fibroneetin [4], colligin [5], a 47 kDa fibroblast heat-shock glycoprotein [6,7] and the gelatinase/type IV procoUagenase proteins secreted by granuiocytes and macrophages [8,9]. A 46 kDa protein, probably coiligin, has been purified from muscle cell extracts

Correspondence to: DJ. Law, Department of Physiological Science, University of California, 405 Hilgard Ave., Los Angeles, CA 900241527, USA.

using gelatin affinity [10], and a collagen-binding integrin also known as COLL-CAM [11] has been purified from hepatocytes by collagen ! affinity chromatography, and immunolocalized at cardiac myocyte cell surfaces [12,13]. In the present study, we use gelatinagarose affinity chromatography as the first step in the isolation of a 95 kDa collagen-binding protein from chicken skeletal muscle, and we evaluate the possibility that this 95 kDa protein may serve in muscle-extracellular matrix attachments in vivo. Preliminary evidence indicated that the 95 kDa protein binds tightly to gelatin and type IV collagen, and is not extracted from intact muscle fibers under a variety of conditions, suggesting that the protein plays a structt:ral role. However, subsequent amino-acid analysis and N-terminal sequencing of proteolytic fragments of the. 95 kDa protein demonstrated that the protein is the chicken skeletal muscle form of glycogen phosphorylase, and not a structural protein. Further characterization of the protein indicates that it may exist in two isoforms in chicken skeletal muscle, that it is immunologically distinct from the phosphorylases found in other avian tissues, and that its function may depend on a tight association with one or more structural elements of the muscle fiber in vivo.

226 Methods

Tissue preparation. Adult chickens ( Gallus domesticus) were anesthetized with chloroform and killed by cervical dislocation. Pectoralis major muscles were dissected out and either used fresh for protein purification or stored at -70°C for later use. Alternatively, anterior and posterior iatissimus dorsi (ALD, PLD) muscles were dissected separately, rinsed in Duibecco's PBS (137 mM NaCi, 0.8 mM CaCI 2, 0.5 mM MgCI 2, 2.7 mM KCI, 1.5 mM KH2PO 4, 8 mM Na2HPO 4 (pH 7.6)), and cut into specimens for immunohistological processing. These specimens were mounted in OCT compound and immersed in freezing isopentane cooled in liquid nitrogen. Frozen specimens were stored at - 7 0 ° C in liquid isopentane until sectioned on a cryostat at 8-12 # m section thickness. Protein isolation. Affinity chromatography was performed at 7°C using a column (6 × l cm diameter) containing gelatin-agarose (Sigma, St. Louis, MO). Alternatively, an identical column containing type IV collagen coupled to cyanogen bromide-activated Sepharose was used. The column was equilibrated with 10 mM Tris (pH 7.0), containing 200 mM NaCl, 10 mM ethylene diamine tetraacetic acid (EDTA), 7 mM flmercaptoethanol, and 0.02% NaN 3 (buffer A). Approx. 5 g of chicken pectoral muscle, freshly dissected or thawed from -70°C, was homogenized in 10 volumes of buffer A in a Sorvail Omnimixer for 2 min at setting 7. The homogenate was centrifuged at 40000 × g for 45 min at 4°C. The supernatant was applied to the affinity column at 0.5 ml/min, then the column was washed with buffer A until the Bradford assay [14] showed the absorbance at 595 nm to reach baseline. Proteins were eluted from the column with buffer A containing 1 M NaCl, 3 M NaCl or 6 M urea. The column was eluted in 2.5 ml fractions that were either stored at -20°C, or dialysed extensively against deionized w a t e r ( d H 2 0 ) , lyophilized, then stored at -20°C. Ammonium sulfate was used to precipitate proteins from muscle homogenates. Fresh or frozen chicken pectoral muscle was homogenized in 10 vol. of buffer A and centrifuged 45 min at 46 500 X g. Solid ammonium sulfate was slowly added to the supernatant in consecutive steps until concentrations of 20, 40, 60 and 80% were reached. Each salt cut was allowed to equilibrate for 1 h on ice, then centrifuged at 91000 x g for 30 rain at 4°(2, and the pellets saved for analysis using polyacrylamide gel electrophoresis performed in the presence of sodium dodecyl sulfate (SDS-PAGE) [15]. Gels used in this study consisted of 8 or 10% acrylamide, 0.13% bisacrylamide, with 5% acrylamide in the stacking gel. Ion-exchange chromatography was performed on a Mono-Q anion-exchange FPLC column (Pharmacia, Piscataway, N J). Ammonium sulfate precipitates con-

