Tropomyosin from human erythrocyte membrane polymerizes poorly but binds F-actin effectively in the presence and absence of spectrin

Biochimica et Biophysica Acta 912 (1987) 157-166 Elsevier

157

BBA32710

Tropomyosin from human erythrocyte membrane polymerizes poorly but binds F - a c t i n effectively in t h e p r e s e n c e and a b s e n c e of s p e c t r i n A l a n S. M a k , G l e n g o s e b o r o u g h a n d H e a t h e r B a k e r Department of Biochemistry, Queen's University, Kingston, Ontario (Canada) (Received 28 July 1986) (Revised manuscript received 5 December 1986)

Key words: Tropomyosin; F-actin binding; Polymerization; Tyrosine distribution; (Human erythrocyte)

Actin in the human erythrocyte forms short protofilaments which are only long enough to accommodate tropomyosin monomers (Shen, B.W., Josephs, R. and Steck, T.L. (1986) J. Cell Biol. 102, 997-1006). This interaction between actin and tropomyosin monomers is predicted to be weak, since tropomyosin polymerization parallels its affinity for F-actin. We examine the binding of human erythrocyte tropomyosin to actin in the presence and absence of spectrin and its ability to polymerize. The binding of human erythrocyte tropomyosin to F-actin is not affected appreciably by the present of spectrin. Saturating F-actin with erythrocyte tropomyosin, however, weakens the binding of spectrin dimers to actin. Although tropomyosin from human erythrocyte and rabbit cardiac muscle have similar affinity for F-actin, the polymerizability of erythrocyte tropomyosin as determined by viscosity measurements is much reduced relative to muscle tropomyosin. This unusual property of erythrocyte tropomyosin is likely due to differences in its primary structure from other known tropomyosin at the amino and carboxyl terminal regions which are responsible for its head-to-tail polymerization and cooperative binding to F-actin. Analysis of the distribution of tyrosine by 2-dimensional tryptic mapping of t251-labelled erythrocyte tropomyosin shows that tyrosine at positions 162, 214, 221, 261 and 267 in rabbit cardiac tropomyosin are conserved in human erythrocyte tropomyosin but Tyr-60 is absent. This observation suggests that erythrocyte tropomyosin has a carboxyl terminal region similar to its muscle counterparts but its amino terminal region resembles that of platelet tropomyosin which also lacks Tyr-60.

Introduction

The cytoplasmic surface of the human red cell plasma membrane is covered by a 2-dimensional network of cytoskeletal proteins which are responsible for maintaining the characteristic biconcave shape and durability of the red cell memAbbreviations: TPCK, L-tosylamido-2-phenylethylchloromethyl ketone; DFP, diisopropyl fluorophosphate. Correspondence: A.S. Mak, Department of Biochemistry, Queen's University, Kingston, Ontario, Canada K7L 3N6.

brane (for reviews see Refs. 1-3). The proteins essential to this skeleton are spectrin, band 4.1 and actin. The network is anchored to the membrane through the spectrin-ankyrin-band 3 linkage [4] and perhaps through the association between band 4.1 and glycophorin A [5] and glycocono nectin [6]. Although the structural organization of the cytoskeletal proteins are quite well documented, the control mechanism governing the dynamic shape changes of the red cell remains unclear. The recent identification and purification of myosin [10,11] and tropomyosin [12] from the

0167-4838/87/$03.50 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)

158

human red cell has hinted at the presence of an ATP-dependent actomyosin contractile apparatus in the red cell [56]. Although the involvement of tropomyosin and troponin in the regulation of the muscle contractile system is fairly well understood (for review, see Ref. 13), the role of tropomyosin in smooth muscle and non-muscle cells, where troponin-like proteins have not yet been identified, is not known (for reviews, see Refs. 14, 15). Unlike F-actin of nucleated cells which exist as long filaments, actin in the human erythrocyte cytoskeleton forms short protofilaments with 10-19 actin protomers which can only accommodate two tropomyosin monomers, one in each groove [7-9,54]. This interaction between short actin filaments and tropomyosin monomers is unusual and would be predicted to be weak in view of findings that end-to-end interactions between tropomyosin moleculare are necessary for high affinity binding to F-actin [26,47,48]. It is possible that the tropomyosin-actin interaction in the red cell can be stabilized by other actin-binding proteins, spectrin and band 4.1. In addition, the primary structure of the erythrocyte tropomyosin can be sufficiently different from other muscle and non-muscle tropomyosin especially at the amino and carboxyl terminal regions which are responsible for tropomyosin's head-to-tail interaction [49,50] and co-operative binding of tropomyosin to F-actin [26,47,51]. In this paper we report: (1) the interaction of human erythrocyte tropomyosin and F-actin in the presence and absence of spectrin; (2) polymerization of the erythrocyte tropomyosin, and (3) a comparison of the primary structures of tropomyosin from human erythrocyte, rabbit cardiac muscle and equine platelets by studying the tyrosine distributions in the three proteins. Materials and Methods

Materials Outdated packed human erythrocytes were obtained from Kingston General Hospital, Kingston, Ontario. Equine platelet tropomyosin was a gift from Drs. L. Burtnik (University of British Columbia) and D. Stewart (University of Alberta). Other essential materials and their suppliers are listed below: frozen rabbit hearts and skeletal

muscle, Pelfreeze Biological; sodium 125I, New England Nuclear; X-ray film, X-Omat AR2, Kodak; enhancing screen, Dupont; cellulosecoated thin-layer plates, 20 x 20 cm, Eastman; Bradford protein determination kit, Bio-Rad; DE52 cellulose, Whatman; TPCK-treated trypsin, lactoperoxidase, Sigma; ODS-PTH column, Altex; all other chemicals were of reagent grades.

