THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 19, Issue of May 13, pp. 14199-14204, 1994 Printed in U.S.A.

Mapping of the Immunophilin-Immunosuppressant Site of Interaction on Calcineurin*

(Received for publication, September 7, 1993, and in revised form, January 11, 1994)

Holger HusiS, Marcel A. LuytenO, and Mauro G. M. ZuriniS From the Laboratories of $Biochemistry and $Molecular Biology, Preclinical Research, Drug Design Group, Sandoz Pharma Ltd., 4002 Basel, Switzerland

The interaction of the immunosuppressive complexes cyclosporin A-cyclophilin A and FKSO6 binding protein- FKS06 with the Ca2+- and calmodulin-dependent protein phosphatase calcineurin has been investigated by means of photoaffinity labeling and chemical cross-link- ing. Photolabeling of purified bovine brain calcineurin with the affinity label [0-[4-[4-(1-diazo-2,2,2-trifluoro- ethyl)benzoyl]aminobutanoyll-~-serine~]cyclosporin in the presence of cyclophilin A results, in addition to the labeling of cyclophilin itself, in the transfer of some of the chemical probe to both the catalytic subunit A and the regulatory subunit B of calcineurin. Chemical cross- linking studies with disuccinimidyl suberate in the pres- ence of either cyclophilin A, B, or C in complex with cyclosporin A or FKSO6 binding protein-FKSO6 result on the other hand in the apparently exclusive and strictly immunosuppressant-dependent formation of covalent immunophilin-calcineurin B subunit products. Cross- linking of immunophilins to calcineurin B subunit re- quires the presence of subunit A. In the present study, using a set of recombinant maltose-binding protein fu- sion products representing different stretches of the catalytic subunit A, we were able to map the minimal calcineurin A sequence necessary for immunophilin- ligand-calcineurin B interaction to occur.

The immunosuppressive effects of both cyclosporin A (Sand- immun) and the structurally unrelated macrolide FK506 arise from their ability to prevent T-cell proliferation by interfering with early Ca’+-dependent events following T-cell stimulation and resulting in the suppression of the transcriptional activa- tion of the interleukin-2 gene (1-3). Main intracellular recep- tors to which these drugs bind with high affinity are a growing family of proteins collectively named immunophilins: cyclophi- lins for cyclosporin and FK506 binding proteins (FKBPs)’ for FK506. Although cyclophilins and FKBPs do not share any sequence and structural homology, members of both protein families have been found to possess peptidyl-prolyl cis-trans- isomerase activity (4, 5 , 7-10)., Binding of CsA or FK506 to their cognate proteins results in the inhibition of such activity. The first and most abundant cyclophilin and FKBP discovered are CypA (18 kDa) (4) and FKBPl2 (12 kDa) (81, respectively.

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “uduertisemnt” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

cyclosporin A , CypA, -B, and -C, cyclophilin A, B, and C; DSS, disuccin- The abbreviations used are: FKBP, FK506-binding protein; CsA,

imidyl suberate; PL-Cs, photolabile [0-[4-[4-(1-diazo-2,2,2-trifluoroeth- yl)benzoyllaminobutaoyll-~-serine~lcyclosporin; PMSF, phenylmethyl- sulfonyl fluoride; Cln, calcineurin; Cter., carboxyl terminus.

Schneider, H., Charara, N., Schmitz, R., Wehrli, S., Mikol, V., Zurini, M. G. M., Quesniaw, V. F. J., and Mowa, N. R. (1993) Biochemistry, in press.

