Biochimica et BiophysicaAaa. 1076(1991)245-251 ~01991ElsevierSciencePublishersB.V.(BiomedicalDivision)0167-4838/91/$03.50 ADONIS 0167483891000547

245

BBAVRO33811

Structural changes in the C-terminal region of human brain creatine kinase studied with monoclonal antibodies N g u y e n thi Man, Alison J. Cartwright, M a r k Osborne a n d G l e n n E. Morris Research Dwision. N.E. Wales Institute, Deeside. Cl~yd (U.K.)

(Received5 Junelgg0) Keywords: Proteinfolding:Creatinekinase.Monoclonalantibody:Epitopemapping;Syntheticpeptide;(Humanbrain) :~pitopes on human brain cre.atine kina~ (B-CK) recognized by three monoclonal antibodies have been located by chemical cleavage methods, followed by peptide synthesis or analysis of specificity for natural variants (isoforms). One antibody, CK-HTB, recognizes a conformational, or assembled, surface epitope on native CK which is also pr~mt on partially unfolded forms, it requires an Asn residue at position 300 in the amino acid scqx~en~ mud will net.._recognize variants with Lys or His in this position. This results in a striking specificity of the antibody, which binds to B-CK only in chicken and man, but to musele4orm (M-CK) only in the rat. The results suggest that Ash-300 is exposed on the enzyme surface as part of a relatively denaturation.resistant region. Two monoclonal antibodies, CK-ENDI and CK-END2, recognise epitopes within 53 amino acids of the C-terminus and bind to a synthetic hexapeptide representing the last six amino acids of human B-CK (Leu-375-Lys.380). The two antibodies show overlapping, but distinct, specifldties in their binding to CK variants. CK-ENDI requires Met-376 and will not tolerate lie in this position, whereas CK-END2 requires Leu-375 and will not tolerate Met. Neither antibody binds to native CK, though both will bind to a folding intermediate and to partially unfolded states. This shows that the C-terminus of CK becomes inaccessible to the antibodies during those later stages of protein folding associated with recovery of enzyme activity and suggests that the protein may 'tuck in its tail' during one of the final steps. Introduction Creatine kinase (CK) (EC 2.7.3.2, ATP:creatine Nphosphotransferase) is a "globular' enzyme of known sequence but unknown lhree-dimensional structure. We have prepared a panel of monoclonal antibodies against CK isoforms and, for many of these antibodies, we have identified amino acids in the CK sequence with which they interact, using epitope mapping techniques [1-3]. These 'mapped' antibodies can be used as structural probes, since they show when the corresponding epitope on CK is conformationally intact and accessible to antibody. Many surface epitopes on native CK are readily lost on denaturation [4]. However, epitopes which are retained on partially unfolded CK, or which are

Abbreviations: CK, creatinekinase (EC 23.3.2, ATP:creatineNphosphotransferase);SDS. sodiumdodecylsulphate: PAGE.polyacrylamide gel electrophoresis: ELLS& enzymedinkedimmunosorbentassay. Cot:espondence: G.E. Morris, ResearchDivision, N.E. Wales Institute. Deeside,CIwydCH5 4BR,U.K.

normally buried and become accessible only when CK is partially unfolded, are more interesting from a structural viewpoint. Although proteins ap~ear to be completely unfolded when they are exposed to 8 M urea or boiled in 2~ SDS, they undergo rapid partial refolding (in seconds or less) when the concentration of denaturant is subsequently reduced. A longer refolding period (minutes to hours) is usually required, however, to recover the native, enzymicaUy-acdve conformation, especially for larger, multi-domain proteins, though even in these cases final recovery can be almost complete. Leaving aside the problems introduced by proline isomerization and disulphide bridges, rapid folding is believed to consist mainly of recovery of secondary and local tertiary structures, while the slower process may involve further t e r t i ~ folding as well as domain and subunit associations [5]. Using monoclonal antibodies, we have defined one 'folding intermediate' as the structure which forms after removal of urea from 8 M urea-unfolded CK by dilution at 0 o C. This urea-treated form of CK is inactive and will not bind to monoclonal antibodies specific for native CK, but will bind to other monoclonal antibodies which do not recognize the native enzyme [2], Recent

