The Small Heat Shock Protein IbpA of Escherichia coli Cooperates with IbpB in Stabilization of Thermally Aggregated Proteins in a Disaggregation Competent State*

The small heat shock proteins are ubiquitous stress proteins proposed to increase cellular tolerance to heat shock conditions. We isolated IbpA, the Escherichia coli small heat shock protein, and tested its ability to keep thermally inactivated substrate proteins in a disaggregation competent state. We found that the presence of IbpA alone during substrate thermal inactivation only weakly influences the ability of the bi-chaperone Hsp70-Hsp100 system to disaggregate aggregated substrate. Similar minor effects were observed for IbpB alone, the other E. coli small heat shock protein. However, when both IbpA and IbpB are simultaneously present during substrate inactivation they efficiently stabilize thermally aggregated proteins in a disaggregation competent state. The properties of the aggregated protein substrates are changed in the presence of IbpA and IbpB, resulting in lower hydrophobicity and the ability of aggregates to withstand sizing chromatography conditions. IbpA and IbpB form mixed complexes, and IbpA stimulates association of IbpB with substrate.

The small heat shock proteins are ubiquitous stress proteins proposed to increase cellular tolerance to heat shock conditions. We isolated IbpA, the Escherichia coli small heat shock protein, and tested its ability to keep thermally inactivated substrate proteins in a disaggregation competent state. We found that the presence of IbpA alone during substrate thermal inactivation only weakly influences the ability of the bi-chaperone Hsp70-Hsp100 system to disaggregate aggregated substrate. Similar minor effects were observed for IbpB alone, the other E. coli small heat shock protein. However, when both IbpA and IbpB are simultaneously present during substrate inactivation they efficiently stabilize thermally aggregated proteins in a disaggregation competent state. The properties of the aggregated protein substrates are changed in the presence of IbpA and IbpB, resulting in lower hydrophobicity and the ability of aggregates to withstand sizing chromatography conditions. IbpA and IbpB form mixed complexes, and IbpA stimulates association of IbpB with substrate.
The proper conformation of proteins is challenged by stress conditions. Exposure to extreme heat shock conditions results in a massive aggregation of proteins inside both prokaryotic and eukaryotic cells (1)(2)(3). Chaperones from the Hsp100 family, that is ClpB in Escherichia coli and Hsp104 in the yeast Saccharomyces cerevisiae, were implicated in the disaggregation reaction, because aggregated proteins were not eliminated in either clpB or HSP104 deletion strains (1)(2)(3). Additionally, the clpB and HSP104 gene products were identified as factors conferring thermotolerance in E. coli and S. cerevisiae (4 -7). However, in vitro studies on the reactivation of aggregated proteins showed that chaperones from the Hsp100 family alone are not sufficient for disaggregation and refolding. Other chaperone proteins are also involved in this process. E. coli Hsp70 (DnaK) and its cochaperones (DnaJ and GrpE) cooperate with ClpB and form a bi-chaperone system capable of efficient disaggregation of aggregated proteins (3,8,9). Analogous Hsp100-Hsp70 bi-chaperone systems able to disaggregate denatured protein substrates in vitro were established using chaperones from other bacterial species (10), as well as from yeast cytosol (11) and mitochondria (12,13). However, the efficiency of refolding reaction catalyzed by these bichaperone systems depends strongly on the physical properties of protein aggregates. It was proposed that small heat shock proteins (sHsps) 1 associate with aggregated proteins and change their physical properties in such a way that chaperone-mediated disaggregation and refolding become much more efficient (14 -19). However, little is known about the molecular mechanism of these processes.
Small heat shock proteins are widely distributed both in prokaryotes and eukaryotes. Members of this diverse protein family are characterized by relatively low monomeric molecular masses  and a conserved stretch of ϳ100 amino acid residues (reviewed in Refs. 20 and 21). This so-called ␣-crystallin domain displays sequence similarities to the vertebrate eye lens protein ␣-crystallin, which prevents protein precipitation and cataract formation in the eye lens.
One of the most striking features of sHsps is their organization in large oligomeric structures (reviewed in Refs. 20 and 21). The interactions between sHsp subunits in these oligomers are highly dynamic, and under physiological conditions, rapid exchange between subunits of these oligomers has been observed (22)(23)(24). Moreover, the temperature induces changes in the oligomerization state of sHsps complexes (20, 21) essential for their chaperone activity (25)(26)(27).
