IbpB‐bound substrate release in living cells as revealed by unnatural amino acid‐mediated photo‐crosslinking

Small heat shock proteins (sHSPs) are known to bind partially folded proteins under stress conditions. In this study, by using photo‐crosslinking, we showed that the substrate binding of sHSP IbpB in living cells is reversible upon short‐time exposure at elevated temperature, providing in vivo evidence on the existence of substrate protein release at nonstress conditions.

Small heat shock proteins (sHSPs) are known to bind non-native substrates and prevent irreversible aggregation in an ATP-independent manner. However, the dynamic interaction between sHSPs and their substrates in vivo is less studied. Here, by utilizing a genetically incorporated crosslinker, we characterized the interaction between sHSP IbpB and its endogenous substrates in living cells. Through photo-crosslinking analysis of five Bpa variants of IbpB, we found that the substrate binding of IbpB in living cells is reversible upon short-time exposure at 50°C. Our data provide in vivo evidence that IbpB engages in dynamic substrate release under nonstress conditions and suggest that photo-crosslinking may be a suitable method for investigating dynamic interaction between molecular chaperones and their substrates in living cells.
Molecular chaperone proteins play key roles in maintaining proteostasis in vivo by assisting the folding of substrate proteins, preventing or reversing misfolding proteins, and the degradation of their misfolded forms [1,2]. Small heat shock proteins (sHSPs), as a conserved family of molecular chaperone with a low molecular mass of 12-43 kDa, are present in all forms of life [3]. The ability of sHSPs to confer resistance on cells under various stress conditions has been widely reported [4][5][6][7]. They are in proteostasis network referred to as first line stress defenders, binding non-native substrate proteins and holding them in a folding-competent state as 'holdases', which might be subsequently refolded with the assistance of other ATPdependent molecular chaperones such as Hsp60, Hsp70, and Hsp100 [3,[8][9][10][11][12][13]. It has been shown that the sHSPs associate with protein aggregates, altering their biochemical properties, and subsequently facilitate efficient disaggregation and refolding [9,[14][15][16][17]. However, the basic understandings about the interaction between sHSPs and their substrate proteins have been obtained mostly from in vitro studies [3], raising the question of their unique in vivo characteristics. Such unresolved scientific questions regarding the unique properties as regard to mode of operation of sHSPs in vivo include among others, what kind of dynamic changes take place in the interaction between sHSPs and their endogenous substrate proteins and, whether such interaction is reversible or irreversible? The answers to these questions may further our understanding of the molecular mechanisms for sHSPs to function in vivo.
To address these questions, we chose the inclusion body binding protein IbpB, a sHSP from Escherichia coli (E. coli) as a model to investigate the interaction characteristics of sHSPs and their substrate proteins in living cells. IbpB, which was initially identified as a component of inclusion bodies [18,19] and later reported to be present in heat shock formed aggregates [20], confers E. coli cells resistance against stresses [21,22]. In our previous studies, substrate-binding residues of IbpB were characterized by using in vivo site-specific photo-crosslinking as mediated by the genetically incorporated unnatural amino acid pbenzoyl-L-phenylalanine (Bpa) [23], and the substrate proteins captured by Bpa were identified to have remarkable preference for translation-related proteins and metabolic enzymes [24]. In the present study, we chose a total of five IbpB Bpa variants to investigate the interaction features of IbpB-substrate in E. coli cells subjected to both stress and nonstress conditions by adopting in vivo site-specific photo-crosslinking, as it is capable of covalently capturing transiently or weakly interacting proteins [25]. Our results revealed that the IbpB-substrate interaction in living cells is reversible upon short-time exposure to 50°C, providing the in vivo evidence on the existence of substrate protein release at nonstress conditions.

Materials and methods
Bacterial strains, plasmid construction, and protein expression Escherichia coli BW25113-DibpB strain was obtained from Nara Institute of Science and Technology in Japan. E. coli DH5a cells were used for gene manipulation. The recombinant plasmids, expressing wild-type or Bpa variants of IbpB with a tag of six histidine residues being added at the C terminus, were constructed as described previously [23]. The pSup-BpaRS-6TRN plasmid, expressing the orthogonal aminoacyl-tRNA synthetase/tRNA pair for the incorporation of Bpa into IbpB, was cotransformed with the recombinant plasmid into E. coli DibpB cells. Cells were cultured at 30°C in the presence of appropriate antibiotics (final concentrations of 100 lgÁmL À1 ampicillin, 50 lgÁmL À1 kanamycin, and 50 lgÁmL À1 chloramphenicol; Sigma, St Louis, MO, USA), 1 mM Bpa (Bachem AG, Bubendorf, Switzerland), and 0.02% arabinose to induce protein expression.