Preparation of tropomyosin from human erythrocytes The methods of C6t6 and Smillie [16] and Fowler and Bennett [12] were used with modifications. Erythrocyte ghosts were prepared from outdated human erythrocytes with the inclusion of 2 mM MgC12 in the lysing buffer [12]. The Mgtreated ghosts were washed with 5 mM NaPO4/1 mm E D T A / 1 mm DFP (pH 7.5) and freeze dried. The freeze-dried ghosts can be stored for months at - 2 0 ° C without decrease in the yield of tropomyosin. Using freeze-dried ghosts allows us to carry out larger scaled preparations of tropomyosin. 4 g freeze-dried Mg-treated ghosts were stirred in 1.2 1 of extraction buffer (1.0 M KC1/20 mM Tris-HC1/1 mM E D T A / 1 mM dithiotbreitol (pH 7.5)) for 90 min at 4°C. The crude extract was centrifuged at 14000 rpm for 2 h at 4°C in a Beckman JA 14 rotor. The supernatant was adjusted to pH 4.5 with 1 M HC1, stirred for 1 h at 4°C and centrifuged at 14000 rpm for 20 min at 4°C in a Beckman JA 14 rotor. The pellets were solubilized in 160 ml of the extraction buffer by stirring for 1 h keeping the pH at 7.5. The solution was centrifuged at 17 000 rpm in a Beckman JA 20 rotor for 20 min. The supernatant was adjusted to pH 4.5 and the precipitate was dissolved in 50 ml of 0.1 M KC1 DEAE buffer (0.1 M K C I / 2 0 mM Tris-HC1/0.5 mM dithiothreitol (pH 7.5)) and stirred for 16 h at 4°C. The solution was heated to 85°C and cooled to room temperature followed by centrifugation at 17000 rpm for 1 h in a Beckman JA 20 rotor. The supernatant was applied to a DEAE-cellulose column, 1.5 x 25 cm pre-equilibrated with the 0.1 M KC1 DEAE buffer. A gradient of 0.1-0.4 M KC1 in the DEAE buffer was applied at a flow rate of 20 ml per h. A yield of 4-10 mg tropomyosin were obtained from 4 g of freeze-dried ghosts.

159

Preparation of other proteins Rabbit cardiac tropomyosin was prepared as described in Ref. 17. Actin was prepared from rabbit skeletal muscle as described in Ref. 18 and stored at 4 ° C as the G-actin form and used within 2 weeks. Spectrin dimers were prepared as described by Cohen and Foley [19]. Troponin was prepared from rabbit skeletal muscle as described in Ref. 20.

Iodination of tropomyosin Tyrosine residues in tropomyosin were labelled with 125I by lactoperoxidase and I'-I202 as described in Ref. 21. Excess sodium 125I was removed from the iodinated protein by gel filtration on a Sephadex G-10 column, 1 × 25 cm, pre-equilibrated with 50 mM N H 4 H C O 3 (pH 8.0). The 125I-labelled tropomyosin were used for actinbinding experiments and studies on the distribution of tyrosine in human erythrocyte tropomyosin and rabbit cardiac tropomyosin.

Location of the tyrosine residues in tropomyosin The a25I-labelled tropomyosin, 1 m g / m l in 50 mM N H 4 H C O 3 (pH 8.0) was digested with trypsin at 37°C for 16 h with an enzyme-to-substrate weight ratio of 1 : 25. The digest was freeze dried twice to remove any trace of N H 4 H C O 3 and dissolved in pH 6.5 buffer (pyridine/acetic a c i d / water, 100 : 3 : 895, v / v ) to give a specific activity of about 3 • 105 cpm//~l. Tryptic maps were prepared on cellulose-coated thin-layer plates. The first dimension was carried out electrophoretically using the pH 6.5 buffer on a Camag thin-layer electrophoretic apparatus running at a constant voltage of 400 volts for 75 min. The current varied from 9 to 12 amp from the start to the end of each run. The second dimension was performed by ascending chromatography in a solvent of butano l / pyridine/acetic a c i d / w a t e r (20 : 25 : 5 : 20, v/v). Autoradiographs were obtained by exposing the tryptic maps to Kodak X-ray films for 6 to 16 h.

