Regulation of Hsp70 Function by HspBP1

Molecular Cell, Vol. 17, 367–379, February 4, 2005, Copyright ©2005 by Elsevier Inc.

DOI 10.1016/j.molcel.2004.12.023

Regulation of Hsp70 Function by HspBP1: Structural Analysis Reveals an Alternate Mechanism for Hsp70 Nucleotide Exchange Yasuhito Shomura,1 Zdravko Dragovic,1 Hung-Chun Chang,1 Nikolay Tzvetkov,1 Jason C. Young,1,4 Jeffrey L. Brodsky,2 Vince Guerriero,3 F. Ulrich Hartl,1 and Andreas Bracher1,* 1 Department of Cellular Biochemistry Max Planck Institute of Biochemistry Am Klopferspitz 18 82152 Martinsried Germany 2 Department of Biological Sciences University of Pittsburgh 274A Crawford Hall Pittsburgh, Pennsylvania 15260 3 Department of Animal Sciences and Department of Molecular and Cellular Biology P.O. Box 210038 University of Arizona Tucson, Arizona 85721

Summary HspBP1 belongs to a family of eukaryotic proteins recently identified as nucleotide exchange factors for Hsp70. We show that the S. cerevisiae ortholog of HspBP1, Fes1p, is required for efficient protein folding in the cytosol at 37ⴗC. The crystal structure of HspBP1, alone and complexed with part of the Hsp70 ATPase domain, reveals a mechanism for its function distinct from that of BAG-1 or GrpE, previously characterized nucleotide exchange factors of Hsp70. HspBP1 has a curved, all ␣-helical fold containing four armadillo-like repeats unlike the other nucleotide exchange factors. The concave face of HspBP1 embraces lobe II of the ATPase domain, and a steric conflict displaces lobe I, reducing the affinity for nucleotide. In contrast, BAG-1 and GrpE trigger a conserved conformational change in lobe II of the ATPase domain. Thus, nucleotide exchange on eukaryotic Hsp70 occurs through two distinct mechanisms. Introduction Molecular chaperones of the Hsp70 family have essential functions in cellular protein quality control, including the folding of newly synthesized polypeptides and the refolding of stress denatured proteins (Bukau et al., 2000; Frydman, 2001; Hartl and Hayer-Hartl, 2002). Hsp70 proteins consist of a regulatory N-terminal ATPase domain and a C-terminal peptide binding domain, which recognizes hydrophobic peptide segments exposed in non-native substrate proteins (DeLuca-Flaherty et al., 1990; Zhu et al., 1996). The ATP bound state of Hsp70 is characterized by high on and off rates for substrate. Hsp40 (DnaJ-type) cochaperones (and pep*Correspondence: [email protected] 4 Present address: Department of Biochemistry, McGill University, 3655 Promenade Sir William Osler, Montre´al, QC H3G 1Y6, Canada.

tide substrate itself) accelerate ATP hydrolysis on Hsp70, resulting in tight binding of substrate. Bound peptide is released after dissociation of ADP and rebinding of ATP. In the presence of Hsp40, ADP dissociation from Hsp70 is the rate-limiting step in the ATPase cycle (Bukau and Horwich, 1998). In the case of bacterial Hsp70 (DnaK) this step is accelerated by the nucleotide exchange factor GrpE (Harrison et al., 1997; Liberek et al., 1991). The structurally unrelated BAG (Bcl2-associated athanogene) domain may functionally replace GrpE in the eukaryotic cytosol (Ho¨hfeld and Jentsch, 1997). Surprisingly, the structures of the GrpE-DnaK and the BAG-1Hsc70 complex indicated a nucleotide exchange mechanism based on a common conformational switch in the ATPase domain, a 14⬚ outwards rotation of subdomain IIb about a hinge at the subdomain boundary (Harrison et al., 1997; Sondermann et al., 2001). The existence of several, otherwise unrelated BAG domain-containing proteins in the cytosol of mammalian cells suggests that these factors recruit Hsp70 for various specialized functions, including proteolysis by the ubiquitin/proteasome machinery (Demand et al., 2001; Lu¨ders et al., 2000). The low abundance of these proteins (Murphy et al., 2001) and the absence of soluble BAG domain proteins in S. cerevisiae (Sondermann et al., 2002) argue further against a general role of BAG proteins as Hsp70 nucleotide exchange factors (NEFs). Human HspBP1 (Hsp70 binding protein 1) was recently characterized as a relatively abundant cytosolic Hsp70/Hsc70 NEF unrelated to GrpE and BAG-domain proteins (Kabani et al., 2002b; Raynes et al., 2003; Raynes and Guerriero, 1998). An inhibitory effect of HspBP1 in Hsp70-assisted refolding reactions was originally suggested based on in vitro experiments employing high concentrations of HspBP1 relative to Hsp70 (Raynes and Guerriero, 1998). Sequence homologs of HspBP1 were identified throughout the eukaryotic domain. The S. cerevisiae ortholog Fes1p functions with Ssa-class Hsp70s (Kabani et al., 2002a), and the paralogous mammalian proteins BAP (BiP-associated protein) and S. cerevisiae Sls1p/Sil1p represent Hsp70 NEFs in the endoplasmic reticulum (ER) that function with the ER lumenal Hsp70, BiP (Chung et al., 2002; Kabani et al., 2000). However, the molecular interactions between these NEFs and Hsp70s have not been examined in detail. Here, we show that HspBP1 homologs are required for efficient Hsp70-mediated protein folding in the cytosol in vivo and report the three-dimensional structures of the conserved core domain of HspBP1 alone and in complex with a fragment of the ATPase domain of mammalian Hsp70. Our results indicate that HspBP1 catalyzes nucleotide exchange on Hsp70 by a mechanism distinct from that of GrpE and BAG-1. Results and Discussion Fes1p Is Necessary for the Efficient De Novo Folding of Luciferase In S. cerevisiae, the single BAG-1 homolog, Snl1p, is a transmembrane protein of the ER and nuclear mem-

Molecular Cell 368

Figure 1. Firefly Luciferase Folding in the Cytosol of S. cerevisiae (A) Firefly luciferase (FL) was expressed in the FES1-deleted S. cerevisiae strain, JY053 (⌬fes1), and in an isogenic wild-type strain, YPH499 (WT). FL expression and folding were monitored as described (Agashe et al., 2004). Values for wild-type yeast grown at 30⬚C were used as an internal reference. The lower panel shows an immunoblot used for quantitation of FL protein. The bands correspond to total (T), soluble (S), and insoluble FL (P). (B) Analysis of apyrase-treated yeast lysates expressing FL by Superdex-200 (Amersham) size exclusion chromatography. The upper and lower panels show wild-type and ⌬fes1 lysates, respectively. White bars denote normalized FL activity per fraction; black bars represent FL protein as determined by quantifying the immunoblots