Surprisingly, in many cases deletion of genes encoding sHsps does not result in strong phenotypic effects, making it difficult to clearly assign them a function in the chaperone network. It has been reported, however, that overproduction of sHsps conveys thermotolerance in a number of organisms and cell types, suggesting the involvement of sHsps in the control of protein aggregation and disaggregation processes upon heat shock (28 -31). Several in vitro studies have reported that sHsps can prevent the aggregation of heat-denatured proteins (32)(33)(34). However, in vivo studies (2,35,36) show that sHsps are rather localized in the insoluble protein fractions of stressed cells, which questions their preventive role during heat stress conditions. Two members of the sHsp family, IbpA and IbpB, are present in E. coli. The IbpA and IbpB proteins are 48% identical at the amino acid sequence level (37), and both were identified as proteins associated with inclusion bodies (35). IbpA/IbpB were also found in aggregated protein fractions following heat stress (2). The deletion of the ibpAB genes results in a weak phenotype manifested only at extreme temperatures (38). Prolonged incubation of such mutant cells at 50°C results in a decrease of viability, which correlates with an increase in protein aggregation (38).
To date, only IbpB has been studied in detail. Purified IbpB forms 2-to 3-MDa oligomers, which, upon exposure to high temperature, dissociate into smaller ϳ600-kDa structures (39,40). The importance of these structural changes is not well understood. IbpB protein was shown in vitro to suppress thermal aggregation of model substrate proteins in a concentrationdependent manner (39). Moreover, IbpB was also tested for the ability to cooperate with chaperones from the Hsp70 and Hsp100 families in substrate disaggregation and refolding (16,18). The presence of IbpB during the heat inactivation of both malate dehydrogenase (MDH) and lactate dehydrogenase results in more efficient reactivation of these enzymes by the Hsp70 system (DnaK/DnaJ/GrpE) (16) as well as by the bichaperone Hsp70-Hsp100 system (ClpB-DnaK/DnaJ/GrpE) (18,19). However, the details, if any, of cooperation between IbpB and the ClpB-DnaK/DnaJ/GrpE system are unknown.
No such detailed studies have been performed for the other E. coli sHsp, IbpA. Only recently, IbpA with a hexahistidine tag at its N terminus was isolated and shown to form multimers (40). Similarly to IbpB, it also has a protective effect on several tested substrate proteins subjected to thermal and oxidative stress (40). However, it was not determined if the presence of IbpA during thermal inactivation of substrate protein results in subsequent efficient disaggregation by the ClpB-DnaK/ DnaJ/GrpE chaperones. It is also not known whether IbpA interacts with IbpB and if these small chaperones cooperate in the modification of aggregates during thermal denaturation of substrate proteins.
Here, we present results which answer the above questions. We observed that the presence of purified IbpA alone during thermal inactivation very weakly influences the ability of the bi-chaperone system (ClpB-DnaK/DnaJ/GrpE) to disaggregate aggregated substrate. A similar minor effect was observed for purified IbpB alone. However, when both IbpA and IbpB are simultaneously present during substrate inactivation, efficient stabilization of thermally aggregated proteins in a disaggregation competent state was observed, indicating that both sHsps work together as an integrated system.

EXPERIMENTAL PROCEDURES
Protein Purification-The IbpA protein was overproduced in the MC4100⌬ibpA/B strain transformed with the pUC18-derived pCA plasmid bearing the ibpA gene, under control of the ptac promoter (38). Bacteria overproducing IbpA protein were lysed in a French press (Aminco) in buffer A (50 mM Tris-HCl, pH 7.4, 10% (v/v) glycerol, 1 mM dithiothreitol, 100 mM KCl). The bacterial lysate was clarified for 30 min at 26,000 rpm in a Beckman 30.5 rotor. The supernatant was discarded, and the pellet was resuspended in buffer A containing 2 M urea. After 1-h incubation on ice followed by centrifugation (30 min at 26,000 rpm in a Beckman 30.5 rotor), the pellet was resuspended in buffer A containing 6 M urea. After another 1-h incubation on ice and centrifugation (30 min at 26,000 rpm in a Beckman 30.5 rotor), the protein extract was loaded onto a Q-Sepharose column equilibrated with buffer A containing 6 M urea. The proteins bound to this column were eluted with a KCl gradient (100 -300 mM) in buffer A supplemented with 6 M urea. Fractions containing pure IbpA, eluted at the end of the gradient, were pooled together. The urea present in the protein preparation was slowly dialyzed out by sequential dialysis to buffer A containing 4 M, 2 M, 1 M, and no urea.