Both proteins are ubiquitous and located in the cytosol. Inter- estingly, the inhibition of their isomerase activity by their cog- nate ligands is not sufflcient to suppress the immune response (11, 12). It is now known that these different immunophilin- ligand complexes commonly interact with the Ca2+/calmodulin- dependent serinekhreonine protein phosphatase calcineurin, thereby inhibiting its activity (13). A direct correlation between immunosuppression and modulation of calcineurin activity has been found (14-16). The three-dimensional structures of CypA and FKBPl2 are known (17-20). Other known, closely related, cyclophilins include CypB (22 kDa) (51, resident in the endo- plasmic reticulum, and the more tissue-specific CypC (23 kDa), found to be particularly abundant in kidney (21). Recently, a number of proteins containing within their sequence Cyp- or FKBP-like domains have been described, most notably, Cyp-40 and p59, respectively, both of which might be components of the inactive steroid receptor complex (7, 10, 22). Although the ul- timate function of immunophilins or of immunophilin-like do- mains remains uncertain, enhancement of the rate of folding of model proteins through their peptidyl-prolyl isomerase activity has been reported (23,241. Of evident pharmacological interest is, however, their involvement in mediating the action of the macrocyclic immunosuppressants CsA and FK506. The finding that calcineurin is a common target for both the very dissimilar complexes CsA-CypA and FK506-FKBP12 (13) has provided a quantum leap in our understanding of the mechanism of action of these immunosuppressants. Calcineurin is a heterodimer consisting of a large calmodulin-binding catalytic subunit A (61 kDa) and a smaller, NH, terminus myristylated, Ca2+-binding subunit B (19 kDa) showing 35% identity to calmodulin. To date, sequence stretches corresponding to the calmodulin-bind- ing site (25), an autoinhibitory domain (261, which is a typical feature of calmodulin-dependent enzymes, and a putative sub- unit B binding site (27) have been identified within the large subunit A. As with many calmodulin-dependent enzymes, cal- cineurin can be permanently activated by a mild proteolytic treatment with trypsin (28) resulting in the removal of the calmodulin-binding and autoinhibitory domains. Recent work has shown that such a shortened enzyme is still sensitive to the inhibitory actions of immunophilin-ligand complexes (29) and that for inhibition to occur, the presence of subunit B is re- quired (30). Moreover, using chemical cross-linking, Li and Handschumacher (31) have demonstrated immunosuppres- sant-dependent immunophilin interaction with subunit B and that such interaction is absolutely dependent on the presence of subunit A. In this study we further characterize the nature of the immunophilin-ligand-calcineurin interaction and present evidence indicating that immunophilin-bound immunosup- pressants recognize a sequence within subunit A.

EXPERIMENTAL PROCEDURES Materials-Human recombinant cyclophilin A (321, B (5), and C,2

human recombinant FKBP (331, bovine brain calcineurin (34) and cal-

14199

14200 Immunophilin-Ligand Complex Interaction with Calcineurin modulin (Sf,), polyclonal sera against the different immunophilins and calcincurin, and anti-cyclosporin A monoclonal antihodies were prr- pared and purified in house hy established procedures. CsA and CsH, FK506. and the photolabile 10-14-144 l-diazo-2,2,2-trifluoroethyl )henzo- yllaminoI,utanoyll-~~-serine”lcyclosporin (PL-Cs) were ohtaincd from Sandoz’s immunology department. Disuccinimidyl suheratc tDSS) was from Pierce Chemical Co. The ECL Western blotting detection system and secondary horseradish peroxidase-linked antibodies were from Am- ersham, UK. All other reagents were of the highest purity availnhle.

Cnlcinrurin R Scrhunil-Calcineurin B suhunit was prepared by heating purified bovine brain calcineurin for 5 min a t 95 ’C. The heat- labile calcineurin A suhunit was then removed by centrifugation in an Eppendorf microcentrifuge at full speed for 10 min. Alternatively the subunit was prepared hv gel filtration of purified bovine brain cal- cineurin in the presence of 3 Y KSCN.

I%otonffinit.v I,nhrling-Rovinc hrain calcineurin was photolaheled in the presence of humnn recomhinant cyclophilin A essrntially as de- scribed in Ref. 36. Qpically, equimolar concentrations of calcineurin and cyclophilin A (1 par) were incuhated for 15 min a t 4 “C in the presence of 1 p~ I’I,-Cs in a huffer containing 20 m.1 Hepes pH 7.4, 100 m51 NaCI, and 50 par CaCI, and exposed then to a UV light source ( U V lamp model R-1004, Kontron, Zurich, Switzerland) positioned a t a dis- tance of 10 cm for 5 min. Aliquots were then separated on 15% SDS- polyacrylamide gels according to Laemmli (37). Proteins were trans- ferred onto nitrocellulose by diffusion (38). and hlots were prohcd with monoclonal anti-CsA antibodies. Signals wrre revealed with the ECI, system according to the manufacturer’s instructions on Kodak XAR-5 films.

Cross-linhing-Cross-linking was performed in a final volume of 100 pl in a buffer consisting of 20 mu Hepes pH 7.2, 100 mM NaCI, and 0.5 mar CaCI,. The amounts of proteins and ligands are given in the cap- tions of the different figures. The protein mixtures (with or without ligands) were incuhated for 30 min a t room temperature. Cross-linking was initiated by the addition of DSS to a final concentration of 0.1 mu (stock 100 mar in dimrthyl sulfoxide), and the mixture underwent reac- tion for 15 min at 37 “C. The reaction was stopped by the addition of Tris-HCI. pH 6.8, to a final concentration of 100 mbr. Aliquots of the reaction mixture were then separated hy 15Q SDS-polyacrylamide gel electrophoresis (37) followed hy Western blotting (38) and detection with polyclonal anti-irnrnunophilin and/or anti-calcineurin antihodies with the ECI, system on Kodak XAR-5 films.