Purification and assignment of tyrosine-tryptic peptides Although tryptic maps of x25I-labelled tropomyosin were routinely prepared on 20 × 20 cm cellulose-coated thin-layer plates for comparison of

x25I-labelled peptides, in order to obtain larger amounts of peptides for reliable amino acid analyses digest of 20 mg of rabbit cardiac tropomyosin were purified by paper electrophoresis using a combination of pH 1.8, 3.5 and 6.5 buffer systems [22]. Tyrosine peptides were located by chemical staining with the Pauly reagent [23]. Positive assignment of each tyrosine residue in the tyrosine peptides was based on their amino acid compositions and electrophoretic mobilities relative to aspartic acid [24] compared to theoretical values obtained from the known amino acid sequence of rabbit cardiac tropomyosin [25]. Since the electrophoretic mobilities of the iodinated tyrosine peptides differed only marginally from that of their unmodified counterparts, the identity of the 125I-labelled peptides in the 2-dimensional tryptic map could be assigned with confidence.

Tropomyosin, spectrin and actin binding studies The binding studies were performed essentially as described in Ref. 26 except that a Beckman airfuge was used. Stock protein samples were dialysed for 16 h at 4°C against the actin-binding buffer (20 mM Tris-HC1/0.1 M KCI/10 mM MgC12/0.01% NAN3/1 mM dithiothreitol (pH 7.0)). Spectrin dimers and 125I-labelled tropomyosin were centrifuged at 100000 × g for 1 h at 4°C to remove any aggregates. Appropriate amounts of the stock protein solutions and buffer were mixed to give a final volume of 150 btl and incubated at 25 ° C for 1 h before centrifugation in a Beckman airfuge at 22 l b / i n 2 for 30 min. The supernatant was carefully removed from the pellet. Amount of bound and free tropomyosin was calculated by counting the radioactivity of the pellet and supernatant. The ratios of bound tropomyosin to actin and bound spectrin to actin were also determined by scanning the Coomassie blue-stained bands of the pellet on SDS gels using LKB Laser Densitometer. Standard curves were constructed using known amounts of proteins to correct for dye-binding ability of the three proteins. The ratio of dye b o u n d / w e i g h t of protein for a c t i n / t r o p o m y o s i n / s p e c t r i n was found to be 1 : 2.2 : 1.7. Amount of protein used for scanning never exceeded the linear portion of the standard curve. Control experiments were included where actin was absent from the binding mixture.

160

Viscosity measurements The relative viscosity of human erythrocyte tropomyosin and rabbit cardiac tropomyosin solutions in 10 mM Tris-HC1, 1 mM dithiothreitol (pH 7.0) with varying amounts of KC1 were measured using a Cannon-Manning semi-micro type A50 viscometer at 22°C. The effect of troponin on the viscosity of human erythrocyte tropomyosin and rabbit cardiac tropomyosin was studied by inclusion of varying amount of troponin in 0.5 mg/ml of human erythrocyte tropomyosin and rabbit cardiac tropomyosin. Details of protein concentrations are shown in the figure legends. Other methods SDS-polyacrylamide gel electrophoresis was performed as described in Ref. 28. Amino acid analyses were performed on a Durrum D-500 analyser in Dr. L.B. Smillie's laboratory, University of Alberta. Tropomyosin concentrations were 1% determined by absorbance measurement, a280n m = 3.3, spectrin were determined by the Bradford method [30]. F-actin concentration was d e termined using the equation [31]: [F-actin] mg/ml

,429o - 1.34 A320 0.69

A

B

C

D

30K 27K

Fig. 1. SDS-polyacrylamide gel electrophoretic analysis of tropomyosin from human erythrocyte, rabbit cardiac muscle and equine platelet, and spectrin from human erythrocyte. Lane A, 10 ~g of human erythrocyte tropomyosin. Lane B, 5 /~g of rabbit cardiac tropomyosin. Lane C, 5 /~g of equine platelet tropomyosin. Lane D, 12 ~g of spectrin. The gel was 0.1% SDS and 12% acrylamide as described in Ref. 28.

Results

Purification and gel pattern of erythrocyte tropomyosin The isolation and purification of human erythrocyte tropomyosin follows essentially the methods of Crt6 and Smillie [16] for the preparation of equine platelet tropomyosin and Fowler and Bennett [12] who first purified and characterized tropomyosin from human erythrocytes. Little difference in the yield and purity of tropomyosin was observed using either freeze-dried or fresh erythrocyte ghosts. We routinely incorporate an additional pH 4.5 isoelectric precipitation step which improves the purity of the tropomyosin appreciably. Fig. 1. shows the relative mobility and purity of rabbit cardiac, equine platelet and human erythrocyte tropomyosin and spectrin on SDS gel. Human erythrocyte tropomyosin is separated into two subunits of apparent M r of approx. 27000

and 30000 and has similar mobility as equine platelet tropomyosin. Both non-muscle tropomyosin migrate faster than rabbit cardiac tropomyosin which has a M r of 33 000. The weight ratio of the M r 27 000 and 30 000 subunits if approx. 1 : 3 by gel scanning. The same two subunits were obtained in all preparations as reported by Fowler and Bennett [12]. The spectrin used in the binding studies showed no contamination by band 4.1, as shown in Fig. 1D.

