Small Heat Shock Protein IbpB Acts as a Robust Chaperone in Living Cells by Hierarchically Activating Its Multi-type Substrate-binding Residues*

Background: Small heat shock proteins are ubiquitous molecular chaperones. Results: A total of 20 and 48 substrate-binding residues in IbpB were identified to function at 30 and 50 °C, respectively, in living cells. Conclusion: The substrate-binding residues of IbpB are hierarchically activated in a temperature-dependent manner. Significance: This is the first systematic identification of substrate-binding residues of a small heat shock protein in living cells. As ubiquitous molecular chaperones, small heat shock proteins (sHSPs) are crucial for protein homeostasis. It is not clear why sHSPs are able to bind a wide spectrum of non-native substrate proteins and how such binding is enhanced by heat shock. Here, by utilizing a genetically incorporated photo-cross-linker (p-benzoyl-l-phenylalanine), we systematically characterized the substrate-binding residues in IbpB (a sHSP from Escherichia coli) in living cells over a wide spectrum of temperatures (from 20 to 50 °C). A total of 20 and 48 residues were identified at normal and heat shock temperatures, respectively. They are not necessarily hydrophobic and can be classified into three types: types I and II were activated at low and normal temperatures, respectively, and type III mediated oligomerization at low temperature but switched to substrate binding at heat shock temperature. In addition, substrate binding of IbpB in living cells began at temperatures as low as 25 °C and was further enhanced upon temperature elevation. Together, these in vivo data provide novel structural insights into the wide substrate spectrum of sHSPs and suggest that sHSP is able to hierarchically activate its multi-type substrate-binding residues and thus act as a robust chaperone in cells under fluctuating growth conditions.

(1). Small heat shock proteins (sHSPs), 4 present in all forms of life (2), function as molecular chaperones to protect a wide spectrum of substrate proteins (3) and also stabilize cell membranes (4,5) and thus increase the tolerance of organisms for such harsh conditions as heat shock and oxidative stress (6 -8).
In addition, sHSPs have been linked to cell differentiation, apoptosis, animal longevity, and such diseases as cancer, cataracts, and neurodegenerative disorders (9 -12).
sHSPs are characterized as having a conserved ␣-crystallin domain of ϳ100 amino acids (2), which is flanked by a highly variable and structurally disordered N-terminal arm and a short C-terminal extension (13,14). Numerous in vitro studies indicate that the structurally disordered N-terminal arm of sHSPs is essential for sHSPs to bind model substrate proteins, although the ␣-crystallin domain and C-terminal extension might also be involved (15)(16)(17)(18)(19), and that their in vitro chaperone activities are enhanced upon exposure to heat shock conditions (13, 20 -22). Significant but unresolved scientific questions regarding the functions of sHSPs in living cells include which residues of sHSPs mediate the binding of the wide spectrum of natural substrate proteins (17) and how they play a role in the normally observed (although under in vitro conditions) heat shock-enhanced substrate-binding capacity of sHSPs (23).
To address these questions, we chose the inclusion bodybinding protein IbpB, a sHSP from Escherichia coli (24 -30), as a model to investigate the substrate-binding features of sHSPs in living cells. For this purpose, we incorporated the unnatural amino acid p-benzoyl-L-phenylalanine (Bpa) into a large number of individual positions in IbpB and performed in vivo photocross-linking (31,32). In contrast to conventional in vitro studies (e.g. truncation of sHSPs) of our own (15,25,26) and of others (16,18,19,33,34), the in vivo photo-cross-linking approach apparently has such advantages as causing minimal structural disturbances on IbpB and facilitating analysis of the participation of individual residues of IbpB in binding natural substrate proteins in E. coli. Notably, Bpa-mediated in vitro photo-cross-linking was applied recently to determine the substrate-binding residues of Hsp18.1, a pea sHSP (17). The use of two model substrate proteins therein, however, might not reflect the diversity of natural substrate proteins in cells.
