Identification of Actin and HSP 70 as Cyclosporin A Binding Proteins by Photoaffinity Labeling and Fluorescence Displacement Assays*

A novel family of cyclosporin A (CsA) binding proteins was identified by using the biologically active, radioiodinated photoaffinity probe [D-Lys-N"(4-azido-3-['251]iodophenyl)propionyl)]8-CsA. In addition to cyclophilin, proteins with molecular masses of 43 kDa and approximately 50-55 kDa were labeled in Jurkat extracts and bovine calf thymus. Sequence analysis of the 43-kDa protein purified from calf thymus and subsequent Western analysis of CsA affinity- purified material from Jurkat extracts identified the 43-kDa component as actin. [~-Lys-N"(5-dimethyla-mino-l-naphthalenesulfonyl)]8-CsA, a fluorescent an- alogue of CsA, was prepared and used to measure the binding constants of cyclosporin derivatives to actin by means of a new fluorescence displacement assay. [D -Lys-N"(5-dimethylarnino-l-naphthalenesulf0nyl)]~ -CsA and [W-t-butoxycarbonyl diaminobutyryl)I8-CsA bind to bovine actin at physiologically relevant concentrations, with dissociation constants of 60 f 33 and 570 f 380 nM, respectively. Because the ATPase fragment of heat shock cognate 70 (HSC

to immunosuppression. Cyclosporin A prevents activation of lymphokine genes essential for T cell proliferation such as IL-2, IL-4, and interferon-y by disrupting calcium-dependent signalling transduction pathways in T cell activation (Crabtree et al., 1989;Lin et al., 1991;Flanagan et al., 1991). This process was thought originally to result from the binding of the drug to cyclophilin (Handschumacher et al., 1984). Subsequently, cyclophilin was discovered to be a peptidyl prolyl cis-trans isomerase (Fischer et al., 1984), an enzyme that is believed to facilitate protein folding i n vivo by catalyzing proline isomerization in newly synthesized proteins (Brandts et al., 1975). Although inhibition of the peptidyl prolyl cistrans isomerase activity of cyclophilin by CsA is competitive with small synthetic substrates (Kofron et al., 1991(Kofron et al., , 1992, and the inhibition constant ( K , 6 nM) is in overall agreement with doses of the drug that cause immunosuppression, structure-activity data have demonstrated that inhibition of cyclophilin's peptidyl prolyl cis-trans isomerase activity does not always correlate with immunosuppression (Sigal et al., 1991). Most notably, the weakly immunosuppressive analogue [MeAla]'-CsA (0.4% relative to CsA activity), strongly inhibits cyclophilin (Ki 2 nM) whereas the moderately (30%) immunosuppressive analogue [MeBm2t]'-CsA (Aebi et al., 1990;Rich et al., 1990) poorly inhibits cyclophilin ( K , 690 nM).
These exceptions may be explained by the recent discovery (Liu et al., 1991) that the complex formed between CsA and cyclophilin inhibits the calmodulin-dependent phosphatase, calcineurin (Klee et al., 1988).
In spite of this remarkable progress toward understanding the molecular basis for cyclosporin's mode of action, it is not yet clear whether inhibition of the phosphatase activity of calcineurin by the drug-cyclophilin complex is sufficient for causing immunosuppression, or if any of these biochemical events are important for the other biological activities of cyclosporin (Scolari et al., 1987;Borel et al., 1989). Two particularly serious side effects of this drug are nephrotoxicity and chronic rejection response, and the possibility exists that other CsA binding proteins are involved in these poorly understood processes.

Novel Cyclosporin
A Binding Proteins 22055 2010 Ci/mmol and retains partial biological activity of CsA (12%). We have utilized this radioiodinated probe to identify novel cyclosporin A binding proteins in the immune competent human cell line Jurkat and in bovine calf thymus. Herein, we determined the dissociation constants for protein-CsA analogues by using a new fluorescence displacement assay.

