- Email: [email protected]
Interactions of Peptides with DnaK and C-Terminal DnaK Fragments Studied Using Fluorescent and Radioactive Peptides
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS
Vol. 356, No. 2, August 15, pp. 177–186, 1998 Article No. BB980784
Interactions of Peptides with DnaK and C-Terminal DnaK Fragments Studied Using Fluorescent and Radioactive Peptides Jundong Zhang1 and Graham C. Walker2 Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
Received March 2, 1998, and in revised form May 26, 1998
Monocysteine derivatives of Peptide C (KLIGVLSSLFRPK) were modified with N-((2-(acetoxy)ethyl)-Nmethyl)amino-7-nitrobenz-2-oxa-1,3-diazole (ANBD) to introduce a fluorescent probe. Five Peptide C derivatives—PepC-V5C-ANBD, PepC-L6C-ANBD, PepC-S7CANBD, PepC-S8C-ANBD, and PepC-L9C-ANBD—were then used to investigate the peptide-binding properties of DnaK. Introduction of the ANBD moiety at positions 8 and 9 of Peptide C yields peptides that bind to DnaK with a high affinity similar to unmodified peptide C. In contrast, the derivative carrying ANBD at position 6, PepC-L6C-ANBD, bound to DnaK with a binding affinity 470 times lower than that of PepCL9C-ANBD. Peptide C derivatives carrying ANBD at positions 5 or 7 have intermediate DnaK binding affinities. By assaying the binding affinities of PepC-L9CANBD and PepNR-S6C-[1-14C]acetamide to DnaK and three C-terminal fragments of DnaK, DnaK 384 – 638 (residues 384 to 638), DnaK 389 – 607 (residues 389 to 607) and DnaK 386 –561 (residues 386 to 561), we found that the last 31 residues of DnaK (residues 607– 638) do not have a significant effect on the peptide binding to DnaK. However, residues 561 to 607, which form the C, D, and E a-helices directly adjacent to the peptide binding pocket of DnaK [X. Zhu, X. Zhou, W. F. Burkholder, A. Gragerov, C. M. Ogata, M. E. Gottesman, and W. A. Hendrickson, Science 272, 1606 –1614, 1996], play important roles in stabilizing the DnaK/peptide complex. The kinetics of PepC-L9C-ANBD binding to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 – 561 also support this observation. © 1998 Academic Press
1 Present address: RepliGen Corporation, 117 4th Avenue, Needham, MA 02194. 2 To whom correspondence should be addressed at 68-633, Department of Biology, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139. Fax: 617-253-2643.
0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
Key Words: DnaK; Peptide C; Peptide NR; peptide binding; fluorescent peptides.
DnaK protein of Escherichia coli is a well-characterized member of the family of highly conserved heatshock proteins with a molecular mass of 70 kDa (Hsp70)3 (1, 2). Working in conjunction with DnaJ and GrpE as part of a molecular chaperone machine, DnaK plays a variety of physiological roles by functioning as a molecular chaperone (3). DnaK, like other members of the Hsp70 family, consists of a highly conserved N-terminal domain that encodes a weak ATPase activity, a peptide-binding domain that is directly adjacent to the ATPase domain, and an extreme C-terminal region (4 – 6). Like other Hsp70s, the ATPase domain of DnaK and the peptide-binding domain of DnaK are functionally coupled, and this coupling of the domains of DnaK is essential for the biological functions of DnaK (7, 8). The interaction of ATP with the N-terminal domain of DnaK induces a conformational change in the peptide-binding domain of DnaK and leads to peptide release, whereas the binding of peptides to the peptide-binding domain of DnaK affects the conformation of the N-terminal domain and stimulates its ATPase activity (9 –11). Phage-displayed peptide libraries and synthetic peptide libraries have been used to study the substrate specificity of DnaK (5) and other Hsp70s (12–14). These studies have suggested that peptides composed of hydrophobic amino acids are likely to have high affinity to DnaK and other Hsp70s. The high selectivity 3 Abbreviations used: Hsp70, heat-shock protein with a molecular mass of 70 kDa; PepNR, Peptide NR; I-ANBD, N-((2-(iodoacetoxy)ethyl)-N-methyl)amino-7-nitrobenz-2-oxa-1,3-diazole; TFA, trifluoroacetic acid; I-ABASA, N-iodoacetamidobutyl-4-aziobsalicylic acid.
177
178
ZHANG AND WALKER
of DnaK and other Hsp70s toward hydrophobic peptides provides a biochemical basis for understanding their molecular chaperoning mechanism in protein folding and unfolding. The hydrophobic regions of a protein are usually buried in the core of the structure of the protein. In a denatured protein or a newly synthesized polypeptide, the exposed hydrophobic region could provide targets for interacting with Hsp70s. By binding to Hsp70s, a protein could be protected from forming aggregates or be kept in a denatured form to be translocated (15). The recently reported (16) crystal structure of part of the C-terminal domain of DnaK (residues 389 to 607) complexed with Peptide NR (NRLLLSG, PepNR) showed that the peptide-binding site consists of two sheets of eight b-strands followed by one long a-helix. The bound PepNR has extensive interactions with the amino acid residues located on the loops that connect the b-sheets. We have used peptide photocrosslinking to study the peptide-binding site of DnaK (17). It was shown that a region of DnaK, from residue 518 to residue 545, plays an important role in peptide binding. Photocrosslinker-modified Peptide C (KLIGVLSSLFRPK) derivatives, PepC-ASA, PepC-S7C-ABASA, and PepC-S8C-ABASA, crosslinked to Arg-536, Arg527, and His-541, respectively (17). The DnaK fragment identified by peptide crosslinking corresponds to the long a-helix that sits on the top of the peptidebinding sites as indicated in the crystal structure of the C-terminal domain of DnaK. We have proposed that this long a-helix may act as a lid to stabilize the bound peptides, and its movement could regulate peptide binding and release (17). To examine the nature of peptide binding to DnaK and extend our photocrosslinking experiments, we have examined the binding properties of Peptide C monocysteine derivatives that are modified with a fluorescent probe. We also studied the effects of the Nterminal domain and the extreme C-terminal domain on the peptide-binding properties of DnaK by analyzing the binding affinities and binding kinetics of fluorescent peptides to various C-terminal fragments of DnaK. MATERIALS AND METHODS DnaK protein was purified as previously reported (18). The DnaK C-terminal fragments DnaK 384 – 638 (residues 384 to 638), DnaK 389 – 607 (residues 389 to 607), and DnaK 386 –561 (residues 386 to 561) were kindly provided by William Burkholder (Columbia University), Xun Zhao (Columbia University), and Gregory Flynn (University of Oregon), respectively. Bovine serum albumin was purchased from New England Biolabs. Protein concentration was determined by Bio-Rad Assay with IgG as the standard. Peptides were synthesized at the Biopolymer Lab at MIT with a peptide synthesizer (Applied Systems, 430A). N-((2-(Iodoacetoxy)ethyl)-N-methyl)amino-7-nitrobenz-2-oxa-1,3-diazole (I-ANBD) was purchased from Molecular Probes. Iodo[1-14C]acetamide was purchased from Amer-
sham. Fluorescence measurements were carried out on a Perkin– Elmer fluorescence spectrometer (Model LS50). Peptide-binding kinetics were studied on a stop-flow apparatus from Applied PhotoPhysics. Preparation of fluorescent probe ANBD-labeled peptides. The procedure used to prepare PepC-L9C-ANBD is described in detail below. The other fluorescent-modified peptides, PepC-V5C-ANBD, PepCL6C-ANBD, PepC-S7C-ANBD, and PepC-S8C-ANBD, were prepared by the same method. Two and a half milligrams (6.2 mmol) I-ANBD in 0.5 ml acetonitrile was added to 10 mg (7.1 mmol) PepC-L9C-ANBD in 0.5 ml of 50 mM Tris buffer (pH 8.0). The resulting solution was incubated at 37°C for an hour. PepC-L9C-ANBD was then purified by HPLC with a Waters Delta-pak DnaK 384 – 638 cartridge column. The reverse-phase column was run at 2 ml/min started with 100% buffer A [0.1% trifluoroacetic acid (TFA)] to 100% buffer B (80% acetonitrile, 0.085% TFA) within 60 min. PepC-L9C and PepC-L9C-ANBD were eluted at 25.5 and 28.3 min, respectively. PepC-L9C-ANBD peak was collected and lyophilized. Preparation of [1-14C]acetamide-labeled peptides. In order to dissolve the radiochemical, 100 ml of 50 mM Tris buffer (pH 8.0) was pipetted into the ampule the manufacture used to store 50 mCi iodo[1-14C]acetamide. Iodo[1-14C]acetamide solution was then transferred to 5.0 mg (6.3 mmol) PepNR-S6C in 0.5 ml of 50 mM Tris buffer (pH 8.0) followed by incubation at 37°C for half an hour. Unlabeled iodoacetamide (2.3 mg, 12.7 mmol) was then added to the reaction solution, and the resulting buffer was further incubated at 37°C for half an hour. [1-14C]Acetamide-modified peptide was purified by HPLC with a Waters Delta-pak DnaK 384 – 638 cartridge column. The reverse-phase column was run at 2 ml/min started with 100% buffer A (0.1% TFA) to 100% buffer B (80% acetonitrile, 0.085% TFA) within 60 min. PepNR-S6C and PepNR-S6C-[1-14C]acetamide were eluted at 21.9 and 20.1 min, respectively. The radioactive peak was collected and lyophilized. The specific radioactivity of PepNR-S6C[1-14C]acetamide was 6.3 mCi/mmol. Peptide binding assay using fluorescent peptides. Emission spectra of fluorescent peptides were recorded on a fluorescence spectrometer upon incubation with relevant proteins at 37°C for 1 h with excitation at 480 nm. The excitation slit was set to 5.0 nm and the emission slit was set to 10.0 nm. Peptide-binding assays were carried out in 0.5 ml of 50 mM Tris (pH 7.4), 50 mM KCl, 10 mM MgCl2, and 5 mM b-mercaptoethanol buffer containing relevant peptides and various concentrations of proteins. The maximum fluorescence intensities were measured after incubation at 37°C for 1 h. Data were fit by nonlinear least squares analysis, using Igor software for the Macintosh from Wavemetrics. Dissociation constants from direct titration of the fluorescent peptides with proteins were calculated using the Langmuir isotherm: U 5 [protein]/[protein] 1 Kd z, where U is the fraction of peptide bound proteins and Kd is the apparent dissociation constant. Measurements of peptide binding kinetics. The kinetics of PepCL9C-ANBD binding to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 were studied on a stopped flow apparatus with equal volume mixing at room temperature (22°C). The peptide was excited at 480 nm, and the fluorescence intensity was measured with a filter cutoff at 495 nm. The data were analyzed on Spectrakinetic Workstation V4.09 provided by Applied Photophysics. Peptide binding assay with radiolabeled peptides. A gel filtration column, Biospin-30 (Bio-Rad), was used to separate free peptides with DnaK-bound peptides. Then, 50 ml of 50 mM Tris (pH 7.4), 50 mM KCl, 10 mM MgCl2, and 5 mM b-mercaptoethanol buffer containing relevant peptides and various concentrations of proteins was incubated at 37°C for 1 h before being loaded into the Biospin-30 columns, followed by centrifugation on a Fisher Centrific centrifuge at its maximum speed for 5 min. The radioactivity of the flowthrough, which contains the DnaK-bound peptide, was measured
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
FIG. 1. ANBD.