branes with functions apparently specialized for these organellar compartments (Sondermann et al., 2002). Fes1p, the S. cerevisiae ortholog of HspBP1, is the only known cytosolic NEF for Hsp70 chaperones in this species. Deletion of the FES1 gene in S. cerevisiae moderately impaired growth compared to wild-type cells at 37⬚C as tested by serial dilution onto rich media (YPD). Growth of the ⌬fes1 cells at 30⬚C was identical to wildtype (data not shown), and this temperature-sensitive behavior is consistent with the previously reported phenotype (Kabani et al., 2002a). To investigate the effect of FES1 deletion on Hsp70 function, we analyzed the efficiency of de novo folding in the cytosol of wild-type and ⌬fes1 cells with firefly luciferase (62 kDa) as a sensitive model protein (Agashe et al., 2004). Luciferase has been shown to require the Hsp70 chaperone system for refolding in vitro and for de novo folding upon translation (Frydman et al., 1994, 1999). Efficient folding of this protein in yeast is dependent on the Ssa-class Hsp70, the equivalent of mammalian Hsc70/Hsp70, and on its Hsp40 cochaperone Ydj1p (data not shown) (Li et al., 2003; Lu and Cyr, 1998). Upon expression in ⌬fes1 cells at 37⬚C, the specific activity of luciferase was reduced by about 50% relative to wildtype cells without affecting the expression level and translation efficiency of the protein (Figure 1A). Similar results were obtained in wild-type and ⌬fes1 mutant yeast when firefly luciferase expression was driven by a different inducible promoter (Ahner et al., 2005). Formation of insoluble aggregates of luciferase was not observed in ⌬fes1 yeast, but size exclusion chromatography of cell lysates revealed that luciferase eluted as a broad peak centered around 160 kDa. In contrast, luciferase protein and activity produced in wild-type cells coeluted as a sharp symmetrical peak at ⵑ70 kDa (Figure 1B). The inactive, high molecular weight form of luciferase may be chaperone bound. Indeed, coprecipitation experiments from cell lysates using the C-terminal His-tag of luciferase showed that a substantial fraction of total Ssa1p (ⵑ5%) was associated with luciferase in ⌬fes1 cells at 37⬚C, but not in wild-type cells (Figure 1C). Likewise, ⵑ10% of total Ydj1p was associated with luciferase in ⌬fes1 cells. Although ⌬fes1 cells expressed approximately 2-fold more Ydj1p than wild-type cells under these conditions, the amount of Ydj1p associated with luciferase in ⌬fes1 cells was at least 10-fold greater than in the wild-type (Figure 1C). These results indicate that Fes1p is required for fully efficient Hsp70-mediated folding of some proteins at 37⬚C, consistent with a role of Fes1p as a NEF in regulating substrate binding and release by Hsp70. Because eukaryotic Hsp70s are not strictly dependent on NEFs, unlike bacterial DnaK (Minami et al., 1996), it is plausible that Fes1p function becomes limiting only at elevated

shown below. The positions of molecular mass markers are indicated. (C) Association of luciferase with Ssa1p and Ydj1p in ⌬fes1 cells at 37⬚C. Apyrase-treated yeast lysates expressing luciferase with a C-terminal His-tag (input) were incubated with Ni-NTA agarose beads and bound proteins (FL-bound) probed with antisera specific for Ssa1p and Ydj1p. The intensity of the bands was analyzed by densitometry.

Crystal Structure of an HspBP1-Hsp70 Complex 369

growth temperatures. Previous findings of an inhibitory effect of HspBP1 on Hsp70 function in vitro (Raynes and Guerriero, 1998) may have been caused by the use of unphysiologically high concentrations of HspBP1 relative to Hsp70. The Core Domain of HspBP1, BP1c, Retains the Hsp70 Nucleotide Exchange Activity The conserved core domain of human HspBP1 (BP1c, residues 84–359) (McLellan et al., 2003), which is shared with Fes1p, moderately stimulated the ATPase activity of porcine Hsc70 in the presence of Hsp40 (Figure 2A). BP1c accelerated the apparent koff rate of the fluorescent ADP-analog MABA-ADP from Hsp70 by two orders of magnitude while only moderately affecting the dissociation rate for MABA-ATP (Figures 2B and 2C; Supplemental Table S1 available at http://www.molecule.org/cgi/ content/full/17/3/367/DC1/). In an assay monitoring the ATP-dependent release of a non-native substrate protein, the ligand binding domain of glucocorticoid receptor from Hsc70·ADP, BP1c triggered the dissociation of Hsc70 from the complex with an efficiency similar to the BAG-domain of BAG-1 (BAG, residues 151–264) (Figure 2D). Thus, BP1c retains the function of full-length HspBP1 as an Hsp70/Hsc70 NEF (Kabani et al., 2002b).

Figure 2. Functional Characterization of HspBP1(84-359), BP1c (A) BP1c stimulates the ATPase activity of Hsc70. Hsc70 (3 ␮M), Hsp40 (1.5 ␮M), and BP1c (3 ␮M) were incubated at 30⬚C in the presence of 2 mM ATP as indicated. ATPase rates were determined as described (Liberek et al., 1991). (B and C) Effect of BP1c on ADP dissociation rates from Hsp70. Dissociation of the fluorescent MABA-analog of ADP from Hsp70 was monitored by stopped flow measurements as described (Brehmer et al., 2001). (C) shows a blow up of the early reaction phase. Averages of at least five experiments are shown. (D) BP1c stimulates Hsc70 release from non-native substrate protein in a nucleotide-dependent manner. 35S-methionine-labeled Hsc70 was bound to immobilized and partially denatured ligand binding domain (LBD) of glucocorticoid receptor (Young and Hartl, 2000). Hsc70 release from LBD-beads was measured after incubation with BAG or with BP1c in the presence or absence of ATP (2 mM).

Crystal Structure of BP1c The crystal structure of BP1c was solved at 2.1 A˚ resolution by single wavelength anomalous diffraction with a selenomethionine-labeled derivative. The asymmetric unit contains a crystallographic dimer of almost perfect 2-fold symmetry (rms deviation 0.248 A˚ for C␣-positions). The structure shows that BP1c has an elongated shape and consists entirely of ␣-helical repeats (Figure 3A). The central four repeating units, helices ␣3–␣14, are each composed of three helices arranged in an open triangle and are structurally most similar to Armadillo repeats. The four helical repeats can be superposed with rms deviations from 0.84 to 1.56 A˚; the rms deviations from the canonical ␤-catenin Arm repeat 7 (Huber et al., 1997) are between 0.86 and 1.41 A˚. Consequently, the Armadillo repeat proteins ␤-catenin (Huber et al., 1997) and importin-␣ (Conti et al., 1998) and the related subunit H of the V-type ATPase (Sagermann et al., 2001) were identified as the closest structural homologs of the Armadillo repeat fragment of BP1c with a DALI search against the PDB database (Holm and Sander, 1996). Similar to the homologous proteins identified, the succession of repeats in BP1c exhibits a right-handed superhelical twist resulting in the formation of a slight curvature and a concave surface. BP1c has a continuous hydrophobic core that is capped at both ends by slightly longer ␣ helices deviating from the regularity of the central repeats. Based on the conservation of small and hydrophobic residues mediating contacts within and between helical repeats, HspBP1 and Fes1p as well as the

Supernatant fractions (S) contain released Hsc70 and pellet fractions (P) contain LBD beads with bound Hsc70. Supernatants and pellets were analyzed by SDS-PAGE and the radioactive signal was quantified by phosphoimager analysis.