Through in vivo photo-cross-linking analysis of 71 Bpa variants of IbpB, we identified 20 and 48 substrate-binding residues at normal and heat shock temperatures, respectively. It is striking that of the 48 residues, about half are polar residues (charged or uncharged). Significantly, these residues generally appear to be hierarchically activated in a temperature-dependent manner, with a number of them switching from mediating oligomerization to mediating substrate binding upon temperature elevation. Together, these features collectively allow IbpB, or sHSPs in general, to function as a robust molecular chaperone to act upon a large diversity of substrate proteins in living cells growing under fluctuating conditions.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmid Construction, and Protein Expression-E. coli DH5␣ cells were used for gene manipulation. The ibpB gene was amplified from E. coli genomic DNA and cloned into the pBAD plasmid at the NcoI and HindIII sites, with a tag of six histidine residues being added at the C terminus, yielding pBAD-ibpB-His 6 . Plasmids for expressing Bpa variants of IbpB were generated from the template plasmid pBAD-ibpB-His 6 using the QuikChange site-directed mutagenesis kit (Transgene, Beijing, China). The pSup-BpaRS-6TRN plasmid, expressing the orthogonal aminoacyl-tRNA synthetase/tRNA pair for the incorporation of Bpa into IbpB, was a gift from Dr. Peter Schultz (The Scripps Research Institute, La Jolla, CA) and was cotransformed with the pBAD-ibpB-His 6 plasmid into the ⌬ibpB mutant E. coli strain (kindly provided by the Nara Institute of Science and Technology, Ikoma, Nara, Japan). Cells were normally cultured in the presence of appropriate antibiotics (final concentrations of 100 g/ml ampicillin, 50 g/ml kanamycin, and 50 g/ml chloramphenicol (Sigma)), 1 mM Bpa (Bachem AG, Bubendorf, Switzerland), and 0.02% arabinose to induce protein expression.
Chaperone Activity Assay for Bpa Variants of IbpB-The chaperone activity of each Bpa variant was assayed as its capacity to suppress the heat-induced protein aggregation of EF-Tu present in the whole cell extract. Briefly, to prepare the whole cell extract, the ⌬ibpB mutant E. coli cells expressing each Bpa variant of IbpB-His 6 were cultured overnight at 30°C, harvested by centrifugation, washed twice with and resuspended in 20 mM Tris-HCl buffer (pH 8.0), lysed by sonication, and centrifuged at 13,000 ϫg for 30 min at 4°C to remove the cell debris. The whole cell extract (with its total protein concentration adjusted to 1 mg/ml) was then incubated at 50°C for 1 h, after which the soluble proteins and insoluble protein aggregates were separated by centrifugation at 13,000 ϫ g for 30 min at 4°C. The whole cell extract, soluble proteins, and insoluble protein aggregates for each Bpa variant were then analyzed by immunoblotting using EF-Tu-specific antisera (prepared by us) and anti-His tag antibody (Zhongshan Goldenbridge Biotechnology Co., Ltd.).
Bpa-mediated in Vivo Photo-cross-linking-To perform photo-cross-linking at 30, 37, 42, 45, or 50°C, the ⌬ibpB mutant E. coli cells transformed with pBAD-ibpB-His 6 carrying one site-specific mutation of IbpB were initially grown at 30°C to A 600 ϭ 0.4; induced by arabinose for 2 h to express the IbpB variant protein; incubated at 30, 37, 42, 45, or 50°C for 30 min; and transferred to 24-well plate (on ice) before being subjected to UV irradiation at 365 nm for 5 min using a Hoefer UVC 500 cross-linker (32). The cells were then lysed with SDS sample loading buffer before analysis by 10% Tricine/SDS-PAGE and immunoblotting with anti-His tag monoclonal antibody. To perform photo-cross-linking at 20, 25, or 30°C, cells were grown overnight in the presence of arabinose at 20, 25, or 30°C.