Chemical structures of I-ANBD and hydroxyethylthio-
by scintillation counting. Dissociation constants of radiolabeled peptides were calculated following the same procedure as the determination of dissociation constants of the fluorescent peptides.
RESULTS
Binding of ANBD-Modified Monocysteine Peptide C Derivatives to DnaK In the course of a previous study in which we used peptide photocrosslinking to identify elements of the peptide binding site of DnaK, we synthesized monocysteine variants of Peptide C in which a single cysteine was substituted for Val5, Leu6, Ser7, and Ser8, respectively (17). These Peptide C derivatives, PepC-V5C, PepC-L6C, PepC-S7C, and PepC-S8C, were modified with a cysteine specific photocrosslinker N-iodoacetamidobutyl-4-azidosalicylic acid (I-ABASA) to yield the modified peptides PepC-V5C-ABASA, PepC-L6C-ABASA, PepC-S7C-ABASA, and PepC-S8C-ABASA. When these peptides were used in crosslinking studies with DnaK, we noticed that the efficiencies of PepC-V5C-ABASA and PepC-L6C-ABASA crosslinking to DnaK were significantly weaker than that of PepC-S7C-ABASA and PepC-S8C-ABASA (unpublished data). And, as a result of this relatively weak crosslinking, we were unable to identify the points of attachment of PepC-V5C-ABASA and PepC-L6C-ABASA to DnaK. Since the N-terminal residues of peptide C had been implicated in the binding to DnaK (19), it seemed likely that modifications at residue Val5 and Leu6 might be interfering with peptide binding to DnaK. To examine this observation further, a cysteine-specific fluorescence probe, I-ANBD (Fig. 1), was reacted with all of the four monocysteine derivatives of Peptide C to generate the fluorescent peptides PepC-V5CANBD, PepC-L6C-ANBD, PepC-S7C-ANBD, and PepCS8C-ANBD. In addition, we prepared another monocysteine variant of Peptide C, PepC-L9C, and also modified it with I-ANBD to yield PepC-L9C-ANBD. Peptides carrying fluorescent probes, such as acrylodan, dansyl, fluorescein, and ANBD, had been previously used to study the peptide binding properties of DnaK and Hsc70 (20 –23). These fluorescent probes were either attached to the N-terminal residue of the
179
peptides by reacting with an N-terminal cysteine residue (22) or by reacting with the N-terminal amino group (21, 23) or attached to the C-terminal of the peptide by reacting with a C-terminal cysteine residue (20). Upon incubation with DnaK at identical concentrations, the fluorescence intensity of these modified Peptide C derivatives, PepC-V5C-ANBD, PepC-L6CANBD, PepC-S7C-ANBD, PepC-S8C-ANBD, and PepCL9C-ANBD, increased to various degrees (Fig. 2). Upon incubation with DnaK, the fluorescence intensity of PepC-L9C-ANBD and PepC-S8C-ANBD increased 2.6and 2.5-fold, respectively. Moderate increases (1.9- and 1.5-fold) were observed for PepC-S7C-ANBD and PepC-V5C-ANBD, while PepC-L6C-ANBD gave no increase in fluorescent intensity upon incubation with DnaK. Several pieces of evidence indicated that the observed increases in the fluorescence intensity of these ANBD-modified peptides were due to specific binding to the peptide-binding site of DnaK. First, the incubation of DnaK with 2-hydroxyethylthio-ANBD which had been prepared by reacting b-mercaptoethanol with I-ANBD did not change its fluorescence spectrum, suggesting ANBD itself does not interact with DnaK (Fig. 3a). Second, inclusion of 5 mM Peptide C in a buffer containing 2 mM DnaK and 1 mM PepC-L9C-ANBD led to a 45% decrease of the fluorescence intensity of DnaK-bound PepC-L9C-ANBD, indicating that unmodified Peptide C competes with PepC-L9C-ANBD for binding to DnaK (Fig. 3b). Third, in the presence of 1.0 mM ATP, the fluorescence intensity of PepC-L9CANBD decreased 63%, suggesting ATP can release PepC-L9C-ANBD from DnaK as it does to other DnaKbound peptides (Fig. 3c). Fourth, incubation of PepCL9C-ANBD with BSA did not affect its fluorescence intensity indicating the fluorescence intensity increase is due to specific interaction with DnaK (Fig. 3c). Similar phenomena were also observed for PepC-V5CANBD, PepC-L6C-ANBD, PepC-S7C-ANBD, and PepCS8C-ANBD (data not shown). Taken together, these observations indicated that the Kd’s of these fluorescent peptides can be determined by measuring their increases in fluorescence intensity upon incubation with various concentrations of DnaK. A Langmuir curve of PepC-L9C-ANBD binding to DnaK is shown in Fig. 4. The dissociation constants of various fluorescence-labeled peptides to DnaK determined by this method are listed in Table I. As indicated by the dissociation constants of the modified peptides, the presence of the bulky ANBD substitutent at position 6 resulted in the lowest binding affinity, indicating residue 6 is critical for Peptide C binding to DnaK. ANBD modification at position 5 less severely but still significantly interferes with peptide binding to DnaK, while modification at positions 7, 8,
180
ZHANG AND WALKER
FIG. 2. Fluorescence intensity changes of 1.0 mM PepC-V5C-ANBD (a), PepC-L6C-ANBD (b), PepC-S7C-ANBD (c), PepC-S8C-ANBD (d), and PepC-L9C-ANBD (e) upon incubation with 2.0 mM DnaK at 37°C for 1 h.
or 9 had relatively minor effects on peptide binding to DnaK. Starting from position 6, the further away the fluorescent probe was from the N-terminus of the peptide, the less the interference of the ANBD modification on the peptide binding.
Studies of the Peptide Binding Properties of DnaK and Various C-Terminal Fragments Like other members of the Hsp70 class, DnaK consists of a 45-kDa N-terminal domain which binds ATP
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
181
FIG. 3. (a) A stock solution of 100 mM 2-hydroxyethylthio-ANBD was prepared by treating 100 mM I-ANBD in dimethyl sulfoxide with 500 mM b-mercaptoethanol at 37°C for 1 h. The emission spectra of 2.0 mM 2-hydroxyethylthio-ANBD (■) and 2.0 mM 2-hydroxyethylthio-ANBD with 5.0 mM DnaK (1) upon incubation at 37°C for 1 h were recorded with excitation at 480 nm. (b) The emission spectra of 1.0 mM PepC-L9C-ANBD (■) and 1.0 mM PepC-L9C-ANBD (1) that was incubated at 37°C with 2.0 mM DnaK for 1 h. (c) The emission spectra of 1.0 mM PepC-L9C-ANBD (■), 1.0 mM PepC-L9C-ANBD (1) that was incubated at 37°C with 2.0 mM DnaK for 1 h, 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK that was treated with 2.0 mM ATP for 0.5 h (E), and 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK that was treated with 5.0 mM Peptide C (‚) at 37°C for 0.5 h.
and ADP and has an ATPase activity, an 18-kDa peptide-binding domain which is directly adjacent to the ATPase domain, and a 7-kDa extreme C-terminal region whose function is not well understood (4 – 6, 16). To explore the nature of the extensive interactions between the various domains of DnaK, we compared the peptide-binding properties of full-length DnaK with those of three C-terminal fragments of DnaK (Fig.