taining 95 kDa protein were desalted using prepacked Sephadex G-25 gel filtration columns equilibrated with 20 mM Tris (pH 7.6) (buffer B). The Mono-Q column was equilibrated with buffer B, the sample loaded onto the column at 0.5 ml/min, then washed to baseline (absorbance of eluate measured at 280 nm). Proteins were eluted with a linear gradient of buffer B containing 1 M NaCl, collected in 1.0 ml fractions and stored at - 20°C. Antibodies. Fractions eluted from the gelatin affinity column that were enriched in a 95 kDa m~scle protein were pooled and lyophilized. These samples were dissolved in SDS-PAGE sample buffer comprising 80 mM Tris (pH 6.8), 100 mM dithiothreitol (DTI'), 70 mM sodium dodecyl sulfate (SDS) and 1.0 mM Bromphenoi blue. The samples were boiled 1-2 min, centrifuged to clear, then separated electrophoretically on a polyacrylamide slab gel [15]. The 95 kDa band was excised from the gel, homogenized in 20 vol. of dH20, and stirred for 14 h at 7°C. The solution was centrifuged to remove gel fragments, and the supernatant dialyzed extensively against d H 2 0 and then lyophilized. An emulsion of the lyophilized protein in Freund's complete adjuvant was injected subcutaneously and intramuscularly into an adult rabbit. The rabbit was boosted after 1 month, and blood samples were obtained 1 week later and each subsequent week for 1 month. The specificity and immunoreactivity of the antisera were checked on Western blots of purified 95 kDa protein and whole skeletal muscle samples [16]. Antisera were diluted in 50 mM Tris (pH 7.5) containing 150 mM NaCl and 0.1% NaN 3 (buffer C), with the addition of 0.2% gelatin, 0.05% Tween-20, and 5% inactivated horse serum. Protein samples were separated using SDS-PAGE, and electrophoretically transferred to nitrocellulose sheets for 3 h at 1.0 A [17]. Transblots were blocked overnight at 7°C with buffer C containing 0.2% gelatin, 0.05% Tween-20, and 3% non-fat dry milk. Blots were incubated 90 rain in the antisera, then rinsed 6 × 10 min in the above blocking buffer with the non-fat dry milk concentration reduced to 0.3%. Blots were then incubated 90 min in radioiodinated goat anti-rabbit IgG, rinsed 6 x 10 min in the above wash buffer, air dried, and exposed to X-omat AR film (Kodak). Antisera were purified based on previously published procedures [17,18]. The 95 kDa protein purified by ammonium sulfate precipitation, anion-exchange chromatography, and SDS-PAGE was used for affinity-purification of anti-95 kDa, and of pre-immune serum, as a control. Immunohistochemistry. Cryosections of adult chicken ALD and PLD were hydrated for 5 min in buffer C, blocked for 20 min in buffer C including 0.2% gelatin, 0.05% Tween-20, and 3% non-fat dry milk, then washed 5 min in buffer C. Sections were then incubated for 6 h