Formaldehyde-mediated in Vivo Chemical Cross-linking-To perform cross-linking at 30, 42, or 50°C, the ⌬ibpB mutant E. coli cells transformed with the pBAD-ibpB-His 6 plasmid were initially grown at 30°C to A 600 ϭ 0.4; induced by arabinose for 2 h to express IbpB; incubated at 30, 42, or 50°C for 30 min; cooled at 30°C for 30 s; and mixed with formaldehyde (at a final concentration of 1% (v/v)) for cross-linking (2.5 min) before quenching with glycine (at a final concentration of 0.5 M). Cells were lysed overnight with SDS sample loading buffer before being subjected to SDS-PAGE and immunoblot analysis. To perform cross-linking at 20, 25, or 30°C, cells were grown overnight in the presence of arabinose.
Protein Purification-All of the in vivo cross-linked products of IbpB were purified by nickel-nitrilotriacetic acid chromatography (GE Healthcare) in the presence of 8 M urea to extensively remove the noncovalently bound proteins. The Y45Bpa variant protein of IbpB used for the in vitro photo-cross-linking was successively purified by nickel-nitrilotriacetic acid chromatography (in the absence of urea) and size exclusion chromatography (Superdex 200, GE Healthcare). Protein concentrations were determined using the BCA assay (Pierce).
Protein Identification by MALDI-TOF-MS-After purification, the formaldehyde-mediated cross-linked products of IbpB were desalted with Amicon Ultra-15 centrifugal filter units (Millipore), boiled for 20 min to break the cross-linkages, separated by 10% Tricine/SDS-PAGE, and visualized by Coomassie Blue staining (see Fig. 2C). The two most predominant substrate protein bands were cut and subjected to MALDI-TOF-MS analysis using an Ultraflex system (Bruker Daltonics). After purification, the photo-cross-linked products of the Y45Bpa variant of IbpB (formed at 30°C) were further resolved by 10% Tricine/SDS-PAGE (see Fig. 3D, lane 4), and the dimeric and trimeric bands were subjected to MALDI-TOF-MS analysis.
Semiquantification of Relative Substrate Binding-The relative levels of photo-cross-linked substrates of each Bpa variant were calculated as the percentage of IbpB cross-linked to substrates based on the immunoblot results (the portion of monomeric, dimeric, and trimeric forms of IbpB was subtracted from the total cross-linked protein products) (see Fig. 3, A-C). Gel density analysis was performed using the NIH ImageJ program. For those variants in which the cross-linked IbpB-substrate products were barely visible, their relative substrate-binding levels were all arbitrarily set as zero.
Molecular Modeling of IbpB-To model the dodecameric structure (12 subunits) of IbpB, we used the three-dimensional structures of Hsp16.9 (Protein Data Bank code 1GME), which has six repeats of chains A and B. Although the sequences of chains A and B are identical to each other, they have ϳ80% structural similarity between them. The structures of chains A and B were separately modeled, and they were put together into the dodecameric structure. To model each chain of IbpB, we employed a recently proposed high accuracy template-based modeling method (35)(36)(37). Three additional Protein Data Bank structures of sHSPs (Hsp16.5 (code 1SHS), HspA (code 3GLA), and Hsp14.0 (code 3AAB)) were used as templates to model each monomeric structure. The dodecameric structure was modeled by fitting 12 monomeric structures into the quaternary structure of Hsp16.9 by optimizing the MODELLER energy function comprising intermolecular terms by Modeler CSA (48) using the sequence alignment between IbpB and Hsp16.9.