The IbpB protein was overproduced in the MC4100⌬ibpA/B strain transformed with the pUC18 derived pCB plasmid bearing the ibpB gene under control of the ptac promoter (38). IbpB was purified using the protocol described previously (39).
To obtain the IbpB and IbpA proteins carrying a hexahistidine tag at the N terminus, pETHis-B and pETHis-A plasmids were constructed by cloning the ibpB or ibpA genes into the NdeI and BamHI sites of the pET28bϩ vector (Novagen). The His-IbpB and His-IbpA proteins encoded by these plasmids have an additional 20-amino acid residues, which includes the hexahistidine tag, at the N terminus. These fusion proteins were overproduced in the BL21(DE3) ⌬ibpA/B strain. The purification of His-IbpB and His-IbpA was based on the interaction of the His tag with Ni-NTA-agarose under denaturing conditions.
Published protocols were used for the purification of Escherichia coli DnaK, DnaJ, GrpE (41), and ClpB (42). Firefly luciferase (E 1701) was purchased from Promega. Malate dehydrogenase (MDH) was purchased from Sigma (410-13). Protein concentrations were determined with the Bradford (Bio-Rad) assay system, using bovine serum albumin as a standard. Molar concentrations are given assuming a hexameric structure for ClpB and a monomeric structure for the rest of the proteins. Luciferase Inactivation and Refolding Experiments-For inactivation, luciferase (1.5 M) in buffer B (50 mM Tris-HCl, pH 7.4, 150 mM KCl, 20 mM magnesium acetate, 5 mM dithiothreitol) was incubated at different temperatures for 10 min in the presence of IbpA and IbpB as stated in the figure legends. For disaggregation and refolding, the inactivated luciferase (42 nM) was incubated at 25°C in buffer B supplemented with ATP (5 mM), ATP regeneration system (10 mM phosphocreatine and 100 g/ml phosphocreatine kinase), and chaperone proteins (3 M DnaK, 0.24 M DnaJ, 0.3 M GrpE, and 3 M ClpB) as indicated. Following 1 h of incubation, luciferase activity was determined in a Beckman scintillation counter using the Luciferase Assay System (Promega). In the control experiment it was shown that luciferase activity increases linearly with time during refolding for at least 1 h.
Sizing Chromatography of Protein Complexes-Luciferase or MDH, following denaturation at different temperatures in the presence of IbpA and IbpB when indicated, was subjected to sizing chromatography on a Sepharose 4BCl column (0.5 ϫ 15 cm) equilibrated in buffer C. Proteins present in fractions (100 l) following sizing chromatography were separated by SDS-PAGE and stained with Coomassie Brilliant Blue. Chromatography resin from the sizing column was removed and discarded after each separation to avoid any contamination of the resin with the aggregated proteins.
Isolation of Protein Complexes by Interaction with Ni-NTA-agarose Resin-His-IbpB (6 M) and IbpA (6 M), or His-IbpA (6 M) and IbpB (6 M), and MDH (8.6 M), as indicated, were incubated in buffer C lacking dithiothreitol at different temperatures in a 35-l reaction volume. After 20 min of incubation, Ni-NTA-agarose (20 l) was added to the reaction mixture, and the mixture was incubated for an additional 10 min. Next, the supernatant was discarded, and the Ni-NTA-agarose was washed several times with buffer containing 50 mM Tris-HCl, pH 8.0, 300 mM NaCl, and 20 mM imidazole. Proteins bound to the resin were eluted with the same buffer containing 300 mM imidazole and subjected to SDS-PAGE followed by staining with Coomassie Brilliant Blue.

RESULTS
Purification of IbpA Protein-Two small heat shock proteins, namely IbpA and IbpB, were identified in E. coli. IbpB has been purified and studied previously (39), but, until now, no detailed biochemical studies had been performed on IbpA. IbpA was overproduced in E. coli and purified to homogeneity (Fig. 1). Previous studies showed that IbpA, either induced by heat shock conditions (38) or overproduced (39), was found in the fast sedimenting fraction of cellular extracts containing membranes and aggregated proteins, but not in the soluble fractions. Therefore, we extracted IbpA from these aggregates by dissolving them in 6 M urea. Further steps of the purification were performed under denaturing conditions. At the final step of the purification urea was slowly dialyzed out from the IbpA preparation. Following such a procedure, soluble IbpA protein was obtained. The purified IbpA formed high molecular weight oligomeric structures that deoligomerized into smaller structures following incubation at 48°C (results not shown).
