The Fe/S Assembly Protein IscU Behaves as a Substrate for the Molecular Chaperone Hsc66 from Escherichia coli *

,

IscU, a NifU-like Fe/S-escort protein, binds to and stimulates the ATPase activity of Hsc66, a hsp70-type molecular chaperone. We present evidence that stimulation arises from interactions of IscU with the substrate-binding site of Hsc66. IscU inhibited the ability of Hsc66 to suppress the aggregation of the denatured model substrate proteins rhodanese and citrate synthase, and calorimetric and surface plasmon resonance measurements showed that ATP destabilizes Hsc66⅐IscU complexes in a manner expected for hsp70-substrate complexes. Studies on the interaction of IscU with Hsc66 truncation mutants further showed that IscU does not bind the isolated ATPase domain of Hsc66 but does bind and stimulate a mutant containing the ATPase domain and substrate binding ␤-sandwich subdomain. These results support a role for IscU as a substrate for Hsc66 and suggest a specialized function for Hsc66 in the assembly, stabilization, or transfer of Fe/S clusters formed on IscU.
Hsp70 1 proteins comprise a widespread family of molecular chaperones that are composed of an N-terminal ATP-binding domain exhibiting weak intrinsic ATPase activity and a Cterminal peptide-binding domain that reversibly binds peptide substrates (1)(2)(3)(4). Proteins from this family have been implicated in a variety of processes including stress response, de novo protein folding, protein degradation, protein trafficking, and disassembly of protein complexes (for reviews, see Refs. [5][6][7]. Central to all hsp70 activities is their ability to undergo association/dissociation cycles with peptide substrates and to couple ATP binding and hydrolysis with conformational changes that modulate their peptide binding affinity. ATP binding results in conformational changes leading to destabilization of hsp70⅐peptide substrate complexes, whereas subsequent ATP hydrolysis to ADP results in conformational changes leading to complex stabilization (8 -13).
Many bacteria contain a constitutively expressed hsp70 designated Hsc66 in addition to the prototypical DnaK (14,15). No specific peptide substrates have been identified for Hsc66, and its exact cellular function(s) are not known. It is also unclear whether Hsc66 functions as a general chaperone with broad substrate specificity as is the case for DnaK (2,6,16), or if Hsc66 has evolved to function with a specific substrate. Localization of the gene encoding Hsc66, hscA, to a gene cluster (iscSUA-hscBA-fdx) encoding proteins thought to function in the biogenesis of iron-sulfur clusters suggests that Hsc66 may play a specialized role in the folding of Fe/S proteins (17)(18)(19). The hscB gene product from this cluster encodes a 20-kDa protein designated Hsc20 that exhibits sequence similarities with the N-terminal J-domain of DnaJ-type auxiliary co-chaperones, and in vitro studies indicate that Hsc20 stimulates the ATPase activity of Hsc66 and regulates the rate of conversion of Hsc66 between its different peptide affinity states (20). Hsc66 and Hsc20 appear to comprise a distinct chaperone system with nonoverlapping functions with the DnaK/DnaJ/ GrpE system since physiologically relevant levels of DnaJ or GrpE do not affect Hsc66 ATPase activity, and Hsc20 does not stimulate DnaK ATPase activity (18).
Recently, we have found that the iscU gene product, IscU, also stimulates the ATPase activity of Hsc66 (21). IscU is a novel Fe/S-binding protein believed to function as a scaffold for Fe/S cluster assembly (21)(22)(23). The nature of the interaction of IscU with Hsc66 and the mechanism by which it stimulates Hsc66, however, is unclear. IscU stimulation could arise from actions as an auxiliary, regulatory co-chaperone in addition to Hsc20. Alternatively, stimulation could result from interactions of IscU with the peptide-binding domain of Hsc66, i.e. with IscU serving as a substrate for Hsc66. To better understand IscU regulation of Hsc66, we have investigated the ability of IscU to compete with model peptide substrates for binding to Hsc66 and have characterized the effects of nucleotides on Hsc66 and IscU binding. In addition, we investigated the ability of IscU to bind Hsc66 truncation mutants lacking specific subdomains of the C-terminal peptide-binding region.

EXPERIMENTAL PROCEDURES
Materials-Escherichia coli DH5␣FIQ cells were from Life Technologies, Inc. Enzymes for DNA manipulation were obtained from Roche Molecular Biochemicals, New England Biolabs, Inc., or U. S. Biochemical Corp. Synthetic nucleotides were obtained from Genosys. Bacterial growth media components were from Difco, and other reagents were from Sigma.