IbpB Chaperone Activity Is Barely Affected by Incorporation of Unnatural Amino Acid Bpa at 71 Selected Individual Residue
Positions-For the determination of substrate-binding residues of IbpB in living cells, the unnatural amino acid Bpa was incorporated into IbpB at 71 selected individual positions (red sites in Fig. 1), replacing 36 of 39 residues in the N-terminal arm, 19 of 81 residues in the ␣-crystallin domain, and 16 of 22 residues in the C-terminal extension. A high proportion of residues from the N-terminal arm and the C-terminal extension were selected for Bpa incorporation because these two regions were found to be indispensable for the chaperone activity of IbpB in our earlier in vitro truncation studies (26). For the convenience of immunoblot analysis and protein purification, these Bpa variants were all expressed as C-terminally His-tagged forms; to avoid the interference of endogenous IbpB, the Bpa variants were expressed in ibpB-deleted E. coli cells. In addition to these 71 Bpa variants, another seven variants (Phe-54, Leu-59, Leu-63, Leu-68, Gly-72, Phe-93, and Arg-137) were generated but not subjected to further examination due to their extremely low expression levels.
Before performing in vivo photo-cross-linking, we first demonstrated that these 71 Bpa variants of IbpB exhibited a level of chaperone activity largely comparable with that of wild-type IbpB. The chaperone activities were assayed in the context of the whole cell extract by measuring their ability to suppress heat-induced aggregation of endogenous EF-Tu protein, which is a translation elongation factor and which was identified by mass spectrometry as being one of the two most predominant substrate proteins that are cross-linked to IbpB by formaldehyde in living cells (Fig. 2C). The immunoblot data displayed in Fig. 2A indicate that most of the Bpa variants of IbpB, when properly expressed, effectively suppressed the heat-induced aggregation of a significant amount of EF-Tu. Notably, many of the Bpa variants were also present in the pellet fraction, which most likely resulted from the formation of co-aggregates between IbpB and its substrate proteins (including EF-Tu), as suggested by our in vitro studies (38). These observations also demonstrate that the introduced Bpa residue has a negligible effect on the structure and function of IbpB. It should be pointed out that the Bpa variants of the IbpB proteins were expressed in ibpB-deleted cells at a level highly comparable (or even lower) with that of endogenous IbpB in wild-type E. coli cells (as shown by immunoblotting) (Fig. 2D).
The Number of Residues Involved in Substrate Binding for IbpB Is More than Doubled at Heat Shock Temperature Compared with Low Temperature in Living Cells-We subsequently probed the interaction of each Bpa variant of IbpB with substrate proteins in cells cultured at 50 or 30°C, given that our earlier in vitro study revealed that the chaperone activity of IbpB is maximal at 50°C but almost undetectable at 30°C (26). The immunoblot results displayed in Fig. 3 (A and B) demonstrate that an overwhelming number of photo-cross-linked products (appearing as smears on the gels) were formed in cells cultured at 50°C for most of the Bpa variants but also at 30°C for some of them. Such cross-linked products were apparently made of both IbpB homo-oligomers and IbpB-substrate complexes. In the case of the Y45Bpa variant, it was predominantly cross-linked as homo-oligomers at 30°C as indicated by the positions (Fig. 3, B, lane 2b, and D, lane 4) and as verified by mass spectrometry analysis, but largely as IbpB-substrate complexes at 50°C (Fig. 3B, lane 2c). The non-homo-oligomer product bands are considered to be mostly IbpB-substrate complexes for each Bpa variant, with a small number likely being complexes of IbpB and other molecular chaperones (28 -30).
The semiquantitative analysis results (displayed in Fig. 4) of these immunoblot data reveal that more residues in IbpB were involved in binding substrate proteins at 50°C than at 30°C: 48 and 20 residues, respectively. Further detailed analysis indicated that for the N-terminal arm, although 17 residues already participated in substrate binding at 30°C, an additional 16 did at 50°C. Similarly, for the ␣-crystallin domain, although three residues participated at 30°C, an additional 11 did at 50°C. For the C-terminal extension, no residues participated at 30°C, but one residue (Ala-139) did at 50°C. Significantly, among the additional 28 residues that exclusively became active in substrate binding at 50°C, seven of them (Phe-32, Phe-33, Lys-39, Tyr-45, Lys-71, Lys-78, and Ala-139) participated primarily in homo-oligomerization at 30°C and switched to substrate binding at 50°C (Fig. 3, A-C). On the other hand, many of the 20 residues involved in substrate binding at 30°C (e.g. Arg-2, Ser-7, and Phe-97) significantly increased the amount of sub-  APRIL 26, 2013 • VOLUME 288 • NUMBER 17 strates bound to them at 50°C (Fig. 4). In addition, the substrate-binding residues of IbpB, especially at 30°C, appeared to be located predominantly in the N-terminal arm (Fig. 4), in line with in vitro observations of our own (15) and others (16 -18).