Presence of Both IbpA and IbpB during Thermal Denaturation of Substrate Protein Is Required for Subsequent Reactivation by ClpB-K/J/E System-Previous work (14 -19) defined the sHsps as factors, which when present during substrate protein denaturation, increase the ability of chaperones from the Hsp100 and Hsp70 families to refold these aggregates into active protein. We used this assay to investigate the chaperone activity of purified IbpA. Firefly luciferase was thermally denatured at 48°C in the presence or absence of IbpA, followed by the addition of ClpB (Hsp100) and DnaK/DnaJ/GrpE (Hsp70 and cochaperones). Following 1-h reactivation at 25°C, the luciferase activity was tested. The presence of IbpA during luciferase denaturation did not increase the efficiency of its refolding by the ClpB-K/J/E bi-chaperone system, although a wide range of IbpA concentrations was tested for a possible stabilizing effect ( Fig. 2A). Because both ibpA and ibpB are located in one operon in E. coli (37) and it was shown that sHsps from other bacteria form heterooligomers (34), it is plausible that IbpA and IbpB cooperate in substrate stabilization during thermal stress. Therefore, we purified IbpB (Fig. 1) and included it in our in vitro reactivation experiments. The presence of IbpB alone during luciferase denaturation only slightly increased (from 14% to 33%) the efficiency of luciferase refolding by the ClpB-K/J/E chaperone system (Fig. 2B). Moreover, this effect was observed only for high concentrations of IbpB. However, when both IbpB and IbpA were added simultaneously to the reaction, a significant increase in luciferase activity was observed following reactivation by the ClpB-K/J/E chaperones (Fig. 2). To find the optimal concentrations of IbpA and IbpB necessary for this stabilizing effect, titrations of both small chaperones were performed. The most efficient reactivation was observed for ϳ2 M IbpA ( Fig. 2A) and ϳ7.8 M IbpB (Fig. 2B). Therefore these concentrations of sHsps were used throughout this report in reactivation experiments.
We also tested at which step of the denaturation/reactivation process the simultaneous presence of IbpA and IbpB was required. IbpA/IbpB when present during either the thermal inactivation step or the reactivation step, and in the absence of the ClpB-K/J/E bi-chaperone system, were not able to protect or to recover luciferase activity (Fig. 2C). Moreover, thermal denaturation followed by addition of IbpA/B and bi-chaperone system to the reactivation mixture did not result in an efficiency of reactivation higher than that observed for the bichaperone system alone. Only when the IbpA/B proteins were present during the thermal denaturation step, followed by subsequent addition of the bi-chaperone system, was the efficient recovery of luciferase activity observed (Fig. 2C). Thus we concluded that the presence of both IbpA and IbpB during thermal inactivation of luciferase is required for efficient reactivation mediated by the bi-chaperone Hsp70-Hsp100 system.
Simultaneous Presence of IbpA and IbpB Results in Stabilization of the Substrate in a Disaggregation Competent State-Because in vivo studies identified the IbpA/B proteins as factors responsible for increasing E. coli thermotolerance (31), we decided to test the ability of IbpA/B to protect substrate proteins over a wide range of temperatures.
First, we tested if IbpA/B are able to maintain the enzymatic activity of luciferase exposed to different temperatures. To this end, luciferase was incubated at the indicated temperatures for 10 min in the presence of IbpA/B, and its activity was assayed directly after incubation. No protective effect of IbpA/B was observed at high temperatures (above 42°C). However, under mild heat shock conditions (36 -40°C), the minor level of luciferase protection was observed, and the luciferase activity was   ) were added, and the mixture was incubated for 1 h. Luciferase activity was measured using a liquid scintillation counter. C, presence of IbpA and IbpB is required during the thermal denaturation step. Luciferase was thermally inactivated at 48°C in the presence or absence of IbpA and IbpB, followed by subsequent addition during the reactivation step of DnaK, DnaJ, GrpE, ClpB, and/or IbpA and IbpB as stated above.