Materials
Bovine calf thymus was purchased from Pel Freeze. All column resins were obtained from Pharmacia LKB Biotechnolog Inc. Buffer components were purchased from commercially available sources.
Cyclosporin A was obtained from Sandoz. SSAl was a generous gift from Jeff Shilling and Dr. Elizabeth Craig, University of Wisconsin, Madison. The actin monoclonal antibody was obtained from Chemicon.

Methods
Purification of the 43-kDa Protein-All steps were performed at 4 "C unless otherwise specified. Bovine calf thymus (250 g) was homogenized in a Waring Blender in 250 ml of buffer that contained 20 mM Tris-C1, pH 7.6, 100 mM NaC1, 5 mM dithiothreitol (DTT), and 20 mg of phenylmethanesulfonyl fluoride (PMSF). After saturation with ammonium sulfate to 40%, the supernatant was cleared by ultracentrifugation a t 45,000 rpm using a Ti-50 rotor. The resuspended pellet was dialyzed overnight against 20 mM HEPES, pH 7.6, containing 1 mM PMSF. This material was applied to a Q-Sepharose Fast Flow column (2.6 X 60 cm) which was pre-equilibrated with 20 mM HEPES, pH 7.6. The column was washed with one column volume of the equilibrating buffer before running a 600-ml linear gradient from 0 to 0.6 M NaCl. Fractions containing the highest specific activity (See "Photolysis Experiments") were combined, and the protein was precipitated by addition of ammonium sulfate to 75% saturation. After centrifugation and dialysis against 20 mM HEPES, pH 7.6, 100 mM NaC1, the protein was applied to a Sephacryl S-200 column (2.6 X 100 cm) which had been pre-equilibrated with the same buffer. The fractions with the best specific activity were combined, dialyzed against 20 mM HEPES, applied to a Mono Q column, and eluted with a 0-500 mM NaCl gradient. Sequencing was performed on fractions containing the highest specific activity protein from the Sephacryl S-200 and Mono Q column chromatographies.
The G-actin used in the fluorescence displacement assay was purified by a slightly modified procedure. To increase the stability of the protein during purification and improve the overall yield of actin, the homogenization buffer was changed to 20 mM Tris-C1, pH 7.6, 0.5 mM DTT, 0.2 mM ATP, and 0.1 mM CaC12. All purification steps were the same as previously described through the Q-Sepharose chromatography. Fractions containing actin from the Q-Sepharose column were combined and dialyzed against the homogenization buffer. The actin was polymerized after dialysis by the addition of 50 mM KC1 and 2 mM MgCl, to the buffer. The F-actin was pelleted by centrifugation at 100,000 X g for 4 h. The pellet was resuspended in the homogenization buffer and then poured into a Dounce homogenizer and subjected to several gentle strokes with an A-size pestle. This suspension was frozen in liquid nitrogen and stored at -72 "C.
Sequencing of the 43-kDa Protein-Automated Edman Degradation was performed with Applied Biosystems Models 470A and 473A protein sequencers equipped with on-line phenylthiohydantoin analyzers. Sequence interpretation was performed on a VAX 8750 as described by Henzel et al. (1987). Cyanogen bromide (CNBr) was used to generate peptide fragments. The protein was solubilized in 0.1 ml of 7% formic acid, and the reaction was initiated by addition of a small crystal of CNBr. The vessel was capped and kept at room temperature for 17 h. Peptides obtained from the cleavage aeaction were separated on a Synchrom column (2 X 100 mm, 4000 A) from Synchrompac by using a 50-min linear gradient of 0.1% trifluoroacetic acid to 70% 1-propanol a t a flow rate of 0.2 ml per min.
Photolysis Experiments-Typically, 5-50 pg of protein was brought to a final volume of 50 p1 with 20 mM HEPES, pH 7.6, and 100 mM NaCl and was treated with 3 pl (150,000 cpm of [ D -L~S ( A '~~I P P S ) ]~-CsA in EtOH) for 2 h at room temperature in a borosilicate conical centrifuge tube. The tubes were placed in an ice bath for 5 min, and then a solution (10% by volume) that contained 3% activated charcoal and 0.3% dextran was added. After low speed centrifugation to pellet the charcoal and dextran, the material was photolyzed for 5 min at 4 "C with a Rayonet photochemical reactor from Southern N.E.