227 in (a) affinity purified anti-95 kDa (10 /zg/ml), (b) affinity purified pre-immune serum, or (c) buffer C only. Sections were rinsed 5 rain in buffer C, l0 rain in the above blocking buffer, then 5 rain in buffer C, followed by a 1 h incubation in FITC-conjugated goat anti-rabbit IgG diluted 1 : 20 in buffer C. Sections were rinsed 5 rain in buffer C, 10 min in blocking buffer, 5 rain in buffer C, and 5 rain in dHzO before being mounted in aqueous medium and stored at 4°C. Specimens were viewed and photographed using an Olympus BH-2 microscope equipped with epifluorescence and Nomarski optics. Peptide mapping. Peptide maps of the 95 kDa protein were generated using V8 proteinase (Sigma, St. Louis, MO), according to established methods [19]. Partially purified protein samples were separated by SDS-PAGE, and the 95 kDa band excised from the gel and cut into small chips. The gel chips were equilibrated in 125 mM Tris, 1 mM EDTA, 0.1% SDS (pH 6.8) (buffer D) with 20% glycerol, and loaded into the lanes of a second, 12.5% acrylamide gel. Approx. 5/.tl of a 5 0 / z g / m l solution of V8 proteinase in buffer D with 10% glycerol was added to each lane of the gel, and the samples were electrophoresed at 85 V until the dye front reached the interface between stacking gel and running gel. The sample was incubated at room temperature for 1 h, then electrophoresed at 150 V. Gels were either stained with Coomassie blue or silver nitrate, or they were electrophoretically transferred to polyvinylidine difluoride (PVDF; lmmobilon, Bedford, MA) membranes for subsequent analysis. Amino-acid analysis and amino terminus sequencing. Intact 95 kDa protein or fragments generated by V8 proteinase digestion were transferred to PVDF membrane, stained with Coomassie blue for 10 rain, then destained for 10 min in 10% methanol with 5% acetic acid in water, rinsed in dH 2O, and air dried. Analyses were performed by the UCLA Protein Sequencing Facility. Prior to amino-acid analysis, cysteins were derivatized by carboxymethylation, and the samples were then subjected to acid hydrolysis. Sequencing was performed on a Porton 1090-E sequencer with in-line HPLC. Muscle cell extr,Jctions. Because of the relative ease of dissecting small muscle fiber bundles from frog skeletal muscles, compared with muscles from adult chickens, extractions of intact muscle fibers were performed on samples from frog muscle [17]. Frogs (Rana pipiens) were killed by decapitation and spinal pithing, and their semitendinosus muscles excised. Small bundles of intact muscle fibers were dissected free, with their attachments to the tendons left intact. The samples were incubated for 30 rain at 70C in one of the following extraction solutions: (a) PBS (pH 7.0) with 50 mM NaCI, (b) PBS with 200 mM NaCl, (c) PBS with 1% Triton X-100, (d) PBS with 0.5 mM EDTA, (e) 1 : 1

200

116 97 68

43

A

B

C

D

E

F

G

Fig. I. (A) SDS-PAGE of pooled, lyophilized fractions eluted from initial gelatin-affinity column. Major band present at approx. 95 kDa used for production of polyclonal antisera. (B-D) SDS-gel of skeletal muscle samples homogenized in PAGE sample buffer and (E-G) immunoblot of identical gel with anti-95 kDa protein. Lanes B and E, chicken skeletal muscle; lanes C and F, frog skeletal muscle; lanes D and G, mouse skeletal muscle. Molecular mass standards (in kDa): myosin, 200; fl-galactosidase, 116; rabbit phosphorylase b, 97.4; BSA, 68; ovalbumin, 43.

PBS with chloroform-methanol, (f) PBS at pH of 4.0, 5.0, 6.0, 7.0 or 8.0. Fiber bundles were then rinsed in PBS (pH 7.0), homogenized in SDS-PAGE sample buffer and assayed for 95 kDa protein in immunoblots. Results

Initial attempts to isolate collagen-binding proteins from chicken skeletal muscle used a gelatin-agarose affinity column. Fractions eluted with 1 M NaCl that contained protein were pooled and lyophilized, and a sample was made for SDS-PAGE. That sample was greatly enriched in a protein with apparent molecular mass of 95000 daltons (Fig. 1). Polyclonal antisera against this 95 kDa protein specifically labelled the protein in immunoblots of crude skeletal muscle extracts from chicken, frog, and mouse (Fig. 1). Subsequent experiments, in which the gelatin-agarose column was first eluted with 1 M NaCI, then 3 M NaCI, then 6 M urea, showed that most of the bound 95 kDa protein eluted only in 6 M urea (Fig. 2). The experiment was repeated using a collagen IV-Sepharose affinity column; attempts to elute the protein with 0.1% Triton X- 100 in column buffer also showed that most of the 95 kDa protein remained bound, eluting only in 6 M urea (data not shown). Thus, the binding of