Cloning nnrl E.rprrssion of (,‘nIcinrurin A Fmgmrnts-The short frag- ments of calcineurin used in this study were prepared hy polymerase chain reaction from a functional, full-length human calcineurin A c r l clone. Primers were derived from the puhlished partial human cal- cineurin A n l cDNA (39). Standard polymerase chain reaction condi- tions with Amplitaq were used as recommended by the supplier. The DNA fragments were purified by agarose gel rlectrophoresis. digested with EcoRI and HindIll, ligated into the pMnI-c2 vector (New England Riolahs). and electroporated into Eschrrichio coli EC682 cells. For pro- tein expression, crlls were grown in rich yeast extract medium (20 glliter), induced with 1 mar isopropyl-l-thio-B-l,-galactopyranoside a t 0.5 ANa, and grown for an additional 5 h.

I’urifirntion of Mnltosr-hinrling Protein Fusion Protrins-Maltose- binding protein fusion products were purified hy amylose aFfnitv chro- matography followed hy gel filtration on Superdex 200 H i h a d ( P h a r - macia LKR Biotechnology Inc.). Briefly, 10 g of recomhinant E. coli pellet were resuspended in 50 ml of 50 mar Tris, pH 7.4, 1 mu EDTA, 2 ma4 ~-mercaptoethanol. and 1 mM phenylmethylsulfonyl fluoride, and the cells were disrupted hy sonication. After centrifugation, the clear supernatant was passed over an amylose resin (1.6 x 20 cm; New Eng- land RioLahs. Beverly, MA). The column was then washed to base line with the above huffer and eluted with the same huffer containing 10 mM maltose. The eluted material was then concentrated hy ultrafiltration (AmiconYM-10 membranes) and gel-filtered in 20 r n Y Hepes pH 7.4 and 100 mar NaCI.

RESULTS Photoaffinity Lahrling of Calcinrurin-The use of the PL-Cs

for the characterization of cyclosporin-binding proteins has been described (36). The recent structural studies on the cy- closporin A-cyclophilin A complex (18) showed position 8 of cyclosporin not to be buried within the cyclophilin molecule, and provided a long flexible side chain is introduced at this position, this latter will he available for interaction with either cyclophilin itself or with binding partners recognizing the im-

k On 1 2 3 4 5 6 7

97 - 66 -

46 -

3 1 -

- =* *4-

FIG. 1. Photoaffmity laheling of purified calcineurin. Exprri- mental conditions are givrn undvr “F:xprrimrntnl I’rocedurr*s” and in the text. Signals wwr drtectcd with mcrnocltrnal ;mti-(‘sA :~ntilx)rlirs. IAnr 1 , 1 p~ rach of calcinrurin, cyclophilin A, cnlmndulin. and P I A ’ S : photolysis was done in the prrsencc’ of 100 pv E(;TA. 1 ~ n r . v 2-1. 1 p v rach of calcineurin, cyclophilin A. and calmodulin: phot~rlysis w a s clone in the presence of 1, 3 , and 10 p~ P I A ’ S , rrsprctivc.ly. 1,crnr.s 5 :~nrl i;. 1 phi each of cyclophilin A. calmodulin, and I’l,-(*s: photolysis W:IS (111nv in the presence of3 and 10 p~ calcinrurin, rcqwctivrly. fnrw 7. I pv cwch of calcinrurin, calmodulin. and P I A k ; photolysis was donr in t h v ah- sence of cyclophilin A. Moleculnr mass st;~ntlnrds arv shown trr th r lrft.

munophilin-ligand complex. When CypA is photolnbc~l(d in thc, presence ofcalcineurin, the main product is laheled cyclophilin. Minor amounts of Iahel are, however. specifically incorporated into calcineurin in a Ca”- and cyclophilin-dc,pc.ndcnt way. In the experiment shown in Fig. 1. an equimolar mixture of cal- cineurin, cyclophilin A, and calmodulin ( 1 pv clach was photo- labeled in the presence of increasing amounts of PI,-Cs ( lanrs 241, or alternatively, increasing amounts of calcincwrin wcrc’ photolabeled in the presence of fixed amounts of cyclophilin. calmodulin, and PL-Cs (1 p>r each; lnncs 5 and 6 ). In both cases. Western blot analysis of the reaction products hy monoclonal anti-cyclosporin antihodies reveals. besides the strong signal for cyclophilin a t about 20 kDa, covalent cyclosporin incorpo- ration in both calcineurin suhunits (arrows ). Interestingl.v, con- centration-dependent Iaheling is evident only for subunit A, whereas subunit R labels in a comparatively strongcv- and rc1la- tively concentration-independent manner. This is particularly evident in lnnr 2. Both signals are, however. completc.ly quenched when photolabeling is performed in the. prcwnce of 100 p~ EGTA (lonr I ). Under these conditions the lahel incor- porates only into cyclophilin. The incorporation of PLCs into both calcineurin subunits is absolutely dependent on the pres- ence of cyclophilin because no signals can he detected after photolysis in the absence of the latter at equimolar concrntra- tion (1 p” of all other reactants (lanr 7 1. These results suggest that the interaction site for the solvent-c.xposcd domain of thc. immunophilin-bound cyclosporin lies at or very near the inter- face between the two calcineurin subunits. Subunit R seems to react more readily with the PL-Cs.