Chaperone-like activity assay for Bpa variants of IbpB
The chaperone-like activity of each Bpa variant was measured to determine its capacity to suppress the heat-induced aggregation of whole cell extract of E. coli DibpB cells. Briefly, the E. coli DibpB cells overexpressing each Bpa variant of IbpB-His 6 were cultured overnight at 30°C. Cells were harvested by centrifugation, washed twice, and resuspended in 20 mM Tris/HCl buffer (pH 8.0), lysed by sonication, and centrifuged at 13 000 g for 30 min at 4°C to remove the cell debris. The resultant whole cell extract was incubated at 50°C for 1 h. After heat shock treatment, 400 lL cell extract was divided into two parts and 20 lL was taken from one part as the whole proteins. The other part was centrifuged at 13 000 g for 30 min at 4°C; then, 20 lL was taken from the supernatant as the soluble proteins fraction. After discarding all the supernatants, the pellet was washed twice and resuspended using 200 lL Tris/ HCl; then, 20 lL was taken as the insoluble proteins fraction. The whole cell extract, soluble proteins, and insoluble protein aggregates were subjected to 10% Tricine/SDS/ PAGE and Coomassie blue staining analysis. The relative chaperone-like activity was defined by the percentage of the soluble protein of the whole cell extract according to the semiquantification results based on the corresponding Coomassie blue staining results. The mean gel density analysis was measured using IMAGEJ software [26] Bpa-mediated in vivo photo-crosslinking The E. coli DibpB cells transformed with pBAD carrying the gene of IbpB Bpa variant and pSup-BpaRS-6TRN plasmid were initially grown at 30°C to A 600 = 0.4; induced by arabinose for 2 h to express the IbpB variant protein; washed twice using fresh LB to remove arabinose; incubated at 50°C for 10 min; and transferred back to 30°C for a prolonged incubation. Cultures were taken out at indicated time points and immediately transferred to 24-well plate before being subjected to UV irradiation at 365 nm for 10 min using a Hoefer UVC 500 crosslinker. The cells were lysed, analyzed by 10% Tricine/SDS/PAGE, and then immunoblotted with anti-His tag monoclonal antibody.

Semiquantification of relative substrate binding
The relative levels of photo-crosslinked substrates of each Bpa variant were calculated as the percentage of IbpB crosslinked to substrates based on the immunoblotting results. It should be mentioned that the portion of monomeric, dimeric, and trimeric forms of IbpB Bpa variant was subtracted from the total crosslinked protein products during image processing using IMAGEJ software.

Results
Incorporation of the unnatural amino acid Bpa into the N-terminal arm of IbpB at five selected individual position The interaction between IbpB and its substrate proteins in E. coli cells at 30°C and 50°C has been investigated by utilizing genetically incorporated photocrosslinker Bpa [23]. Given that the substrate protein binding of IbpB was shown to be enhanced upon temperature elevation, it is of interest to determine whether a decrease in growth temperature will reversely weaken such interactions. To address this point, five individual residue positions (Phe-4, Leu-6, Trp-13, Ala-20, and Phe-32 as shown in Fig. 1A) from the Nterminal arm of IbpB were selected for Bpa incorporation because these IbpB Bpa variants have been shown to participate primarily in homo-oligomerization at 30 o C and switch to substrate binding at 50°C [23], making it easy to compare the changes in substrate binding of IbpB when temperature rises from 30°C to 50°C and then falls back to 30°C. To avoid the interference of endogenous IbpB, E. coli BW25113-DibpB strain was used to express the Bpa variants. Besides, these Bpa variants were all expressed with a tag of six histidine residues being added at the C terminus for the convenience of western immunoblot analysis. As displayed in Fig. 1B, the expression of five IbpB variants cannot be detected unless by the addition of Bpa, while the expression of wild-type IbpB is independent of Bpa, confirming the successful incorporation of unnatural amino acid Bpa into the selected position.
The chaperone-like activity of IbpB is barely affected by incorporation of Bpa at five selected individual residue positions To confirm that the insertion of Bpa at the selected residue positions does affect the known chaperone function of IbpB, we first determined the chaperonelike activity of five Bpa variants of IbpB by measuring their abilities of suppressing heat-induced aggregation of the whole cell extract prior to performing in vivo photo-crosslinking. Wild-type IbpB or its Bpa variants were overexpressed in E. coli BW25113-DibpB strain and cultured at 30°C, and then, the whole cell extract was isolated and incubated at 50°C for 1 h. After centrifugation, the soluble proteins and insoluble proteins were analyzed by Coomassie blue staining. As for DibpB cells, only about half of the cell extract proteins remained soluble upon heat treatment ( Fig. 2A, Lane  3), while almost all the cell extract proteins were kept in the soluble state when wild-type IbpB was overexpressed ( Fig. 2A, lane 6). This probably could be due to the high level of expression of wild-type IbpB protein, which protected the cell extract proteins from aggregation, therefore unable to detect insoluble proteins in the pellet fraction. The protein level of Bpa variants in the cell extract varied with the Bpa incorporation site and is much lower than that of the wild- type IbpB. However, most of the cell extract proteins were still maintained in the supernatant fraction ( Fig. 2A, lanes 8, 11, 14, 17, 20). We further compared the relative chaperone-like activity of the Bap variants by using semiquantitative analysis. Despite the relative low level of protein expression, all the Bpa variants exhibited over 75% chaperone-like activities (Fig. 2B), indicating that the incorporation of Bpa has a negligible effect on the molecular chaperone function of IbpB.