228

200 --

116 - 97--

68-

,,,.i "'

• i

I"

43--

A

B

C

.D

E

F

Fig. 2. SDS-gel (A-C) and immunoblot (D-F) Of peak fractions eluted successively from a single gelatin-affinity column run. (A, D) peak eluted with 1 M NaCI. (B, E) peak eluted with 3 M NaCI. (C, F) peak eluted with 6 M urea. Molecular mass standards as in Fig. 1.

the 95 kDa protein to gelatin-agarose and collagen IV-Sepharose appeared to be strong, and unaccounted for by non-specific hydrophobic interactions. A typical gelatin-affinity experiment yielded approx. 200 /.tg of 95 kDa protein from each gram of starting mass.

1 2

34

5 6

7 8 910

The 95 kDa protein was further characterized by testing the protein's resistance to extraction from intact muscle cells. Bundles of intact skeletal muscle fibers from frog semitendinosus were extracted in buffers that spanned a range of ionic strength and pH, and included non-ionic detergent. Immunoblots of samples from the fiber bundles show that the 95 kDa protein was not extracted from the cells under any of the conditions used (Fig. 3), indicating a tight association between the 95 kDa protein and other relatively stable structural elements of the muscle. Purification of milligram quantities of the 95 kDa protein was carried out using ammonium sulfate precipitation and anion-exchange chromatography of chicken skeletal muscle samples. Immunoblots demonstrate that the 95 kDa protein is enriched in the 20-40% ammonium sulfate cut, and to a lesser extent in the 40-60% cut (data not shown). Pellets from 20-40% ammonium sulfate precipitation were resuspcnded in 20 mM Tris (pH 7.6) and desalted into the ~ame buffer using Sephadex G-25 columns. These samples were loaded onto a Mono-Q anion exchange column, and bound proteins were eluted in a linear NaC! gradient. Two peaks were eluted that contained 95 kDa proteins (Fig. 4). The typical yield of the purification consisting of salt precipitation followed by ion-exchange chromatography was approx. 6 mg of 95 kDa protein from each gram of starting mass. Antisera to the 95 kDa protein were affinity purified against the 95 kDa protein isolated by ion exchange, as described above. These purified antisera recognized a

12345678910

Fig. 3. (A) SD:;-gel of frog skeletal muscle fiber bundles extracted in the following solutions: (1) PBS (pH 7.0) with 50 mM NaCI; (2) PBS with 200 mM NaCI; (3) PBS with 1% Triton X-100; (4) PBS with 0.5 mM EDTA; (5) i : 1 PBS with chloroform-methanol; (6-10) PBS with 150 mM NaCI, at pHs of 4.0, 5.0, 6.0, 7.0 and 8.0. (B) Immunoblot with anti-95 kDa protein of samples identical to those in (A). Molecular mass standards in lane at far left.

229

i;~i;~ i. - ~

4

:;:.

•

,

%/

. :

- .¢.

7 ': "

A

A

t 10o T 90 -I-

~

•- - '

II

6o+

I

sos

:

1 ,T 300 ¢" / ",~

fi

r= 70 f :

t

./

Jr2oo

/\

/ S f;so_ 50

20 -Jl I -

10 0 . B

.

.

.

.

.

~ . . . . .

0

Fraction Number

Fig. 4. (A) SDS-gcI, stained with silver nitrate, of chicken skeletal muscle 40% ammonium sulfate precipitate, applied to a Mono-Q anion-exchange column and eluted with a linear gradient of NaCI from 0-300 mM. Fractions containing the highest amount of 95 kDa protein are indicated by arrows. (13) Graph of relative concentration of 95 kDa protein (% of maximum value, determined by densitometry of 95 kDa bands in the gel) in each eluted fraction. Plot of [NaCll in each fraction is superimposed.