Next, we tested how different temperatures affect the ability of IbpA and IbpB alone or in combination to maintain luciferase in a state competent for bi-chaperone mediated disaggregation and refolding (Fig. 3B). First we incubated luciferase alone at 36 -52°C. Following its thermal inactivation, the Hsp100-Hsp70 bichaperone system was added for reactivation. Bi-chaperone-dependent reactivation was not efficient when luciferase was inactivated at a temperature higher than 40°C (Fig. 3B). Addition of only IbpA to luciferase during the inactivation step did not increase the ability of the ClpB-K/J/E system to reactivate luciferase (Fig. 3B). When a similar experiment was performed with IbpB instead of IbpA, we observed an increase in luciferase activity following the reactivation step. However, this effect was significant only when luciferase was inactivated at 44°C or lower temperature. At higher temperatures, IbpB alone was not able to stabilize luciferase in a refolding competent state (Fig. 3B). Ad-dition of both IbpA and IbpB to luciferase during inactivation resulted in its much more efficient reactivation by the bi-chaperone system. Moreover this effect was pronounced over the entire range of inactivation temperatures (up to 50°C) (Fig. 3B). We conclude that IbpA and IbpB are not capable of protecting luciferase activity; however, their presence during luciferase thermal denaturation is required to maintain the inactivated luciferase in a reactivation competent state.
To characterize differences between the refolding competent luciferase aggregates formed in the presence of IbpA/B and aggregates formed in the absence of IbpA/B, we performed sizing chromatography experiments (Fig. 3C). Native luciferase eluted as a monomer in fractions 8 -10 (Fig. 3C). Luciferase inactivated at 48°C in the absence of IbpA/IbpB and subjected to sizing chromatography was not detected in the eluate from the column (Fig. 3C). Instead, we quantitatively recovered the luciferase from the chromatographic resin at the top of the column by boiling the resin in SDS containing buffer (Fig. 3C). The latter indicated that the physical properties of the luciferase aggregate prevented it from flowing through the chromatographic resin. Similar sizing chromatography performed with luciferase inactivated in the presence of both IbpA and IbpB, resulted in quantitative recovery of luciferase in the void volume (fractions 6 and 7) of the sizing column (Fig. 3C). Moreover, IbpA and IbpB were eluted in the same fractions as luciferase (result not shown). This experiment shows that the presence of small heat shock proteins during the thermal inactivation step changes the nature of luciferase aggregates in such a way that the aggregates withstand sizing chromatography. To further characterize the properties of this aggregate, we tested its hydrophobicity using the bis-ANS binding technique. The fluorescence of bis-ANS substantially increases upon binding to hydrophobic protein surfaces. When luciferase was thermally inactivated in the presence of IbpA and IbpB, we recorded a lower bis-ANS fluorescence signal as compared with the reaction in which luciferase and IbpA/B were separately incubated at high temperature (results not shown). These results suggest that luciferase aggregates formed in the presence of IbpA/B expose less hydrophobic surface to the solution than aggregates formed in the absence of sHsps.
To test if the observed cooperation between IbpA and IbpB is not restricted to one particular substrate, we performed similar experiments using MDH. Purified IbpB was previously shown to maintain MDH in a refolding competent state (16,18). First, we thermally denatured MDH alone by a10-min incubation at 47°C as described previously (16,18). Following such thermal inactivation and reactivation by the bi-chaperone ClpB-K/J/E system, only ϳ20% of the initial MDH activity was recovered. When IbpB, IbpA, or both IbpA and IbpB were present during the thermal inactivation step, the efficiency of ClpB-K/J/E-dependent MDH reactivation increased 3-to 5-fold (Fig. 4). The observed stabilizing effect by IbpB is in good agreement with previously published results (16,18). However, because our experiments performed with luciferase indicate that the ability of IbpB alone to maintain the refolding competent state is strongly dependent on the temperature of inactivation (Fig. 3), we decided to test the effect of higher temperature on the maintenance of MDH in a refolding competent state. When MDH was inactivated by incubation at 50°C, IbpB alone was no longer able to efficiently maintain MDH in a refolding competent state. Moreover, in this case, IbpA alone was more efficient than IbpB in maintaining MDH in a refolding competent state. However, as it was shown before for luciferase, the presence of both IbpA and IbpB is needed to maintain MDH in a state allowing recovery of over 70% of the initial enzyme activity following reactivation by the bi-chaperone system. To find the differences between MDH aggregates formed during incubation at 47 and 50°C, we subjected the reaction mixtures following such inactivations to sizing chromatography (Fig.  5). Each panel in Fig. 5 represents a separate sizing chromatography experiment. Lane V corresponds to the fractions collected from the void volume of the column, lane M corresponds to fractions in which the monomeric form of MDH was eluted, and finally lane R represents protein aggregates that did not enter the resin. The fraction R, was obtained by collecting resin from the top of the chromatographic column and boiling it in SDS containing buffer. Fresh sizing chromatography columns were set up for each gel filtration experiment to avoid any cross contamination with aggregated proteins. In these experiments IbpB with a hexahistidine tag at the N terminus (His-IbpB) was used instead of wild-type IbpB. This allowed us to distinguish between IbpA and IbpB in SDS-PAGE. The activity of His-IbpB was tested separately and is indistinguishable from IbpB lacking the hexahistidine tag (results not shown).