5). These different classes of C-terminal fragments of DnaK have been used to gain insight into the nature of the interaction between DnaK with various peptide substrates. One of these (DnaK 384 – 638) consists of the whole C-terminal domain of DnaK from residue 384 to residue 638 with a histidine tag at the N-terminus (24). The second (DnaK 389 – 607), whose crystal structure has been solved, begins at residue 389 and
182
ZHANG AND WALKER
FIG. 4. The Langmuir curve of PepC-L9C-ANBD binding to DnaK. 1.0 mM PepC-L9C-ANBD was incubated with various concentrations of DnaK at 37°C for 1 h. The fluorescence intensity at 540 nm was measured, and the maximum fluorescence intensity increase was treated as 1.
ends at residue 607 (16). The third (DnaK 386 –561) consists of a fragment from residue 386 to residue 561 with a histidine tag at the N-terminus. The solution structure of DnaK 386 –561 has been studied by NMR and is similar to the corresponding fragments of Hsc70 (25, E. Zuiderweg and G. Flynn, personal communication). DnaK 386 –561 contains the eight b-strands that form the peptide binding pocket and the long a-helix that sits on the top of the peptide-binding pocket (16), but it lacks the final 77 amino acids of DnaK. The fluorescence intensity of PepC-L9C-ANBD (1.0 mM) increased 2.6-fold upon incubation with DnaK (2.0 mM) and DnaK 384 – 638 (2.0 mM). Interestingly, the maximum emission wavelength of bound PepCL9C-ANBD to DnaK shifted from 539.5 to 534.5 nm, whereas the maximum emission wavelength of
TABLE I
A List of Dissociation Constants Proteins DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK
384–638 389–607 386–561 384–638 389–607 386–561
Peptides PepC-V5C-ANBD PepC-L6C-ANBD PepC-S7C-ANBD PepC-S8C-ANBD PepC-L9C-ANBD PepC-L9C-ANBD PepC-L9C-ANBD PepC-L9C-ANBD PepNR-S6C-[1-14C]acetamide PepNR-S6C-[1-14C]acetamide PepNR-S6C-[1-14C]acetamide PepNR-S6C-[1-14C]acetamide
Kd (mM) 11.8 299 3.7 1.1 0.63 0.93 0.85 2.6 3.2 2.8 2.9 29.6
bound PepC-L9C-ANBD to DnaK 384 – 638 shifted from 539.5 nm to 535.5 nm (Fig. 6). The difference in the emission blue shift indicates that the N-terminal domain of DnaK does cause a change in the chemical environment of the peptide binding domain that can be sensed by a fluorescent probe at position 9 of Peptide C. The fluorescence intensity of PepC-L9CANBD increased to a comparable level upon incubation with DnaK 389 – 607 as with DnaK and DnaK 384 – 638 (Fig. 6). The maximum emission wavelength of DnaK 389 – 607-bound PepC-L9C-ANBD shifted from 539.5 to 534.5 nm, similar to the shift with full-length DnaK. In contrast, the fluorescence intensity of PepC-L9C-ANBD (1.0 mM) increased much less upon incubation with DnaK 386 –561 (2.0 mM) compared to the fluorescence increase observed upon incubation with DnaK, DnaK 384 – 638, or DnaK 389 – 607. In addition, there was no change in the maximum emission wavelength of PepC-L9CANBD upon incubation with DnaK 386 –561 (Fig. 8). By measuring fluorescent enhancements as a function of DnaK concentration, the Kd’s of PepC-L9CANBD to DnaK 384 – 638 and DnaK 389 – 607 were determined to be 0.93 and 0.85 mM, respectively, comparable to that of DnaK (0.63 mM). The Kd of PepCL9C-ANBD to DnaK 389 – 607 was determined to be 2.6 mM (Table 1), which is four times larger than that of DnaK. To investigate the peptide-binding properties of DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 further, the dissociation constants of a radiolabeled Peptide NR (NRLLLSG) derivative, PepNR-S6C-[1-14C]acetamide (NRLLLCG), were also assayed. Peptide NR was selected from a phagedisplayed peptide library (5) as a peptide that has a high affinity to DnaK and was the peptide that was bound to the C-terminal of DnaK in the crystal structure by Zhu et al. (16). A Langmuir curve of PepNRS6C-[1-14C]acetamide binding to DnaK is shown in Fig. 7 and the Kd’s of PepNR-S6C-[1-14C]acetamide to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 are listed in Table I. The Kd’s of PepNRS6C-[1-14C]acetamide to DnaK, DnaK 384 – 638, and DnaK 389 – 607 are comparable, but the affinity of
FIG. 5. Schematic representation of DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 fragments.
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
183
FIG. 6. The emission spectra of 1.0 mM PepC-L9C-ANBD (■), 1.0 mM PepC-L9C-ANBD that was incubated at 37°C with 2.0 mM DnaK (1) for 1 h, 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK 384 – 638 (Œ) at 37°C for 1 h, 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK 389 – 607 (3) at 37°C for 1 h, and 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK 386 –561 (h) at 37°C for 1 h.
PepNR-S6C-[1-14C]acetamide to DnaK 386 –561 is about 10 times weaker than that of DnaK. Taken together with the measurements using the ANBDmodified peptides, these observations indicate that the amino acids between 562 and 607 influence the stability of peptide binding. Kinetics of PepC-L9C-ANBD Binding to Various C-Terminal Fragments of DnaK Besides determining the equilibrium binding constants, we were interested in studying the kinetics of
peptide binding by the DnaK C-terminal fragments, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561, so that we could compare their properties to DnaK. The kinetics of PepC-L9C-ANBD binding to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 were studied using a stop-flow apparatus. The kinetic curves shown here represent an average of at least five consistent measurements, and one preparation was used in the kinetic measurements of the relevant protein. When 2.0 mM DnaK was mixed with 10.0 mM PepCL9C-ANBD, the time course of the fluorescence intensity increase showed a two-phase process as previously
184
ZHANG AND WALKER
respectively. The peptide offrate of DnaK 386 –561 is about twice as fast as that of DnaK, DnaK 384 – 638, or DnaK 389 – 607, but the onrate of DnaK 386 –561 is about six to seven times slower. Therefore, the weaker affinity of peptides to DnaK 386 –561 is primarily due to a slow binding rate. DISCUSSION
FIG. 7. The Langmuir curve of PepNR-S6C-[1-14C]acetamide binding to DnaK. 5.0 mM PepNR-S6C-[1-14C]acetamide was incubated with various concentrations of DnaK at 37°C for 1 h followed by separating the DnaK-bound PepNR-S6C-[1-14C]acetamide with the free PepNR-S6C-[1-14C]acetamide using Biospin columns. The radioactivity of the bound PepNR-S6C-[1-14C]acetamide was counted.
reported (20, 22, 26). The half-time (t1/2) of the first phase was determined to be 10.0 s, and the t1/2 of the slow phase was 130 s (Fig. 8a). By varying the protein concentration, the k1 and k2 of the first phase of PepCL9C-ANBD binding to DnaK were calculated to be 1.56 3 1013 M21 s21 and 0.98 3 1023 s21, respectively. The time course of PepC-L9C-ANBD binding to DnaK 384 – 638 showed a pattern very similar to that of PepC-L9C-ANBD binding to DnaK. It is a two-phase binding process with t1/2’s of 10.3 and 132 s (Fig. 8b). The k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK 384 – 638 were calculated to be 1.16 3 1013 M21 S21 and 1.08 3 1023 S21, respectively. Similarly, the binding curve of 10.0 mM of PepC-L9CANBD to 1.0 mM DnaK 389 – 607 showed a two-phase reaction with t1/2’s of 8.9 and 129 s (Fig. 8c). The k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK 389 – 607 were calculated to be 1.30 3 1013 M21 s21 and 1.11 3 1023 s21, respectively. Thus, the kinetics of peptide binding were largely unaffected by the removal of the N-terminal domain of the protein and by the removal of both the N-terminal domain and also the C-terminal 31 amino acids. In contrast, the kinetics of PepC-L9C-ANBD binding to DnaK 386 –561 are very different from those of DnaK, DnaK 384 – 638, and DnaK 389 – 607. Although it showed a characteristic two-phase binding curve, the binding of PepC-L9C-ANBD to DnaK 386 –561 is slower, and the net fluorescence intensity increase is smaller. The t1/2’s of the two phases are 209 and 311 s, respectively (Fig. 8d). The k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK 386 –561 were calculated to be 196 M21 s21 and 2.47 3 1023 s21,
A variety of peptides that are labeled with fluorescent probes have previously been used to study the interactions of peptides with Hsp70s, including DnaK (20 –23). In those studies, the fluorescent probes were attached either to the N-terminal residue or to the C-terminal residue. By attaching a fluorescent molecule to various locations in the middle of the amino acid sequence, we prepared five fluorescent monocysteine derivatives of Peptide C. They are PepC-V5C-ANBD, PepC-L6C-ANBD, PepC-S7C-ANBD, PepC-S8C-ANBD, and PepC-L9C-ANBD. Introduction of the bulky fluorescent ANBD moiety at different locations of Peptide C results in peptides with different binding affinities for DnaK. The degree of the change in DnaK binding affinity of a particular modified peptide is likely to be related to the relative importance of the residue at that position to the interactions of Peptide C with DnaK. Introduction of the ANBD moiety at positions 8 and 9 affords peptides with similar high binding affinities to DnaK, suggesting Ser8 and Leu9 may not play a critical role in peptide binding. However, introduction of ANBD at position 6 yields a peptide, PepC-L6C-ANBD, with a DnaK binding affinity 470 times lower than that of PepC-L9C-ANBD. This observation implies that Leu6 plays a central role in the binding of Peptide C to DnaK. A substitution of a cysteine–ANBD moiety at position 5 had the next strongest effect, reducing the binding affinity by 18-fold. The introduction of ANBD at position 7 had a less severe effect on peptide binding, reducing the binding affinity by 6-fold. It is interesting to relate our data to the structure of the DnaK C-terminal fragment bound to the peptide PepNR (NRLLLSG) which has been reported by Zhu et al. (16). The authors showed that Leu4 of PepNR (NRLLLSG) is the most critical residue for peptide binding and is completely buried in the peptide-binding pocket of DnaK. Leu3 of PepNR is also completely buried, while Leu5 is only moderately buried. Together, Leu3 and Leu4 provide most of the side chain contacts. Thus, it seems reasonable to suggest that Peptide C is positioned in the binding pocket of DnaK in a manner related to that of peptide PepNR. By this interpretation, one residue of Peptide C would be the residue that is completely buried into the peptide binding pocket of DnaK. Any modification at this residue would be predicted to lead to peptides with dramatically decreased DnaK binding affinity. Several
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
185
FIG. 8. The time courses of 10.0 mM PepC-L9C-ANBD binding to 2.0 mM DnaK (a), 10 mM DnaK 384 – 638 (b), 1.0 mM DnaK 389 – 607 (c), and 10 mM DnaK 386 –561 (d).