Cross-linking of Immr~nophilin-I,i~nntt Complr.rrs t o Cal- cinmrrin-The nature of the calcineurin-immunophilin-ligand interaction was further investigated by means of chemical cross-linking with the homohifunctional reagent DSS. In thc experiment shown in Fig. 2, 50 pmol of pur i f id bovine hrain calcineurin were incuhated with equimolar concentrations of either CypA, CypR, CypC, or FKRP in the presence of their cognate ligands CsA or FK506 and treated with 0.1 mv DSS as described under “Experimental Procedures.” Immunodetcction with anti-calcineurin antibodies revealed bands at . X 3 (Innr 2 ), 37 (lonr 3 ) , 38 (lnnr 4 ) . and 27 kDa (lonr, .=I I. which can only

Immu.nophilin-Ligand Complex Interaction with Calcineurin

1 2 3 4 5

31 t c . . -

FIG. 2. Cross-linking of different immunophilins to cineurin. Calcineurin (0.5 p\!) W:IS incuhntrd with rquimolar concrn- tr:ltions of e i thr r CypA. CypR. CypC. or FKRP in the prrsrncr o f sntu- rating amounts of thrir cognatr ligands CsA or FK5OG I 1 0 V\I rach , and cross-linkctl with 1)SS as drscrihrd undrr "Exprrimcntnl I'rocrtlurrs." Signals wrrr tlrt.rct.rd with anti-calcinrurin antihodirs. 1,nnr 1 . cal- cinrurin and FKHI' cross-linking in thr nhsrncr of FK5OG; lnnrs 2-5, calcinrurin cross-linking with C'vpA. Q p R . CypC, and FKRP in the prrsrncr o f t h r i r cognatc, ligands. Molrcul:~r mass s tandards arr shown to t hc lrfi .

correspond to the cross-linking products of CypA (18 kDa), CypR (22 kDa), CypC (23 kDaI, and FKRP (12 kDa) to the B subunit of calcineurin. The calcineurin R subunit itself (19 kDa) runs on SDS gels with an apparent molecular mass of 15 kDa. The formation of these immunophilin-calcineurin €3 cross- linking product,s is absolutely dependent on the presence of the specific immunophilin ligands, as is best exemplified in lone I , where FKRP and calcineurin were cross-linked in the absence of FK506. It is not clear whether the different intensities ob- served for the different immunophilin-calcineurin R cross-link- ing products (FKBP >> CypB = CypC > CypA) are the result of differential distribution of reactive amino groups or rather the consequence of different affinities. It has, however, been re- ported that the complex CypR-CsA is 3-5-fold more potent than its CypA counterpart in inhibiting the phosphatase activity of calcinrurin (29). On the other hand. the exact quantification of the cross-linking efficiency of the different immunophilins to the subunit R of calcineurin and the direct comparison between the different intensities observed are very difficult because these cross-linking products are transient and their amounts variable from experiment to experiment as a result of parallel cross-linking of calcineurin subunit A to subunit R. Fig. 2 shows, in fact, extensive cross-linking products of high molecu- lar masses r a n ~ n g from 80 to above 100 kDa. Additional cross- linking of immunophilin-ligand complexes to suhunit A of cal- cineurin cannot be ruled out by the experimental approach used because the cross-linking product of the two calcineurin subunits would run at about the same molecular weight. Higher cross-linking products correspond to the expected tri- meric subunit A-subunit R-immunophilin complex and/or to higher aggregates.