No complex between MDH and sHsps was detected following incubation of MDH with any combination of sHsps at 0°C (Fig.  5, left panels). In each case MDH was eluted in the fraction M characteristic to native MDH, and both IbpB and IbpA were eluted in the void volume of the column, as expected for proteins forming oligomeric structures. When MDH, in the absence of sHsps, was denatured at 47°C and subjected to sizing chromatography (Fig. 5, middle panels), it was not recovered in the column eluate either in the V or M fractions. Instead, it was found to remain at the top of the column (fraction R). The presence of IbpB or both IbpB/IbpA during MDH thermal inactivation at 47°C changed the properties of inactivated MDH, and it was found in the void volume of the column, suggesting that complexes between MDH and sHsps were formed. The presence of IbpA alone partially protected MDH, as ϳ50% of MDH was in the void volume and the rest remained with the resin (fraction R).
Next, we increased the temperature of MDH denaturation to 50°C (Fig. 5, right panels). MDH denatured in the absence of any sHsps associated with the resin as observed previously at 47°C. The presence of IbpB also did not result in the stabilization of MDH aggregates at 50°C. MDH was found exclusively in fraction R. Similarly, inactivation of MDH in the presence of IbpA results in nearly all the MDH associated with resin (fraction R). Only small amounts of MDH were detected in the void volume of the column (fraction V). However, when the thermal inactivation of MDH at 50°C was performed in the presence of both IbpA and IbpB, the properties of inactivated enzyme changed and nearly all the MDH was found in the void volume of the column (fraction V) (Fig. 5).
The sizing chromatography experiments (Fig. 5) and reactivation experiments (Fig. 4) gave corresponding results. The ability of IbpA/B to maintain MDH in the state competent for refolding by the bi-chaperone (ClpB-K/J/E) system correlates with the ability of these sHsps to change the nature of inactivated MDH in such a way that it withstands sizing chromatography and is recovered in the void volume of the column together with the IbpA and IbpB proteins. The fact that the small Hsps were recovered in the same fraction as denatured MDH indicates that under heat shock conditions they most likely are associated with the substrate protein.
IbpA Interacts with IbpB and Stimulates Its Association with a Substrate Protein-The sizing chromatography experiment did not directly address the question of whether a complex between inactivated MDH and sHsps is formed and which sHsps are present in this complex, because the gel filtration chromatography does not allow differentiation between oligomers formed by sHsps themselves and heterooligomeric complexes between aggregated MDH and sHsps. To solve this problem we took advantage of the hexahistidine-tagged IbpB. We started our analysis by investigating interactions between His-IbpB and IbpA. His-IbpB was incubated in the presence or absence of IbpA at different temperatures followed by addition of Ni-NTA-agarose. Proteins associated with the resin were analyzed by SDS-PAGE. Surprisingly, when His-IbpB alone was incubated with Ni-NTA-agarose, no His-IbpB was found to associate with the resin at 30°C and 42°C (Fig. 6A, lanes 1 and  3) and only a small amount of His-IbpB was bound at 47°C (Fig. 6A, lane 5). This suggests that IbpB alone is in a conformation that prevents direct interaction between the N-terminal His tag and the Ni-NTA resin. However, addition of IbpA to the reaction mixture containing His-IbpB resulted in recovery of both IbpA and His-IbpB associated with Ni-NTA-agarose but only when both proteins were incubated at higher temperatures (42 and 47°C) (Fig. 6A, lanes 4 and 6). In contrast, no binding of either His-IbpB or IbpA to Ni-NTA resin was observed after incubation of the proteins at 30°C (Fig. 6A, lane 2). In a control experiment, no binding of IbpA protein to the Ni-NTA-agarose was observed when His-IbpB was missing (data not shown). One of the possible interpretations of these results is that, because His-IbpB forms oligomers at low temperatures, the hexahistidine tags could be buried inside these structures. Increasing the temperature results in deoligomerization of His-IbpB, and IbpA, through its interaction with His-IbpB, stabilizes the structures that have exposed hexahistidine tags promoting the association of His-IbpB-IbpA complexes with Ni-NTA-agarose. As a consequence, both IbpA and His-IbpB associate with Ni-NTA-agarose at high temperature.