A ctin-tropomyosin-spectrin interaction Fig. 2A compares the binding of human erythrocyte tropomyosin to F-actin in the presence and absence of spectrin dimers. The concentration of F-actin and spectrin, when present, were kept constant at 8 /~M and 3 #M, respectively, while human erythrocyte tropomyosin was increased from 0 to 4 ~tM at which concentration F-actin was saturated with human erythrocyte

161

0.14 .EE 0.12 F 0.10 ~0.08

o.o~

• •

0.0~

•

o •

g 0.02 ro ~ ~'mO

i L 1.0 2.0 ~ r y t h r o c y t e TM~ /uM

i 3.0

4.0

0.16 0.14

I • --~---------I---

•

|

i

E E 0.12

=-

~ 0.10 L 0.08 o

i

0.06

o

u

1.0

2.0

3,0

4.0

[ ~ p e c t r i ~ ~M

Fig. 2. A. Effects of spectrin dimers on the binding of human erythrocyte tropomyosin (TM) to F-actin. Binding of human erythrocyte tropomyosin alone (O); human erythrocyte tropomyosin+3 I~M spectrin (O); the amount of spectrin bound to F-actin in the presence of human erythrocyte tropomyosin (mR).The F-actin concentration was kept constant at 8/~M. B. Effects of human erythrocyte tropomyosin on the binding of spectrin dimers to F-actin. Binding of spectrin alone (O); spectrin+3 /LM human erythrocyte tropomyosin (O); amount of human erythrocyte tropomyosin bound to F-actin in the presence of spectrin (11). F-actin was kept constant at 8 /~M.

tropomyosin at a molar ratio of 1 tropomyosin for every 6-7 actin monomers. Under the present experimental conditions at 10 mM Mg 2+, human erythrocyte tropomyosin binds F-actin with similar affinity as rabbit cardiac tropomyosin and equine platelet tropomyosin [12,33]. The Kapp was approx. 1.5 /~M. The binding of human erythro-

cyte tropomyosin to actin was not weakened appreciably when spectrin was present. The spectrin dimers, however, were displaced significantly from the actin filaments as the amount of human erythrocyte tropomyosin was increased. The bound spectrin/actin was reduced from approx. 0.06, when human erythrocyte tropomyosin was absent, to approx. 0.03 when the F-actin was saturated with human erythrocyte tropomyosin. The inhibition of the binding of spectrin to F-actin by human erythrocyte tropomyosin was also investigated by comparing the spectrin-actin binding curves in the presence and absence of human erythrocyte tropomyosin (Fig. 2B). Actin and human erythrocyte tropomyosin were kept constant at 8 and 3 ffM, respectively, and spectrin dimers were increased from 0 to 4 ffM. Binding of spectrin to F-actin in the absence of human erythrocyte tropomyosin was stronger than that observed by Cohen and Foley [19] and Ohanian et al. [34]. This stronger binding is not due to band 4.1 contamination of the spectrin preparation as shown in Fig. 1D. When human erythrocyte tropomyosin was present in sufficient amount to saturate the F-actin, the spectrin binding to F-actin is again weakened as shown by the shifting of the spectrin-action binding curve to the right. The amount of human erythrocyte tropomyosin displaced from the F-actin, however, was not significant, consistent with the results presented in Fig. 2A.

Viscosity measurements of human erythrocyte and rabbit cardiac tropomyosin solutions Fig. 3A shows the effect of KC1 on the relative viscosity of human erythrocyte and rabbit cardiac tropomyosin solutions. The relative viscosity of the human erythrocyte tropomyosin solution at 1 mg/ml remained close to 1.0 even at low ionic strength, 10 mM KC1, when the relative viscosity of rabbit cardiac tropomyosin was increased by approx. 1.5 times. As shown in Fig. 3B, inclusion of troponin in the solution containing 0.5 mg/ml of rabbit cardiac tropomyosin at 10 mM KC1 increases the relative viscosity of the rabbit cardiac tropomyosin solution by 4 fold when troponin was present at a troponin/tropomyosin molar ratio of 1.6. Identical experiments showed that troponin did not increase the viscosity of the human

162 1.5

A 1.4 1.3

O

> ...: 1.2

O

~H

1.1

,°--,h-----~, 1,0 0

20

,

o

40 60 r'KCI'I mM

,

?