2

3

4

5

S

1

2

3

4

5

Fig. 5. (A) SDS-PAGE and (B) immunoblot with affinity-pt, rified anti-95 kDa protein. Lanes 1, chicken skeletal muscle homogenate. Lanes 2, first peak from ion-exchange column (see fraction 13, Fig. 4). Lanes 3, second ion-exchange peak (fraction 17, Fig. 4). Lanes 4, peak eluted from gelatin affinity column. Lanes 5, peak eluted from collagen IV affinity column. S, molecular mass standards.

acid composition and amino-terminus sequence of the protein, for comparison with existing known sequences from other proteins. The immunoblot (Fig. 6) shows that antisera to the 95 kDa protein recognize polypeptides only in skeletal and cardiac muscle, and not in brain or liver. The band labelled in the cardiac muscle sample is labelled much less strongly than the band in the skeletal muscle sample. The amino-acid composi-

2 0 0 ""

116 -"

95 kDa protein in immunoblots of samples from whole chicken skeletal muscle, ion exchange, and gelatin or collagen IV affinity columns (Fig. 5), indicating that the same polypeptide is present in both the affinity column fractions and the peaks eluted from the MonoQ column. This is confirmed by a comparison of peptide maps generated by V8 proteinase digestion of the gelatin-binding 95 kDa protein, and the 95 kDa proteins in each of the peaks from the Mono-Q column (data not shown). No significant difference is apparent between the three peptide maps, indicating that the same protein is present in each sample. The possibility that the 95 kDa protein is a unique skeletal muscle protein was addressed (a) in immunoblots of gel samples from several different chicken tissues, and (b) by attempting to determine the amino-

97-"

68-"

43-1

2

3

4

1

2

3

4

Fig. 6. (A) SDS-gel and (B) immunoblot with affinity-purified anti-95 kDa protein of samples from adult chicken. (l) skeletal muscle; (2) cardiac muscle; (3) brain: (4) liver. Tissue samples were homogenized in PAGE sample buffer and gel Ioadings were adjusted to minimize differences in total protein in each lane. Standards as in Fig. 1.

230 TABLE ! Amino-acid composition of 95 kDa protein

Amino a c i d

Residues/molecule

Cys Asp Glu Ser Gly His Arg Thr Ala Pro Tyr Val Met Be Leu Phe Lys OH-Pro OH-Lys

9 106 103 39 99 13 60 36 73 33 25 46 24 39 72 3I 34 3 0

tion (Table I) indicates that the 95 kDa protein is very acidic, the two most prominent residues being Asp and Glu. T h e r e are not enough Gly residues for the protein to be a collagen, although the presence of trace hydroxyproline indicates that there may be a collagen-like domain in the molecule. T h e amino terminus of the intact polypeptide is blocked from sequence analysis, so a V8 proteinase digest o f the 95 kDa protein was performed, the proteolytic fragments were transferred to P V D F membrane, and the amino terminus sequence was determined for three fragments with apparent molecular masses of 37, 30, and 29 kDa. Comparison o f the three partial sequences with those published for known proteins shows that the highest degree o f sequence similarity for each of the three 95 kDa protein fragments is with the rabbit skeletal muscle form of glycogen phosphorylase (Table 11). Immunolocalization of the 95 kDa protein in transverse sections of adult chicken A L D and PLD muscles (Fig. 7) shows a diffuse, perimyofibrillar labelling that is absent in controls. No a p p a r e n t labelling is seen at

Fig. 7. Nomarski and immunofluorescence images of frozen sections of adult chicken skeletal muscle. (A-C) transverse sections. (D-F) longitudinal sections. (B) and (E) are fluorescence images of ALD muscles shown by Nomarski optics in (A) and (D). Comparable regions of PLD muscles are shown in (C) and (F). Magnifications:400 x (transverse sections), 650 x (longitudinal sections).