residues of Peptide C would interact less with DnaK and their side chains would be predicted to project away from the binding pocket. Therefore, modifications with ANBD at these positions of Peptide C would result in peptides with smaller decreases in their DnaK binding affinity than the derivative modified at the completely buried residue. Since NMR data (19) has shown that the N-terminal part of Peptide C interacts with DnaK and it has been shown (9) biochemically that DnaK binds to short hydrophobic cores of four to five residues length, it is reasonable to suggest that residues 2–7 of Peptide C consists of the region that binds to the peptide-binding pocket of DnaK. Modifications of certain residues in this region severely hindered the DnaK binding affinity of the modified peptides, like ANBD modification at Leu6. Modifications of certain residues hindered their DnaK binding moderately, like ANBD modification at Val5 and Ser7. Consistant with this interpretation, substitution at position 9 does not interfere with DnaK binding. Furthermore, our observation that the DnaK 386 –561 fragment of DnaK, which lacks the C-terminal 77 amino acids, does not cause a blue shift when PepCL9C-ANBD binds suggests that these terminal resi-
dues influence the environment that is experienced by the extreme C-terminal of the peptides. Using both a fluorescent peptide and a radiolabeled peptide, we have shown that full-length DnaK, DnaK 384 – 638, and DnaK 389 – 607 are equally effective in forming a peptide complex. Upon incubation with DnaK, DnaK 384 – 638, and DnaK 389 – 607, the changes in the fluorescence intensity and the maximum emission wavelength of PepC-L9C-ANBD are very similar. The binding affinity of PepC-L9C-ANBD and PepNR-S6C-[1-14C]acetamide to DnaK, DnaK 384 – 638, and DnaK 389 – 607 are very close, and the binding kinetics of PepC-L9C-ANBD are comparable. All those observations suggest the last 31 residues of DnaK (residues 607– 638) do not have a significant effect on the peptide binding to DnaK, at least in this type of assay, which does not involve a cycle of ATP hydrolysis. On the other hand, the DnaK 386 –561 fragment (residues 386 –561), has significantly different peptide-binding properties compared to DnaK, DnaK 384 – 638, and DnaK 389 – 607. DnaK 386 –561 binds to PepC-L9C-ANBD 4 times more weakly and it binds to PepNR-S6C-[1-14C]acetamide 10 times more weakly. Kinetically, PepC-L9C-ANBD binding to DnaK 386 –
186
ZHANG AND WALKER
561 showed a slower onrate and a faster offrate. Those observations indicate residues 561 to 607, which form the C, D, and E a-helixes that are directly adjacent to the peptide binding pocket of DnaK (16), play important roles in stabilizing the DnaK/peptide complex. Interestingly, Hu and Wang (27) have reported that the last 10-kDa domain of Hsc70 plays a role in stabilizing the Hsc70/CMLA complex. Several papers (21, 22, 26) have reported that in the presence of ADP, peptide binding to DnaK and other Hsp70s, such as Hsc70, is a two-step process. The first step is relatively rapid, and the second step is slow, but results in “capture” of the peptide. The binding of PepCL9C-ANBD to DnaK behaved just like the other peptides reported. The half-time (t1/2 of the first phase was measured at 10.0 s, and the t1/2 of the slow phase was 130 s, and the k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK were calculated to be 1.56 3 1013 M21 and 0.98 3 1023 s21, respectively. Schmid et al. (22) measured the t1/2 of the first and second step of the binding of a peptide, a-pp, to DnaK at 27 and 200 s, respectively, and the k1 and k2 of the first phase of a-pp binding to DnaK were determined to be 9.4 3 1013 M21 s21 and 4 3 1023 s21, respectively. Takeda and McKay (21) reported the k1 and k2 of the first phase of peptide faf1 binding to DnaK to be 1.21 3 1013 M21 s21 and 0.017 s21, respectively, and they also calculated the k1 and k2 of the second phase to be 0.013 and 0.0038 s21. The rate constant in the forward direction (binding) of the second phase is of the order 0.01 sec21 as reported (21); this is the same magnitude as the rate constants of the second phase of the fluorescence kinetics of PepC-L9C-ANBD binding to DnaK, suggesting the second phase of the fluorescence change corresponds to the second step of peptide binding. ACKNOWLEDGMENTS We are grateful to Paul Schimmel for the use of their fluorescence spectrometer and to Robert Sauer for the use of their stop-flow apparatus. We thank Max Gottesman, Wayne Hendrickson, and Gregory Flynn for generously providing us with the relevant Cterminal fragments of DnaK. This work was supported by U.S. Public Health Service Grant GM28988 to G.C.W.
REFERENCES 1. Bardwell, J. C., and Craig, E. (1984) Proc. Natl. Acad. Sci. USA 81, 848 – 852. 2. Lindquist, S., and Craig, E. A. (1988) Annu. Rev. Genet. 22, 631– 677.
3. Craig, E. A., Gambill, B. D., and Nelson, R. J. (1993) Microbiol. Rev. 57, 402– 414. 4. Chappell, T. G., Konforti, B. B., Schmid, S. L., and Rothman, J. E. (1987) J. Biol. Chem. 262, 746 –751. 5. Gragerov, A., Zeng, L., Zhao, X., Burkholder, W., and Gottesman, M. (1993) J. Mol. Biol. 235, 848 – 854. 6. Wang, T.-F., Chang, J.-H., and Wang, C. (1993) J. Biol. Chem. 268, 26049 –26051. 7. Churchich, J. E. (1995) Eur. J. Biochem. 231, 736 –741. 8. Fung, K. L., Hilgenberg, L., Wang, N. M., and Chirico, W. J. (1996) J. Biol. Chem. 271, 21559 –21565. 9. Flynn, G. C., Chappell, T. G., and Rothman, J. E. (1989) Science 245, 385–390. 10. Jordan, R., and McMacken, R. (1995) J. Biol. Chem. 270, 4563– 4569. 11. Buchberger, A., Valencia, A., McMacken, R., Sander, C., and Bukau, B. (1994) EMBO J 13, 1687–1695. 12. Blond-Elguindi, S., Cwirla, S., Dower, W., Lipshutz, R., Sprang, S., Sambrook, J., and Gething, M.-J. (1993) Cell 75, 717–728. 13. Flynn, G. C., Pohl, J., Flocco, M. T., and Rothman, J. E. (1991) Nature 353, 726 –730. 14. Takenaka, I. M., Leung, S.-M., McAndrew, S. J., Brown, J. P., and Hightower, L. E. (1995) J. Biol. Chem. 270, 19839 –19844. 15. Frydman, J., Nimmesgern, E., Ohtsuka, K., and Hartl, F. (1994) Nature 370, 111–117. 16. Zhu, X., Zhou, X., Burkholder, W. F., Gragerov, A., Ogata, C. M., Gottesman, M. E., and Hendrickson, W. A. (1996) Science 272, 1606 –1614. 17. Zhang, J., and Walker, G. C. (1996) J. Biol. Chem. 271, 19668 – 19674. 18. McCarty, J. S., and Walker, G. C. (1991) Proc. Natl. Acad. Sci. USA 88, 9513–9517. 19. Landry, S. J., Jordan, R., McMacken, R., and Gierasch, L. M. (1992) Nature 355, 455– 457. 20. Theyssen, H., Schuster, H.-P., Packschies, L., Bukau, B., and Reinstein, J. (1996) J. Mol. Biol. 263, 657– 670. 21. Takeda, S., and McKay, D. B. (1996) Biochemistry 35, 4636 – 4644. 22. Schmid, D., Baici, A., Gehring, H., and Christen, P. (1994) Science 263, 971–973. 23. Farr, C. D., Galiano, F. J., and Witt, S. N. (1995) Biochemistry 34, 15574 –15582. 24. Burkholder, W. F., Zhao, X., Zhu, X., Hendrickson, W. A., Gragerov, A., and Gottesman, M. E. (1996) Proc. Natl. Acad. Sci. USA 93, 10632–10637. 25. Morshauser, R., Wang, H., Flynn, G., and Zuiderweg, E. (1995) Biochemistry 34, 6261– 6266. 26. Banecki, B., and Zylicz, M. (1996) J. Biol. Chem. 271, 6137– 6143. 27. Hu, S.-M., and Wang, C. (1996) Arch. Biochem. Biophys. 332, 163–169.