246 evidence suggests that some antibodies of the latter category recognize epitopes at domain interfaces which are buried in the native state, but accessible in the folding intermediate [3]. Further antibody studies suggest that the 'intermediate' is related to, though probably not identical with, the structure of the inactive CK formed after SDS treatment (e.g., on Western blots of SDS-PAGE) or after attachment to plastic surfaces (e.g., microtitre plates for ELISA) [2,3]. Although most monoclonal antibodies against CK bind either to native form only or only to partially unfolded forms, two antibodies bind equally well to native and urea-treated CK (CK-ART and CK-JOE, specific for the muscle-form of chick CK) and they recognize an assembled, or conformational, epitope on the C-terminal one-third of the CK molecule [2]. We now show that a monoclonal antibody against the brain-form of human CK, CK-HTB, also binds to native and various denatured states and also recognises an LI.,.I ~1 as~,,,o,~ . . . .P . .~-- nn-an,~, ~itope, similar, if not homo!~ gous, to the CK-ART epitope. We also use two other monocional antibodies to demonstrate a significant difference between native CK and the folding intermediate in the C-terminal region; the extreme C-terminus is accessible to antibody in the partially refolded state, but not in the native, enzymically active state. . . . . .

:

Materials and Methods

Materials CK-HTB monoclonal antibody against human BBCK was obtained from Hybritech (Nottingham, U.K.) or F~ow Laboratories (Glasgow). This mouse monoclonal antibody (IgGl) is one component of the 'Tandem-E' two-site immunoassay kit for MB-CK. Peroxidase-labelled rabbit anti-(mouse Ig) antibody was obtained from DAKO (High Wycombe, Bucks, U.K.) and ultra-pure urea from Schwartz-Mann (Rahway, N J, U.S.A.). Peroxidase substrates, o-phenylene diamine and diaminobenzidine (Sigma) were handled as possible carcinogens. Microtitre plates for ELISA (M25) were obtained from Dynatech (Billinghurst, Sussex, U.K.). MM-CK from chick breast muscle and BB-CK from chick heart were purified to homogeneity as described in Ref. 6. Human MM-CK was a gift from Dr. E. James Milner-White (Department of Biochemistry, Glasgow). Rat CKs were purified by Dr. Graham Pay (CIBA/ Geigy, Horsham) and an electric organ membrane preparation from Torpedo marmorata containing CK was a Oft from Dr. 3ohn Cavanagh and Professor Eric Barnard (MR(:: Centre, Cambridge). Rabbit and ox CKs and lobster arginine kinase were obtained from Sigma. Peptide synthesis materials (pentafluorophenyl esters of fiuorenylmethoxycarbonyl-protected (F-moc) amino acids and polyamide (Pepsyn KA) solid-phase resin) were obtained from M~llipore (Milligen).