231 the cell surface. The PLD, composed primarily of fast twitch muscle fibers, is labelled much more intensely than the ALD, which comprises slower-contracting tonic fibers [20.21]. This distribution of labelling is consistent with that of glycogen phosphorylase, which is more concentrated in fast, glycolytic fibers than in slower, more oxidative cells [22]. Longitudinal sections are labelled with a sarcomeric periodicity, indicating a concentration of glycogen phospho.,ylase in discrete bands along the length of the fiber (Fig. 7). Discussion

The goal of this investigation was to identify collagen-associated proteins in skeletal muscle through gelatin-affinity chromatography, a technique that has been used successfully in initial purifications of proteins with physiologically relevant collagen-binding properties [4,11-13]. We found that the most prominent gelatin-binding protein in chicken skeletal muscle was a 95 kDa polypeptidc that also bound strongly to collagen type IV. However, our subsequent studies, reported here, show that the 95 kDa protein is the chicken skeletal muscle form of glycogen phosphorylase. This conclusion is chiefly supported by the extremely high sequence similarity between three proteolytic fragments of the 95 kDa protein and three distinct domains of the published sequence for rabbit muscle glycogen phosphorylase. In addition, the 95 kDa protein is within the range of published subunit molecular masses fo~ phosphorylases from other sources [23]. Accepted protocols for the purification of glycogen phosphorylase include ammonium sulfate precipitation followed by anion exchange using DEAE-cellulose [24-27], consistent with the purification steps used here to isolate the 95 kDa protein. The amino-acid composition of the 95 kDa protein is very similar to

those of vertebrate glycogen phosphorylases and, like the latter, the amino terminus of the 95 kDa protein is blocked from sequencing [24,26]. The weight of the above evidence, particularly the sequence comparisons, makes it unnecessary to perform enzymatic assays to confirm the identity of the 95 kDa protein. Our identification of the 95 kDa protein as glycogen phosphorylase leads us to conclude that the protein's binding affinity for gelatin and collagen IV are not biologically-relevant properties of the protein. Several previous studies have isolated proteins from a variety of tissue sources by using either gelatin or native forms of collagen as affinity substrates, e.g., Refs. 4 and 5. In many of these preparations, conditions of tissue extraction and elution from affinity substrates have closely resembled those in the present study, and some eluted fractions have included polypeptides of apparent molecular mass from 90-100 kDa [8,10,12,28-31]. In light of the present study, future attempts to characterize collagen-binding proteins, especially those in the above size range, should include experiments to eliminate the possibility that glycogen phosphorylase is a contaminant of the preparation. The present study has elucidated several previously unreported characteristics of the chicken skeletal muscle glycogen phosphorylase that will be useful for studies on the evolution of the function of the enzyme in vertebrates, and for investigation of the possibility that the protein's function in vivo depends on its localization in one or more subcellular compartments in skeletal muscle. The two separate peaks eluted from the Mono-Q anion exchange column in this study both contain 95 kDa protein; immunoblots and peptide maps identify both as the glycogen phosphorylase. Tile difference in charge between the two may reflect isotorm-speciflc differences in sequence, indicating that chickens have more than one skeletal muscle phospho-

TABLE I!

Comparison of N-terminal sequences of proteolytie fnagments of 95 kDa protebr with sequence of rabbit skeletal muscle glycogen phosphorylase Identity (%)

E

37 kDa fragment

x °

I'