Ultraviolet Co using a wavelength of 350 nM. A 40-pl aliquot was removed, added to 10 pl of loading dye, and analyzed directly on a discontinuous 12% SDS-polyacrylamide gel electrophoresis Laemmli system. The gels were stained with Coomassie Brillant Blue and subjected to autoradiography for 1-2 days with an intensifying screen. Specific activities were determined by cutting out the 43-kDa band from the gel and determining the number of cpm bound to the 43-kDa protein with a Cobra y counter from Packard. The specific activity was defined as: cpm in the 43-kDa protein/total cpm/mg of protein.
Purification of CsA Binding Proteins from Jurkat Extracts-Jurkat cells (1 X 10') grown in RPMI 1640 medium with 5% fetal calf serum and antibiotics were washed with phosphate-buffered saline and then resuspended in 10 ml of ice-cold lysis buffer (35 mM HEPES, pH 7.3, 50 mM NaF, 0.1 mM DTT, 0.015% Triton X-100, and 1 mM PMSF). After breakage of the cells with a polytron, the cell debri was pelleted by centrifugation at 2000 rpm for 10 min followed by ultracentrifugation for 30 min at 37 K in a Ti-50 rotor. One half of the extract was incubated with Affi-Gel-CsA resin prepared by coupling [~-L y s ]~-c s A to cyanogen bromide-activated Affi-Gel 10, while the remaining extract was mixed with a control resin that had been capped with ethanolamine (Affi-Gel-EA). The resin was washed, and the proteins were eluted and run on a 10 or 12% SDS polyacrylamide gel. Proteins were detected by either Coomassie Blue staining or silver staining as indicated.
Fluorescence Binding Experiments-The fluorescence binding experiments were carried out on a Greg PC fluorometer (ISS, Inc., Champaign, IL). Excitation was at 350 nm, and the emission was monitored at 460 nm using a Y-42 cut-off filter (Hoya Optics, Inc., Freemont, CA) to eliminate any scattering of light. Data for the heat shock protein fluorescence binding experiments was collected on a Perkin-Elmer MPF-4 fluorimeter with a Perkin-Elmer 150 xenon power supply. The excitation and emission wavelengths were set to 350 and 500 nM, respectively. G-actin was prepared from frozen stocks of F-actin by dialysis overnight against 10 mM Tris-C1, pH 7.6, 0.5 mM dithiothreitol, 0.2 mM CaCI2, and 0.1 mM ATP. Actin (3.2-7 mg/ml) was diluted 10and 100-fold into the dialysis buffer. Aliquots from these stock solutions along with dialysis buffer (50 p1 total volume) was added to 1.93 ml of 10 mM Tris-C1, pH 7.6, immediately before addition of [D-Lys(Dns)]*-CsA 4 (40 pl, 10 p~ stock solution dissolved in 0.47 M dry THF/LiCl).
The yeast cytoplasmic heat shock protein 70 (Hsp 70), a mixture of SSAl and SSA2, was provided as a 1.2 mg/ml solution dissolved in 20 mM HEPES, pH 7.6,50 mM KC1, and 10% glycerol. The stock solution was diluted 10-and 100-fold into this same buffer. The protein and buffer (118 pl total volume) were added to 1.86 ml of 10 mM Tris-C1, pH 7.6, immediately before addition of the fluorophore 4 (40 pl, 10 p~ stock solution dissolved in 0.47 M dry THF/LiCl). Fluorescence intensity values were measured 50 s after addition of [D-Lys(Dns)18-CsA 4 to actin or Hsp 70.