Molecular Cell 370

Figure 3. Structure of BP1c and of the BP1c-Hsp70 Complex (A) Ribbon diagram of BP1c. The helical repeat units are highlighted in rainbow colors. The termini and secondary structure elements are labeled. (B) Subdomain composition of the ATPase domain of Hsp70 (Flaherty et al., 1990). The subdomains Ia, Ib, IIa, and IIb are indicated in brown, blue, yellow, and white, respectively. Bound ADP·Mg is shown in ball-and-stick representation. The hinge region for the BAG-induced rotation of subdomain IIb is indicated by arrows. The positions of Trp90 and of the chymotrypsin cleavage sites after residues Phe68 and Tyr183 are also indicated as orange spheres. Phe68 faces the cleft between subdomains Ib and IIb. (C) Superposition of free and Hsp70 bound BP1c. The corresponding C␣ traces are shown in red and blue, respectively. The orientation

Crystal Structure of an HspBP1-Hsp70 Complex 371

homologous BAP proteins of the ER are likely to share a common fold in this region (Figure 4A). Crystal Structure of the BP1c-Hsp70 Complex Although BP1c forms a stable complex with the ATPase domain of Hsp70 in the absence of adenine nucleotide, attempts to crystallize this complex failed, suggesting that BP1c induces additional flexibility in Hsp70 compared with BAG. Limited proteolysis experiments revealed that upon binding to HspBP1, lobe I (subdomains Ia and Ib [Flaherty et al., 1990], see Figure 3B) of the ATPase domain becomes susceptible to proteolytic degradation (McLellan et al. [2003] and discussed below). We crystallized the remaining fragment comprising residues 184 to 371 (lobe II, i.e., subdomains IIa and IIb) in complex with BP1c in the presence of the ATP analog AMP-PNP and solved the structure by molecular replacement at 2.9 A˚ resolution. The asymmetric unit contains two almost identical copies of the complex (rms deviation 0.462 A˚). In the complex, BP1c wraps around subdomain IIb of the Hsp70 ATPase domain (Figure 3C). No contacts were found to subdomain IIa. The concave face of Armadillo repeats 1–3 of BP1c, which exhibits the highest degree of surface conservation (Figures 3D and 4), surrounds most of ␣ helix 8 and a ␤ hairpin composed of strands 16 and 17 in Hsp70 (numbered according to the fulllength ATPase domain) (Figure 3E). These contacts form the major interaction site between the two proteins (Figure 3E). In addition, salt bridge contacts to ␣ helix 7 of Hsp70 form a minor interaction site, causing the C-terminal third of BP1c to curve inwards compared to the structure of free BP1c (Figure 3C). Overall, the contacts between BP1c and Hsp70 bury 2267 A˚2 of accessible surface area and reach almost 180⬚ around subdomain IIb. Complementarity in shape and polarity leads to a tight and extended interface between BP1c and Hsp70 at the major interaction site stabilized by numerous hydrogen bonds and van der Waals interactions. A hydrogen bond network formed by the invariant surface residues Gln174BP1c, Ser213BP1c, and Arg217BP1c contacts the Hsp70 backbone at residues 283 and 281 and the side chain of Glu283Hsp70, promoting hydrogen bonds from Asn175BP1c to Ser281Hsp70 and to Thr273Hsp70 (Figure 5A). A second conserved hydrogen bond network is organized around N⑀ of Lys249BP1c. The side chain of Arg269Hsp70 makes a tight contact with BP1c by reaching into a pocket on the surface of BP1c that is lined by the conserved residues Ala137BP1c and Met134BP1c (Figure 5B). These interactions are summarized in Figure 3F. The

high degree of conservation of the residues involved in contacts at the major interaction site suggests that all members of the HspBP1 family bind the corresponding Hsp70 homolog in a similar manner to human HspBP1 and, thus, are likely to catalyze nucleotide exchange by a common mechanism (Figure 4). Indeed, substitution of Ala79 in Fes1p (Ala137 in HspBP1) by Arg, which is expected to block the binding pocket for Arg265 in Ssa1p (Arg269Hsp70), strongly reduced Ssa1p binding in vitro (Figure 5C). Because Fes1p(A79R) possessed residual NEF activity (data not shown), an additional mutation, R195A (Lys249 in HspBP1), was introduced. This additional mutation caused the essentially complete loss of Fes1p binding to Ssa1p (Figure 5C). Importantly, expression of Fes1p(A79R,R195A) failed to rescue the ⌬fes1 temperature-sensitive growth phenotype in contrast to plasmid-encoded wild-type Fes1p (Figure 5D). Thus, a direct interaction between Fes1p and Ssa1p is required for Fes1p function in vivo in line with the role of Fes1p as a NEF. However, an additional function of Fes1p in Hsp70-assisted folding reactions cannot be ruled out. The conformation of subdomain IIb in the ATPase domain is unaltered compared to the known structures of the domain in complex with BAG and in complex with nucleotide (Flaherty et al., 1990; Sondermann et al., 2001). Binding of AMP-PNP was required for crystallization, presumably, to stabilize the subdomain arrangement of the Hsp70 lobe II. Indeed, the relative position of the subdomains in the complex was indistinguishable from that in the ADP bound form of the ATPase domain (rmsd 0.64 A˚ over 165 residues) (Flaherty et al., 1990). However, ATP competed efficiently with BP1c binding to the complete ATPase domain of Hsp70, as indicated by an ⵑ120 times lower affinity of the ATPase domain for BP1c in the presence of 5 mM ATP compared to that in the absence of nucleotide (Supplemental Table S2). Comparison of the BP1c-Hsp70 and the BAG-Hsc70 Complexes BP1c and BAG catalyze nucleotide exchange on Hsp70/ Hsc70 by distinct mechanisms as revealed by a structural comparison of the respective complexes. The interaction surface of Hsp70 with BAG overlaps partially with that for BP1c and both proteins have a very similar binding affinity for the Hsp70 ATPase domain (Supplemental Table S2). However, whereas BAG binds to the central cleft of the ATPase domain from the top, making contacts with subdomains IIb and Ib, HspBP1 ap-

corresponds to a 150⬚ rotation around the x axis compared to Figure 3A (top). The C␣ traces of Hsp70 subdomains IIa and IIb are indicated in yellow and gray, respectively. For the superposition, 228 out of 264 C␣ positions were matched with an rms deviation of 1.43 A˚ by using the DP algorithm as implemented in the program LSQMAN (Kleywegt and Jones, 1994). (D) Side view of the BP1c-Hsp70 complex. The similarity score for the alignment of cytosolic HspBP1 sequences was plotted onto the van der Waals surface of BP1c with a scale from green (identical) to light red (no conservation) to indicate the surface conservation of HspBP1. Surface residues involved in interactions with Hsp70 are indicated. Yellow and white strands denote the backbone of subdomains IIa and IIb of the Hsp70 ATPase domain, respectively. The ordered fragment of the bound AMP-PNP molecule is shown in ball-and-stick representation. (E) Side view of the BP1c-Hsp70 complex indicating the areas defined as major and minor interaction site as red and blue surfaces, respectively. The backbone of Hsp70 is shown as ribbon diagram. Secondary structure elements of Hsp70 involved in interactions with HspBP1 are indicated. (F) Schematic representation of the interactions between BP1c and Hsp70. Residues involved in polar and hydrophobic interactions are connected by blue and red lines, respectively. Main chain contacts are indicated with dotted lines. Magenta and red characters indicate residues that are conserved and invariant among cytosolic HspBP1 homologs, respectively.