HTVLPEAL

TN

HTVLPEAL

72.2 Residues 350-385 30 kDa fragment Residues 123-150

L

V

R

DWDKANEV

~']Q~EE I E E D A G L G N G G L G R L A i E

I

K l

i

C ~

i

~

89.3

E E D A 6 L G N G G L G R L A

29 kDa fragment

GAVKRINR

Residues 435-462

G A V K R I N H

x A

L C

X A

U

G S HAVNGV

71.4 R

E

232 rylase isoform. Two glycogen phosphorylase isoforms have been identified in rat cardiac muscle [25,32], but it appears that only one form exists in mammalian skeletal muscle [23]. There may be functional differences between the two rat heart isozymes based on differences in their affinity for various allosteric effector ligands [25,32], although such functional differences have not yet been demonstrated in vivo. In addition, the antiserum to the 95 kDa protein recognizes the protein in blots of skeletal muscle from frogs, mice, and chicken, but fails to label any bands in blots of chicken brain or liver, tissues that normally express glycogen phosphorylase in other vertebrates [23]. This indicates that the structure, and therefore function, of the enzyme is conserved in skeletal muscle from different species, but that the protein has evolved tissue-specific structure and function within the same species. Previous studies have also demonstrated antigenic differences in glycogen phosphorylases from rat muscle, liver, and brain [33], and closer sequence homologies among muscle forms from different species than among tissue-specific forms from the same species Finally, the immunofluorescence labelling of frozen skeletal muscle sections shows a diffuse perimyofibrillar distribution in transverse section, but labels longitudinal sections in transverse bands with sarcomeric periodicity. Thus, glycogen phosphorylase in striated muscle may be preferentially associated with periodic, extramyofibrillar membranous structures such as transverse tubules or terminal cisternae of the sarcoplasmic reticulum. Such association may be quite strong, as indicated by the protein's resistance to extraction from intact muscle fibers (Fig. 3). Rabbit skeletal muscle phosphorylase kinase has been immunolocalized to the SR [34], and a form of phosphorylase phosphatase has been isolated from a skeletal muscle microsomal membrane preparation [35]. A 97.5 kDa protein thought to be glycogen phosphorylase was present in preparations of purified sarcoplasmic reticulum [36,37], although the identity of the protein was not confirmed by immunoblot. Alternatively, the phosphorylase may remain bound to intraceUular glycogen particles rather than to internal membranous structures. Glycogen particles are distributed inhomogeneously in the muscle cytoplasm, with a concentration in and around myofibrillar I-bands [38], and the intensity of histochemical staining for glycogen phosphorylase is known to depend on the amount of intracellular glycogen [39]. Histochemical localization of glycogen phosphorylase in muscle has demonstrated a periodic labelling, possibly corresponding to the locations of sarcomeric I-bands [40-42]. However, previous immunolocalization of phosphorylase in muscle revealed no periodicity in longitudinal sections [43]. The immunolocalization of muscle glycogen phosphorylase in the present study

confirms the periodic distribution of the enzyme in longitudinal sections. This result, coupled with the resistance of the protein to extraction from intact muscle fibers, encourages pursuit of the possibility that phosphorylase function in skeletal muscle may be mediated in part by association with specific structural components within the muscle fiber.

Acknowledgements The authors thank Susan Wei for technical assistance. This investigation was supported by NIH grant AR40343 to J.G.T. Protein sequencing performed at the UCLA Protein Microsequencing Facility was aided by BRS Shared Instrumentation Grant 1 S10RR0555401 from the NIH.