Preparation of human brain CK All steps were performed at 4 ° C and in the presence of 50 mM 2-mercaptoethanol to preserve BB-CK activity. Frozen, post-mortem human brain tissue (200 g) was homogenised in 3 vols. of buffer A (50 mM Tris. HCI, pH 7.5, 50 raM 2-mercaptoethanol, 1 raM EDTA and 0.1 mM PMSF) using a Silverson blender and stirred for 1 h. After centrifugation at 10000 × g for 2 h, the pellet was re-extracted for 30 min. An equal volume of ethanol at -20 °C was added to the com. bined supematants slowly, without allowing the temperature to rise above 8°C and the mixture was stirred overnight. After removal of the precipitate by centrifugation at 12000 x g for 10 rain, further ethanol was added to 70% and the mixture stirred overnight. The pellet was resuspended in 1/10 the original volume of buffer A and dialysed for 48 h. The solution was clarified by centrifugation (100000×g for 1 h) and applied to a column (85 x 1.2 cm) of DEAE-Sephadex A-SO in buffer A. The column was washed overnight with the same buffer before applying a gradient of 0.1-0.45 M NaCl in the same buffer. CK eluted near the end of the gradient, so it was often necessary to wash fu,rther with the 0.45 M NaCI buffer to recover the enzyme. Enolase co-ehites with CK (actually at a slightly higher salt concentration) and is the main component. Fractions containing CK activity were adjusted to pH 7.0 with H3PO4 and dialysed overnight against buffer B (50 mM sodium phosphate, pH 7.0, 5 raM MgCI 2 and 5 mM 2-mercaptoethanol). The dialysed solution was adjusted to pH 6.0 with 50 mM H3PO4 immediately before applying it to a column (50 x 1 cm) of Blue Sepharose CL-tB (l=harmacia) in buffer B, pH 6.0 (CK precipitates on prolonged e::posure to pH 6). After washing with the same buffer until all enolase was removed (A2s0), the bound CK was eluted with buffer B (pH 8.0). Since 2-mercaptoethanol is toxic to mice and interferes with cleavage by NTCB, it may be omitted from the elution buffer and added immediately afterwards at 50 mM to those fractions requiring preservation of enzymic activity. The average yield was about 8 mg of CK ( > 95~ pure on SDS-PAGE) and recovery of enzyme activity was about 25~, most of the 19sses occurring during the final separation from enolase.

Monoclonal antibody production Baib/c mice were immunised with 100 ~g of human BB-CK in Freund's complete adjuvant intraperitopeally and a further 100 ?~g in incomplete adjuvant 1 month later. After a rest period of 1-2 months, mice were given 400 #g of BB-CK in PBS 4, 3 and 2 days before fusion of spleen cells with Sp2/0 or NS0 mouse myeloma lines using polyethylene glycol. Hybridoma lines were cloned twice by limiting dilution. CK-END1 is an lgG2a and CK-END2 an IgG2b.

247 Dige.~'tions For cleavage at cysteine residues, creatine kinase (0.5 rag) in 0.5 ml of 5 M guanidine-HCl/0.2 M Tds-acetate (pH 8.0) was reduced with 1 mM dithiocrythritol for 30 rain at room temperature. Nitrothiocyanobenzoic acid (NTCB; 30 ~tl of a 25 mg/ml solution) was added [7] and the mixture incubated at 37°C for 15 rain. The mixture was then loaded immediately onto a Sephadex G-15 column equilibrated at 4°C with 8 M urea/0.2 M Tris-acetate (pH 9.0)/1 mM PMSF to remove excess reagent and to raise the pH. Cleavage was completed by incubation at 37°C for 16 h. The pH was adjusted to 6.8 with HCI and SDS (1~), 2-mercaptoethanol (5%) and sucrose (10~) were added before boiling for 2 rain for dectrophoresis. Digestion of CK (20 mg/ml) with proteinase K (4/tg/ml) for 30 rain at 30 °C was carded out as described in Ref. 8 and the digest was used directly for electrophoresis after adding SDS (l~g), 2mercaptoethanol (5~) and sucrose (10~) and boiling for 2 min.

oxidase-labelled second antibody and diaminobenzidine as described previously [9,10]. Pepttde synthesis and ELISA Peptides were synthesized using a solid-phase method and apparatus [11]. Efficiency of peptide bond formation was monitored by A~13of the 9-fluorenylmethoxycarbonyl protecting group at each step and was similar for all peptides. Deprotection and cleavage from the solid phase was performed with trifluoroacetic acid in the presence of methionine as 'scavenger' (95 : 5). After rotary evaporation and repeated ether washing, the dry peptide (plus methionine) was dissolved in water. To measure antibody binding by peptides or CK, microtitre plates were coated with different concentrations of peptide or 5 #g/ml CK (0.1 ml per well in PBS (0.9~, NaCI/25 mM sodium phosphate, pH 7.2)) for 1 h at room temperature, blocked and incubated successively with monoclonal antibody (1/100 dilution of culture supernatants), peroxidase-labelled rabbit anti(mouse lg) and o-phenylenediamine substrate as described previously [4,12].