The experiments shown in Fig. 3 confirm tha t the cross- linking products of intermediate molecular weight are indeed the result of covalent immunophilin-subunit B interaction and that this interaction only occurs in the presence of the appro- priate immunophilin ligand. This is demonstrated by probing replica blots with both anti-calcineurin (A and C ) and anti- immunophilin (h' and D ) antibodies. The band of about 33 kDa visihle in Fig. 3 ( A and I?. lone 1 ) corresponds to the cross- linking product of CypA with calcineurin R (note that the dif- ference in intensities between Fig. 2, lonc. 2 and Fig. 3A is simply due to longrr exposure). Analogously, the band a t 27 k n a visible in Fig. 3 ( C and D, l o w 1 ) corresponds to the

* O m

97

- 46

- 66

-

31 -

21 -

l a -

k O n

97 - 66 - a6 -

31 -

21 - l a -

1 2 3

L

. I

A B

1 2 3

C

cross-linking product of FKRP with c:llcincwrin R. The tlrpt'nd- ence of the reaction on the presrncr of ('SI\ or FIG06 is clrm- onstrated in Fig. 3. A and /{ rlonc 2 I antl (' :lnd /I c / n n r , 2 1,

respectively, where these were omittrd. Fig. 3 ( A :lnd I { , / o n ( , .? I demonstrates the inabilitv of purified calcintwrin I3 suhtlnit tn interact with CypA-CsA in the ahsrncr of t h r catnlytic suhunit A, as independently drmonstratrcl bv I,i and Hnndschum;lchrr in a recent work (31 J. The same was found to hr tnw for FKl3P- FK506 (result not shown,. Finally, Fig. 3 I ( ' nnd I ) . I n r r r . .'I I

emphasizes the positive effect of calmodulin on thr rx t rn t nf FKRP-calcineurin R cross-linking, in analom to thr findings of Li and Handschumacher (31 I for CypA. T h r intrnsivcb tnntls at 18 and 12 kDa in 13 and D represent uncomplrxtd :Inti FKRP, respectively, whrrras the hand at about 24 k1)n in / I corresponds to FKRP dimers. Again, ciirrct c.vidcncc. of addi- tional cross-linking h t w r r n immunophilins a n d thf. l:lrKrr tal- cineurin subunit A is lacking. becausr no particular rrlrvant signals are found at the corresponding molrculnr tvchight usinc anti-immunophilin antibodies. On the othrr hand. poor tr:lns- fer efficiency for high moltwJlar wright sprcic.?; cannot t ) < b t o - tally ruled out.

Cross-linking nf Co/rinc~rrrin R to ~ r l ~ r l ~ r r r ~ o p h i / i t ~ . s I n t h o /'rvs- cnce of Short Rccomhinnnt C'olcinvurIn A S l r c ~ l ~ ~ i ~ c ~ . ~ - - l ~ ~ ~ c ; l u . s c ~

14202 Immrtnophilin-Ligand Complex Interaction with Calcinertrin

both calcineurin subunits are seemingly needed in order to obtain subunit R-immunophilin cross-linking products (Fig. 3, A and R, Ian(. 3 1, the role of subunit A has been further inves- tigated with the help of a number of maltose-binding protein fusion products representing different stretches of the large catalytic subunit. Different fusion proteins were incubated with purified calcineurin R subunit, obtained either by heat denaturation or gel filtration in KSCN, and FKBP-FK506. The mixture was then treated with DSS as described. Of the differ- ent constructs tried, only those containing the minimal subunit A sequence comprised between residues 332 and 390 gave a subunit R-FKBP cross-linking product (for sequence number- ing see Ref. 39). This sequence starts with the putative cal- cineurin R-binding site (residues 332-357; see Ref. 27) and ends just at the beginning of the published sequence for the calmodulin-binding site (2.5). Fig. 4, lane 3 , shows anti-cal- cineurin immunodetection of the calcineurin R-FKRP cross- linking product (27 kDa) obtained by treating with DSS a maltose-binding protein fusion product with the minimal cal-

stretch in the presence of the FKRP-FK506 complex and purified calcineurin B. A comparatively stronger signal is obtained with a fusion product corresponding to the calcineurin A stretch beginning with residue 332 and spanning all the way to the carboxyl terminus (lane 2 ) . The latter con- struct contains the calmodulin-binding site (25) and gives even stronger signals in the presence of calmodulin (result not shown). The ahsence of any cross-linking product in lanc I (as compared with lane 2 ) demonstrates the absolute requirement of FK506 for subunit B-FKBP interaction.