Vol. 356, No. 2, August 15, pp. 177–186, 1998 Article No. BB980784
Interactions of Peptides with DnaK and C-Terminal DnaK Fragments Studied Using Fluorescent and Radioactive Peptides Jundong Zhang1 and Graham C. Walker2 Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
Received March 2, 1998, and in revised form May 26, 1998
Monocysteine derivatives of Peptide C (KLIGVLSSLFRPK) were modified with N-((2-(acetoxy)ethyl)-Nmethyl)amino-7-nitrobenz-2-oxa-1,3-diazole (ANBD) to introduce a fluorescent probe. Five Peptide C derivatives—PepC-V5C-ANBD, PepC-L6C-ANBD, PepC-S7CANBD, PepC-S8C-ANBD, and PepC-L9C-ANBD—were then used to investigate the peptide-binding properties of DnaK. Introduction of the ANBD moiety at positions 8 and 9 of Peptide C yields peptides that bind to DnaK with a high affinity similar to unmodified peptide C. In contrast, the derivative carrying ANBD at position 6, PepC-L6C-ANBD, bound to DnaK with a binding affinity 470 times lower than that of PepCL9C-ANBD. Peptide C derivatives carrying ANBD at positions 5 or 7 have intermediate DnaK binding affinities. By assaying the binding affinities of PepC-L9CANBD and PepNR-S6C-[1-14C]acetamide to DnaK and three C-terminal fragments of DnaK, DnaK 384 – 638 (residues 384 to 638), DnaK 389 – 607 (residues 389 to 607) and DnaK 386 –561 (residues 386 to 561), we found that the last 31 residues of DnaK (residues 607– 638) do not have a significant effect on the peptide binding to DnaK. However, residues 561 to 607, which form the C, D, and E a-helices directly adjacent to the peptide binding pocket of DnaK [X. Zhu, X. Zhou, W. F. Burkholder, A. Gragerov, C. M. Ogata, M. E. Gottesman, and W. A. Hendrickson, Science 272, 1606 –1614, 1996], play important roles in stabilizing the DnaK/peptide complex. The kinetics of PepC-L9C-ANBD binding to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 – 561 also support this observation. © 1998 Academic Press
1 Present address: RepliGen Corporation, 117 4th Avenue, Needham, MA 02194. 2 To whom correspondence should be addressed at 68-633, Department of Biology, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139. Fax: 617-253-2643.
0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
Key Words: DnaK; Peptide C; Peptide NR; peptide binding; fluorescent peptides.
DnaK protein of Escherichia coli is a well-characterized member of the family of highly conserved heatshock proteins with a molecular mass of 70 kDa (Hsp70)3 (1, 2). Working in conjunction with DnaJ and GrpE as part of a molecular chaperone machine, DnaK plays a variety of physiological roles by functioning as a molecular chaperone (3). DnaK, like other members of the Hsp70 family, consists of a highly conserved N-terminal domain that encodes a weak ATPase activity, a peptide-binding domain that is directly adjacent to the ATPase domain, and an extreme C-terminal region (4 – 6). Like other Hsp70s, the ATPase domain of DnaK and the peptide-binding domain of DnaK are functionally coupled, and this coupling of the domains of DnaK is essential for the biological functions of DnaK (7, 8). The interaction of ATP with the N-terminal domain of DnaK induces a conformational change in the peptide-binding domain of DnaK and leads to peptide release, whereas the binding of peptides to the peptide-binding domain of DnaK affects the conformation of the N-terminal domain and stimulates its ATPase activity (9 –11). Phage-displayed peptide libraries and synthetic peptide libraries have been used to study the substrate specificity of DnaK (5) and other Hsp70s (12–14). These studies have suggested that peptides composed of hydrophobic amino acids are likely to have high affinity to DnaK and other Hsp70s. The high selectivity 3 Abbreviations used: Hsp70, heat-shock protein with a molecular mass of 70 kDa; PepNR, Peptide NR; I-ANBD, N-((2-(iodoacetoxy)ethyl)-N-methyl)amino-7-nitrobenz-2-oxa-1,3-diazole; TFA, trifluoroacetic acid; I-ABASA, N-iodoacetamidobutyl-4-aziobsalicylic acid.
177
178
ZHANG AND WALKER
of DnaK and other Hsp70s toward hydrophobic peptides provides a biochemical basis for understanding their molecular chaperoning mechanism in protein folding and unfolding. The hydrophobic regions of a protein are usually buried in the core of the structure of the protein. In a denatured protein or a newly synthesized polypeptide, the exposed hydrophobic region could provide targets for interacting with Hsp70s. By binding to Hsp70s, a protein could be protected from forming aggregates or be kept in a denatured form to be translocated (15). The recently reported (16) crystal structure of part of the C-terminal domain of DnaK (residues 389 to 607) complexed with Peptide NR (NRLLLSG, PepNR) showed that the peptide-binding site consists of two sheets of eight b-strands followed by one long a-helix. The bound PepNR has extensive interactions with the amino acid residues located on the loops that connect the b-sheets. We have used peptide photocrosslinking to study the peptide-binding site of DnaK (17). It was shown that a region of DnaK, from residue 518 to residue 545, plays an important role in peptide binding. Photocrosslinker-modified Peptide C (KLIGVLSSLFRPK) derivatives, PepC-ASA, PepC-S7C-ABASA, and PepC-S8C-ABASA, crosslinked to Arg-536, Arg527, and His-541, respectively (17). The DnaK fragment identified by peptide crosslinking corresponds to the long a-helix that sits on the top of the peptidebinding sites as indicated in the crystal structure of the C-terminal domain of DnaK. We have proposed that this long a-helix may act as a lid to stabilize the bound peptides, and its movement could regulate peptide binding and release (17). To examine the nature of peptide binding to DnaK and extend our photocrosslinking experiments, we have examined the binding properties of Peptide C monocysteine derivatives that are modified with a fluorescent probe. We also studied the effects of the Nterminal domain and the extreme C-terminal domain on the peptide-binding properties of DnaK by analyzing the binding affinities and binding kinetics of fluorescent peptides to various C-terminal fragments of DnaK. MATERIALS AND METHODS DnaK protein was purified as previously reported (18). The DnaK C-terminal fragments DnaK 384 – 638 (residues 384 to 638), DnaK 389 – 607 (residues 389 to 607), and DnaK 386 –561 (residues 386 to 561) were kindly provided by William Burkholder (Columbia University), Xun Zhao (Columbia University), and Gregory Flynn (University of Oregon), respectively. Bovine serum albumin was purchased from New England Biolabs. Protein concentration was determined by Bio-Rad Assay with IgG as the standard. Peptides were synthesized at the Biopolymer Lab at MIT with a peptide synthesizer (Applied Systems, 430A). N-((2-(Iodoacetoxy)ethyl)-N-methyl)amino-7-nitrobenz-2-oxa-1,3-diazole (I-ANBD) was purchased from Molecular Probes. Iodo[1-14C]acetamide was purchased from Amer-
sham. Fluorescence measurements were carried out on a Perkin– Elmer fluorescence spectrometer (Model LS50). Peptide-binding kinetics were studied on a stop-flow apparatus from Applied PhotoPhysics. Preparation of fluorescent probe ANBD-labeled peptides. The procedure used to prepare PepC-L9C-ANBD is described in detail below. The other fluorescent-modified peptides, PepC-V5C-ANBD, PepCL6C-ANBD, PepC-S7C-ANBD, and PepC-S8C-ANBD, were prepared by the same method. Two and a half milligrams (6.2 mmol) I-ANBD in 0.5 ml acetonitrile was added to 10 mg (7.1 mmol) PepC-L9C-ANBD in 0.5 ml of 50 mM Tris buffer (pH 8.0). The resulting solution was incubated at 37°C for an hour. PepC-L9C-ANBD was then purified by HPLC with a Waters Delta-pak DnaK 384 – 638 cartridge column. The reverse-phase column was run at 2 ml/min started with 100% buffer A [0.1% trifluoroacetic acid (TFA)] to 100% buffer B (80% acetonitrile, 0.085% TFA) within 60 min. PepC-L9C and PepC-L9C-ANBD were eluted at 25.5 and 28.3 min, respectively. PepC-L9C-ANBD peak was collected and lyophilized. Preparation of [1-14C]acetamide-labeled peptides. In order to dissolve the radiochemical, 100 ml of 50 mM Tris buffer (pH 8.0) was pipetted into the ampule the manufacture used to store 50 mCi iodo[1-14C]acetamide. Iodo[1-14C]acetamide solution was then transferred to 5.0 mg (6.3 mmol) PepNR-S6C in 0.5 ml of 50 mM Tris buffer (pH 8.0) followed by incubation at 37°C for half an hour. Unlabeled iodoacetamide (2.3 mg, 12.7 mmol) was then added to the reaction solution, and the resulting buffer was further incubated at 37°C for half an hour. [1-14C]Acetamide-modified peptide was purified by HPLC with a Waters Delta-pak DnaK 384 – 638 cartridge column. The reverse-phase column was run at 2 ml/min started with 100% buffer A (0.1% TFA) to 100% buffer B (80% acetonitrile, 0.085% TFA) within 60 min. PepNR-S6C and PepNR-S6C-[1-14C]acetamide were eluted at 21.9 and 20.1 min, respectively. The radioactive peak was collected and lyophilized. The specific radioactivity of PepNR-S6C[1-14C]acetamide was 6.3 mCi/mmol. Peptide binding assay using fluorescent peptides. Emission spectra of fluorescent peptides were recorded on a fluorescence spectrometer upon incubation with relevant proteins at 37°C for 1 h with excitation at 480 nm. The excitation slit was set to 5.0 nm and the emission slit was set to 10.0 nm. Peptide-binding assays were carried out in 0.5 ml of 50 mM Tris (pH 7.4), 50 mM KCl, 10 mM MgCl2, and 5 mM b-mercaptoethanol buffer containing relevant peptides and various concentrations of proteins. The maximum fluorescence intensities were measured after incubation at 37°C for 1 h. Data were fit by nonlinear least squares analysis, using Igor software for the Macintosh from Wavemetrics. Dissociation constants from direct titration of the fluorescent peptides with proteins were calculated using the Langmuir isotherm: U 5 [protein]/[protein] 1 Kd z, where U is the fraction of peptide bound proteins and Kd is the apparent dissociation constant. Measurements of peptide binding kinetics. The kinetics of PepCL9C-ANBD binding to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 were studied on a stopped flow apparatus with equal volume mixing at room temperature (22°C). The peptide was excited at 480 nm, and the fluorescence intensity was measured with a filter cutoff at 495 nm. The data were analyzed on Spectrakinetic Workstation V4.09 provided by Applied Photophysics. Peptide binding assay with radiolabeled peptides. A gel filtration column, Biospin-30 (Bio-Rad), was used to separate free peptides with DnaK-bound peptides. Then, 50 ml of 50 mM Tris (pH 7.4), 50 mM KCl, 10 mM MgCl2, and 5 mM b-mercaptoethanol buffer containing relevant peptides and various concentrations of proteins was incubated at 37°C for 1 h before being loaded into the Biospin-30 columns, followed by centrifugation on a Fisher Centrific centrifuge at its maximum speed for 5 min. The radioactivity of the flowthrough, which contains the DnaK-bound peptide, was measured
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
FIG. 1. ANBD.