8O

100

x

B 5.0

>-

Y

4.C 3.C > ~: 2.C

o

1.0

oi

0'5

1!0 TN/TM M/M

I i

|¢ tFlo 0l" °

tP

X

115

C

Fig. 3. Viscosity studies on tropomyosin from human erythrocyte and rabbit cardiac muscle• A. Effects of ionic strength on the viscosity (Rel. vis.) of human erythrocyte tropomyosin and rabbit cardiac tropomyosin solutions. Rabbit cardiac tropomyosin (O); human erythrocyte tropomyosin (O). The tropomyosin concentration was 1 mg/ml. B. Effects of troponin (TN) on the viscosity of human erythrocyte tropomyosin and rabbit cardiac tropomyosin solutions. Rabbit cardiac tropomyosin (O); human erythrocyte tropomyosin (©). The tropomyosin (TM) concentration was 0.5 mg/ml.

e"

tt x -

erythrocyte tropomyosin solution. The viscosity of the human erythrocyte tropomyosin solution, whether in the presence or absence of troponin, was not altered when 10 m M MgC12 were included in the buffer (not shown).

Assignment of tyrosine-containing tryptic peptides in rabbit cardiac and human erythrocyte tropomyosin A typical a u t o r a d i o g r a p h of 125I-labelled tyrosine peptides from r a b b i t cardiac t r o p o m y o s i n is shown in Fig. 4A. Eight m a j o r radioactive peptides are consistently present in the autoradiograph when the t r y p s i n / t r o p o m y o s i n ratio was 1 : 1 0 0 w / w a n d the digestion was carried out at 3 7 ° C for 5 h. W h e n the t r y p s i n / t r o p o m y o s i n ratio was

ELECTROPHORE818

+

Fig. 4. Autoradiographs of 125I-labelled tyrosine-containing tryptic peptides from rabbit cardiac muscle and human erythrocyte tropomyosin. Conditions for electrophoresis and chromatography are described under 'Materials and Methods'. A. rabbit cardiac tropomyosin digested at a trypsin/ tropomyosin ratio of 1:100, w/w, at 37°C, 5 h. B. and C. rabbit cardiac tropomyosin and human erythrocyte tropomyosin respectively, digested at a trypsin/tropomyosin ratio of 1:25, w/w at 37°C for 16 h.

raised to 1 : 2 5 w / w a n d the digestion was increased to 16 h at 3 7 ° C , the relative intensity of spots A, D a n d G increased with c o n c o m i t a n t weakening of spots B a n d E, as shown in Fig. 4B. The a m i n o acid composition a n d properties of these peptides are shown in Tables I a n d II. The

163

TABLE I A M I N O ACID COMPOSITIONS TROPOMYOSIN

OF

TYROSINE-CONTAINING

TRYPTIC

PEPTIDES

OF

RABBIT

CARDIAC

Amino acid composition is expressed as number of residue/molecule of peptide. Integers in parentheses are theoretical values obtained from the published sequence of rabbit cardiac tropomyosin [25]. When the yield of the amino acid is less than 0.1 m o l / m o l of the parent peptide, the content of the amino acid is not reported. Amino acid

Peptide A

Lys Arg Asx Thr Ser Glx Gly Ala Ile Leu Tyr Val

B

C

D

E

F

G

H

1.0 (1)

2.6 (3)

1.9 (2)

3.5 (3)

1.0 (1) 2.5 (1)

1.1 (1)

1.0 (1)

2.6 (3)

2.8 (3)

1.1 (1)

1.2 (1)

0.8 (1) 1.7 (2)

1.1 (1) 2.8 (3) 0.1 (0) 1.0 (1) 0.9 (1) 2.0 (2) 0.9 (1)

2.6 (3) 1.0 (1) 1.2 (1) 4.6 (5) 1.1 (1) 2.0 (2)

0.9 (1) 2.0 (1) 1.0 (1)

1.0 (1) 5.0 (5) 0.4 (0) 0.8 (0) 0.9 (1) 0.8 (0) 1.3 (2)

3.9 (4)

1.0 (1) 2.5 (2) 1.0 (1)

1.0 (1)

assignment of tyrosine residues in these peptides are based on the published amino acid sequence of rabbit cardiac tropomyosin [25]. All six tyrosine residues in rabbit cardiac tropomyosin at positions 60, 1'62, 214, 221,261 and 267 can be identi-

0.9 (1) 0.9 (1)

4.8 (2) 2.4 (1)

1.0 (1) 1.0 (1)

1.0 (1)

1.0 (1)

fled on the tryptic map as shown in Table II. Peptide A is a breakdown product of B due to cleavage on the COOH-side of tyrosine-261 by chymotrypsin which contaminates the trypsin sample. Peptide E contains both tyrosine-214 and

TABLE II PROPERTIES OF THE TYROSINE-CONTAINING TRYPTIC PEPTIDES FROM RABBIT CARDIAC TROPOMYOSIN Peptide name

Tyrosine residue contained

Assigned sequence

A

Tyr-261

B

Tyr-261

C

Tyr-60

D E

Tyr-221 Tyr-214 and -221 Tyr-162 Tyr-214 Tyr-267

S-I-D-D-L-E-D-EL-Y S-I-D-D-L-E-D-EL-Y-A-Q-K E-T-E-D-E-L-D-KY-S-E-A-L-K-D-AQ-E-K E-D-K-Y-E-E-E-I-K Y-S-Q-K-E-D-K-YE-E-I-K K-Y-E-E-V-A-R Y-S-Q-K Y-K

F G H

Residue No.