Substrate-binding Residues of IbpB in Living Cells
Substrate-binding Residues of IbpB Are Not Necessarily Hydrophobic-Remarkably, many of the identified substratebinding residues of IbpB are actually polar amino acids, with no apparent preferences for hydrophobic ones (Table 1). This result seems to contradict the conventional view that sHSPs use mainly hydrophobic surfaces to bind substrate proteins. In retrospect, such hydrophobic surfaces, often being examined with hydrophobic probes under in vitro conditions, were only assumed to participate in substrate binding, with the detailed amino acid composition unexamined (15, 18, 39 -41). In line with our observations reported here, non-polar residues in the ␣-crystallin domain of Hsp16.5 were recently suggested to be involved in substrate binding based on cryo-EM studies (42).
Substrate-binding Residues Are Activated at Characteristic Temperatures and Can Be Categorized into Three Types-The aforementioned in vivo photo-cross-linking data demonstrate   (lanes 10a, 10b, and 10c in B) were used as controls. The dotted lines in B and C indicate the region (Glu-104 -Ala-139) within which the IbpB dimers and trimers appear as doublets, most likely reflecting one or two cross-linkages within the dimers and two or three within the trimers. the existence of three types of substrate-binding residues in IbpB (as indicated in Fig. 1): type I (20 residues, represented by Phe-8, Phe-16, and Gln-24), located generally in the N-terminal arm and capable of mediating substrate binding at low temperature (30°C); type II (21 residues, represented by Phe-4, Met-10, Trp-13, Arg-46, Glu-57, and Arg-67), exclusively able to mediate substrate binding at high temperature (50°C); and type III (seven residues, represented by Phe-32 and Tyr-45), involved in oligomerization at low temperature (30°C) but switching to substrate binding at high temperature (50°C).
We subsequently determined the characteristic lower temperatures at which the type I residues were incapable, as well as the higher temperatures at which the type II and III residues became capable of binding substrate proteins. For this purpose, cells expressing the Bpa variants of type I residues were grown at 20, 25, or 30°C before being subjected to photo-cross-linking on ice; cells expressing the Bpa variants of type II and III residues were first grown at 30°C and then treated at 30, 37, 42, 45, or 50°C for 30 min before being subjected to photo-cross-linking.
The immunoblot analysis results (presented in Fig. 5, A-C) demonstrate the following. Type I residues (as represented by Phe-8, Phe-16, and Gln-24) were able to significantly bind substrate proteins at 25°C (albeit at lower levels than at 30°C) but became incapable at 20°C (Fig. 5A). Type II residues became capable of binding substrates at the normal growth temperature of 37°C, with the level of substrate proteins bound consecutively increased upon further temperature elevation (as represented by Phe-4 and Trp-13 (Fig. 5B) and similarly for Met-10 and Arg-67 (data not shown)). Type III residues (as represented by Phe-32 and Tyr-45) largely mediated homo-oligomerization at lower temperatures but switched to binding primarily sub-strate proteins at particular heat shock temperatures of 42°C for the N-terminal arm residue Phe-32 and 50°C for the ␣-crystallin domain residue Tyr-45 (Fig. 5C). Thus, type I residues apparently start binding substrate proteins at temperatures as low as 25°C, whereas type II residues do so at the normal growth temperature of 37°C and type III residues in the heat shock temperature range of 42-50°C.