Chemical structures of I-ANBD and hydroxyethylthio-
by scintillation counting. Dissociation constants of radiolabeled peptides were calculated following the same procedure as the determination of dissociation constants of the fluorescent peptides.
RESULTS
Binding of ANBD-Modified Monocysteine Peptide C Derivatives to DnaK In the course of a previous study in which we used peptide photocrosslinking to identify elements of the peptide binding site of DnaK, we synthesized monocysteine variants of Peptide C in which a single cysteine was substituted for Val5, Leu6, Ser7, and Ser8, respectively (17). These Peptide C derivatives, PepC-V5C, PepC-L6C, PepC-S7C, and PepC-S8C, were modified with a cysteine specific photocrosslinker N-iodoacetamidobutyl-4-azidosalicylic acid (I-ABASA) to yield the modified peptides PepC-V5C-ABASA, PepC-L6C-ABASA, PepC-S7C-ABASA, and PepC-S8C-ABASA. When these peptides were used in crosslinking studies with DnaK, we noticed that the efficiencies of PepC-V5C-ABASA and PepC-L6C-ABASA crosslinking to DnaK were significantly weaker than that of PepC-S7C-ABASA and PepC-S8C-ABASA (unpublished data). And, as a result of this relatively weak crosslinking, we were unable to identify the points of attachment of PepC-V5C-ABASA and PepC-L6C-ABASA to DnaK. Since the N-terminal residues of peptide C had been implicated in the binding to DnaK (19), it seemed likely that modifications at residue Val5 and Leu6 might be interfering with peptide binding to DnaK. To examine this observation further, a cysteine-specific fluorescence probe, I-ANBD (Fig. 1), was reacted with all of the four monocysteine derivatives of Peptide C to generate the fluorescent peptides PepC-V5CANBD, PepC-L6C-ANBD, PepC-S7C-ANBD, and PepCS8C-ANBD. In addition, we prepared another monocysteine variant of Peptide C, PepC-L9C, and also modified it with I-ANBD to yield PepC-L9C-ANBD. Peptides carrying fluorescent probes, such as acrylodan, dansyl, fluorescein, and ANBD, had been previously used to study the peptide binding properties of DnaK and Hsc70 (20 –23). These fluorescent probes were either attached to the N-terminal residue of the
179
peptides by reacting with an N-terminal cysteine residue (22) or by reacting with the N-terminal amino group (21, 23) or attached to the C-terminal of the peptide by reacting with a C-terminal cysteine residue (20). Upon incubation with DnaK at identical concentrations, the fluorescence intensity of these modified Peptide C derivatives, PepC-V5C-ANBD, PepC-L6CANBD, PepC-S7C-ANBD, PepC-S8C-ANBD, and PepCL9C-ANBD, increased to various degrees (Fig. 2). Upon incubation with DnaK, the fluorescence intensity of PepC-L9C-ANBD and PepC-S8C-ANBD increased 2.6and 2.5-fold, respectively. Moderate increases (1.9- and 1.5-fold) were observed for PepC-S7C-ANBD and PepC-V5C-ANBD, while PepC-L6C-ANBD gave no increase in fluorescent intensity upon incubation with DnaK. Several pieces of evidence indicated that the observed increases in the fluorescence intensity of these ANBD-modified peptides were due to specific binding to the peptide-binding site of DnaK. First, the incubation of DnaK with 2-hydroxyethylthio-ANBD which had been prepared by reacting b-mercaptoethanol with I-ANBD did not change its fluorescence spectrum, suggesting ANBD itself does not interact with DnaK (Fig. 3a). Second, inclusion of 5 mM Peptide C in a buffer containing 2 mM DnaK and 1 mM PepC-L9C-ANBD led to a 45% decrease of the fluorescence intensity of DnaK-bound PepC-L9C-ANBD, indicating that unmodified Peptide C competes with PepC-L9C-ANBD for binding to DnaK (Fig. 3b). Third, in the presence of 1.0 mM ATP, the fluorescence intensity of PepC-L9CANBD decreased 63%, suggesting ATP can release PepC-L9C-ANBD from DnaK as it does to other DnaKbound peptides (Fig. 3c). Fourth, incubation of PepCL9C-ANBD with BSA did not affect its fluorescence intensity indicating the fluorescence intensity increase is due to specific interaction with DnaK (Fig. 3c). Similar phenomena were also observed for PepC-V5CANBD, PepC-L6C-ANBD, PepC-S7C-ANBD, and PepCS8C-ANBD (data not shown). Taken together, these observations indicated that the Kd’s of these fluorescent peptides can be determined by measuring their increases in fluorescence intensity upon incubation with various concentrations of DnaK. A Langmuir curve of PepC-L9C-ANBD binding to DnaK is shown in Fig. 4. The dissociation constants of various fluorescence-labeled peptides to DnaK determined by this method are listed in Table I. As indicated by the dissociation constants of the modified peptides, the presence of the bulky ANBD substitutent at position 6 resulted in the lowest binding affinity, indicating residue 6 is critical for Peptide C binding to DnaK. ANBD modification at position 5 less severely but still significantly interferes with peptide binding to DnaK, while modification at positions 7, 8,
180
ZHANG AND WALKER
FIG. 2. Fluorescence intensity changes of 1.0 mM PepC-V5C-ANBD (a), PepC-L6C-ANBD (b), PepC-S7C-ANBD (c), PepC-S8C-ANBD (d), and PepC-L9C-ANBD (e) upon incubation with 2.0 mM DnaK at 37°C for 1 h.
or 9 had relatively minor effects on peptide binding to DnaK. Starting from position 6, the further away the fluorescent probe was from the N-terminus of the peptide, the less the interference of the ANBD modification on the peptide binding.
Studies of the Peptide Binding Properties of DnaK and Various C-Terminal Fragments Like other members of the Hsp70 class, DnaK consists of a 45-kDa N-terminal domain which binds ATP
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
181
FIG. 3. (a) A stock solution of 100 mM 2-hydroxyethylthio-ANBD was prepared by treating 100 mM I-ANBD in dimethyl sulfoxide with 500 mM b-mercaptoethanol at 37°C for 1 h. The emission spectra of 2.0 mM 2-hydroxyethylthio-ANBD (■) and 2.0 mM 2-hydroxyethylthio-ANBD with 5.0 mM DnaK (1) upon incubation at 37°C for 1 h were recorded with excitation at 480 nm. (b) The emission spectra of 1.0 mM PepC-L9C-ANBD (■) and 1.0 mM PepC-L9C-ANBD (1) that was incubated at 37°C with 2.0 mM DnaK for 1 h. (c) The emission spectra of 1.0 mM PepC-L9C-ANBD (■), 1.0 mM PepC-L9C-ANBD (1) that was incubated at 37°C with 2.0 mM DnaK for 1 h, 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK that was treated with 2.0 mM ATP for 0.5 h (E), and 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK that was treated with 5.0 mM Peptide C (‚) at 37°C for 0.5 h.
and ADP and has an ATPase activity, an 18-kDa peptide-binding domain which is directly adjacent to the ATPase domain, and a 7-kDa extreme C-terminal region whose function is not well understood (4 – 6, 16). To explore the nature of the extensive interactions between the various domains of DnaK, we compared the peptide-binding properties of full-length DnaK with those of three C-terminal fragments of DnaK (Fig.