Mobility a observed

calculated b

Net charge at pH 6.5

Purification c steps

252-261

- 0.92

> 0.88

- 5

6.5, 1.8

252-264

-0.70

-0.72

-4

6.5, 1.8

52- 70

-0.55

-0.55

-4

6.5, 1.8

218-226 214-226

- 0.68 -0.31

- 0.68 -0.34

- 3 -2

6.5, 1.8 6.5, 1.8, 3.5

161-167 214-217 267-268

0 + 0.31 + 0.41

0

0 +1 +1

6.5, 1.8, 3.5 6.5, 1.8 6.5, 1.8

a Electrophoretic mobility relative to Asp at pH 6.5. b Electrophoretic mobility calculated by the method of Offord [24]. c Purification by high voltage paper electrophoresis using pH 1.8, 3.5 and 6.5 buffer systems as indicated.

164 221 resulting from uncleaved internal lysines at positions 217 and 220 which are adjacent to acidic residues, glutamate-218 and aspartate-219. Since simpler and more reproducible tryptic maps were obtained using the higher trypsin./tropomyosin ratio of 1:25 w/w, at 37°C for 16 h, these digestion conditions were used to produce tryptic maps of rabbit cardiac and human erythrocyte tropomyosin for comparison as shown in Figs. 4B and 4C, respectively. Peptides A, D, E, F, G and H are unmistakably present but peptide C is always absent in the human erythrocyte tropomyosin map. Thus, based on the assignment of tyrosine residues in these tryptic peptides as shown in Table II, it is clear that tyrosine-162, 214, 221,261 and 267 in rabbit cardiac tropomyosin are conserved in human erythrocyte tropomyosin, but tyrosine-60 is absent in human erythrocyte tropomyosin. On the other hand, amino acid sequence data [32] show that only tyrosine-162, 214 and 221 are present in equine platelet tropomyosin. Discussion

It is known that the actin-binding ability of tropomyosin parallels its polymerizability. Thus, tropomyosin from non-muscle tissues has weak head-to-tail interaction and has reduced affinity for actin [14,33,52]. Consistent with this generalization is the observation that removal of 11 residues from the carboxyl terminus of rabbit cardiac tropomyosin abolishes its polymerization and actin-binding abilities [48]. These observations raise a few questions concerning the cytoskeletal structure of the human erythrocyte membrane in which short actin protofilament - tropomyosin monomers complexes are linked by spectrin tetramers and band 4.1 [1,7,54]. How does human erythrocyte tropomyosin monomer bind actin protofilaments? Does spectrin and/or band 4.1 strengthen this human erythrocyte tropomyosin actin interaction? Is human erythrocyte tropomyosin structurally similar to other known tropomyosin? Our observation that human erythrocyte tropomyosin does not polymerize even at low salt is surprising, since Fowler and Bennett [12] found that human erythrocyte tropomyosin binds to F-

actin as strongly as muscle tropomyosin even at low Mg 2+ concentration. The parallel relationship between polymerizability and actin-binding ability apparently does not apply to human erythrocyte tropomyosin. The binding of human erythrocyte tropomyosin to F-actin is not affected significantly by spectrin dimers (Fig. 2). Saturating the F-actin with human erythrocyte tropomyosin reduces the binding of spectrin to F-actin by approx. 50%, which is consistent with other reports [36,43]. Furthermore, since saturating amounts of tropomyosin cannot displace the spectrin dimers completely from the F-actin suggests ternary complex of spectrin-actin-tropomyosin can be formed in vitro. Whether band 4.1, which enhances the binding of spectrin to actin [19,34], can affect the human erythrocyte tropomyosin-actin interaction is not known at present. To correlate the observed functional properties of human erythrocyte tropomyosin to its structure, we have compared the distribution of tyrosine in human erythrocyte tropomyosin with those in the well-characterized tropomyosin from equine platelet and rabbit cardiac tropomyosin. Inspection of the known primary structures of tropomyosin from rabbit skeletal and cardiac muscle [25], chicken gizzard [37,38], equine platelet [32] and rat uterus [39] allows one to divide the tropomyosin molecules into three domains based on the numbering system in rabbit cardiac tropomyosin [25,38]; i.e., the NHz-terminal domain (residues 1-80), the central domain (residues 81-258) and the COOH-terminal domain (residues 259-284). The central domain is essentially conserved in all types of tropomyosin sequenced so far and it is believed to bind to the COOH-terminal region of troponin T and possibly troponin I [40,41]. In contrast, the amino acid sequence of the NH 2and COOH-terminal domains are highly variable among different tropomyosin types. Thus, a and fl rabbit skeletal tropomyosin and gizzard fltropomyosin possess similar NH2-terminal domains which are quite different from that of gizzard 7-tropomyosin, rat uterus tropomyosin and equine platelet tropomyosin [38]. The COOHterminal domain in gizzard and platelet tropomyosin is essentially identical but it is distinctly different from that of rabbit skeletal tropomyosin which is known to interact with the

