Domain Structure, Stability and Interactions in ... - Wiley Online Library

Eur. J. Biochem. 239, 333-339 (1996) 0 FEBS 1996

Domain structure, stability and interactions in streptokinase Leonid V. MEDVED’.’, Dmitry A. SOLOVJOV2 and Kenneth C. INGHAM’

’

J. Holland Laboratory, American Red Cross, Rockville MD, USA

’ Institute of Biochemistry, National Academy of Sciences of Ukraine, Kiev, Ukraine (Received 26 Februaryl26 April 1996)

-

EJB 96 0267/3

The structural organization of streptokinase was established through detailed study of its denaturation by differential scanning calorimetry. Streptokinase exhibited a complex endotherm whose shape was sensitive to changing pH. In all cases the endotherms were easily described by four two-state transitions indicating unambiguously the presence of four independently folded domains in the molecule. Two of them were slightly destabilized by lowering pH from 7.0 to 3.8 while the other two were stabilized in this pH range. Two proteolytic fragments of streptokinase were examined, a 37-kDa fragment beginning at Ilel with a cleavage following Phe62, and a 17-kDa fragment beginning at Lys147. At pH 8.5, three two-state transitions were observed in the former and two in the latter indicating this many domains in each and suggesting that the fragments are formed by a step-wise removal of individual domains from the parent molecule. Comparison of the melting of these fragments with that of streptokinase allowed the first two transitions in the parent protein to be assigned to the melting of two NH,-terminal domains and the two higher-temperature transitions to the melting of the two COOH-terminal domains. The latter two domains strongly interact with each other since the absence of the most stable extreme COOH-terminal domain in both fragments resulted in a strong destabilization of its neighbor whose melting occurred with a midpoint near room temperature. The two NH,-terminal domains seem to be more independent. One of them melts similarly in the parent protein and both fragments while the other, formed by the 1 - 146 region, is less stable in the 37-kDa fragment. This destabilization is most probably due to the cleavage after Phe62 which, based on the sequence similarity of streptokinase with serine proteases, may be part of a surface-oriented loop. Keywords: streptokinase; fragments ; domains ; denaturation: scanning calorimetry.

Streptokinase is a 47-kDa protein produced by various strains of hemolytic streptococci. It has been widely used for fibrinolytic therapy for more then three decades. In many cases streptokinase is as effective as other popular fibrinolytics such as urokinase-type and tissue-type plasminogen activators (Gaffney, 1994) but lower cost makes it more attractive for therapeutic use. In contrast to the plasminogen activators, streptokinase is not an enzyme and activates plasminogen by a different mechanism. It forms a 1: 1 stoichiometric complex with plasminogen resulting in conformational changes that expose the active site of the latter (Castellino, 1979). The altered plasminogen in the complex is then able to activate other plasminogen molecules into plasmin. To establish the mechanism of this process requires detailed knowledge about the structure of both plasminogen and streptokinase. Plasminogen has been intensively studied resulting in substantial understanding of its structural organization (Ponting et al., 1992). Less is known about the structure of streptokinase. Streptokinase consists of one polypeptide of 415 amino acid residues with no cysteine or carbohydrate (Jackson and Tang, 1982). In spite of numerous physico-chemical studies of streptokinase, there is no clear idea about how this polypeptide is folded and arranged in space. Based on the similarity of the Correspondence to L. V. Medved, American Red Cross, J. Holland Laboratory, 15601 Crabbs Branch Way, Rockville, MD 20855, USA Fax: + I 301 738 0194. ADDreviations. DSC, differential scanning calorimetry ; AC,,cxL,excess heat capacity function.

NH,-terminal 230-residue portion of streptokinase with trypsin and other serine proteases and on the internal similarity between its NH,-terminal and COOH-terminal halves, it was suggested that its three-dimensional structure likely contains two independently folded domains, each similar to serine proteases (Jackson and Tang, 1982). An attempt to characterize the domain structure of streptokinase using differential scanning calorimetry (DSC) was performed by Radek and Castellino (1989). They found that, in different conditions, streptokinase undergoes a single two-state thermal transition reflecting denaturation of one domain, although in the alkaline region they observed evidence for melting of a second region of the molecule suggesting an additional domain. Welfle et al. (1992) using the same method demonstrated that at neutral pH streptokinase denatures in two distinct temperature peaks suggesting a two-domain structure. In another study the domain structure of streptokina$e in solution was analyzed by dynamic light scattering and small-angle X-ray scattering (Damaschun et al., 1992). In contrast to the DSC data, the results of that study suggested that streptokinase consists of four or possibly five compact separately folded domains linked by mobile segments. At least three domains which have independent stability were resolved when thermal and chemical denaturation of streptokinase was studied by IH-NMR spectroscopy (Teuten et al., 1993). Thus knowledge about the domain structure of this molecule is still controversial. This issue is becoming important in view of recent studies attempting to localize functional sites on streptokinase using different proteolytic (Rodriguez et al., 1994; Shi et al., 1994) and recombinant fragments