Western blouing After SDS-PAGE, gels were blotted for 18 h onto two ni!rocellu!~s~sheets by diffusion in 25 ram Tris-192 mM glycine to give two identical mirror-image blots. Prestained marker proteins (GIBCO/BRL) were also transferred to enable precise identiEcation of antibodybinding bands. Markers are ovalbumin (43000), chymotrypsinogen (25700), lactoglobulin (18400), lysozyme/ cytochrome c (13300), bovine trypsin inhibitor (6200) and insulin (3000). The blots were stained for protein with Amido Black or blocked and incubated successively with monoclonal antibody (1/100 dilution of culture supematant), per-

I;.i i

A A A A

All three antibodies bind to the C-termi~lalregion and two bind to the last 53 amino acids Cleavage of CK at its four Cys residues with nitrothiocyanobenzoic acid (NTCB) produces a complete 'ladder' on SDS-PAGE of partial digestion fragments from which each monoclonal antibody picks out a subset containing its unique epitope. For example, monoclonal antibodies with epitopes within C-terminal fragment E will pick out five bands: E, DE, CDE,

;~K

t

CPSHLGTGLR G

R ~

I I

5.3K

I I

ProteinaseK fragments

VKL~KLS~Hp~FEEILHRLRL~RGTGGVDTAAVGA~FD~SHADRLGF~EVEQ~H~VDGV~L~VE~EK~LEQN~DDM~p~ AI! I~ T A $ V 5 L KG S

I D] I I

GK E Gg E H G E H 1< E I H CE E

G V R SDV K $ V~ S V R S V R T

£

G G G G SIY

V V ~ V

L L L L

LI LI LI LI

R QR Q~ QR

}~G S A G l~ NGKS

t NM-CHICK * RH-HUMAN

BB-RABBIT * BB-RAT * TORPE]30

Fi8.1. Locationof ~itopeswithinthe amino-acidsequenceof a CqerminalNTCBfragmez,~tof ¢reatinckinasc,(a) The relativesizesof fragn~nts generatedby proteinaseK and NTCB.(b) Arrowsindicatethe positionsof aminoacidsimportantfor antibodybinding.Theessentialrcsidu~for CK-HTBantibody(Asn-300)and CK-END1/END2(Leu-375-Met-376)are boxed.The Ala-A|abondcleavedby proteinaseK is indicatedby * Sourcesof the sequencesweee:chickMM[20],humanMM[21],rabbitMM[22],rat MM[23l,chickBB[24],humanBB[25],rabbitIIB[26],rat BB [27]and Torpedomarmorata[28].

248

ART

Mr i

M

HTB

B,B

END1

END2

26-

END1

END2

HTB

JAC

.

.

U

CK 38

- ABCDE BCDE

43•

Mr

M, B i B ]

~w

- CDE

18-

,~

I i

.

- DE E

5.3

1

2

3

4

5

6

7

Fig. 2. Monoclonal antibody binding to NTCB cleavagefragmenlsof

chick M-CK and human B-CK. Western blots are shown of 15% polyacrylamidc gels in the SDS-Tris buffer system of Lacmmli 129] loaded with prestained Mr markers (lane 1), NTCB-treated chick MM-CK (20 ~ag;lanes 2 and 5) or NTCB-treated human BB-CK(20 /~g; lanes 3, 4. 6 and 7) after development with four monoclonal antibodies (1/I00 dilution of culture supe.rnatant). The letters A-E correspond to NTCB segments in Fig. la.