,,ineurin A:I:<2-:Vl~I

k O n 1 2 3

46 -

3 1 -

2 1 -

1 4 -

FIG. 4. Cross-linking of purified calcineurin B subunit to FKRP-FK506 in the presence of short recombinant calcineurinA stretches. 1 0 pg c;lch of two different maltose-hintling protein fusion products representing calcineurin A residues 3.72-390 I C I ~ A " ' " I " ) and 332 through the carboxyl terminus (ClnA'"""' '~ were incubated in a final volume of 100 pl with 2 pg of purified calcineurin R suhuni t and 2 pg of FKRP in thr presence or ahsence of 2 p51 FK506. Signals were detected hy anti-calcineurin nntihodies. I ~ n r 1 , cross-linking of CRI- cineurin R , FKBP, and ClnA:"'2 '"cr in the ahsence of added FK506. Lane 2, Cross-linking of calcineurin R , FKBP, and ClnA''2' I"' in the presence of FK506. Lnrw 3 , Cross-linking of calcinrwrin B, FKBP, and ClnA'" ,'""'

lrft . in the presence of FK506. Molcmhr mass s tandards are shown to the

DISCUSSION

In the current study we have used two indrpc.ndrnt ap- proaches, photoaffinity labeling and chemical cross-linking, in order to explore the exact topology of the immunophilin-ligand interaction with calcineurin. The first approach, affinity lahrl- ing with a photolabile cyclosporin derivative in the prrsrnce of CypA, clearly shows specific immunophilin- and Ca"-depend- ent insertion of the probe into both calcineurin subunit A and R but with markedly higher reactivity toward the smaller sub- unit. This uneven distribution of the label betwern the two subunits is certainly not due to a differential distribution of reactive side chains, because the carbene radical generated upon irradiation of the photolabel can react even with aliphatic carbon atoms. Rather, the geometry of the interaction is likely to be responsible, indicating that residue 8 of cyclophilin-hound photolabile cyclosporin is oriented somewhat nearer to subunit R. Unfortunately no other photoaffinity lahel congenrrs of CsA could be tried, because derivatization of CsAwith photoactivat- able groups other than in position 8 results in drastic decreases in affinity toward either the immunophilin or the calcineurin moiety. Photolabeling was also attempted on the recombinant maltose-binding protein-ClnA."'" 'v"' reconstituted with purified subunit R. Unfortunately UV treatment of the fusion protein leads to its partial oligomerization and to unsprcific. non- quenchable photolabel insertion.

Cross-linking experiments, on the other hand, rrsult primar- ily in the covalent attachment of immunophilins to calcineurin R. This cross-linking is, however. ahsolutrly dependent on the presence of a minimal calcineurin R-interacting strrtch of suh- unit A, identified as residues 332-390. In the recent work of Li and Handschumacher ( 3 1 ~ , using a similar cross-linking tech- nique, the authors conclude that the primary site of intrraction of immunophilin-ligand complexes resides on calcineurin suh- unit R and hypothesize the possibility that the larger A subunit might be needed in order to assure the proper conformation of subunit B. We favor a different interpretation on the basis of the following points. f i ) Photolabile cyclosporin inserts into both calcineurin subunits. ( i i ) In our effort to further map the minimal calcineurin A stretch needed for immunophilin-suh- unit B cross-linking to occur. we synthesized n number of dif- ferent peptides covering the whole identified minimal stretch (ClnA"" :""'). None of the peptides was able to induce cross- linking, even though a peptide that corresponded to the puh- lished putative minimal subunit R binding site (271 efficiently cross-linked to purified subunit B, pointing possibly to a func- tional conformation of the latter. Interestingly, peptides con- taining the region immediately adjacent to the putative R hind- ing site also cross-linked to subunit R although to a lesser extent (Table I,. This seems to imply that the suhunit R binding stretch is a little hit longer than tentatively proposed earlier (27) and could comprise most of the region hetwern the *cata- lytic core" of the enzyme and the calmodulin-binding domain (6). ( i i i ) The immunophilin-ligand complex cross-links to a

TAIII.E I SlrhrnnppinK o r t h e minimnl sfretrh ClnAJ"'

Four different peptities representing different regions of the minimal stretch ClnA'" ""I were probed for promotion of immunrrphilin-ClnR cross-linking (column A ) or f i r direct cross-linking to ClnB (column 13). The experimental conditions are essrntinlly as d r s c r i h d ~n thv i(yend for Fig. 4 except that 2 pg of thr different peptides were used instead of the fusion protein. Only the full CInA''" "*' st r r tch 1s ahlr to promote cross-linking of ClnB to the immunophiiin. Peptidr 1 represents the putative ClnH-binding sequence ,271. The peptidrs wrre synthesized hy convention:~l solid-phase chemistry and purified by reverse phase chromatogrnphv. Other dcstails arc* g i v m in the tcaxt.

~~~ ~ ~~~ ~~~ ~ ~ ~~~

Srqucmcr . ~-~~ ~~ ~~ ~~ ~ ~ ~~ ~~ ~ ~

A H ~~~~~~~ ~- ~~

1 2 QFNASPHPYWLPNFMDVFTWSLPFVGEK 3 KVTEMLVNVLSIASDDELMTEG 4 KVTEMLVNVLSIASDDELMTEGEDQFDGSAAARK - + 5 DVFTWSLPFVGEKVTEMLVN - +

~~

(MBP)-QFNCSPHPY~PNFMDVFTWSLPFVGEKVTEMLVNVLSICSDDELMTEGEDQFDGSAAA + ++ ++ +

- -