Molecular Cell 372

Figure 4. Evolutionary Conservation of HspBP1 Sequence alignment of the core domains of HspBP1 and BAP homologs (A) and of their putative Hsp70 binding regions (B). Secondary structure elements in HspBP1 and Hsp70 are indicated above the sequences. In each group (indicated by numbers), similar residues are

Crystal Structure of an HspBP1-Hsp70 Complex 373

Figure 5. Molecular Interactions between BP1c and Hsp70 (A) HspBP1 residue Arg217 is at the center of an extended hydrogen bond network in the interface region with Hsp70. All residues indicated are strictly conserved in HspBP1 homologs. (B) Hsp70 residue Arg269 reaches into a conserved pocket in the surface of HspBP1 making intimate van der Waals contacts. At the bottom of the binding pocket, the overall dipole of helix ␣5 points toward the guanidinium group of Arg269. The molecules are shown as ball-and-stick models in stereo representation; carbon atoms in HspBP1 and Hsp70 are shown in beige and gray, respectively. Residues in HspBP1 and Hsp70 are indicated by black and red lettering, respectively. Composite omit electron density contoured at 1.0 ␴ is shown as meshwork in cyan. Hydrogen bond interactions are indicated by dotted lines. (C) Ssa1p binding to Fes1p is strongly reduced by the combined mutation A79R (corresponding to Ala137 in HspBP1) and R195A (K249 in HspBP1). His-tagged Fes1p, Fes1p(A79R), and Fes1p(A79R,R195A) were incubated with purified Ssa1p at a ratio of 1:1 and applied to Ni-affinity resin. After being washed with three column volumes, bound protein was eluted and analyzed on a Coomassie-stained SDS-PAGE gel. Bound Ssa1p was quantified by densitometry and corrected for nonspecific binding in the absence of Fes1p. Amounts of Ssa1p bound relative to that observed with wild-type Fes1p are shown. (D) Fes1p(A79R,R195A) does not restore thermotolerance to ⌬fes1 cells. The FES1deleted S. cerevisiae strain JY053 was transformed with centromeric plasmids encoding Fes1p(WT), Fes1p(A79R,R195A), or a plasmid without insert (vector). Serial dilutions of overnight cultures were inoculated on agar plates containing selective medium at 37⬚C for two days. No growth differences were observed at 30⬚C (data not shown).

shown in red and identical residues in bold lettering on red background. Blue frames indicate similarity between cytosolic and ER homologs. The HspBP1 residue closest to the nucleotide binding site of the Hsp70/Hsc70 ATPase domain is indicated by a blue circle. Amino acid residues involved in van der Waals and polar interactions with Hsp70 are indicated by green triangles and red asterisks, respectively. Blue triangles and black asterisks indicate Hsc70 residues involved in the interaction with the BAG-domain (Sondermann et al., 2001).

Molecular Cell 374

Figure 6. Comparison of BP1c-Hsp70 and BAG-Hsc70 Complexes (A–D) Top and side views of the BP1c-Hsp70 (A and C) and BAG-Hsc70 (B and D) complexes, respectively, shown in ribbon representation and aligned for superposition of subdomain IIb in the ATPase domain. The binding areas of BP1c and BAG on lobe II of the ATPase domain partially overlap. Secondary structure elements of Hsp70 involved in interactions are indicated. The color scheme for Hsp70 corresponds to that in Figure 3B; BP1c and BAG are shown as green ribbons. The arrow indicates the shift of subdomain IIb in the complex with BAG. Blue shapes indicate the approximate position of the parts of the Hsp70 ATPase domain missing in the complex structure with BP1c. (E) Regions of potential steric conflict between the N-terminal segment of BP1c and Hsp70/Hsc70 shown for the BAG bound conformation of the ATPase. The Hsp70 ATPase domain is depicted in the same conformation as in panel B with its accessible surface shown as a transparent hull. A red arrow indicates the region of potential steric conflict: helices ␣1 and ␣3 of BP1c would make the closest approach to subdomain Ib, and the extreme N terminus of BP1c (residues 87 to 91, shown as red ribbon) would intersect with the BAG bound, open conformation of the ATPase.

proaches lobe II of the ATPase sideways (Figure 6). Although subdomain IIb of the ATPase provides the main contact area for both proteins, the binding of BP1c to subdomain IIb is stronger than that of BAG (Supplemental Table S2). The additional contacts of BAG with subdomain Ib lock the ATPase domain in an “open” conformation in which subdomain IIb is rotated 14⬚ outwards about a hinge at the subdomain boundary (Figures 3B and 6B) (Sondermann et al., 2001). This conformation is incompatible with nucleotide binding (Sondermann et al., 2001). In contrast to the highly conserved binding surface of BP1c for subdomain IIb, the N-terminal region of BP1c exhibits only a low degree of surface conservation (Figure 3D), suggesting rather poorly defined interactions with subdomain Ib. The adjacent region of fulllength HspBP1, residues 1 to 83, may contribute to the interactions with lobe I. This segment however is poorly conserved (data not shown) and exhibits almost unaltered protease sensitivity in the complex with the ATPase domain of Hsp70 compared with HspBP1 alone, arguing against a significant contribution to binding (McLellan et al., 2003). Interestingly, superpositioning of the ADP bound form

of the ATPase domain with the structure in complex with BP1c indicates a severe steric conflict of the N-terminal segment of BP1c (helices ␣1 and ␣3; residues 87–95 and 133–137) with subdomain Ib (Figure 6E). Furthermore, the open conformation of the ATPase domain observed in the complex with BAG would still be in steric conflict with the first turn of helix ␣1 in BP1c (residues 87–91) (Figure 6E). Thus, BP1c would be expected to induce a substantial distortion of the ATPase domain not seen in the complex with BAG, possibly a mechanical separation of the opposite lobes of the ATPase domain with the tightly bound subdomain IIb serving as a support for BP1c (Figure 7C). This disruption of both lobes of the ATPase domain would promote the release of bound nucleotide. Such a mechanism would also be consistent with the flexibility of lobes I and II relative to each other noted by solution NMR (Zhang and Zuiderweg, 2004). Mechanism of Nucleotide Exchange on Hsp70/Hsc70 by HspBP1 Limited proteolysis experiments were performed to explore the predicted effect of BP1c on the structure of

Crystal Structure of an HspBP1-Hsp70 Complex 375

Figure 7. Effect of HspBP1 on the Structure of the Hsp70 ATPase Domain (A) Lobe I of the ATPase domain of Hsp70 is progressively degraded by protease in presence of BP1c, but not of BAG. Proteins were incubated in the combinations indicated with increasing amounts of chymotrypsin and analyzed by SDS-PAGE. Fragments of the ATPase domain are indicated. Fragment 1 begins at residue 69. An otherwise identical experiment with trypsin as protease gave a similar fragmentation pattern (data not shown). BAG and BP1c are insensitive to protease treatment (data not shown). (B) Tryptophan fluorescence spectra of the ATPase domain of Hsp70 alone and in complex with BP1c(W163A) and BAG. BP1c(W163A) binding increases Trp90 fluorescence in accordance with a more “open” structure of the ATPase domain. (C) Schematic models of the nucleotide exchange mechanism of HspBP1and BAG/GrpE. (Top) Binding of HspBP1 is suggested to result in a mechanical separation of the two lobes of the ATPase domain, causing ADP dissociation. A rearrangement of subdomain IIb similar to that induced by BAG may contribute to this effect. (Bottom) The BAG domain docks onto the central cleft of the ATPase domain and locks subdomain IIb in an “open” orientation, employing a switch mechanism that is also used by GrpE.

the Hsp70 ATPase domain. In complex with BP1c, the entire lobe I was progressively cleaved by chymotrypsin and trypsin from the ATPase domain up to its joints with lobe II (McLellan et al., 2003) (Figure 7A). In contrast, the ATPase domain remained stable upon binding of BAG or in the absence of bound cofactor. Chymotrypsin treatment in the presence of BP1c, but not of BAG, transiently produced an ⵑ33 kDa fragment of the ATPase domain, which resulted from cleavage at residue 69 (Figure 7A, Hsp70[fragment 1]), as indicated by N-terminal sequencing. Importantly, this cleavage site is located in the cleft between lobes I and II of the domain (Figure 3B) and would only become accessible to protease through local unfolding of subdomain I or a substantial separation of the two lobes. Difference CD measure-

ments revealed no significant change in secondary structure composition (data not shown), arguing in favor of the latter possibility. The possibility that BP1c binding induces a conformational distortion of the Hsp70 ATPase domain was further tested by tryptophan fluorescence studies: the ATPase domain contains only a single Trp residue, Trp90, which is in close proximity to Phe68 and the nucleotide binding cleft but distant from the BP1c and BAG contact areas and, thus, can probe conformational changes in subdomain Ib (Figure 3B) (Theyssen et al., 1996). A mutant form of BP1c, W163A, which binds normally to the ATPase domain but lacks tryptophan was used. Upon binding of BP1c(W163A), the tryptophan fluorescence emission of the ATPase domain shifted