BCDE and CK itself (ABCDE) (Fig. la), producing a pattern on SDS-PAGE/Western blots which is recognisably different from that produced by antibodies with epitopes in other fragments [2]. Fig. 2 (lane 4) shows that CK-HTB binds to the C-terminal 14 kDa fragment (DE), but not the smaller 10.6 kDa fragment (E). For comparison, the binding of C K - A R T antibody to NTCB fragments of chick MM-CK on the same Western blot is also shown in Fig. 2 (lane 2) and the pattern is similar, except that C K - A R T will also bind to fragment E. The comparison also shows the complete lack of cross reaction of each monoclonal antibody with the other isoenzyme and the phenomenon of anomalously fast migration of M-CK subunits on SDS-PAGE [13], a property shared by the largest of the NTCB fragments. CK-END1 and CK-END2 bind to the same NTCB fragments of human B-CK as CK-HTB, except that they also bind well to the smallest fragment, E (Fig. 2, lanes 6 and 7). Proteinase K 'nicks" native C K at Ala-328 [14] and subsequent SDS-PAGE separates the N-terminal M r 38000 and the C.terminal M r 5300 fragments. Fig. 3 shows that CK-END1 and CK-E.ND2 both bind to the smaller fragment, while CK-HTB binds to neither fragment. The last lane in Fig. 3 shows a control antibody (CK-JAC) which binds to the larger fragment.

CK-HTB re.cognizes an assembled, surface epitope which requires an Ash residue at position 300

Table I shows th,~t of seven isoforms (or sequence variants) of the human BB-CK immunogen, which were

1

2

3

4

5

Fig. 3. Monoclonal antibody binding to proteinase K fragments. Western blots are shown of 20% polyacrylamidegels in the SDS-phosphate buffer system of Weber and Osborn [30] with 6 M urea in the gel and 7 M urea in the sample. This buffer syo,tem gave greatly improved resolution of small protein fragments. The gel was loaded with prestained Mr markers (lane 1) and proteinase K-treatea human BB*CK(6/zg; lanes 2-5). Blots were developed with CK-END1 (lane 2), CK-END2 (lane 3), CK-HTB (lane 4) or control antibody, CK-JAC (lane 5). In addition to the main 38000 and 5300 M~ fragments, traces of undigested CK are also visible on some blots.

tested when attached to a plastic microtitre plate in an ELISA test, only chick BB and rat MM bound to the monoclonal antibody, CK-HTB. Rat BB, human MM and chick MM did not bind and neither did rabbit BB

TABLE I Monoclonal antibody binding to CK isoforms coated onto microtitre plates

Bound antibody was detected with pero~dase-labelled second antibody and the A492 of the substrate is shown (see Materials and Methods). Two separate experiments for CK-HTB and CK-END1 are shown and. in each, CK-JAC, which hinds to all CK isoforms, was used as a control for differences between the various CK coatings. In brackets, the figures are e~pressed as a percentage of the A492 for CK-JAC. (n.d., not determined). .,4492

A492

(CK-HTB)

(CK-ENDI)

Human B Chick B Rabbit B Rat B

n.d. 1.69 (85) 0.08 (4) 0.04 (2)

0.82 (100) 1.15 (107) 0.64 (119) 0.99 (108)

Human M Chick M Rabbit M Rat M

0.07 (4) 0.04 (2) 0.07 (4) 1.65 ( 8 3 )

0 0 0

(0) (0) (0)

0

(0)

Torpedo

n.d.

0

(0)

249 80,

a competition ELISA (Fig. 4). The apparent binding affinities of CK-HTB for rat MM-CK and chick BB-CK are lower than that for the human BB-CK immunogen, but of the same order of magnitude. Throughout the amino acid sequence of the 9 CKs tested, there is only one 'point mutation', or single amino acid change, that can account for th/.s striking specificity. All three CKs which bind CK-HTB antibody have an Ash residue at position 300 while chick MM-CK has a Lys and the others ha-,e a His residue (Fig. lb). The change from an uncharged Ash to a basic, charged residue is a significant one in terms of protein-protein (or antibody-antigen) interactions.