~~~~- . ~~ ~ ~~

Immunophilin-Ligand Complex Interaction with Calcineurin 14203

FIG. 5. Model of immunophilin-im- munosuppressant mode of interac- tion with calcineurin. Both calcineurin subunits contribute to the binding of im- munophilin-ligand complexes. The bind- ing site resides a t or very near the inter- face between the two subunits. The main interacting feature is the solvent-exposed and conformationally stabilized immuno- suppressant moiety. bdg, binding; CaM, calmodulin.

Immunosuppressant

higher extent to subunit B of calcineurin in the presence of calmodulin or if the enzyme is permanently activated by lim- ited proteolysis with trypsin (31); a direct effect of calmodulin- calcineurin A interaction on the conformation of subunit B is not very likely. We postulate instead that both calcineurin sub- units contribute to the binding of immunophilin-ligand com- plexes and that the binding site resides at, or very near, the interface between the two subunits.

Even though a few authors have reported that a number of immunophilin surface residues are involved in the formation of the immunophilin-ligand-calcineurin complex, as studied in the case of the FKBP-FK506 model (40, 411, it is known from structural work (17-20) that binding of both CsA and FK506 to their cognate proteins does not induce any major conforma- tional change. Because immunophilins alone do not interact noticeably with calcineurin it is fair t o assume that the main interacting feature is the solvent-exposed and conformationally stabilized immunosuppressant moiety itself. The inhibition of calcineurin phosphatase activity might be explained as the re- sult of a major conformational stress induced by a wedge-like insertion of immunophilin-ligand complexes, with the immu- nosuppressant on its tip, between the two subunits (Fig. 5). The lack of detection, in our study, of any direct immunophilin- subunit A cross-linking product, might be the result of the type of cross-linker used (DSS; about 11 A span). In the work of Li

and Handschumacher (311, however, theoreagent ethylene gly- colbis(succinimidy1 succinate) (about 16 A span) was used, and in their cross-linking experiment on trypsinized calcineurin a fairly intense doublet in the region spanning between 50 and 80 kDa is visible. The lower of the two bands can only correspond to a cross-linking product between CypA and calcineurin sub- unit A, though the authors do not comment on this. Co-crystal- lization trials on the complex calcineurin B-CypNCsA-cal- cineurin A332990 will be initiated in our laboratory. Structural information on this complex will elucidate the mode of interac- tion of immunophilin-ligand complexes with calcineurin and can potentially lead to the design of novel and more potent immunosuppressants.

REFERENCES 1. Cohen, D. J., Loertscher, R., Ruhin, M. F., Tilney, N. L., Carpenter, C. B., and

2. Thomson, A. W. (1989) Immunol. Today 10, 6-9 3. Kahan, B. D., Chang, J. Y., and Sehgal, S. N. (1991) Dansplantation 52,

4. Fischer, G., Wittrnann-Liebold, B., Lang, K., Kiefhaber, T., and Schmid, F. X. 185-191

5. Price, E. R., Zydowsky, L. D., Jin, M., Baker, C. H., McKeon, I? D., and Walsh, (1989) Nature 337, 476-478

6 . Ueki, K., and Kincaid, R. L. (1993) J. Biol. Chem. 268, 6554-6559 C. T. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 1903-1907

7. Kieffer, L. J., Thalhammer, T., and Handschumacher, R. E. (1992) J. Bid.

8. Harding, M. W., Galat, A,, Uehling, D. E., and Schreiber, S. L. (1989) Nature

Strom, T. B. (1984) Ann. Intern. Med. 101, 667482

Chem. 267,5503-5507

341, 758-760

14204 Immunophilin-Ligand Complex Interaction with Calcineurin 9. Jin, Y.-J., Albers, M. W., Lane, W. S., Bierer, B. E., Schreiber, S. L., and

Burakoff, S. J. (1991) Proc. Natl. Acad. Sci. U. S . A . 88,6677-6681 10. Tai, P.-K. K.,Albers, M. W., Chang, H., Faber, L. E., and Schreiber, S. L. (1992)

Science 266,1315-1318 11. Schreiber, S. L. (1991) Science 261, 283-287

13. Liu, J., Farmer, J. D., Jr,, Lane, W. S., Friedman, J., Weissman, I., and 12. Sigal, N. H., and Al, E. (1991) J. Ezp. Med. 173, 619-628