Molecular Cell 376

slightly to a longer wavelength and increased ⵑ12% in fluorescence intensity. In contrast, binding of BAG, which lacks tryptophan, resulted in an ⵑ10% decrease in fluorescence intensity (Figure 7B). The observed changes in fluorescence quantum yield are likely to reflect changes in the proximity and orientation of adjacent aromatic side chains, such as Phe68 and His227, relative to Trp90. In any case, these results support the conclusion that the average conformations of the Hsp70 ATPase domain in complex with BP1c and BAG differ. In addition to its effect on lobe I of the ATPase domain, BP1c may also cause a conformational shift in lobe II resembling that affected by BAG, because the steric conflict of the N terminus of BP1c with lobe I should suffice to stabilize lobe II of the ATPase domain in an “open” state (Figure 7C). Distortion of lobe I and its separation from lobe II, on top of the opening of lobe II, appear to strongly lower the affinity of the ATPase domain for adenine nucleotide (Figure 2B, Supplemental Table S1). Coexistence of HspBP1 and BAG Domain Proteins in the Eukaryotic Cytosol The existence of two classes of evolutionarily distinct NEFs for Hsp70 in the eukaryotic cytosol has interesting implications. BAG-domain proteins and HspBP1-homologs are functionally equivalent with respect to catalyzing nucleotide exchange on Hsp70/Hsc70 but differ in mechanism. Presumably, the two types of protein have partially overlapping functions. It is possible that each class of NEF arose independently in different primordial branches of the eukaryotic domain, and the common ancestor of the crown eukaryotes received a second class of NEF by lateral gene transfer. However, remote homologs of these proteins are difficult to identify, because both classes use ubiquitous folds as a scaffold— the three-helical bundle (BAG) and successions of armadillo repeats (HspBP1) are found in many proteins with greatly differing functions. The diversity of NEFs in eukaryotes contrasts with the situation in bacteria where GrpE homologs have coevolved with DnaK. The active site cleft in the ATPase domain of DnaK appears to be stabilized by additional contacts not present in eukaryotic Hsp70s, which also form part of the GrpE interface (Brehmer et al., 2001). During the evolution of certain eukaryotic species or phyla, one of the NEF classes might have been lost and replaced by isoforms of the other class. For example, no ortholog of HspBP1 has been identified in the genome of C. elegans so far, although a putative homolog of BAP is present as is at least one putative BAG-domain protein. In contrast to S. cerevisiae where only a membrane bound isoform of BAG was identified, Snl1p, S. pombe contains two BAG-domain proteins in addition to a Fes1p homolog. It is possible that in some cases, BAG isoforms and HspBP1 participate in different but partially overlapping pathways of protein quality control. Considering their functional redundancy, it is remarkable however that practically all BAG-domain proteins identified contain various additional domains, whereas all homologs of HspBP1 contain no recognizable domains other than the conserved core. This suggests that HspBP1 can act in multiple contexts, whereas the BAG-

domain proteins may be more specialized, being linked to particular cellular functions through different proteinprotein interactions. Therefore, HspBP1 is likely to function more generally as a NEF in Hsp70-mediated protein folding than the BAG-domain proteins. Indeed, evidence for an in vivo role of BAG proteins in de novo protein folding is still missing. The conformation of Hsp70 induced in the complex with HspBP1 may influence the interdomain communication in the chaperone and modulate Hsp70-substrate interactions beyond regulating nucleotide exchange. Numerous residues implicated in interdomain regulation were found on the surface of lobe I of the ATPase domain of Hsp70 (Davis et al., 1999). Additional cochaperones might recognize the conformation of Hsp70/Hsc70 induced by HspBP1. The ubiquitin ligase CHIP (C-terminus of Hsp70-interacting protein), for example, was reported to form ternary complexes with both HspBP1-Hsc70 and BAG-1-Hsc70 (Alberti et al., 2004; Demand et al., 2001). CHIP contains a specialized tetratricopeptide repeat domain that interacts with the extreme C terminus of Hsp70/Hsc70 but does not form stable binary complexes with either NEF. Whereas BAG-1 enhances the degradation function of CHIP (Demand et al., 2001), HspBP1 inhibits CHIP-induced degradation (Alberti et al., 2004). These different physiological effects of the two NEFs may arise at least in part from their divergent biochemical mechanisms. Future studies will certainly uncover further biological distinctions between the HspBP1 and BAG families of proteins. Experimental Procedures Construction of FES1-Deficient S. cerevisiae JY053 Strain JY022 (fes1::Kan) was constructed by direct replacement of the FES1 gene by the KanMX4 cassette (Wach et al., 1994) in the wild-type MAT␣ strain YPH500, isogenic to wild-typeYPH499 (MATa ade2–101 his3⌬200 leu2⌬1 lys2–801 trp1⌬63 ura3–52) (Sikorski and Hieter, 1989). JY022 was mated with another strain isogenic to YPH499, the diploid was sporulated, and a ⌬fes1 progeny was mated with YPH499 again to eliminate any second site mutations. JY053 (MATa fes1::Kan ade2–101 his3⌬200 leu2⌬1 lys2–801 trp1⌬63 ura3– 52) was obtained from the final sporulation. JY053 and JY022 behaved identically under the growth conditions tested. De Novo Folding of Firefly Luciferase in Yeast S. cerevisiae strains JY053 and YPH499 were transformed with an expression plasmid for firefly luciferase (FL) containing a c-Myc and a His6 tag at the C terminus under control of the copper promoter p425Cup and grown in SC-Leu medium to an OD600 nm ⫽ 0.8 at 30⬚C. Protein expression was induced by addition of 0.25 mM CuSO4 for 3 hr either at 30⬚C or 37⬚C. Spheroplasts were prepared by zymolyase treatment and lysed in luciferase dilution buffer (25 mM Trisphosphate [pH 7.8], 2 mM DTT, 2 mM CDTA, 10% glycerol, and 0.1% Triton X-100) containing EDTA-free protease inhibitors (Complete, Roche) (Agashe et al., 2004). Samples were fractionated into supernatant and pellet by centrifugation (20,000 ⫻ g for 20 min). FL activities were determined with the Luciferase Assay System (Promega #E1501). Protein quantitations were performed by immunoblotting with the anti-c-Myc 9E10 monoclonal antibody followed by densitometry. After addition of apyrase (20 U/ml, Sigma), fresh lysates were separated by size exclusion chromatography on Superdex 200 PC 3.2/30 equilibrated at room temperature in luciferase dilution buffer and analyzed as described above. Alternatively, apyrase-treated lysate was diluted 1:1 with 2⫻ binding buffer (50 mM HEPES-KOH [pH 7.5], 200 mM KOAc, and 10% glycerol) containing 2% BSA and 15 ␮l Nickel-NTA agarose beads

Crystal Structure of an HspBP1-Hsp70 Complex 377

Table 1. Data Collection and Refinement Statistics SeMet-BP1c

BP1c-Hsp70184–371

ID14-4 C2

ID14-2 P212121

76.7, 84.3, 90.0 90, 96.0, 90 0.9795 (peak) 49.4–2.1 (2.21–2.10) 98.9 (99.2) 33,305 (4,826) 7.4 (7.4) 0.073 (0.236) 5.9 (2.8)