[~-Lys(Dns)]~-CsA 4 was prepared by dissolving [D-Lys]'-CsA (Colucci et al., 1990) and 1.3 equivalents of 5-dimethylamino-lnaphthalenesulfonyl chloride in dry DMF. After one equivalent of triethylamine was added, the reaction was allowed to proceed at room temperature for 3 h. The [~-Lys(Dns)]~-CsA 4 was purified by preparative silica gel thin-layer chromatography (chloroform/methanol 20:l). A 1 mM stock solution of [o-Lys(Dns)18-CsA 4 was made up in dry THF and stored for no longer than 3 days. Dilutions were made into dry THF/LiCl, 0.47 M, immediately before use.

RESULTS
Photoaffinity Labeling of Crude Extracts-Extracts from calf thymus and the human leukemic cell line Jurkat were treated separately with the photolabel [D-L~S(A'~'IPP)]~-CSA 2 ( Fig. 1) to screen for novel cyclosporin A binding proteins. Three distinct bands were previously identified in mouse thymocyte cytosols with molecular masses of 55, 43, and 37 kDa, respectively (Tung et ul., 1989). In addition to cyclophilin, primarily proteins with M , of 50-55,000 and 43,000 were detected in Jurkat extracts (Fig. 2, lune I). Minor components a t both 37 and 90 kDa were also identified (Fig. 2, lane 1 ). Bovine calf thymus contained almost exclusively the 43-kDa protein and cyclophilin (lune 3 ) . In addition, all bands could be displaced by nonradioactive cyclosporin A (lane 2 ) .
Diethylamide-A"'1PP reagent 7, prepared as a control com- the salt gradient (Fig. 3B). After further purification of the 43-kDa protein by Sephacryl S200 chromatography, fractions with the highest specific activity were chosen for sequence analysis. When the protein was found to be N-terminally blocked, CNBr was used to generate peptide fragments. These were purified, sequenced, and analyzed as described under "Methods." Two proteins were identified. The sequenced peptide fragments XKAGRAGDDAPRAVFPSIVGRPR and GQKDSYVGDEAQSKRGILTL (where Xequals cysteine not identified during sequencing) established that the major component was y-cytosolic actin, 42 kDa (actin migrates with an apparent M, of 43,000 after photolysis). The identity of the second protein was determined when sequence analysis revealed that peptides MADGTXQDAAIVGYKDSPSVW-AAVPG and DLRTKSTGGAP were identical to portions of profilin, 15-kDa. Homogeneous fractions from the final purification step (Mono Q chromatography) were sequenced after cyanogen bromide cleavage and found to contain only ycytosolic actin.
Purification of the 43-and 55-kDa Proteins from Jurkat Extracts-A cyclosporin A affinity resin was used to purify CsA binding proteins in a one-step purification procedure from Jurkat cells. Whole cell extracts were prepared and " I " -.
Purification and Identification of the 43-kDa Protein-The incubated with either the Affi-Gel-CsA affinity resin or a 43-kDa protein was purified from bovine calf thymus by control Affi-Gel resin that was capped with ethanolamine. fractionation with ammonium sulfate followed by chromatog-The eluted proteins from the column were further separated raphy over Q-Sepharose. A typical elution profile from the Q-by SDS-polyacrylamide gel electrophoresis. In addition to Sepharose chromatography is shown in Fig. 3A. Cyclophilin cyclophilin, proteins with M, of 43,000, 50,000-55,000, and emerged from the column in the first protein peak, whereas approximately 90,000 were detected (Fig. 4). Western analysis the 43-kDa CsA-binding protein was eluted near the end of revealed that the 43-kDa protein is actin (data not shown). Binding of Cyclosporin A to Actin-related Proteins-SSAl: the cytosolic Hsp 70 protein from yeast that is structurally related to actin (Kabsch et al., 1990), was photolyzed in the presence of either [D-L~s(A'""IP)]~-CSA 2 or the diethylamide control photoreagent

SSAl was labeled by [D-
L~S(A""IPP)]~-CSA 2 in the presence of a 3-fold excess of soybean trypsin inhibitor or bovine serum albumin, which were added to serve as control proteins that do not specifically bind CsA (data not shown). These results establish that SSAl is labeled by CsA photoaffinity reagent 2 in a specific manner. However, when an excess of Hsp 70 was photolyzed in the presence of G-actin, only actin was radiolabeled (Fig. 5). In both experiments, diethylamide-AIPP 7 was not incorporated into the proteins, which established that the CsA moiety was needed for efficient labeling of actin and Hsp 70. Whereas addition of nonradioactive CsA (10 p~) to to the photoaffinity labeling reaction mixture prevented incorporation of the radiolabel 2 into 1 p~ cyclophilin, a 50 p~ or 25-fold excess of cyclosporin was required for protection of SSAl or actin. In contrast, a 2-fold excess (2 p~) of cold photoreagent 2 could protect actin completely (1 p~) from being labeled by [D-L~S(A""IPP)]~-CSA 2 (data not shown). When the binding assay was performed in the presence of a 2-fold excess of bovine serum albumin, actin was the only detectably labeled protein (data not shown).