5). These different classes of C-terminal fragments of DnaK have been used to gain insight into the nature of the interaction between DnaK with various peptide substrates. One of these (DnaK 384 – 638) consists of the whole C-terminal domain of DnaK from residue 384 to residue 638 with a histidine tag at the N-terminus (24). The second (DnaK 389 – 607), whose crystal structure has been solved, begins at residue 389 and
182
ZHANG AND WALKER
FIG. 4. The Langmuir curve of PepC-L9C-ANBD binding to DnaK. 1.0 mM PepC-L9C-ANBD was incubated with various concentrations of DnaK at 37°C for 1 h. The fluorescence intensity at 540 nm was measured, and the maximum fluorescence intensity increase was treated as 1.
ends at residue 607 (16). The third (DnaK 386 –561) consists of a fragment from residue 386 to residue 561 with a histidine tag at the N-terminus. The solution structure of DnaK 386 –561 has been studied by NMR and is similar to the corresponding fragments of Hsc70 (25, E. Zuiderweg and G. Flynn, personal communication). DnaK 386 –561 contains the eight b-strands that form the peptide binding pocket and the long a-helix that sits on the top of the peptide-binding pocket (16), but it lacks the final 77 amino acids of DnaK. The fluorescence intensity of PepC-L9C-ANBD (1.0 mM) increased 2.6-fold upon incubation with DnaK (2.0 mM) and DnaK 384 – 638 (2.0 mM). Interestingly, the maximum emission wavelength of bound PepCL9C-ANBD to DnaK shifted from 539.5 to 534.5 nm, whereas the maximum emission wavelength of
TABLE I
A List of Dissociation Constants Proteins DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK DnaK
384–638 389–607 386–561 384–638 389–607 386–561
Peptides PepC-V5C-ANBD PepC-L6C-ANBD PepC-S7C-ANBD PepC-S8C-ANBD PepC-L9C-ANBD PepC-L9C-ANBD PepC-L9C-ANBD PepC-L9C-ANBD PepNR-S6C-[1-14C]acetamide PepNR-S6C-[1-14C]acetamide PepNR-S6C-[1-14C]acetamide PepNR-S6C-[1-14C]acetamide
Kd (mM) 11.8 299 3.7 1.1 0.63 0.93 0.85 2.6 3.2 2.8 2.9 29.6
bound PepC-L9C-ANBD to DnaK 384 – 638 shifted from 539.5 nm to 535.5 nm (Fig. 6). The difference in the emission blue shift indicates that the N-terminal domain of DnaK does cause a change in the chemical environment of the peptide binding domain that can be sensed by a fluorescent probe at position 9 of Peptide C. The fluorescence intensity of PepC-L9CANBD increased to a comparable level upon incubation with DnaK 389 – 607 as with DnaK and DnaK 384 – 638 (Fig. 6). The maximum emission wavelength of DnaK 389 – 607-bound PepC-L9C-ANBD shifted from 539.5 to 534.5 nm, similar to the shift with full-length DnaK. In contrast, the fluorescence intensity of PepC-L9C-ANBD (1.0 mM) increased much less upon incubation with DnaK 386 –561 (2.0 mM) compared to the fluorescence increase observed upon incubation with DnaK, DnaK 384 – 638, or DnaK 389 – 607. In addition, there was no change in the maximum emission wavelength of PepC-L9CANBD upon incubation with DnaK 386 –561 (Fig. 8). By measuring fluorescent enhancements as a function of DnaK concentration, the Kd’s of PepC-L9CANBD to DnaK 384 – 638 and DnaK 389 – 607 were determined to be 0.93 and 0.85 mM, respectively, comparable to that of DnaK (0.63 mM). The Kd of PepCL9C-ANBD to DnaK 389 – 607 was determined to be 2.6 mM (Table 1), which is four times larger than that of DnaK. To investigate the peptide-binding properties of DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 further, the dissociation constants of a radiolabeled Peptide NR (NRLLLSG) derivative, PepNR-S6C-[1-14C]acetamide (NRLLLCG), were also assayed. Peptide NR was selected from a phagedisplayed peptide library (5) as a peptide that has a high affinity to DnaK and was the peptide that was bound to the C-terminal of DnaK in the crystal structure by Zhu et al. (16). A Langmuir curve of PepNRS6C-[1-14C]acetamide binding to DnaK is shown in Fig. 7 and the Kd’s of PepNR-S6C-[1-14C]acetamide to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 are listed in Table I. The Kd’s of PepNRS6C-[1-14C]acetamide to DnaK, DnaK 384 – 638, and DnaK 389 – 607 are comparable, but the affinity of
FIG. 5. Schematic representation of DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 fragments.
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
183
FIG. 6. The emission spectra of 1.0 mM PepC-L9C-ANBD (■), 1.0 mM PepC-L9C-ANBD that was incubated at 37°C with 2.0 mM DnaK (1) for 1 h, 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK 384 – 638 (Œ) at 37°C for 1 h, 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK 389 – 607 (3) at 37°C for 1 h, and 1.0 mM PepC-L9C-ANBD with 2.0 mM DnaK 386 –561 (h) at 37°C for 1 h.
PepNR-S6C-[1-14C]acetamide to DnaK 386 –561 is about 10 times weaker than that of DnaK. Taken together with the measurements using the ANBDmodified peptides, these observations indicate that the amino acids between 562 and 607 influence the stability of peptide binding. Kinetics of PepC-L9C-ANBD Binding to Various C-Terminal Fragments of DnaK Besides determining the equilibrium binding constants, we were interested in studying the kinetics of
peptide binding by the DnaK C-terminal fragments, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561, so that we could compare their properties to DnaK. The kinetics of PepC-L9C-ANBD binding to DnaK, DnaK 384 – 638, DnaK 389 – 607, and DnaK 386 –561 were studied using a stop-flow apparatus. The kinetic curves shown here represent an average of at least five consistent measurements, and one preparation was used in the kinetic measurements of the relevant protein. When 2.0 mM DnaK was mixed with 10.0 mM PepCL9C-ANBD, the time course of the fluorescence intensity increase showed a two-phase process as previously
184
ZHANG AND WALKER
respectively. The peptide offrate of DnaK 386 –561 is about twice as fast as that of DnaK, DnaK 384 – 638, or DnaK 389 – 607, but the onrate of DnaK 386 –561 is about six to seven times slower. Therefore, the weaker affinity of peptides to DnaK 386 –561 is primarily due to a slow binding rate. DISCUSSION
FIG. 7. The Langmuir curve of PepNR-S6C-[1-14C]acetamide binding to DnaK. 5.0 mM PepNR-S6C-[1-14C]acetamide was incubated with various concentrations of DnaK at 37°C for 1 h followed by separating the DnaK-bound PepNR-S6C-[1-14C]acetamide with the free PepNR-S6C-[1-14C]acetamide using Biospin columns. The radioactivity of the bound PepNR-S6C-[1-14C]acetamide was counted.
reported (20, 22, 26). The half-time (t1/2) of the first phase was determined to be 10.0 s, and the t1/2 of the slow phase was 130 s (Fig. 8a). By varying the protein concentration, the k1 and k2 of the first phase of PepCL9C-ANBD binding to DnaK were calculated to be 1.56 3 1013 M21 s21 and 0.98 3 1023 s21, respectively. The time course of PepC-L9C-ANBD binding to DnaK 384 – 638 showed a pattern very similar to that of PepC-L9C-ANBD binding to DnaK. It is a two-phase binding process with t1/2’s of 10.3 and 132 s (Fig. 8b). The k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK 384 – 638 were calculated to be 1.16 3 1013 M21 S21 and 1.08 3 1023 S21, respectively. Similarly, the binding curve of 10.0 mM of PepC-L9CANBD to 1.0 mM DnaK 389 – 607 showed a two-phase reaction with t1/2’s of 8.9 and 129 s (Fig. 8c). The k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK 389 – 607 were calculated to be 1.30 3 1013 M21 s21 and 1.11 3 1023 s21, respectively. Thus, the kinetics of peptide binding were largely unaffected by the removal of the N-terminal domain of the protein and by the removal of both the N-terminal domain and also the C-terminal 31 amino acids. In contrast, the kinetics of PepC-L9C-ANBD binding to DnaK 386 –561 are very different from those of DnaK, DnaK 384 – 638, and DnaK 389 – 607. Although it showed a characteristic two-phase binding curve, the binding of PepC-L9C-ANBD to DnaK 386 –561 is slower, and the net fluorescence intensity increase is smaller. The t1/2’s of the two phases are 209 and 311 s, respectively (Fig. 8d). The k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK 386 –561 were calculated to be 196 M21 s21 and 2.47 3 1023 s21,
A variety of peptides that are labeled with fluorescent probes have previously been used to study the interactions of peptides with Hsp70s, including DnaK (20 –23). In those studies, the fluorescent probes were attached either to the N-terminal residue or to the C-terminal residue. By attaching a fluorescent molecule to various locations in the middle of the amino acid sequence, we prepared five fluorescent monocysteine derivatives of Peptide C. They are PepC-V5C-ANBD, PepC-L6C-ANBD, PepC-S7C-ANBD, PepC-S8C-ANBD, and PepC-L9C-ANBD. Introduction of the bulky fluorescent ANBD moiety at different locations of Peptide C results in peptides with different binding affinities for DnaK. The degree of the change in DnaK binding affinity of a particular modified peptide is likely to be related to the relative importance of the residue at that position to the interactions of Peptide C with DnaK. Introduction of the ANBD moiety at positions 8 and 9 affords peptides with similar high binding affinities to DnaK, suggesting Ser8 and Leu9 may not play a critical role in peptide binding. However, introduction of ANBD at position 6 yields a peptide, PepC-L6C-ANBD, with a DnaK binding affinity 470 times lower than that of PepC-L9C-ANBD. This observation implies that Leu6 plays a central role in the binding of Peptide C to DnaK. A substitution of a cysteine–ANBD moiety at position 5 had the next strongest effect, reducing the binding affinity by 18-fold. The introduction of ANBD at position 7 had a less severe effect on peptide binding, reducing the binding affinity by 6-fold. It is interesting to relate our data to the structure of the DnaK C-terminal fragment bound to the peptide PepNR (NRLLLSG) which has been reported by Zhu et al. (16). The authors showed that Leu4 of PepNR (NRLLLSG) is the most critical residue for peptide binding and is completely buried in the peptide-binding pocket of DnaK. Leu3 of PepNR is also completely buried, while Leu5 is only moderately buried. Together, Leu3 and Leu4 provide most of the side chain contacts. Thus, it seems reasonable to suggest that Peptide C is positioned in the binding pocket of DnaK in a manner related to that of peptide PepNR. By this interpretation, one residue of Peptide C would be the residue that is completely buried into the peptide binding pocket of DnaK. Any modification at this residue would be predicted to lead to peptides with dramatically decreased DnaK binding affinity. Several
STUDIES OF DnaK AND PEPTIDE INTERACTIONS
185
FIG. 8. The time courses of 10.0 mM PepC-L9C-ANBD binding to 2.0 mM DnaK (a), 10 mM DnaK 384 – 638 (b), 1.0 mM DnaK 389 – 607 (c), and 10 mM DnaK 386 –561 (d).