CK-HTB bound (%of conlxvl)

6(]

10

100

CK-ENDI and CK-END2 recognize two different epitopes involving the last six amino acids of the protein

1000

CK competitor (ng) Fig. 4. Competition for CK*HTBbinding between three native CK isoformsand denatured human III~CKattached to a microtitreplale. The finalconcentrationof CK-HTBantibodywas 0.4 ~g/ml.

or rabbit MM. In a separate experiment (not shown), CK from Torpedo marmorata also failed to bind CKHTB. Western blots from SDS-PAGE confirmed that CK was responsible for CK-HTB binding rather than some minor impurity in the CK preparations (results not shown). CK attached to plastic is enzymically inactive and partially unfolded [1,12] but CK-HTB also binds to the native forms of the three CKs in solution in

2.0

Table I shows that CK-END1 binds to BB-CKs from man, rat, rabbit and chicken when they are coated onto ELISA plates, but not to MM-CKs or Torpedo CK. CK-END2 differs in that it also binds to Torpedo CK. Since both bind to the last 53 amino acids of CK, the specificity of CK-END1 can be accounted for by a requirement for one (or more) of the following amino acids: Gly-332, Leu-348, Leu-359 and Met-376, while CK-END2 may require Arg-365 or Leu-375 (Fig. lb). Peptides containing the last 6, 10 or 15 amino acids of the human B-CK sequence were chemically synthesized and used to coat EHSA plates to test antibody binding. Fig. 5a shows that both monoclonal antibodies bind to the 15-met, whereas two control monoclonal antibodies,

2.0 a

Antibody bound

(A492 )

1,0

1.0 QK

JAC Eft03

%1

.Ol

.1

1

p e p t i d e e o n e n (mg/ml)

,

o

1

,

,

.Ol .1 1 p e p t i d e e o n e n (mg/ml)

lO

Fig. 5. Mot~ocl~nt antibody bLndingto C-terminal syntheticpeptides. (a) Microfitreplates were coated with dilutions in PBS of a peptide correspondingto the last 15 ami,aoacid~of ltuman I~CK. Control monoclonalantibodies were CKJAC, which binds near the centre of the CK molecule[3] and CK-END3,a third antibodywhich also binds to the last 53 amino acids of CK (results not shown).All monoclonalantibodi~, •~'ere i / l ~ dilutionsof culture supernatams.(b) Mictotitreplates werecoatedwith shorter syntheticpep!ides,a 6-met(I.MPAQK)a.d a 10-mer (AIDDLMPAQK),CK-ENDI culturesupernatantwas used at a dilutionof 1/100 and CK-END2gavesimilarresults~The effectof peptidechain len8!h on efficiencyof coating to plastic was not studied, so quantitativecomparisonsbetweenpeptides should be avoided. Control ~mibodyo CK-$AC,did not bind to thesepeptides(resultsnot shown).

250 CK-JAC and CK-END3, do not bind. Fig. 5b shows that binding can also be achieved with the 6-mer, LMPAQK. Although negative results are inconclusive because some peptides may be unable to mimic epitopes, the corresponding 6-mer from the human M-CK sequence, MIPAQK, was synthesized and failed to bind either antibody (resulis not shown). The results show that CK-END1 and CK-END2 recognize epitopes which include the C-ternfinus and that they require Met-376 and Leu-375, respectively. In contrast to CK-HTB, CK-END1 and CK-END2 will b.;nd only to inactive, denatured and partially unfolded CK whether after urea treatment, attached to plastic for ELISA or on Western blots after SDS-PAGE. In competition ELISA using these antibodies, there is no effect of native CK on antibody binding (results not shown) as previously described for the monoelonal antibodies CK-2A7, CK-SH5 and CK-JAC [1,2] (cf. Fig. 4 for CK-HTB). Discussion

The results suggest that the native conformation around the extreme C-terminus of CK is readily disrupted during mild denaturation, but that there are other parts of the C-terminal half of the molecule which are more resistant to disruption. The latter conclusion is based on evidence that one antibody, CK-HTB, recognises an assembled su.q~ce structure ('conformational epitope') on CK, but will also bind to several 'denatured' forms of CK. The most likely explanation for the clear-cut effects of changes at residue 300 on CK-HTB binding (Table I) is that Asn-300 is involved in direct interaction with the at:tibody, in which case it must be exposed on the enzym~ surface. Alternatively, the change from Asn-300 to a charged residue could change the disposition of other amino acids which actually form the epitope, either locally or even at some remote area on the enzyme surface, but this explanation does not stand up well to closer scrutiny. Large-scale structural differences between isoforms produced solely by naturally occurring variations at residue 300 seem unlikely in view of the sensitivity of CK catalytic activity to conformational change [15] and the persistence of the specificity produo:d by Asn-3,~ i,. partially unfolded CKs. Even a local structural effect oF Asn-300 seems unlikely, since there are significant adjacent differences between rat M-CK and human B-CK (Ala-299 to Pro; Set-302 to GIy; Fig. lb) which might also tend to disrupt secondary structures [16] and yet they do not prevent CK-HTB binding. In addition to Asn-300, formation of the CK-HTB epitope requires residues which precede Cys-282 and follow Ala-328, since CK-HTB binds neither to NTCB fragment 'E' nor to the large proteinase K fragment