73.5, 94.8, 155.7 90, 90, 90 0.933 46.6–2.9 (3.06–2.90) 99.8 (100.0) 24,763 (3,574) 3.5 (3.5) 0.075 (0.456) 6.5 (1.6)

Data Collection ESRF beamline Space group Unit cell dimensions a, b, c (A˚) ␣, ␤, ␥ (⬚) Wavelength (A˚) Resolution (A˚)a Completeness (%)a Unique reflectionsa Redundancya Rmergea,b ⬍I/␴⬎a,c SAD phasing statistics Number of Selenium sites Mean Figure of Merit

16 0.36

Refinement Statistics Number of protein atoms Number of water atoms Number of ligand atoms Rwork/Rfreed,e Rms deviations from idealg Bonds (A˚) Angles (⬚) Ramachandran plot (%)h Most favored regions Additional allowed regions

4,118 144 — 21.2/24.8

6,700 49 46f 23.7/29.6

0.021 1.7

0.007 1.3

94.4 5.6

86.7 13.3

a

Values in parentheses are for highest resolution shells. Rmerge ⫽ ⌺(|I ⫺ ⬍I⬎|)/⌺I. c Signal-to-noise ratio of intensities. d R ⫽ ⌺(|Fo ⫺ Fc|)/⌺Fo. e Five percent of reflections were randomly chosen for calculation of Rfree. f AMP moiety of AMP-PNP. g Root mean square deviations from ideal geometry (Engh and Huber, 1991). h As defined in Procheck (Laskowski et al., 1993). b

(Qiagen) and incubated on a rotator at 4⬚C for 2 hr. After three washes with 200 ␮l binding buffer containing 0.1% Triton X-100, bound protein was eluted in an equal volume of warmed SDS sample buffer and subjected to SDS-PAGE analysis. Ssa1p and Ydj1p were detected by Western blotting with specific antisera. Functional Assays ATPase rates were determined as described (Liberek et al., 1991). Combinations of Hsc70 (3 ␮M), Hsp40 (1.5 ␮M) and cBP1 (3 ␮M) were incubated at 30⬚C in the presence of 2 mM ATP containing 0.1 ␮Ci [␣-32P]ATP (Amersham Biosciences). Reactions were terminated by addition of 25 mM EDTA and placed on ice. ADP and ATP were separated by thin layer chromatography with 0.5 M formic acid and LiCl as eluent. Radioactivity was quantified with a phosphoimager. Stopped-flow measurements with MABA nucleotide analogs were performed with a SX.18MV system (Applied Photophysics, Surrey, UK) as described (Brehmer et al., 2001; Ga¨ssler et al., 2001; Theyssen et al., 1996). The complex of full-length Hsp70 and MABA-ADP (1 ␮M) was mixed with an equal volume of 250 ␮M ADP containing no NEF, BP1c, or BAG (both 10 ␮M) at 30⬚C, and the fluorescence of MABA-ADP was monitored. Crystallization, Data Collection, and Structure Determination Native His6-BP1c and SeMet-substituted His6-BP1c were crystallized at 20⬚C with the microbatch method by mixing 8 ␮l of protein solution (15 mg/ml protein, 20 mM HEPES-KOH [pH 7.5], 50 mM KCl, and 2 mM DTT) with 6–7 ␮l of the precipitant solution (30% [w/v] polyethylene glycol 2000 monomethyl ether, 100 mM Tris HCl [pH 8.5], 200 mM K-acetate, 5% [v/v] glycerol, 3% [v/v] iso-propanol,

10 mM EDTA, and 10 mM TCEP). Crystals belong to space group C2 with two monomers per asymmetric unit obeying almost perfect 2-fold symmetry (Table 1). Microseeding was applied to improve the quality and size of the crystals. Crystals were transferred stepwise into cryo-protectant (26% [v/v] glycerol, 25% [w/v] polyethylene glycol 3350, 100 mM Tris HCl [pH 8.5], 160 mM K-acetate, and 2 mM TCEP) and flash cooled in liquid nitrogen. The SeMet SAD data set used for model building and refinement was collected at ESRF beamline ID14-EH4. The crystals were maintained at 100 K by a gaseous nitrogen stream during the data collection. Diffraction data were processed with MOSFLM (Leslie, 1992) and SCALA (Evans, 1997) and subsequently handled with programs of the CCP4 suite (CCP4, 1994). Sixteen selenium sites were determined by direct methods with SnB (Weeks and Miller, 1999). Refinement of Se positions and the calculation of initial phases were carried out with SOLVE (Terwilliger and Berendzen, 1999). Density modification and automated model building were carried out with RESOLVE (Terwilliger, 2000). Subsequent iterative model building and refinement were performed with O (Jones et al., 1991) and CNS (Bru¨nger et al., 1998). A final round of TLS refinement was carried out with REFMAC5 (Murshudov et al., 1997). Each BP1c peptide chain in the final model comprises residues 84–350 plus an N-terminal methionine residue introduced during cloning. In addition, the model contains 147 water molecules. The complex of BP1c and Hsp70184–371 was crystallized at 293 K by the hanging drop vapor diffusion method mixing 1 ␮l of protein solution (15 mg/ml protein, 20 mM HEPES-KOH [pH 7.5], 25 mM KCl, 10 mM MgCl2, 5 mM DTT, and 5 mM AMP-PNP) with 1 ␮l of the precipitant solution (7% [w/v] polyethylene glycol 8000, 100 mM HEPES-KOH [pH 7.5], and 10 mM DTT). For cryo-protection, the

Molecular Cell 378

crystals were stepwise transferred into precipitant solution including 30% [v/v] ethylene glycol and frozen in liquid nitrogen. Diffraction data were collected at ESRF beamline ID14-EH2 and processed with MOSFLM (Leslie, 1992) and SCALA (Evans, 1997). The crystals belong to space group P212121 and contain two complexes per asymmetric unit. The structure was determined by molecular replacement. First, the orientation and positioning of the Hsp70-fragments were determined with BEAST (Read, 2001) and AMoRe (Navaza, 1994), respectively, with the corresponding region of Hsp70 (PDB code, 1S3X [Sriram et al., 1997]) as a model. The positions of BP1c were subsequently determined with MolRep (Vagin and Isupov, 2001) keeping the Hsp70-fragments fixed. Further model correction and structure refinement were iteratively carried out with O (Jones et al., 1991) and CNS (Bru¨nger et al., 1998). NCS restraints were applied during the initial stages of refinement. The final model contains in both complexes residues 87–351 of HspBP1, residues 192– 356 of Hsp70, and one molecule of AMP-PNP. Because of missing electron density, the imidodiphosphate moieties of both nucleotides were omitted, and twelve solvent exposed residues were modeled as alanine. Figures were generated with the programs Molscript (Kraulis, 1991), Bobscript (Esnouf, 1997), Raster-3D (Merritt and Bacon, 1997), ESPript (Gouet et al., 1999), and GRASP (Nicholls et al., 1991).

References

Fes1p-Ssa1p Binding Assay A total of 1.4 ␮M Ssa1p was incubated for 30 min at room temperature with 1.4 ␮M His-tagged Fes1p, Fes1p(A79R), or Fes1p(A79R, R195A) in a final volume of 300 ␮l 20 mM HEPES-KOH (pH 7.4), 200 mM KCl, 4 mM MgCl2, and 5 mM imidazole (buffer 5). Subsequently, ⵑ200 ␮l Ni-chelating Sepharose was added followed by incubation for further 5 min. After being washed with 3 column volumes of buffer 5, bound protein was eluted with 300 ␮l buffer 5 containing 250 mM imidazole and analyzed by SDS-PAGE.