Similar Ni-NTA-agarose binding experiments were performed with His-IbpA instead of His-IbpB. Interestingly, in contrast to His-IbpB, His-IbpA efficiently interacts with Ni-NTA-agarose at all temperatures tested (Fig. 6C). Moreover, addition of IbpB to the reaction mixtures results in the efficient association of IbpB with the Ni-NTA-agarose at high (42 and 47°C) but not low (30°C) temperatures (Fig. 6C). These experiments indicate clearly that temperature-dependent formation of mixed complexes containing both IbpA and IbpB proteins does not depend on the presence of the histidine tag on a particular protein.
Next, we tested how the addition of the substrate protein, MDH, affects interaction of His-IbpB with Ni-NTA-agarose. When His-IbpB was incubated with MDH at 30°C, neither His-IbpB nor MDH was bound to Ni-NTA-agarose (Fig. 6B, lane 1). Moreover, addition of IbpA to the reaction mixture did not result in binding of any proteins to Ni-NTA-agarose. At higher temperature (47°C) incubation of MDH with His-IbpB alone resulted in recovery of limited amounts of both proteins in the fraction associated with Ni-NTA-agarose. Addition of IbpA to the reaction containing both MDH and IbpB increased the amount of MDH associated with the Ni-NTA resin (Fig. 6B, lane 4). When His-IbpB was omitted in the experiment, no proteins were detected in the eluate from the Ni-NTA-agarose (result not shown). These results suggest that at high temperature IbpA facilitates the association of substrate protein with IbpB.
Similar Ni-NTA-agarose binding experiments were performed with His-IbpA in the presence of MDH and IbpB. MDH did not associate with His-IbpA at 30°C, neither alone or in the presence of IbpB (Fig. 6D). Increasing the temperature of incubation to 47°C resulted in the association of MDH with His-IbpA. Addition of IbpB to the reaction mixture did not increase the association of MDH with Ni-NTA-agarose (Fig.  6D). In the control we showed that MDH and IbpB did not interact with the Ni-NTA-agarose at any temperature tested. DISCUSSION In this study we isolated IbpA, an E. coli small Hsp, and investigated its contribution to the ability of the chaperone network to disaggregate and refold substrate proteins subjected to heat shock conditions. Our results point to the importance of cooperation between IbpA and IbpB for the maintenance of aggregated proteins in a refolding competent state following challenge of the system with high temperature. We showed that IbpA alone is not efficient in stabilizing aggregated substrates in a refolding competent state during thermal inactivation. Similarly, the presence of IbpB alone also resulted in limited stabilization of luciferase or MDH. It was reported previously that under in vitro conditions the presence of IbpB alone was sufficient to stabilize MDH and luciferase in a refolding competent state, which allowed its subsequent reactivation by the Hsp100-Hsp70 bi-chaperone or Hsp70 chaperone systems (16,18). These results were obtained, however, for MDH and luciferase that were inactivated at relatively low temperatures (47 and 43°C, respectively) (16,18). Our observations are fully consistent with these reports but additionally point to the fact that IbpB alone is able to stabilize substrate aggregates in a refolding competent state only if thermal inactivation occurs at temperatures slightly above the denaturation threshold temperature for a particular protein substrate. Increasing the inactivation temperature over this limit, 44°C for luciferase and 47°C for MDH, results in the inability of IbpB to stabilize the aggregated substrates in a refolding competent state.
At higher temperatures the simultaneous presence of IbpA and IbpB during thermal inactivation of either luciferase or MDH was required to allow subsequent efficient disaggregation and refolding by the Hsp70-Hsp100 bi-chaperone system. It should be noted that the temperature range at which IbpA/B activity is manifested in vitro corresponds perfectly to the range of temperatures at which these small E. coli Hsps are required in vivo. Recently it was shown that E. coli ⌬ibpAB cells are 100-fold less viable after prolonged cultivation at 50°C as compared with wild-type cells (38). This phenotype is not observed for cells growing under conditions of lower temperature heat stress.