165

NH2-terminal fragment CB1 of troponin T [21,42]. The six tyrosine residues in rabbit cardiac tropomyosin are distributed throughout the polypeptide chain and therefore offer convenient 'markers' for the three different domains, i.e., tyrosine-60, tyrosine-162, 214 and 221, and tyrosine-261 and 267 are unique to the NHz-terminal, central and COOH-terminal domains, respectively [2~]. The absence of tyrosine-60 in both human erythrocyte tropomyosin subunits indicates that they both have similar NHz-terminal domains which probably resemble that of equine platelet tropomyosin [32]. Due to the highly conservative nature of the central domain, it is likely that both human erythrocyte tropomyosin subunits also have similar central regions with tyrosine-162, 214 and 221 conserved. Although tyrosine-261 and 267 are present in human erythrocyte tropomyosin (Fig. 4), it is not known whether one or both human erythrocyte tropomyosin subunits possess tyrosine-261, and 267. Heterogeneity in the COOH-regions of the two human erythrocyte tropomyosin subunits, i.e., a non-muscle-like and a muscle-like COOH-domain, respectively, can offer a plausible explanation for the difference in the affinity of the two subunits for F-actin observed by Fowler and Bennett [12]. The reduced polymerizability of human erythrocyte tropomyosin as determined by viscosity measurements is similar to platelet tropomyosin [44] rather than to muscle tropomyosin [45]. Since the COOH-region of at least one of the two human erythrocyte tropomyosin subunits belongs to a muscle-type and both subunits resemble the non-muscle type in their NH/-terminal regions, the reduced polymerizability of human erythrocyte tropomyosin and equine platelet tropomyosin must be attributable to the changes in sequence of their NH2-termini. This is consistent with observations by Sanders and Smillie [38] that gizzard fl-tropomyosin, which has a non-muscle tropomyosin-like COOH-terminus but a muscle tropomyosin-like NHz-terminus , polymerizes as well as muscle tropomyosin. It appears, then, a muscle tropomyosin-like NHz-terminus can interact with either a non-muscle tropomyosin or muscle tropomyosin-like COOH-terminus resulting in polymerization of the tropomyosin molecules. On the other hand, a non-muscle

tropomyosin-like NH2-terminus interacts poorly with either a muscle or non-muscle tropomyosinlike COOH-terminus. In conclusion, our results demonstrate that human erythrocyte tropomyosin can form stable complexes with F-actin in the presence or absence of spectrin, although it alone does not polymerize in solution. The lack of polymerizability in human erythrocyte tropomyosin is likely due to conservation of a non-muscle-like NH2-terminal region. The reason for its strong affinity for F-actin, however, is not clear. There remains the possibility that end-to-end interaction of the human erythrocyte tropomyosin molecules can be enhanced by binding to F-actin and thus strengthening the human erythrocyte tropomyosin-actin interaction. At present, the nature of the interaction between human erythrocyte tropomyosin monomer and short actin protofilament in the presence of band 4.1 and spectrin remains unclear and needs further investigation.

Acknowledgements This work was supported by the Medical Research Council of Canada and the Dean's Endowment Fund, Faculty of Medicine, Queen's University, Kingston, Ontario, Canada. We thank M. Mattriss in Dr. L.B. Smillie's laboratory, Department of Biochemistry, University of Alberta, Edmonton for performing the amino acid analyses. We also thank Dr. G. Crt6 for reading the manuscript and B. Mundell and M. Webb for expert secretarial work.

References 1 2 3 4 5 6

7 8

Bennett, V. (1985) Annu. Rev. Biochem. 54, 273-304 Cohen, C.M. (1983) Semin. Hematol. 20, 141-158 Marchesi, V.T. (1983) Blood 61, 1-11 Branton, D., Cohen, C.M. and Tyler, J. (1981) Cell 24, 24-32 Anderson, R.A. and Lovrien, R.E. (1984) Nature 307, 655-658 Mueller, T.J. and Morrison, M. (1981) in Erythrocyte Membranes 2: Recent Clinical and Experimental Advances (Kruckeberg, Eaton and Brewer, eds.), pp. 95-112, Liss, New York Shen, B.W., Josephs, R. and Steck, T.L. (1984) J. Cell Biol. 99, 810-821 Pinder, J.C. and Gratzer, W.B. (1983) J. Cell Biol. 96, 768-775