Medved et al. (ELMJ . Biochem. 239)

334

digest was fractionated on a column (1.5 X 87 cm) of with Sephadex G-75SF equilibrated with NaCl/Tris. The major peak contained the 30-kDa band and some lower-molecular-mass impuri30kDa ties revealed by SDSPAGE. It was pooled and rechromatob I-. 7-kDa 37-kDa fragment graphed on the same column. The final product exhibited a 3062-63 C 147 17-kDafragment kDa band and a weak 7-kDa band (Fig. 1, lane b); sequence analysis also revealed two major sequences with approximately equimolar ratio starting at Ilel and Ala63, indicating that the product consists of two non-covalently attached remnants with molecular masses of 7 and 30 kDa; it was denoted as 37-kDa fragment. The cleavage was apparently quantitative in the final preparation since no band is visible in the 37-kDa range by SDS/ PAGE. Solid-phase binding measurements revealed that biotinylated 37-kDa fragment retained its affinity for plastic-bound Glu-plasminogen. Preparation of 17-kDa fragment. The 17-kDa fragment a b c was prepared from the prolonged chymotryptic digest performed Fig. 1. Schematic representation of streptokinase and its fragments in similar conditions as described above. The digest was stopped (top) and their SDSlPAGE analysis (bottom). Lane a represents strepwith diisopropyl fluorophosphate after 90 rnin when the 17-kDa tokinase; lane b, 37-kDa fragment (30-kDa + 7-kDa bands);lane c, 17- fragment was a predominant digestioa product. The digest was kDa fragment; the outer lanes in the gel contain molecular mass markers fractionated on the same column as above and the fraction conof 14.4, 21.5, 31, 45, 66.2 and 92.5 kDa. Numbers indicate position of taining the 17-kDa fragment was pooled and rechromatoNH,- and COOH-termini in streptokinase (a) and NH,-termini in its 37kDa fragment (b) and 17-kDa-fragment ( c ) ; arrow indicates cleavage graphed. The final product exhibited a single 17-kDa band (Fig. 1, lane c); sequence analysis also revealed a single sesite at position 62-63 in the 37-kDa fragment resulting in non-covaquence starting at Lys147. This indicates that the 17-kDa fraglently attached 7-kDa and 30-kDa rcmnants. ment represents the central portion of streptokinase similar to that prepared with trypsin earlier (Misselwitz et a]., 1992). It (Rodriguez et al., 199.5; Reed et al., 1995) whose proper folding should be noted that sequence analysis of the partially purified and domain composition is not obvious. A knowledge of domain 17-kDa fragment revealed the presence of a minor sequence borders and interactions between domains is also essential for starting at Asp289, suggesting that this fragment may end at optimum design and interpretation of experiments utilizing re- Phe288. This is consistent with the specificity of chymotrypsin combinant fragments. In this paper we present the results of a and with the molecular mass of the fragment. Solid-phase bindstudy of the denaturation process of streptokinase and its proteo- ing measurements revealed no affinity of biotinylated 17-kDa lytic fragments that enabled us to make conclusions about their fragment for plastic-bound Glu-plasminogen. Amino acid sequence analysis. NH,-terminal sequence domain structure and domain -domain interactions and to estianalysis was performed with a Hewlett-Packard model GIOOOS mate the borders between domains. sequenator. The NH,-termini in all cases were determined by direct sequencing for 10 cycles. Computer-assisted comparison of the residues in each cycle against the known sequence of MATERIALS AND METHODS streptokinase allowed specific sequences to be identified. Calorimetric study. Differential scanning calorimetry Preparation of streptokinase. Streptokinase was prepared from commercially available Kabikinase (AB Kabi) that con- (DSC) measurements were made with an updated DASM-1 M tained albumin as stabilizer. The albumin was removed by affin- calorimeter (Privalov and Potekhin, 1986) in the temperature ity chromatography on Blue Sepharose CL-6B (Pharmacea range 15- 115"C and at a scan rate of 1"C/min. Protein concenLKB). In a typical preparation 2000000 1U Kabikinase was dis- trations varied over 1.O-2.0 mg/ml. These were determined solved in 2 ml cold water, dialyzed against 2 1 water at +4"C spectrophotometrically using the absorbtion coefficient A;!,'' = 0.98 for streptokinase as determined by Weltle et al. (1992), overnight and then applied to a column (2.5 X 13 cm) of Blue Sepharose equilibrated with 20 mM Tris/HCI, pH 7.2, 0.15 M and A::[: = 0.57 for the 37-kDa fragment and 0.53 for the 17NaCl (NaCVTris). The bound material containing mainly albu- kDa fragment. The latter two were calculated from the amino min was eluted with 8 M urea. The passed-through material con- acid composition of the fragments by the following equation: taining streptokinase was concentrated with a Centriprep-10 A:':,;" = (5690W + 1280Y)/m, where W and Y represent the (Amicon) and rechromatographed on the same column. The pu- number of Trp and Tyr residues, and m represents molecular rity of streptokinase was checked by SDS/PAGE and by se- mass (Edelhoch, 1967; Gill and von Hippel, 1989; Pace et al., quence analysis. The preparation exhibited a single band (Fig. 1, 1995). The DSC curves were corrected for an instrumental baseline obtained by heating the solvent. Melting temperatures, lane a) and a single sequence corresponding to the NH,-terminus calorimetric and van't Hoff enthalpies were determined from the of streptokinase without any detectable impurity. Preparation of 37-kDa fragment. Streptokinase (2.5 mg/ DSC curves using software provided by Dr V. Filimonov (Instiml) in NaCl/Tris was digested at 25°C with a-chymotrypsin tute of Protein Research, Pouschino, Russia). Deconvolution (Sigma) at an enzyme/substrate mass ratio of 1 :500. When ana- analysis was performed according to Privalov and Potekhin lyzed by SDS/PACE, the time-course digestion patterns were (1986) and Filimonov et al. (1982) using the same software. similar to that reported for trypsin (Misselwitz et al., 1992); the The software allows one to perform forward deconvolution of 30-kDa fragment became the predominant degradation product an endotherm, i.e. from lower to higher temperatures, as well as after 20 min and persisted after 90 min when accumulation of a reverse deconvolution, i.e. from higher to lower temperature. substantial amount of a 17-kDa fragment was observed (not Both methods give similar results but reverse deconvolution is shown). The digestion was stopped after 30 niin by the addition useful when the beginning of the endotherm is not obvious. The of diisopropyl fluorophosphate to final concentration 2 mM. The software also allows the analysis to be performed by either indeStreptokinase