(Fig. 1). This suggests that the partially unfolded and inactive forms of CK to which CK-HTB binds have retained more than just local folding in their C-terminal regions. The extent of similarity to the native conformation is not clear, however, since antibodies can accelerate local refolding by stabilising the epitopes they recognise [17], so the HTB epitope need not be perfectly reformed on partially unfolded CK until after CK-HTB antibody is added. On the other hand, CK-HTB does bind well to partially u:lfolded CK even after its attachment to solid phase (nitrocellulose or polyvinylchloride) which may well restrict antibody-induced conformational changes; certainly those conformationai changes involved in catalytic activity seem to be prevented, since CK is completely inactivated when attached to microtitre plates [12]. The idea of a C-terminal region which is mere stable or refolds more readily than the rest of the molecule is consistent with trypsin studies. Native CK is very resistant to proteolysis but trypsin rapidly degrades the N-terminal part of the urea-treated folding intermediate, leaving a 20 kDa C-terminal resistant 'core' [2]. However, digestion of SDS-treated CK by staphylococcal proteinase V8 occurs most rapidly at Glu residues in the centre of the molecule and differences between N-terminal and C-terminal regions are less clear-cut [3]. We have not yet encountered any instances of the N-terminal region showing greater proteinase-resistance than the C-terminal region. The other monoclonal antibody which recognises both native and denatured CK, CK-ART, is specific for chick M-CK which has a Lys residue at position 300, one of only three unique residues which could account for CK-ART specificity [2[ Even though the ART epitope is retained on smaller fragments of CK than is the HTB epitope (Fig. 2). it remains an intriguing possibility that the two antibodies are recognising homologous surface structures, but this cannot be tested directly since there is no CK isoform which both antibodies recognise. Proteinase K cleavage at Ala-328 has a similar effect on both ART and HTB epitopes; namely, that the epitope is retained while the two fragments remain together but is lost completely when the two fragments are separated by SDS or urea treatment (Ref. 2 and results not shown). We have shown that the END1 and END2 epitopes near the C-terminus of CK are inaccessible to antibody in native CK but are easily exposed. Studies with carboxypeptidase are also consistent with an inaccessible C-terminus. This enzyme can remove the last two amino acids from native CK, without any effect on enzyme activity, but further digestion is difficult [18]. We have suggested that four other buffed epitopes on CK may be located at a possible interface between N-terminal and C-terminal domains and become inaccessible only at a late stage of refolding involving

251 domain association [4]. It is not inconceivable that CK may 'tuck in its tail' as part of such a domain association process. In phosphoglycerate kinase, for example, a functionally analogous enzyme of known three-dimensional structure, the last ten amino acids cross-over from the C-terminal domain to interact with the Nterminal domain [19] and such an interaction could depend on domain association or even be an essential part of it. Acknowledgements W e thank the Muscular D y s t r o p h y G r o u p of G r e a t Britain a n d N o r t h e r n Ireland a n d C a m b r i d g e Life Sciences plc for financial s u p p o r t and Celia M. Yates ( M R C Brain M e t a b o l i s m Unit, Edinburgh) for help a n d advice on the isolation o f brain enzymes.