Bukau, B., and Horwich, A.L. (1998). The Hsp70 and Hsp60 chaperone machines. Cell 92, 351–366.

Limited Proteolysis For limited proteolysis, stoichiometric mixtures of Hsp701–382, BAG, and BP1c (all 28 ␮M) were incubated with chymotrypsin (0, 0.056, 0.11, and 0.22 ␮M) in 25 mM HEPES-KOH (pH 7.5), 100 mM K-acetate, 5 mM Mg-acetate, and 5 mM ␤-mercaptoethanol at 25⬚C for 30 min. The reactions were stopped by addition of 5 mM phenylmethylsulfonyl fluoride and analyzed by SDS-PAGE. Tryptophan Fluorescence Spectroscopy Fluorescence measurements were performed with a LS50 luminescence spectrometer (Perkin Elmer). All measurements were carried out at 25⬚C in a buffer containing 20 mM HEPES-KOH (pH 7.5), 100 mM K-acetate, 5 mM Mg-acetate, and 2.5 mM ␤-mercaptoethanol. The concentrations of Hsp701–382 and the NEFs were 5 ␮M and 50 ␮M, respectively. Samples were preincubated for at least 30 min prior to recording of fluorescence emission spectra resulting from excitation at 295 nm wavelength. Spectra from three equal samples were averaged and Hsp701–382-free reference spectra subtracted. Additional Experimental Procedures are described in the Supplemental Data. Acknowledgments Technical aid and advice was generously provided by Elisabeth Weyher-Stingl (circular dichroism spectroscopy), Reinhardt Mentele (F. Lottspeich, N-terminal sequencing), Dr. Raina Boteva (fluorescence spectroscopy), and Michael Kerner (surface plasmon resonance). The antiserum against Ssa1p was a generous gift by Dr. Katja Siegers. We thank Dr. Gleb Bourenkov at beamline BW-6, DESY, Hamburg, Germany, and Joint Structural Biology Group staff at beamlines ID14-EH4 and ID14-EH2, ESRF, Grenoble, France, for assistance during diffraction data collection and Dr. Jose´ Barral for critically reading the manuscript. V.G. is owner of Desert Genetics, Inc., which holds an issued U.S. patent on HspBP1. Received: August 6, 2004 Revised: November 5, 2004 Accepted: December 23, 2004 Published: February 3, 2005

Agashe, V.R., Guha, S., Chang, H.C., Genevaux, P., Hayer-Hartl, M., Stemp, M., Georgopoulos, C., Hartl, F.U., and Barral, J.M. (2004). Function of trigger factor and DnaK in multidomain protein folding: increase in yield at the expense of folding speed. Cell 117, 199–209. Ahner, A., Whyte, F.M., and Brodsky, J.L. (2005). Distinct but overlapping functions of Hsp70, Hsp90, and an Hsp70 nucleotide exchange factor during protein synthesis in yeast. Arch. Biochem. Biophys., in press. Alberti, S., Bohse, K., Arndt, V., Schmitz, A., and Ho¨hfeld, J. (2004). The co-chaperone HspBP1 inhibits the CHIP ubiquitin ligase and stimulates the maturation of the cystic fibrosis transmembrane conductance regulator. Mol. Biol. Cell 15, 4003–4010. Brehmer, D., Ru¨diger, S., Ga¨ssler, C.S., Klostermeier, D., Packschies, L., Reinstein, J., Mayer, M.P., and Bukau, B. (2001). Tuning of chaperone activity of Hsp70 proteins by modulation of nucleotide exchange. Nat. Struct. Biol. 8, 427–432. Bru¨nger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al. (1998). Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921.

Bukau, B., Deuerling, E., Pfund, C., and Craig, E.A. (2000). Getting newly synthesized proteins into shape. Cell 101, 119–122. Chung, K.T., Shen, Y., and Hendershot, L.M. (2002). BAP, a mammalian BiP-associated protein, is a nucleotide exchange factor that regulates the ATPase activity of BiP. J. Biol. Chem. 277, 47557– 47563. CCP4 (Collaborative Computational Project, Number 4) (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr. 50, 760–763. Conti, E., Uy, M., Leighton, L., Blobel, G., and Kuriyan, J. (1998). Crystallographic analysis of the recognition of a nuclear localization signal by the nuclear import factor karyopherin alpha. Cell 94, 193–204. Davis, J.E., Voisine, C., and Craig, E.A. (1999). Intragenic suppressors of Hsp70 mutants: interplay between the ATPase- and peptidebinding domains. Proc. Natl. Acad. Sci. USA 96, 9269–9276. DeLuca-Flaherty, C., McKay, D.B., Parham, P., and Hill, B.L. (1990). Uncoating protein (hsc70) binds a conformationally labile domain of clathrin light chain LCa to stimulate ATP hydrolysis. Cell 62, 875–887. Demand, J., Alberti, S., Patterson, C., and Ho¨hfeld, J. (2001). Cooperation of a ubiquitin domain protein and an E3 ubiquitin ligase during chaperone/proteasome coupling. Curr. Biol. 11, 1569–1577. Engh, R.A., and Huber, R. (1991). Accurate bond and angle parameters for X-ray protein structure refinement. Acta Crystallogr. A 47, 392–400. Esnouf, R.M. (1997). An extensively modified version of MolScript that includes greatly enhanced coloring capabilities. J. Mol. Graph. Model. 15, 132–134, 112–113. Evans, P.R. (1997). SCALA. Joint CCP4 and ESF-EACBM Newsletter 33, 22–24. Flaherty, K.M., DeLuca-Flaherty, C., and McKay, D.B. (1990). Threedimensional structure of the ATPase fragment of a 70K heat-shock cognate protein. Nature 346, 623–628. Frydman, J. (2001). Folding of newly translated proteins in vivo: the role of molecular chaperones. Annu. Rev. Biochem. 70, 603–647. Frydman, J., Nimmesgern, E., Ohtsuka, K., and Hartl, F.U. (1994). Folding of nascent polypeptide chains in a high molecular mass assembly with molecular chaperones. Nature 370, 111–117. Frydman, J., Erdjument-Bromage, H., Tempst, P., and Hartl, F.U. (1999). Co-translational domain folding as the structural basis for the rapid de novo folding of firefly luciferase. Nat. Struct. Biol. 6, 697–705. Ga¨ssler, C.S., Wiederkehr, T., Brehmer, D., Bukau, B., and Mayer, M.P. (2001). Bag-1M accelerates nucleotide release for human

Crystal Structure of an HspBP1-Hsp70 Complex 379

Hsc70 and Hsp70 and can act concentration-dependent as positive and negative cofactor. J. Biol. Chem. 276, 32538–32544. Gouet, P., Courcelle, E., Stuart, D.I., and Metoz, F. (1999). ESPript: multiple sequence alignments in PostScript. Bioinformatics 15, 305–308. Harrison, C.J., Hayer-Hartl, M., Di Liberto, M., Hartl, F., and Kuriyan, J. (1997). Crystal structure of the nucleotide exchange factor GrpE bound to the ATPase domain of the molecular chaperone DnaK. Science 276, 431–435. Hartl, F.U., and Hayer-Hartl, M. (2002). Molecular chaperones in the cytosol: from nascent chain to folded protein. Science 295, 1852– 1858. Ho¨hfeld, J., and Jentsch, S. (1997). GrpE-like regulation of the hsc70 chaperone by the anti-apoptotic protein BAG-1. EMBO J. 16, 6209– 6216.