ATPase Assays-Steady-state ATPase rates were determined at 23°C in HKM buffer (50 mM Hepes, pH 7.3, 150 mM KCl, and 10 mM MgCl 2 ) containing 1 mM dithiothreitol (DTT) and 0.4 mM ATP as previously reported (18,20,21) by measuring phosphate released using a coupled enzyme assay with the EnzCheck phosphate assay kit (Molecular Probes). * This work was supported by National Institutes of Health Grant GM54264 and Training Grant GM07311. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Rhodanese and Citrate Synthase Aggregation Assays-
The aggregation of bovine rhodanese and porcine citrate synthase were performed in HKM buffer containing 1 mM ADP as described previously (18).
Surface Plasmon Resonance (SPR) Analysis-SPR methods were carried out at 25°C with a Biacore 3000 instrument (Piscataway, NJ) using methods described previously (21). Hsc66 in the presence of MgADP was randomly cross-linked to the surface of the sensor chip using amine coupling as recommended by the manufacturer. IscU was injected over the immobilized Hsc66 in HKM buffer containing 1 mM DTT and either 1 mM ADP or 1 mM ATP. Data are reported as changes in relative response units (RU). Similar results were obtained when Hsc66 was immobilized to the surface of a sensor chip in the presence of MgATP indicating that the differences in binding kinetics observed were not the result of the conditions of Hsc66 immobilization.

IscU Effects on Hsc66
Chaperone Activity-To investigate the possible role of IscU as a specific substrate for Hsc66, we examined the ability of IscU to compete with the model peptide substrates rhodanese and citrate synthase (18). Assays were carried out by measuring the effect of Hsc66⅐ADP on the extent of rhodanese or citrate synthase aggregation in the presence of varying levels of IscU. The effect of IscU on the ability of Hsc66 to suppress the aggregation of chemically denatured rhodanese is shown in Fig. 1A. IscU inhibited Hsc66 suppression of rhodanese aggregation in a concentration-dependent manner with a molar ratio of IscU to Hsc66 of 1:1 resulting in ϳ75% inhibition of Hsc66 chaperone activity (15 min), and a ratio of 5:1 giving Ͼ95% inhibition. IscU alone had no effect on rhodanese aggregation suggesting that the changes in turbidity caused by addition of IscU to reactions containing Hsc66 are not due to IscU directly enhancing rhodanese aggregation.
The effect of IscU on the ability of Hsc66 to suppress the aggregation of thermally denatured citrate synthase is shown in Fig. 1B. IscU inhibited Hsc66 suppression of citrate synthase aggregation, and this effect was dependent on the level of IscU with a molar ratio of IscU to Hsc66 of 2.5:1 inhibiting Hsc66 chaperone activity ϳ85% (40 min). IscU alone had no effect on citrate synthase aggregation. The competition of IscU with both rhodanese and citrate synthase for binding to Hsc66 is consistent with interaction of IscU with the peptide-binding domain of Hsc66 and suggests that IscU may form a stable complex with the high peptide affinity ADP state of Hsc66.
Nucleotide Effects on Hsc66 and IscU Binding-A hallmark of hsp70 proteins is their ability to modulate substrate binding in a nucleotide-dependent manner. The ADP complexes of hsp70 proteins display high affinity for peptide substrates and exhibit slow on and off rates, and exchange of ADP for ATP results in conformational changes leading to a low substrate affinity form with faster substrate association and dissociation rates (8 -13). Because IscU was found to compete with model peptide substrates for binding to Hsc66, it was of interest to determine whether IscU binding to Hsc66 is regulated in a manner similar to that observed for other hsp70 substrates.
In initial experiments, SPR analysis was used to investigate nucleotide effects on the interaction of IscU with Hsc66. Fig. 2 shows the results of titrations in which Hsc66 was randomly cross-linked to the sensor chip and exposed to different concentrations of IscU in the presence of ADP or ATP. The extent of binding of IscU to Hsc66 was similar in the presence of either nucleotide, but the apparent affinity of Hsc66 for IscU was greater in the presence of ADP (K D(app) Х 9 M) compared with that observed in the presence of ATP (K D(app) Х 37 M). Examination of the sensorgrams (insets Fig. 2) also reveals differences in the binding kinetics as a function of nucleotide. IscU association in the presence of ATP is faster than that observed in the presence of ADP, and the half-time for IscU dissociation in the presence of ATP (t1 ⁄2 Х 2 s) is ϳ30-fold faster than that observed in the presence of ADP (t1 ⁄2 Х 60 s). These results indicate ATP destabilizes Hsc66⅐IscU complexes in a manner similar to that observed for other hsp70-substrate interactions (8 -13) and are consistent with interaction of IscU with the peptide-binding site of Hsc66.
Because of possible complications in the SPR binding studies arising from surface and/or immobilization effects, we also used isothermal titration calorimetry (ITC) to more accurately quantitate the interaction of IscU with Hsc66. Fig. 3 shows the enthalpic changes observed in an experiment in which succes- sive additions of IscU were made to Hsc66 in the presence of ADP. The data are plotted as the integrated heats of binding (Q inj ) versus the molar ratio of IscU to Hsc66. The best fit curve to the data indicates the presence of a single, high affinity binding site (K D Х 1.6 M) consistent with specific binding to the Hsc66⅐ADP complex in a manner expected for a substrate.