References 1 Hay, E.D. (ed.) (1981) Cell Biology of Extracellular Matrix, Plenum, New York. 2 Hynes, R.O. (1987) Cell. 48, 549-554. 3 Ruoslahti, E. and Pierschbacher, M.D. (1987) Science. 238, 491493. 4 Engvall, E. and Ruoslahti, E. (1977) Int. J. Cance-. 20, !-5. 5 Kurkinen, M., Taylor, A., Garrels, J.l. and Hogan, B.L.M. (1984) J. Biol. Chem. 259, 5915-5922. 6 Nagata, K. and Yamad~, K.M. (1986) J. Biol. Chem. 26106), 7531-7536. 7 Saga, S., Nagata, K., Chen, W.-T. and Yamada, K.M. (1987) J. Cell Biol. 105, 517-527. 8 Vartio, T. and Hovi, T. (1987) Acta Chem. Scand. Ser. B 41, 754-756. 9 Vartio, T. and Baumann, M. (1989) FEBS Lett. 255, 285-289. l0 Vaillancourt, J.P. and Cates, G.A. (1991) Biochem. J. 274, 793798. I1 Gullberg, D., Terracio, L., Borg, T.K. and Rubin, K. (1989) J. Biol. Chem. 264, 12686-12694. 12 Gullberg, D., Terracio, L. and Rubin, K. (1988) Exper. Cell Res. 175, 388-395. 13 Terracio, L., Gullberg, D., Rubin, K., Craig, S. and Borg, T.K. (1989) Anat. Rec. 223, 62-71. 14 Bradford, M.M. 0976) Anal. Biochem. 72, 248-254. 15 Laemmli, U.K. (1970) Nature 227, 680-685. 16 Burnette, W.N. (1981)Anal. Biochem. 112, 195-203. 17 Tidball, J.G. (1990) Devel. Biol. 142, 103-114. 18 Cox, J.V., Schenk, E.A. and OImsted, J.B. (1983) Cell. 35, 331339. 19 Kennedy, T.E., Gawinowicz, M.A., Barzilai, A., Kandel, E.R. and Sweatt, J.D. (1988) Proc. Natl. Acad. Sci. USA 85, 7008-7012. 20 Hess, A. (1970) Physiol. Rev. 50, 40-62. 21 Close, R.I. (1972) Physiol. Rev. 52, 129-197. 22 Horl, W.H., Jennissen, H.P., GroseheI-Stewart, U. and Heilmeyer, J.L.M.G. (1975) in Calcium Transport in Contraction and Secretion (Carafoli, E. et al., eds.), pp. 535-546, North-Holland. 23 Newgard, C.B., Hwang, P.K. and Fletterick, R.J. (1989) Crit. Rev. Biochem. Mol. Biol. 24, 69-99. 24 Bonamusa, L. and Baanante, l.V. (1990) Comp. Biochem. Physiol. 95B, 295-301. 25 Berndt, N. and Rosen, P. (1984) Arch. Biochem. Biophys. 228, 143-154. 26 Metzger, B., Glaser, L. and Helmreich, E. (1968) Eiochem. 7, 2021-2036.

233 27 Schmidt, H. and Wegener, G. (1990) Am. J. Physiol. 258 (Cell. Physiol. 27), C344-C351. 28 Ogle, R.C. and Little, C.D. (1989) Biosci. Rep. 9, 329-340. 29 Ogle, R.C., Potts, A.J., Yacoe, M. and Little, C.D. (1989) J. Cell Sci. 94, 361-369. 30 Vartio, T., Hovi, T. and Vaheri, A. (1982) J. Biol. Chem. 257. 8862-8866. 31 Vartio, T., Hedman, K., Jansson, S.-E. and Hovi, T. (1985) Blood. 65, 1175-1180. 32 Berndt, N., Neubauer, H.P. and Rosen, P. (1987) Biochim. Biophys. Acta 915, 217-224. 33 Shimada, S., Maeno, M., Akagi, M., Hatayama, I., Sato, T. and Sato, K. (1986) Histochem. J. 18, 334-338. 34 Thieleczek. R., Behle, G., Mes~r. A.. Varsanyi. M., Heilmeyer, J., Ludwig M.G. and Drenckhahn, D. (1987) Eur. J. Cell Biol. 44, 333-34O.

35 Villa-Moruzzi, E. and Heilmeyer, L.M.G., Jr. (1987) Eur. J. Biochem. 169, 659-667. 36 Chu, A, Sumbilla, C., Inesi, G., Jay, S.D. and Campbell, K.P. (1990) Biochem. 29, 5899-5905. 37 Wanson, J.-C. and Drochmans, P. (1972) J. Cell Biol. 54, 206-224. 38 Friden, J., Seger, J. and Ekblom, B. (1989) Aeta Physiol. Scand. 135, 381-391. 39 Martin, D.L. and Engel, W.K. (1972) J. Histochem. Cytochem. 20, 476-479. 40 Takeuchi, T. and Kuriaki, H. (1955) J. Histoehem. Cytoehem. 3, 153-160. 41 Pette, D. (1975) Acta Histochem. Suppl.-Bd. XIV, 47-68. 42 Sigel, P. and Pette, D. (1969) J. Histochem. Cytochem 17, 225237. 43 Dvorak, H.F. and Cohen, R.B. (1965) J. Histochem. Cytochem. 13, 454-460.