14. Clipstone, N. A,, and Crabtree, G. R. (1992) Nature 367, 695-697 15. O'Keefe, S. J., Tamura, J., Kincaid, R. L., Tocci, M. J., and O'Neill, E. A. (1992)

16. Fruman, D. A,, Klee, C. B., Bierer, B. E., and Burakoff, S. J. (1992) Proc. Natl.

17. Kallen, J., Spitzfaden, IC, Zurini, M., Wider, G., Widmer, H., Wiithrich, K., and

18. F'fluegel, G., Kallen, J., Schirmer, T., Jansonius, J. N., Zurini, M. G. M., and

19. Michnick, S. W., Rosen, M. K., Wandless, T. J., Karplus, M., and Schreiber, S.

20. Van Duyne, G. D., Standaert, R. E , Karplus, P. A,, Schreiber, S. L., and Clardy,

21. Friedman, J., and Weissman, I. (1991) Cell 66, 799406 22. Ratajczak, T., Carrello, A,, Mark, P. J., Warner, B. J., Simpson, R. J., Moritz, R.

L., and House, A. K. (1993) J. B i d . Chem. 268, 13187-13192 23. Schonbrunner, E. R., Mayer, S., Tropschug, M., Fischer, G., Takahashi, N., and

Schmid, F. X. (1991) J. B i d . Chem. 266,3630-3635 24. Fransson, C., Freskgard, P., Herbertsson, H., Johansson, A., Jonasson, P.,

Lett. 296, 90-94 Martensson, L., Svensson, M., Jonsson, B., and Carlsson, U. (1992) FEBS

25. Kincaid, R. L., Nightingale, M. S., and Martin, B. M. (1988) Proc. Natl. Acad. Sci. U. S . A . 85,89834987

Schreiber, S. L. (1991) Cell 66, 807-815

Nature 367,692-694

Acad. Sci. U. S. A. 89, 3686-3690

Walkinshaw, M. D. (1991) Nature 363, 276-279

Walkinshaw, M. D. (1993) Nature 361,91-94

L. (1991) Science 262,836439

J. (1991) Science 262,839-842

26. Hashimoto, Y., Perrino, B. A,, and Soderling, T. R. (1990) J. B i d . Chem. 266,

27. Guerini, D., Montell, C., and Klee, C . B. (1992) J. Biol. Chem. 267, 22542-

28. Manalan, A. S., and Klee, C. B. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,

29. Swanson, S. K.-H., Born, T., Zydowsky, L. D., Cho, H., Chang, H. Y., Walsh, C.

30. Haddy, A,, Swanson, S. K.-H., Born, T. L., and Rusnak, F. (1992) FEBS Lett.

31. Li, W., and Handschumacher, R. E. (1993) J. Biol. Chem. 268,14040-14044 32. Liu, J., Albers, M. W., Chen, C.-M., Schreiber, S. L., and Walsh, C. T. (1990)

33. Standaert, R. F., Galat, A,, Verdine, G. L., and Scbreiber, S. L. (1990) Nature Proc. Natl. Acad. Sci. U. S. A. 87, 2304-2308

34. Pallen, C. J., Sharma, R. K., and Wang, J. H. (1988) in Calcium Binding 346,671-674

Proteins, Characterization and Properties (Thompson, M. P., ed) Vol. 1, pp. 51-82, CRC Press, Inc., Boca Raton, FL

1924-1927

22549

4291-4295

T., and Rusnak, F. (1992) Proc. Natl. Acad. Sci. U. S. A . 89, 3741-3745

314,3740

36. Foxwell, B. M. J., Woerly, G., Husi, H., Mackie, A,, Quesniaux, V. F. J., 35. Dedman, J. R., and Kaetzel, M. A. (1983) Methods Enzymol. 102, 1 4

Hiestand, P. C., Wenger, R. M., and Ryffel, B. (1992) Biochim. Biophys. Acta 1138, 115-121

37. Laemmli, U. K. (1980) Nature 227,680-685 38. Lee, C. Y., Huang, Y. S., Hu, P. C., Gomel, V., and Menge, A. C. (1982) Anal.

39. Kincaid, R. L., Gin, P. R., Higuchi, S., Tamura, J., Dixon, S. C., Marietta, C. A,, Bioehem. 111,385-392

40. Aldape, R. A,, Futer, O., DeCenzo, M. T., Jarrett, B. P., Murcko, M. A., and Amorese, D. A., and Martin, B. M. (1990) J. B i d . Chem. 266, 11312-11319

41. Yang, D., Rosen, M. K., and Schreiber, S. L. (1993) J. Am. Chem. SOC. 116, Livingston, D. J. (1992) J. Biol. Chem. 267, 16029-16032

819-820