"The SSAl obtained from Dr. Elizabeth Craig's laboratory is a mixture of SSAl and SSA2.
These displacement studies, which were carried out with nonradioactive CsA, indicate that actin binds more tightly to the AIPP analogue 2 than to CsA 1. T o determine if other analogues would compete for binding to actin more effectively than CsA, displacement studies were performed with a cyclosporin A derivative that contains another hydrophobic side chain in the 8 position, [~-Dab(Boc)]'-CsA 5. This analogue, which binds tightly to cyclophilin ( K , 2 nM) and has 14% of the immunosuppressive activity of CsA in an 11-2 secretion assay, prevented labeling of actin at concentratiors as low as 10 p~. On the other hand, 50 P M CsA was required to protect actin from being labeled by the photoreagent 2 (Fig. 6). It is apparent from these experiments that the presence of a hydrophobic or aromatic residue at the 8 position of CsA increases the affinity for actin in comparison with CsA itself.
Fluorescence Binding Assays-In order to determine the dissociation constants for the complexes formed of various CsA analogues to G-actin and other CsA binding proteins, we devised a competition binding assay. [n-Lys(Dns)IX-CsA 4, a fluorescent analog that has 10-50% the immunosuppressive activity of CsA (Hess et al., 1985) was synthesized and used as the reporting ligand. The dissociation constant for the CsA 4-actin complex first was determined essentially by the method of Inglese et al. (1989) by titrating increasing concentrations of actin against a fixed concentration (200 nM) of the fluorescent derivative (Fig. 7, open circles). Critical to the success of the competition binding assay is the method for

Novel Cyclosporin
A Binding Proteins dissolving the CsA analogues in anhydrous 0.47 M LiCl/ tetrahydrofuran, which is then added to the buffer. This method sets the conformation of the CsA analogue predominantly into one conformation that is close to the bioactive conformation (Kofron et al., 1992). Erratic results were obtained when dansyl analogue 4 was dissolved in 95% ethanol, because cyclosporin A and its derivatives exist as slowly interconverting mixtures of multiple conformations in protic solvents. The binding constant of [~-Lys(Dns)]~-CsA 4 to actin was calculated as 60 & 33 nM (five determinations) by nonlinear least-squares fit of fluorescence intensities.
A new method for data analysis was utilized to determine the dissociation constants for the cyclosporine analogues relative to the fluorescent analogue 4 (Kuzmic et al., 1992). The competition binding assay is illustrated for the complex of The labeling experiments suggested that actin may bind better than SSAl to photoreagent 2 and other 8-position analogues of CsA. To test the binding affinity of the yeast cytoplasmic Hsp 70s to [~-Lys(Dns)]'-CsA 4, a titration experiment was performed with SSA1. The amount of the protein was varied from 0-1 p~ and the fluorophore 4 was kept at a constant concentration of 200 nM. Analysis of the data as previously described for actin gave a Kd of 53 f 48 nM.