166 9 Brenner, S.L. and Korn, E.D. (1980) J. Biol. Chem. 255, 1670-1676 10 Fowler, V.M., Davies, J.Q. and Bennett, V. (1985) J. Cell Biol. 100, 47-55 11 Wong, A.J., Kiehart, D.P. and Pollard, T.D. (1985) J. Biol. Chem. 260, 46-49 12 Fowler, V.M. and Bennett, V. (1984) J. Biol. Chem. 259, 5978-5989 13 Smillie, L.B. (1979) Trends Biochem. Sci. 4, 151-155 14 CSt~, G.P. (1983) Mol. Cell Biochem. 57, 127-146 15 Payne, M.r. and Rudnick, S.E. (1984) Trends Biochem. Sci. 9, 361-363 16 CSt~, G.P. and Smillie, L.B. (1981) J. Biol. Chem. 256, 11004-11010 17 Hodges, R.S. and Smillie, L.B. (1972) Can. J. Biochem. 50. 312-329 18 Spudich, J.A. and Watt, S. (1971) J. Biol. Chem. 246, 4866-4871 19 Cohen, C.M. and Foley, S.F. (1984) Biochemistry 23, 6091-6098 20 Ebashi, S., Wakabayashi, T. and Ebashi, F. (1971) J. Biochem. (Tokyo) 69, 441-445 21 Mak, A.S. and Smillie, L.B. (1981) J. Mol. Biol. 149, 541-550 22 Welinder, K. and Smillie, L.B. (1972) Can. J. Biochem. 50. 63-90 23 Acher, R. and Crocker, C. (1952) Biochim. Biophys. Acta 9, 704 24 Offord, R.E. (1966) Nature 211,591-593 25 Lewis, W.G. and Smillie, L.B. (1980) J. Biol. Chem. 255, 6854-6859 26 Mak, A.S., Golosinska, K. and Smillie, L.B. (1983) J. Biol. Chem. 258, 14330-14334 27 Carlin, R.K., Grab, D,J. and Siekevitz, P. (1981) J. Cell Biol. 89, 449-455 28 Laemmli, U.K. (1970) Nature (Lond.) 227, 680-685 29 Gershoni, J.M. (1985) Trends Biochem. Sci, 10, 103-106 30 Bradford, M.M. (1976) Anal. Biochem. 72, 248-254 31 Jolmson, K.A. and Taylor, E.W. (1978) Biochemistry 17, 3432-3442 32 Lewis, W.G., CSt~, G.P., Mak, A.S, and Smillie, L.B. (1983) FEBS Lett. 156, 269-273 33 C6t6, G.P. and Smillie, L.B. (1981) J. Biol. Chem. 256, 7257-7261 34 Ohanian, V., Wolfe, L.C., John, K.M., Pinder, J.C., Lux, S.E. and Gratzer, W.B. (1984) Biochemistry 23, 4416-4420

35 Brenner, S. and Korn, E. (1979) J. Biol. Chem. 254, 8620-8627 36 Fowler, V.M. and Bennett, V. (1984) Erythrocyte Membranes 3: Recent Clinical and Experimental Advances, pp. 57-71, Liss, New York 37 Lau, S.Y.M., Sanders, C. and Smillie, L.B. (1985) J. Biol. Chem. 250, 7257-7263 38 Sanders, C. and Smillie, L.B. (1985) J. Biol. Chem. 260, 7264-7275 39 Ruiz-Opazo, N., Weinberger, J, and Nadal-Ginand, B. (1985) Nature 315, 67-70 40 Nonomura, Y., Drabikowski, W. and Ebashi, S. (1968) J. Biochem. (Tokyo) 64, 419-422 41 Chong, P.C.S. and Hodges, R.S. (1982) J. Biol. Chem. 257, 9152-9160 42 Pato, M.D., Mak, A.S. and Smillie, L.B. (1981) J. Biol. Chem. 256, 602-602 43 Puszkin, S., Maimon, J. and Puszkin, E. (1978) Biochim. Biophys, Acta 513, 205-220 44 C6t~, G.P., Lewis, W.G. and Smillie, L.B. (1978) FEBS Lett. 91,237-241 45 Tsao, T.C., Bailey, K. and Adair, G.S. (1951) Biochem. J. 49, 27-36 46 Ebashi, S. and Kodama, A. (1965) J. Biochem. (Tokyo) 58, 107 47 Wegner, A. (1979) J. Mol. Biol. 131, 83%853 48 Mak, A.S. and Smillie, L.B. (1981) Biochem. Biophys. Res. Commun. 101,208-214 49 Ooi, T., Mihashi, K. and Kobayahi, H. (1962) Arch. Biochem. Biophys. 98, 1-11 50 Johnson, P. and Smillie, L.B. (1977) Biochemistry 16. 2264-2269 51 Yang, Y.-Z., Korn, E.D. and Eisenberg, E. (1979) J. Biol. Chem. 254, 7137-7140 52 Dabrowska, R., Nowak, E. and Drabikowski, W. (1983) J. Musc. Res. Cell Motility 4, 143-161 53 Nowak, E. and Dabrowska. R. (1985) Biochim. Biophys. Acta 829, 335-341 54 Shen, B.W., Josephs, R. and Steck, T.L. (1986) J. Cell Biol. 102, 997-1006 55 MacLeod, A.R., Houlker, C., Reinach, F.C., Smillie, L.B., Talbot, K., Modi, G. and Walsh, F.S. (1985) Proc. Natl. Acad. Sci. USA 82, 7835-7839 56 Fowler, V.M. (1986) J. Cell. Biochem 31, 1-9