We next examined the overall substrate-binding capacity of IbpB in living cells by applying formaldehyde-mediated in vivo chemical cross-linking. Unlike the position-specific photocross-linking method, this approach would theoretically covalently link all bound substrate proteins with IbpB. The immunoblot results displayed in Fig. 5D demonstrate that the substrate binding of IbpB barely occurred at 20°C (lane 3); significantly occurred at 25°C (lane 5); and further increased successively at 30°C (lane 6), 42°C (lane 9), and 50°C (lane 10). In particular, a significant amount of high molecular mass crosslinked products, which were apparently unable to enter the separating gel but were retained in the stacking gel, was detected in cells after being treated at 50°C (Fig. 5D, lane 10), but not at 30 and 42°C (lanes 8 and 9).

Insights into Why sHSPs Are Able to Bind a Wide Spectrum of
Substrate Proteins in Living Cells-sHSPs are known to nonselectively bind aggregation-prone unfolded model substrate proteins under in vitro conditions (43,44) and a wide range of cellular proteins (3). Our data presented here demonstrate that IbpB, a sHSP from E. coli, is able to bind a wide spectrum of natural substrate proteins in living cells.
The structural basis for the wide substrate spectrum of IbpB might be viewed as follows based on our in vivo data. First, a large  number of residues, spreading across the entire polypeptide chain, participate in substrate binding. Second, many of these residues are spatially dispersed throughout the dodecameric structure of IbpB as modeled on the basis of the determined three-dimensional structure of Hsp16.9 (13) (Fig. 6A), thus potentially creating a variety of substrate-binding surfaces that cope with the diversity of substrate proteins. Third, these residues seem to have no preference for non-polar amino acids, as about half of them are polar (uncharged and charged) ( Table 1), suggesting that hydrogen bonds and ionic interactions (in addition to hydrophobic interactions) are also involved in substrate binding. Additionally, the potentially disordered region in the N-terminal arm of IbpB ending at Phe-32 (as indicated by the dashed line in Fig. 4), which aligns with the defined disordered region of Hsp16.9 (data not shown) and is highly flexible according to our earlier proteolytic cleavage results (26), contributes a large number of substrate-binding residues. This suggests that structural disorderness is a prerequisite for IbpB to bind its substrate proteins, consistent with the general belief that molecular chaperones utilize disordered structures to bind substrate proteins (45)(46)(47)(48). Given that the substrate proteins have to be unfolded before they can bind to sHSPs, we propose that the sHSP-substrate interaction occurs through a "disorder-fittingdisorder" mechanism. Such a mechanism would conceivably also allow IbpB to possess a diversity of geometries for binding a wide spectrum of substrate proteins.
Hierarchical Activation of Multi-type Substrate-binding Residues Enables sHSP to Act as a Robust Molecular Chaperone in Living Cells over a Wide Spectrum of Temperatures-Our data also strongly suggest that, in living cells, IbpB functions as a robust molecular chaperone, with its substrate-binding residues being hierarchically activated upon temperature elevation, as schematically illustrated in Fig. 6B. Briefly, remaining inactive at temperatures as low as 20°C, IbpB becomes moderately active at relatively low temperatures of 25-30°C, highly active at such normal/mild heat shock temperatures of 37-42°C, and maximally active under severe heat shock conditions of 45-50°C. In parallel with this, type I-III residues in IbpB start to bind substrate proteins at a particular temperature (Fig. 6B).