415

a 1

.-

335

Medved et al. (ELKJ. Biochem. 239)

c 0

.-

m

Q

-

d

m a, I

I

I

I

20 30 40

I

I

50

60

I

I

70 80

I

I

I

90 100 110 120

Temperature (C) Fig. 2. Original differential scanning calorimetry curves of streptokinase and their deconvolution analysis. (A) The sample in 20 mM sodium phosphate pH 7.2 was heated to 115 "C (solid curve l ) then cooled and reheated to the same temperature (dotted curve 2). The smooth curve 3 represents the theoretical heat capacity function for the denatured protein obtained as described in Materials and Methods. (B) Both original DSC curves are the same as presented in A. The baseline (broken line) for the upper curve was drawn by connecting the beginning of the transition with the plateau starting at 80°C; that for the lower curve was determined by connecting the beginning of the transition with the second plateau starting at 95°C (see text). The dotted lines represent the component two-state transitions obtained by deconvolution and the best-fit curve which essentially coincides with the experimental one.

pendent or dependent schemes. The former is based on the assumption that each domain unfolds independently, regardless of the state of the neighboring domains. The latter assumes an ordered process in which constituent domains unfold sequentially, implying the occurrence of interactions between domains such that the unfolding of any given domain depends on the status of its neighbors. In the absence of domain-domain interactions, the two schemes give essentially the same results. The relative error of the experimental enthalpy values is estimated at ? 5 % and that of the melting temperatures at 50.2"C. The corresponding errors in the parameters of individual transitions obtained by deconvolution of complex endotherms are estimated at ? 10% and ? 1.O"C. The theoretical heat capacity of the denatured streptokinase was calculated based on its amino acid composition according to Privalov and Makhatadze (1990).

RESULTS Denaturation of streptokinase. The original melting curve of streptokinase in sodium phosphate pH 7.2, recorded in the temperature range between 15- 1 15 "C, is presented in Fig. 2 A (curve 1). The curve exhibits a well defined beginning of denaturation at 35°C and two well resolved heat absorption peaks centered at 48°C and 61 "C similar to those obtained by Welfle et al. (1992) in the same buffer at pH 7.5. In agreement with those authors, we found denaturation to be highly reversible when the protein is heated to only 90°C (not shown). But even after heating up to 115"C, some reversibility was evident upon second heating (dotted curve 2). The data at higher temperature also suggest the presence of a third heat absorption peak above 75°C which is especially obvious when the endotherm is com-

I

--

< .,