DISCUSSION
Our understanding of how cyclosporin A and FK 506 affect specific events during T cell activation has increased greatly in recent years. Liu and coworkers (1991) discovered that the CsA-cyclophilin complex and FK 506-FKBP complex inhibit the phosphatase activity of the calmodulin-dependent phosphatase, calcineurin, and have suggested that this process may be linked to the suppression of transcription of both lymphokine and nonlymphokine genes. Another immunosuppressant rapamycin, which is structurally similar to FK 506, forms a complex with FKBP that does not inhibit calcineurin's phosphatase activity, nor does rapamycin interfere with the transcription of genes such as IL-2, IL-4, and IL-2R (Liu et al., 1991;Dumont et al., 1990). This indicates that inhibition of these transcriptional events is not the only mechanism for causing immunosuppression and that these compounds may disrupt multiple processes required for T cell activation.
Moreover, immunosuppression is only one of a plethora of biological activities exhibited by CsA, and little is known about the biochemical events leading to nephrotoxicity and graft vessel disease caused by chronic rejection, two major side effects produced by this drug (Bore1 et al., 1989). It is possible that cyclosporin produces these effects through binding to other as yet unidentified proteins in a variety of tissues. T o examine this issue, we have synthesized analogue 2, a highly radioactive, biologically active, photoaffinity probe. In our previous report, we demonstrated that photoaffinity reagent 2 selectively labeled cyclophilin, one of the CsA binding proteins thought to be essential to immunosuppression (Tung et al., 1989). In this paper we describe the identification and partial characterization of a novel family of CsA binding proteins.
Novel CsA binding proteins (43, 50-55 kDa) were detected in bovine calf thymus and in Jurkat extracts. The assay conditions were optimized to maximize specific binding of reagent 2 to the 43-kDa protein and used to purify the 43-kDa protein present in calf thymus, a source chosen for its ready availability. Sequence analysis of the purified protein established that the 43-kDa protein is y-cytosolic actin. When only partially purified material was affinity-labeled and sequenced, profilin copurified with y-cytosolic actin. Profilin itself was not labeled, which indicates that CsA and profilin bind to different sites on actin. These experiments were performed at high enough protein concentrations (4-20 p~) so that most of the actin would be tied up in the profilin-actin complex (Kd 2 ~~-1 0 p~; La1 and Korn (1985)). Also, profilin and actin cochromatographed with the specific 43-kDa binding activity of the photoaffinity label, indicating that under certain experimental conditions, the label 2 may have greater affinity for the profilin-actin complex than for actin alone.
In order to compare the relative affinities of the newly identified proteins for CsA with those previously reported for other receptors (Foxwell et al., 1988(Foxwell et al., 1989Palaszynski et al., 1991) Actin clearly binds CsA analogues that contain hydrophobic side chains in the 8 position with higher affinity than cyclosporin A. [~-Lys(Dns)]'-CsA 4 and [~-Dab(Boc)]~-CsA 5 bind to actin with dissociation constants of 60 f 33 and 570 k 380 nM, respectively. The binding of CsA itself to actin is much weaker, with an estimated dissociation constant greater than 1 p~. It is intriguing that like actin, cyclophilin prefers these 8-position analogues compared to the parent drug, cyclosporin A, even though these proteins are not known to be structurally related. Both [~-Lys(Dns)]'-CsA 4 ( K d 5 nM, Kuzmic et al., 1992) and [~-Dab(Boc)]'-CsA 5 ( K , 2 nM, Kofron et al., 1991) bind to cyclophilin an order of magnitude tighter than CsA ( K d 30 nM, Kuzmic et al., 1992). The lower immunosuppressive activity of the Dab(Boc) analog (10% the activity of CsA) is most likely due to its higher affinity for actin, which because of its high cellular concentration (5-20% of the total protein) can compete with cyclophilin for binding to cyclosporin A and its analogues. Actin is important for a variety of cellular processes including mRNA transport and localization (Sundell and Singer, 1991), cell growth, and motility. We stress, however, that it is unclear at this time whether the biological effects produced by CsA are a result of the interaction of this cyclic peptide with actin.