For both luciferase and MDH, neither IbpA nor IbpB alone were able to stabilize either of these aggregated substrates in a refolding competent state. The high hydrophobicity of MDH aggregates formed in the absence of sHsps following incubation at high temperature did not allow them to enter the sizing chromatography resin. Binding of both IbpA and IbpB most likely blocks possible hydrophobic interactions between substrate polypeptide chains exposed to the solution during thermal denaturation, thus allowing them to withstand sizing chromatography conditions. The nature of the cooperation between IbpA and IbpB in substrate stabilization during the aggregation process is not understood.
The cooperation between IbpA and IbpB was observed not only by stabilization of substrate aggregates in a refolding competent state but also by the direct interaction between IbpA and IbpB. We showed that IbpA associates with His-IbpB and, vice versa, IbpB associates with His-IbpA. Both complexes were formed only at high temperature, suggesting that following temperature-induced deoligomerization, IbpA and IbpB do FIG. 6. Temperature-dependent formation of complexes containing sHsps and MDH. A, both IbpA and high temperature increase the binding of His-IbpB to Ni-NTA-agarose. His-IbpB in the presence or absence of IbpA was incubated at the indicated temperatures followed by addition of Ni-NTA-agarose at the same temperature. After washing with binding buffer, proteins associated with the Ni-NTA-agarose were eluted with buffer containing a high concentration of imidazole, run on SDS-PAGE, and stained with Coomassie Brilliant Blue. B, both IbpA and high temperature increase the association of MDH with His-IbpB. His-IbpB and MDH in the presence or absence of IbpA were incubated at the indicated temperatures followed by addition of Ni-NTA-agarose at the same temperature. Proteins were eluted from the resin and visualized as above. C, IbpB interacts with His-IbpA at high but not at low temperature. Proteins were eluted from the resin and visualized as above. D, high temperature promotes the association of MDH with His-IbpA. Proteins were eluted from the resin and visualized as above. physically interact. Additionally, the presence of IbpA greatly enhances binding of His-IbpB to the Ni-NTA-agarose resin, suggesting that the interaction between IbpA and His-IbpB changes the conformation of His-IbpB in such a way that the hexahistidine tag is exposed to solution. IbpA and IbpB cellular localization studies (38) showed that, in the absence of IbpA, IbpB is localized in the cytoplasmic fraction, whereas in the presence of IbpA it is found in fractions containing aggregated protein and membranes. These results suggest that IbpA and IbpB proteins interact in vivo and that IbpA may be responsible for targeting of IbpB to aggregated proteins.
Formation of heterooligomeric sHsp complexes was previously reported for the same class of sHsps isolated from Bradyrhizobium japonicum (34), closely related ␣Aand ␣B-crystallins (44), Hsp27 and ␣B-crystallin (45), and human Hsp20 and Hsp27 (46). The studies performed for B. japonicum sHsps reveal that heterooligomers display chaperone activities indistinguishable from homooligomeric complexes (34). In contrast, the properties of homo-and hetero-oligomeric ␣-crystallin complexes differ, because it was reported that these complexes have different lens plasma membrane-binding properties (47). It was also shown that ␣A and ␣B-crystallin homooligomers possess unique chaperone activities that are additionally differently modulated in response to temperature changes (48,49).
Our studies revealed that homooligomers of IbpA and IbpB exhibited only weak chaperone activities manifested during substrate inactivation at relatively low temperature, whereas simultaneous presence of IbpA and IbpB leads to efficient stabilization of substrate protein aggregates in a refolding competent state at higher temperature. However, our in vitro results do not directly show the formation of the functional ternary complex containing IbpA, IbpB, and the substrate at high temperature. Therefore one has to consider the alternative scenario in which a simultaneous presence of both IbpA and IbpB is required, because IbpA (or IbpB) acts at high temperature as a chaperone to facilitate the conversion of the other small Hsp to an active form. Then both sHsps would interact with the substrate separately.
In summary, data presented in this report establish the cooperation of IbpA with IbpB in stabilizing protein aggregates in disaggregation and refolding competent state. The detailed biochemical analysis of interactions between IbpA and IbpB, which take place at high temperature conditions, will be important for understanding the mechanism of substrate stabilization.