Domain Requirements for IscU Binding to Hsc66 -Hsp70 proteins are composed of two functionally distinct domains, a highly conserved N-terminal ATPase domain ϳ44-kDa and a C-terminal peptide-binding domain ϳ25 kDa (1)(2)(3)(4). Structural studies on the C-terminal domain of DnaK complexed with peptide substrate indicate that this domain can be further divided into two subdomains: a ␤-sandwich region that directly binds model peptide substrates, and a C-terminal helical cap that lies above the peptide binding pocket of the ␤-sandwich (25). To investigate which domain(s) of Hsc66 are required for IscU binding, we constructed two truncation mutants (Fig. 4, A  and B). The first mutant, designated Hsc:2-505, lacked the C-terminal helical cap but contained both the N-terminal ATPase domain and the peptide-binding ␤-sandwich subdomain (residues 2-505). The second mutant, designated Hsc:2-382, contained only the ATPase domain (residues 2-382).
The effect of IscU on the ATPase activity of Hsc66, Hsc:2-505, and Hsc:2-382 was compared to characterize the region(s) of Hsc66 required for IscU binding (Fig. 4C). Both truncation mutants exhibited intrinsic ATPase activity that was slightly greater than that of full-length Hsc66. The ATPase activities of full-length Hsc66 and Hsc:2-505 were stimulated by IscU indicating the C-terminal helical cap is not necessary for IscU binding and activation. The activity of the ATPase domain fragment Hsc:2-382, however, was not affected by IscU indicating that the peptide-binding ␤-sandwich subdomain is required for IscU stimulation of ATPase activity as expected if IscU is a substrate for Hsc66. To determine whether IscU interacts with the isolated ATPase domain Hsc:2-382 in a manner not manifested as effects on ATPase activity, we also used ITC to measure enthalpic changes that might arise as a result of binding. Injection of IscU into a cell containing Hsc: 2-382 and ADP did not result in any detectable enthalpic changes suggesting that IscU does not interact directly with the isolated ATPase domain (data not shown).
Two approaches were used to further investigate the role of the C-terminal helical cap of Hsc66 in interactions with IscU. To investigate interactions of IscU with the ATP-bound state of Hsc:2-505, we examined the concentration dependence of IscU stimulation of Hsc:2-505 ATPase activity (see Fig. 5A). The concentration of IscU required for half-maximal stimulation of Hsc:2-505 activity (K m Х 31 M) is similar to that previously reported for full-length Hsc66 (K m Х 34 M; Ref. 21), although the maximal stimulation observed for Hsc:2-505 (ϳ14-fold) is slightly greater than that of full-length Hsc66 (ϳ8-fold; Ref. 21). These findings indicate that removal of the C-terminal helical cap of Hsc66 has little effect on either the affinity of ATP-bound Hsc66 for IscU or its ability to couple IscU binding with increased ATPase activity.
To investigate interactions of IscU with the ADP-bound state of Hsc:2-505, we measured the thermodynamics of IscU binding to Hsc:2-505 in the presence of ADP. Fig. 5B shows the results of an ITC experiment in which successive additions of IscU were made to a cell containing the Hsc:2-505⅐ADP complex. The binding affinity observed (K D Х 26 M) is ϳ16-fold weaker than that of full-length Hsc66 under similar conditions (K D Х 1.6 M; Fig. 3). Instead, it is similar to the apparent affinity observed for the ATP-bound state of Hsc:2-505 (31 M; Fig. 5A). Thus, the C-terminal helical cap of Hsc66 is required for the increased binding affinity for IscU that occurs following ATP hydrolysis. DISCUSSION In earlier studies we found that both Hsc20 and IscU stimulate the ATPase activity of Hsc66 and regulate the rate of conversion of Hsc66 between its different peptide affinity states (20,21). Whereas Hsc20 shares sequence similarities with DnaJ-type auxiliary co-chaperones and is thought to function in a manner similar to these proteins (18), IscU represents a novel ATPase stimulatory protein whose mechanism of action was unclear. IscU was also found to bind to Hsc20, and in the presence of Hsc20 the concentration of IscU required for halfmaximal stimulation of Hsc66 ATPase activity was reduced (21). The interaction between IscU and Hsc20 resembles the interaction of peptide substrates with DnaJ, a hsp40 co-chaperone. DnaJ decreases the concentration of peptide substrates required for half-maximal stimulation of DnaK ATPase activity (26), and thereby serves to target these substrates to the high peptide affinity ADP state of DnaK (27). Similarities in the actions of Hsc20 and DnaJ suggested that the increased binding affinity of Hsc66 for IscU observed in the presence of Hsc20 could reflect the role of IscU as a substrate.