The fact that the tertiary structure of the nucleotide binding domain of actin is closely related to the tertiary structure of the ATPase domain of the heat shock cognate 70 (HSC 70)

Novel Cyclosporin
A Binding Proteins (Kabsch et al., 1990) prompted us to test if CsA or the photoaffinity labeling analogue 2 would bind to proteins belonging to the 70-kDa heat shock protein family. SSA1, the cytosolic Hsp 70 protein from yeast, was found to be labeled by the AIPP derivative 2 after photolysis. Quantitative analysis of the relative binding by using the fluorescence displacement assay established that 4 binds approximately equally well to actin ( K d 60 nM) and to SSAl (Kd 53 nM). Because the binding experiments with actin and SSAl were performed under slightly different buffer conditions (ATP and calcium were used to stabilize the G actin and were present in the actin binding assay, whereas SSAl was provided to us as a stock solution in 10% glycerol and 50 mM KC1) the two dissociation constants may not be exactly comparable. Since the ATP binding sites of actin and Hsp 70 are structurally related, but not identical, it is possible that the phenyl ring of photoreagent 2 or the naphthalene ring of the dansyl derivative 4 occupy the ATP binding site in the groove between the two subdomains of actin. The Hsp 70 family is believed to catalyze protein folding in vivo (Rothman, 1989), and it will be important to determine if other Hsp 70s exist that will bind to cyclosporin A analogues under physiological conditions.
When a mixture of actin and a 1.4-fold excess of SSAl were photolyzed in the presence of CsA 2, only actin was labeled. Similarly, photoaffinity experiments on cell homogenates labeled actin but not Hsp 70. These results are surprising because the binding studies indicate that the affinity of the closely related DNS-CsA derivative 4 for these proteins is nearly identical so that we would have expected to detect at least some labeling of SSAl (or Hsp 70) in either experiment. One possible explanation for this discrepancy is that photolysis results and the binding assays are not directly comparable (for the reasons cited above) and that the efficiency of labeling SSAl by photolabel 2 may further be affected by differences in labeling efficiencies between proteins for aryl azides (Ruoho et al., 1984). Consequently, the relative labeling of these proteins by the cyclosporin affinity-labeling reagent may not reflect their relative binding affinities. However, an alternative and more provocative explanation is that actin is binding to SSAl so as to prevent access of CsA 2 to its binding site, and thus provides some information about the nature of the complex. If this occurs, it happens both in the purified protein and in the cell homogenate experiments. Further work is needed to test this hypothesis. CsA binding proteins with M , of approximately 45,000 and 50,000 have been reported previously in Jurkat extracts (Foxwell et al., 1988;Palaszynski et al., 1991). We have purified and identified one of these components as actin by CsA affinity purification followed by Western analysis. These findings again establish that actin recognizes and binds to the parent drug, cyclosporin A. The characterization of the 55-kDa protein labeled by 2 has not yet been completed, but several observations merit reporting at this time. While this work was in progress, Hsp 56 was identified as an FK 506 binding protein and in addition was found to have some sequence homology to FKBP 12 and 13 (Yem et al., 1992). Hsp 56, Hsp 70, and Hsp 90 are components of the inactive steroid-receptor protein complexes. In the absence of ligand, steroid receptors are designated as inactive, because they are unable to regulate transcriptional events through binding to specific response elements on DNA. The glucocorticoid and progesterone receptors have been proposed to exist as inactive complexes in the cytoplasm, and upon activation through binding to their respective ligands, migrate to the nucleus where they become competent to bind DNA. We have established that one member of the inactive steroid receptor complexes, Hsp 70, has affinity for immunosuppressive CsA derivatives. Work is in progress to determine if the 55-and 90-kDa proteins we have purified from Jurkat extracts are related to other steroid receptor binding proteins.