Pratt, W.B. (2001). Stoichiometry, abundance, and functional significance of the hsp90/hsp70-based multiprotein chaperone machinery in reticulocyte lysate. J. Biol. Chem. 276, 30092–30098. Murshudov, G.N., Vagin, A.A., and Dodson, E.J. (1997). Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr. 53, 240–255. Navaza, J. (1994). AMoRe: an automated package for molecular replacement. Acta Crystallogr. A 50, 157–163. Nicholls, A., Sharp, K.A., and Honig, B. (1991). Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins 11, 281–296. Raynes, D.A., Graner, M.W., Bagatell, R., McLellan, C., and Guerriero, V. (2003). Increased expression of the Hsp70 cochaperone HspBP1 in tumors. Tumour Biol. 24, 281–285.

Holm, L., and Sander, C. (1996). Mapping the protein universe. Science 273, 595–603.

Raynes, D.A., and Guerriero, V. (1998). Inhibition of Hsp70 ATPase activity and protein renaturation by a novel Hsp70-binding protein. J. Biol. Chem. 273, 32883–32888.

Huber, A.H., Nelson, W.J., and Weis, W.I. (1997). Three-dimensional structure of the armadillo repeat region of beta-catenin. Cell 90, 871–882.

Read, R.J. (2001). Pushing the boundaries of molecular replacement with maximum likelihood. Acta Crystallogr. D Biol. Crystallogr. 57, 1373–1382.

Jones, T.A., Zou, J.Y., Cowan, S.W., and Kjelgaard, M. (1991). Improved methods for building protein models in electron density maps and location of errors in these models. Acta Crystallogr. A 47, 110–119.

Sagermann, M., Stevens, T.H., and Matthews, B.W. (2001). Crystal structure of the regulatory subunit H of the V-type ATPase of Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 98, 7134–7139.

Kabani, M., Beckerich, J.M., and Gaillardin, C. (2000). Sls1p stimulates Sec63p-mediated activation of Kar2p in a conformationdependent manner in the yeast endoplasmic reticulum. Mol. Cell. Biol. 20, 6923–6934. Kabani, M., Beckerich, J.M., and Brodsky, J.L. (2002a). Nucleotide exchange factor for the yeast Hsp70 molecular chaperone Ssa1p. Mol. Cell. Biol. 22, 4677–4689. Kabani, M., McLellan, C., Raynes, D.A., Guerriero, V., and Brodsky, J.L. (2002b). HspBP1, a homologue of the yeast Fes1 and Sls1 proteins, is an Hsc70 nucleotide exchange factor. FEBS Lett. 531, 339–342. Kleywegt, G.J., and Jones, T.A. (1994). A super position. Joint CCP4 and ESF-EACBM Newsletter 31, 9–14. Kraulis, P. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24, 946–950. Laskowski, R.A., MacArthur, M.W., Moss, D.S., and Thornton, J.M. (1993). PROCHECK: a program to check the stereochemical quality of protein structures. J. App. Cryst. 26, 283–291. Leslie, A.G.W. (1992). Recent changes to the MOSFLM package for processing film and image plate data. Joint CCP4 and ESF-EACBM Newsletter 26. Li, J., Qian, X., and Sha, B. (2003). The crystal structure of the yeast Hsp40 Ydj1 complexed with its peptide substrate. Structure 11, 1475–1483. Liberek, K., Marszalek, J., Ang, D., Georgopoulos, C., and Zylicz, M. (1991). Escherichia coli DnaJ and GrpE heat shock proteins jointly stimulate ATPase activity of DnaK. Proc. Natl. Acad. Sci. USA 88, 2874–2878. Lu, Z., and Cyr, D.M. (1998). Protein folding activity of Hsp70 is modified differentially by the hsp40 co-chaperones Sis1 and Ydj1. J. Biol. Chem. 273, 27824–27830. Lu¨ders, J., Demand, J., and Ho¨hfeld, J. (2000). The ubiquitin-related BAG-1 provides a link between the molecular chaperones Hsc70/ Hsp70 and the proteasome. J. Biol. Chem. 275, 4613–4617. McLellan, C.A., Raynes, D.A., and Guerriero, V. (2003). HspBP1, an Hsp70 cochaperone, has two structural domains and is capable of altering the conformation of the Hsp70 ATPase domain. J. Biol. Chem. 278, 19017–19022. Merritt, E.A., and Bacon, D.J. (1997). Raster3D photorealistic graphics. Methods Enzymol. 277, 505–524. Minami, Y., Ho¨hfeld, J., Ohtsuka, K., and Hartl, F.U. (1996). Regulation of the heat-shock protein 70 reaction cycle by the mammalian DnaJ homolog, Hsp40. J. Biol. Chem. 271, 19617–19624. Murphy, P.J., Kanelakis, K.C., Galigniana, M.D., Morishima, Y., and

Sikorski, R.S., and Hieter, P. (1989). A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122, 19–27. Sondermann, H., Scheufler, C., Schneider, C., Ho¨hfeld, J., Hartl, F.U., and Moarefi, I. (2001). Structure of a Bag/Hsc70 complex: convergent functional evolution of Hsp70 nucleotide exchange factors. Science 291, 1553–1557. Sondermann, H., Ho, A.K., Listenberger, L.L., Siegers, K., Moarefi, I., Wente, S.R., Hartl, F.U., and Young, J.C. (2002). Prediction of novel Bag-1 homologs based on structure/function analysis identifies Snl1p as an Hsp70 co-chaperone in Saccharomyces cerevisiae. J. Biol. Chem. 277, 33220–33227. Sriram, M., Osipiuk, J., Freeman, B., Morimoto, R., and Joachimiak, A. (1997). Human Hsp70 molecular chaperone binds two calcium ions within the ATPase domain. Structure 5, 403–414. Terwilliger, T.C. (2000). Maximum-likelihood density modification. Acta Crystallogr. D Biol. Crystallogr. 56, 965–972. Terwilliger, T.C., and Berendzen, J. (1999). Automated MAD and MIR structure solution. Acta Crystallogr. D Biol. Crystallogr. 55, 849–861. Theyssen, H., Schuster, H.P., Packschies, L., Bukau, B., and Reinstein, J. (1996). The second step of ATP binding to DnaK induces peptide release. J. Mol. Biol. 263, 657–670. Vagin, A.A., and Isupov, M.N. (2001). Spherically averaged phased translation function and its application to the search for molecules and fragments in electron-density maps. Acta Crystallogr. D Biol. Crystallogr. 57, 1451–1456. Wach, A., Brachat, A., Pohlmann, R., and Philippsen, P. (1994). New heterologous modules for classical or PCR-based gene disruptions in Saccharomyces cerevisiae. Yeast 10, 1793–1808. Weeks, C.M., and Miller, R. (1999). The design and implementation of SnB v2.0. J. Appl. Crystallogr. 32, 120–124. Young, J.C., and Hartl, F.U. (2000). Polypeptide release by Hsp90 involves ATP hydrolysis and is enhanced by the co-chaperone p23. EMBO J. 19, 5930–5940. Zhang, Y., and Zuiderweg, E.R. (2004). The 70-kDa heat shock protein chaperone nucleotide- binding domain in solution unveiled as a molecular machine that can reorient its functional subdomains. Proc. Natl. Acad. Sci. USA 101, 10272–10277. Zhu, X., Zhao, X., Burkholder, W.F., Gragerov, A., Ogata, C.M., Gottesman, M.E., and Hendrickson, W.A. (1996). Structural analysis of substrate binding by the molecular chaperone DnaK. Science 272, 1606–1614. Accession Numbers Coordinates and structure factor amplitudes were deposited to Protein Data Bank under accession codes 1XQR and 1XQS.