residues of Peptide C would interact less with DnaK and their side chains would be predicted to project away from the binding pocket. Therefore, modifications with ANBD at these positions of Peptide C would result in peptides with smaller decreases in their DnaK binding affinity than the derivative modified at the completely buried residue. Since NMR data (19) has shown that the N-terminal part of Peptide C interacts with DnaK and it has been shown (9) biochemically that DnaK binds to short hydrophobic cores of four to five residues length, it is reasonable to suggest that residues 2–7 of Peptide C consists of the region that binds to the peptide-binding pocket of DnaK. Modifications of certain residues in this region severely hindered the DnaK binding affinity of the modified peptides, like ANBD modification at Leu6. Modifications of certain residues hindered their DnaK binding moderately, like ANBD modification at Val5 and Ser7. Consistant with this interpretation, substitution at position 9 does not interfere with DnaK binding. Furthermore, our observation that the DnaK 386 –561 fragment of DnaK, which lacks the C-terminal 77 amino acids, does not cause a blue shift when PepCL9C-ANBD binds suggests that these terminal resi-
dues influence the environment that is experienced by the extreme C-terminal of the peptides. Using both a fluorescent peptide and a radiolabeled peptide, we have shown that full-length DnaK, DnaK 384 – 638, and DnaK 389 – 607 are equally effective in forming a peptide complex. Upon incubation with DnaK, DnaK 384 – 638, and DnaK 389 – 607, the changes in the fluorescence intensity and the maximum emission wavelength of PepC-L9C-ANBD are very similar. The binding affinity of PepC-L9C-ANBD and PepNR-S6C-[1-14C]acetamide to DnaK, DnaK 384 – 638, and DnaK 389 – 607 are very close, and the binding kinetics of PepC-L9C-ANBD are comparable. All those observations suggest the last 31 residues of DnaK (residues 607– 638) do not have a significant effect on the peptide binding to DnaK, at least in this type of assay, which does not involve a cycle of ATP hydrolysis. On the other hand, the DnaK 386 –561 fragment (residues 386 –561), has significantly different peptide-binding properties compared to DnaK, DnaK 384 – 638, and DnaK 389 – 607. DnaK 386 –561 binds to PepC-L9C-ANBD 4 times more weakly and it binds to PepNR-S6C-[1-14C]acetamide 10 times more weakly. Kinetically, PepC-L9C-ANBD binding to DnaK 386 –
186
ZHANG AND WALKER
561 showed a slower onrate and a faster offrate. Those observations indicate residues 561 to 607, which form the C, D, and E a-helixes that are directly adjacent to the peptide binding pocket of DnaK (16), play important roles in stabilizing the DnaK/peptide complex. Interestingly, Hu and Wang (27) have reported that the last 10-kDa domain of Hsc70 plays a role in stabilizing the Hsc70/CMLA complex. Several papers (21, 22, 26) have reported that in the presence of ADP, peptide binding to DnaK and other Hsp70s, such as Hsc70, is a two-step process. The first step is relatively rapid, and the second step is slow, but results in “capture” of the peptide. The binding of PepCL9C-ANBD to DnaK behaved just like the other peptides reported. The half-time (t1/2 of the first phase was measured at 10.0 s, and the t1/2 of the slow phase was 130 s, and the k1 and k2 of the first phase of PepC-L9C-ANBD binding to DnaK were calculated to be 1.56 3 1013 M21 and 0.98 3 1023 s21, respectively. Schmid et al. (22) measured the t1/2 of the first and second step of the binding of a peptide, a-pp, to DnaK at 27 and 200 s, respectively, and the k1 and k2 of the first phase of a-pp binding to DnaK were determined to be 9.4 3 1013 M21 s21 and 4 3 1023 s21, respectively. Takeda and McKay (21) reported the k1 and k2 of the first phase of peptide faf1 binding to DnaK to be 1.21 3 1013 M21 s21 and 0.017 s21, respectively, and they also calculated the k1 and k2 of the second phase to be 0.013 and 0.0038 s21. The rate constant in the forward direction (binding) of the second phase is of the order 0.01 sec21 as reported (21); this is the same magnitude as the rate constants of the second phase of the fluorescence kinetics of PepC-L9C-ANBD binding to DnaK, suggesting the second phase of the fluorescence change corresponds to the second step of peptide binding. ACKNOWLEDGMENTS We are grateful to Paul Schimmel for the use of their fluorescence spectrometer and to Robert Sauer for the use of their stop-flow apparatus. We thank Max Gottesman, Wayne Hendrickson, and Gregory Flynn for generously providing us with the relevant Cterminal fragments of DnaK. This work was supported by U.S. Public Health Service Grant GM28988 to G.C.W.
REFERENCES 1. Bardwell, J. C., and Craig, E. (1984) Proc. Natl. Acad. Sci. USA 81, 848 – 852. 2. Lindquist, S., and Craig, E. A. (1988) Annu. Rev. Genet. 22, 631– 677.
3. Craig, E. A., Gambill, B. D., and Nelson, R. J. (1993) Microbiol. Rev. 57, 402– 414. 4. Chappell, T. G., Konforti, B. B., Schmid, S. L., and Rothman, J. E. (1987) J. Biol. Chem. 262, 746 –751. 5. Gragerov, A., Zeng, L., Zhao, X., Burkholder, W., and Gottesman, M. (1993) J. Mol. Biol. 235, 848 – 854. 6. Wang, T.-F., Chang, J.-H., and Wang, C. (1993) J. Biol. Chem. 268, 26049 –26051. 7. Churchich, J. E. (1995) Eur. J. Biochem. 231, 736 –741. 8. Fung, K. L., Hilgenberg, L., Wang, N. M., and Chirico, W. J. (1996) J. Biol. Chem. 271, 21559 –21565. 9. Flynn, G. C., Chappell, T. G., and Rothman, J. E. (1989) Science 245, 385–390. 10. Jordan, R., and McMacken, R. (1995) J. Biol. Chem. 270, 4563– 4569. 11. Buchberger, A., Valencia, A., McMacken, R., Sander, C., and Bukau, B. (1994) EMBO J 13, 1687–1695. 12. Blond-Elguindi, S., Cwirla, S., Dower, W., Lipshutz, R., Sprang, S., Sambrook, J., and Gething, M.-J. (1993) Cell 75, 717–728. 13. Flynn, G. C., Pohl, J., Flocco, M. T., and Rothman, J. E. (1991) Nature 353, 726 –730. 14. Takenaka, I. M., Leung, S.-M., McAndrew, S. J., Brown, J. P., and Hightower, L. E. (1995) J. Biol. Chem. 270, 19839 –19844. 15. Frydman, J., Nimmesgern, E., Ohtsuka, K., and Hartl, F. (1994) Nature 370, 111–117. 16. Zhu, X., Zhou, X., Burkholder, W. F., Gragerov, A., Ogata, C. M., Gottesman, M. E., and Hendrickson, W. A. (1996) Science 272, 1606 –1614. 17. Zhang, J., and Walker, G. C. (1996) J. Biol. Chem. 271, 19668 – 19674. 18. McCarty, J. S., and Walker, G. C. (1991) Proc. Natl. Acad. Sci. USA 88, 9513–9517. 19. Landry, S. J., Jordan, R., McMacken, R., and Gierasch, L. M. (1992) Nature 355, 455– 457. 20. Theyssen, H., Schuster, H.-P., Packschies, L., Bukau, B., and Reinstein, J. (1996) J. Mol. Biol. 263, 657– 670. 21. Takeda, S., and McKay, D. B. (1996) Biochemistry 35, 4636 – 4644. 22. Schmid, D., Baici, A., Gehring, H., and Christen, P. (1994) Science 263, 971–973. 23. Farr, C. D., Galiano, F. J., and Witt, S. N. (1995) Biochemistry 34, 15574 –15582. 24. Burkholder, W. F., Zhao, X., Zhu, X., Hendrickson, W. A., Gragerov, A., and Gottesman, M. E. (1996) Proc. Natl. Acad. Sci. USA 93, 10632–10637. 25. Morshauser, R., Wang, H., Flynn, G., and Zuiderweg, E. (1995) Biochemistry 34, 6261– 6266. 26. Banecki, B., and Zylicz, M. (1996) J. Biol. Chem. 271, 6137– 6143. 27. Hu, S.-M., and Wang, C. (1996) Arch. Biochem. Biophys. 332, 163–169.