We noticed that the substrate-binding residues of IbpB include both buried and exposed ones accordingly to our mod- FIGURE 5. Substrate binding of IbpB in living cells starts at 25°C and further increases upon temperature elevation. A-C, immunoblot results with the in vivo photo-cross-linked products of Bpa variants of IbpB using the anti-His tag antibody, with Bpa incorporated at Phe-8, Phe-16, and Gln-24 (type I residues), Phe-4 and Trp-13 (type II residues), or Phe-32 and Tyr-45 (type III residues), respectively. Cells expressing P8Bpa, F16Bpa, and Q24Bpa were cultured overnight at 20, 25, or 30°C before being subjected to UV irradiation for photo-cross-linking. Cells expressing F4Bpa, W13Bpa, F32Bpa, or Y45Bpa were grown at 30°C and treated at the indicated temperatures (30,37,42,45, and 50°C) for 30 min before being subjected to UV irradiation. D, immunoblot results with formaldehydemediated in vivo cross-linked products of IbpB, with the cells grown overnight at 20, 25, or 30°C before incubation at 25°C and addition of formaldehyde (lanes 1-6) or with the cells grown at 30°C, treated at 30, 42, or 50°C for 30 min, and pre-cooled to 30°C for 30 s before addition of formaldehyde (lanes 7-10). HMW, high molecular weight. eling data (Fig. 6A). In line with this, a recent study defined two modes for Hsp16.5, with the substrates being bound either in the core or on the surface of the oligomer (42). During the temperature-dependent activation process of IbpB, significant structural readjustments would have to occur for those buried residues to become accessible to substrate proteins. In addition, oligomeric reorganization apparently occurs in order for the type III residues to be activated, consistent with the common observation that heat shock-induced oligomeric dissociation of sHSPs correlates with the enhancement of their chaperone function (13, 20 -22, 26).
This temperature-dependent hierarchical activation mechanism may well be applicable to other sHSPs whose in vitro chaperone activities have been demonstrated to be enhanced at heat shock temperatures, as reported for yeast Hsp26 (20,23,49), Mycobacterium tuberculosis Hsp16.3 (21), Methanococcus jannaschii Hsp16.5 (22), Archaeoglobus fulgidus Hsp20.2 (50), wheat Hsp16.9 (13), pea Hsp18.1 (17,47), and mammalian ␣-crystallin (51). Our remarkable observation that IbpB signif-icantly binds substrate proteins at temperatures as relatively low as 25 and 30°C strongly suggests the following. In living cells, not only are unfolded substrate proteins commonly present, but also sHSPs are able to effectively capture them. On the other hand, the lack of substrate binding for IbpB at temperatures as low as 20°C indicates the following: either aggregationprone substrate proteins are rarely present, or the structure of IbpB lacks the "disorderness" and thus becomes inert in binding substrate proteins.
Other outstanding unresolved issues regarding the chaperone function of IbpB in living cells that need to be further clarified include the following: the nature and fate of substrate proteins captured by IbpB under different temperatures, the way that IbpB and other protein quality control factors (e.g. IbpA, DnaK, and ClpB) cooperate in dealing with the unfolded substrate proteins, and the specific recognition of non-native but not native substrate proteins by sHSPs conceivably through the disorder-fitting-disorder mechanism. FIGURE 6. Mapping type I-III residues into the modeled dodecameric structure and schematically illustrating the temperature-dependent hierarchical activation of the substrate-binding residues of IbpB. A, the dodecameric structure of IbpB was modeled by reference to the threedimensional structure determined for Hsp16.9, with the N-terminal arm, ␣-crystallin domain, and C-terminal extension colored in blue, white, and green (first panels), respectively. The amino-terminal 31 residues in half of the 12 subunits of IbpB were assumed to be structurally disordered by referring to Hsp16.9 (13) and were thus disregarded during modeling. Type I-III residues (all colored red) were mapped into the modeled IbpB structure (second, third, and fourth panels, respectively). B, schematic illustration of the hierarchical activation of type I-III substrate-binding residues of IbpB at particular temperatures. Inactivated residues are colored dark gray, and activated residues are colored blue. The activation of type III residues involves certain oligomeric reorganization. It is also illustrated here that multiple species of substrate proteins may simultaneously bind to each IbpB oligomer and that each substrate molecule may bind to multiple residues in IbpB.