References 1 Morris. G.E., Frost. L.C., Newport, P.A. and Hudson. N. (1987) Biochem. J. 248, 53-59, 2 Morris, G.E. (1989) Biochem. J, 257, 461-469. 3 Morris, G.E. and Cartwdght. A.J. (1990) Bioehim. Biophys. Acta 1039, 318-322. 4 Nguyen thi Man, Cartwright, AJ.. Andrews, K.M, and Morris. G.E. (1989) J. Immunol. Methods 125. 251-259. 5 Jaenicke, R. (1987) Prog. Biophys. Mot. Biol. 49, 117-237, 6 Eppenberger, H.M., Dawson. D.M. and Kaplan, N.O. (1967) J. Biol. Chem. 242. 204-209. 7 Stark, G. (1977) Methods EnzymoL 47. 139-142. 8 Williamson. J., Greene, J., Cherif, S. and Milner-White, EJ. (1977) Bieehem. J, 167, 731-737. 9 Morris, G.E. and Head, L.P. (1982) FEBS Lett. 145, 163-168. I0 Morris, G.E., Frost, L.C. and Head, L.P. (1987} Biochem. J. 228, 375-381.

11 Atherton, E. and "Sheppard, R.C. (1981) J. Chem. So¢. Chem. Commun. 165-166, 12 Morris, G.E. and Head, L.P, (1983) Biochem. J. 213, 417-425. 13 Perriard, J.-C., Perriard, E. and Eppenberger, H.M. (1978) J. Bid. Chem. 253, 6529-6535. 14 Lebherz, H.G., Burke, T,, Strickler. J.E. and Wilson, KJ. (1986) Biochem.J. 233. 51-56. 15 Watts, D.C. (1973) Enzymes 3rd Edn. 8, 348-455, i6 Chou, P.Y. and Fasman, G.D 11978} Adv. Enzymo]. 47, 45-148. 17 Chavez, L.G. and Benjamin, D.C. (1978) J. Biol. Chem. 253, 8081-8086. 18 Olson, O.E. and Kuby, S,H. (1964) J. Bic! Chem. 239, 460-567. 19 Banks, R,D., Blake, C,CF., Evans, P.R., Haser, R., Rise, D.W,. Hardy, G,W., Merrett, M. and Phillips, A.W. (1979) Nature 279, 773-777. 20 Kwiatkowski, R.W., Sehweinfest. C.W. and Dottin. R.P. (1984) Nucleic Acids Res. 12. 6925-6934. 21 Perryman, M.B., Kerner, S.A., Bohlmeyer, T.J. and Roberts. R. (1986) Biochem. Biophys. Res. Commun. 140. 981-989. 22 Pamey, S., Herlihy, W., Royal, N, Pang. H., Aposhian, H.V., Pickering, L. Belagaje, R., Biemann, B.. Page, D., Kuby, S. and Schimmel, P. (1984) J. Biol. Chem. 259. 14317-14320. 23 Benfield, P.A., Zivin, R.A., Miller. L.S,, Sowder, R., Smythers, G.W., Henderson, L., Oroszlan, S. and Pearson, M.L. (1984) J. BioL Chem. 259, 14979-14984, 24 Hossle, J.P., Rosenberg, U.B., Schafer, B., Eppenberger, H.M. and Perriard, J.-C. (1987) Nucleic Acids Res. 14, 1449-1463. 25 VillarreaI-Levy. G., Ma, T.S,, Kerner. S.A., Roberts, R. and Perryman, M.D, (1987) Biochem. Biophys. Res. Commun. 144, 11161127. 26 Picketing, L, Pang, H., Biemann, K., Munro, H. and Schimmel, P. (1985) Proc, Natl. Acad, Sci. USA 82, 2310-2314. 27 Benfield, P.A., Henderson, L. and Pearson, M.L (1985) Gene 39, 263-267. 28 Giraudat, J., Devillers-Thiery, A., Perriard, J.-C. and Changeux, J.-P. (1984) Proe. Natl, Acad. Sci. USA 81, 7313-7317. 29 Laemmli, U.K. (1970) Nature 227, 680-685. 30 Weber, K. and Osbom, M. (1969) J. Biol. Chem. 244. 4406-4412.