The results described herein provide direct evidence that IscU behaves as a substrate for Hsc66. IscU was shown to compete with model peptide substrates for binding to the Hsc66⅐ADP complex consistent with IscU binding to the high peptide affinity ADP state of Hsc66. Affinity sensor studies also showed that nucleotides regulate Hsc66 binding to IscU in a manner consistent with that observed for nucleotide effects on hsp70⅐peptide substrate interactions (8 -13). ATP-bound Hsc66 exhibited faster association and dissociation rates for IscU compared with ADP-bound Hsc66 indicating that ATP destabilizes Hsc66⅐IscU complexes. In addition, calorimetric studies revealed that IscU binds to Hsc66 at a single site with an affinity (K D Х 1.6 M) ϳ20-fold greater than the concentration of IscU required for half-maximal stimulation of Hsc66 ATPase activity (K m Х 34 M; Ref. 21). Assuming that the K m observed in ATPase stimulation assays reflects the binding affinity of IscU for the ATP state of Hsc66, 2 the differences in affinities of the ADP and ATP-bound states is similar to that observed with other hsp70⅐substrate complexes (9,11,12).
Studies examining the interaction of IscU with truncation mutants of Hsc66 provide further support that IscU behaves as a substrate for Hsc66. IscU stimulated the ATPase activity of a mutant containing the ATPase domain and the peptide-binding ␤-sandwich region (Hsc:2-505), but no interactions were observed with a mutant composed solely of the ATPase domain (Hsc:2-382). These results establish that the ␤-sandwich region of Hsc66 is essential for IscU binding as is the case for substrate binding to other hsp70 proteins (25).
Although the Hsc66 C-terminal helical cap was not required for IscU stimulation of Hsc66 ATPase activity, this subdomain was required for coupling of ATP hydrolysis with conformational changes that increase Hsc66 binding affinity for IscU. In contrast to full-length Hsc66 in which Hsc66⅐ADP complexes exhibit ϳ20-fold higher affinity for IscU compared with ATP complexes, ADP and ATP complexes of Hsc:2-505 displayed similar IscU affinities indicating that both the ␤-sandwich and helical cap subdomains are required for high affinity binding. The C-terminal helical cap does not, however, appear to be required for all hsp70-substrate interactions. A recent report showed that DnaK mutants lacking the cap retain efficient coupling of ATP hydrolysis with increased substrate binding affinity (28). This may reflect structural differences between Hsc66 and DnaK or may arise from differences in the types of substrates recognized. The short peptide used in the DnaK studies (28) would not be expected to make extensive contacts with the C-terminal helical cap (25). The larger IscU protein, in contrast, may interact with both the ␤-sandwich and helical cap subdomains of Hsc66, and interactions of IscU with both subdomains may be required for high affinity binding. There is evidence that the helical cap may be important for regulatory interactions of the eukaryotic homologs of Hsc66. Mutations in the corresponding C-terminal subdomain of the yeast mitochondrial homolog of Hsc66, Ssq1, gives rise to phenotypic effects indicative of reduced chaperone activity (19).
The finding that IscU behaves as a substrate for Hsc66 raises the question of whether the Hsc66/Hsc20 chaperone system has evolved specifically to interact with IscU in Fe/S-cluster assembly or whether it serves a more general chaperone function in the cell. Like other hsp70 chaperones Hsc66 is able to bind denatured proteins (e.g. rhodanese and citrate synthase) and to prevent their aggregation (18). These "model substrates," however, have no effect on the ATPase activity of Hsc66. IscU binding, in contrast, is coupled to the nucleotide state of Hsc66 and regulates the chaperone's ATPase reaction cycle. IscU also exhibits specific interactions with the co-chaperone Hsc20 that is not observed with other proteins or peptides (18,21). These results, together with genetic findings implicating yeast homologs of Hsc66, Hsc20, and IscU in ironsulfur protein biogenesis (19), suggest that the primary cellular function of the Hsc66/Hsc20 chaperone system may be to interact with IscU in Fe/S-cluster generation.
While the studies described herein indicate that IscU binds as a substrate to Hsc66, the specific structural features of IscU required for interaction with Hsc66 are not known. Hsc66 may recognize a structured motif within IscU or may bind an unstructured region of IscU in an extended conformation similar to that recognized by DnaK (25). Further studies are also needed to investigate how Hsc66 and Hsc20 affect the function of IscU in iron-sulfur cluster assembly. The exact role the Hsc66/Hsc20 chaperone system plays in iron-sulfur cluster biogenesis remains unclear. The chaperone system may function in assembly or stabilization of Fe/S clusters on IscU or may facilitate transfer of Fe/S clusters formed on IscU to an apoacceptor protein.