Substrate protein binding of IbpB in living cells is reversible upon short-time exposure to 50°C
Given that the chaperone-like activity of IbpB has been shown to be maximal at 50°C but almost undetectable at 30°C [22], we selected 50°C as the heat shock temperature and 30°C as the recovery temperature. The E. coli cells expressing IbpB Bpa variants precultured at 30°C were subjected to heat shock at 50°C for 10 min and then transferred back to 30°C for a prolonged incubation. The cell cultures were sampled at different time points (i.e., before heat shock, immediately after heat shock and 1, 2, 4 h after returning to 30°C) and subsequently subjected to photocrosslinking analysis. We observed that after experiencing a heat shock treatment, the interaction of the tested residues Bpa with cellular proteins, as reflected by its photo-crosslinked products (exception of crosslinked homo-oligomers), was substantially decreased when returned to incubation for varying length of time at 30°C, and this decrease in the in vivo photo-crosslinked bound proteins showed different patterns, which is dependent on the locations of the residue.   3. Characterization of the IbpB-substrate interaction in living cells during recovery from heat shock. Escherichia coli DibpB cells were grown at 30°C with the presence of 0.02% arabinose to mid-exponential phase, centrifuged at 2504 g for 1 min, resuspended and washed by fresh LB medium. The obtained cell cultures were exposed to heat shock at 50°C for 10 min, cooled down on ice for 30 s and returned to the incubation at 30°C. The cell cultures were sampled at different time points, i.e. before heat shock, right after heat shock and 1, 2, and 4 h after returning to 30°C. All the samples were exposed to UV light irradiation for 10 min, resolved by SDS/PAGE and immunoblotted by anti-His tag antibody. (A-E) The western blotting analysis of the crosslinking of F4Bpa, L6Bpa, W13Bpa, A20Bpa, and F32Bpa with cellular proteins and the semiquantification of relative substrate-binding levels for each of the five Bpa variants of IbpB, respectively. Each experiment was repeated three times. Semiquantitative analysis data are presented as mean AE SEM. Specifically, the photo-crosslinked products in the F4Bpa were not decreased during recovery for 1 and 2 hours until 4 hours (Fig. 3A). In contrast, the photocrosslinked products of other 4 Bpa variants were gradually decreased at indicated time upon a recovery from heat shock treatment (Fig. 3B-E). It could be seen that a 4-h incubation at 30°C enabled all these five IbpB Bpa variants to display almost the same crosslinking profile similar to that of the cells without heat shock. Besides, it should be mentioned that we have removed the arabinose after the induction of Bpa variants, avoiding the interference of newly synthesized variants in the crosslinking results during the prolonged incubation at 30°C. Together, these data suggested that the substrate protein binding of IbpB in living cells was reversible upon short-time exposure to 50°C, which implied that the substrate proteins bound with IbpB upon heat shock stress and are released at nonstress conditions.

Discussion
It has been found that IbpB was able to bind a wide spectrum of natural substrate proteins at 50°C in living cells [23,24]. Although it is believed that the substrate proteins which are bound to sHSPs would be released after the heat shock condition is removed, this interaction between sHSPs and their substrates during the recovery stage has rarely been examined in vivo. Here, by using Bpa-mediated photo-crosslinking, we characterize the interaction between IbpB and its endogenous substrates in living cells. Our data suggested that the substrate protein binding of IbpB in vivo was reversible upon short-time exposure to 50°C, providing the evidence that decreasing temperature leads IbpB to gradually hand off its bound aggregation-prone and partially unfolded proteins. In vitro study demonstrated that IbpB-bound proteins were stabilized in a conformation that can be subsequently released and specifically refolded by the DnaK-DnaJ-GrpE chaperones [15,16]. However, it remains unknown how does the release of substrate take place in vivo. Further research needs to be done to reveal how the ATP-dependent chaperones participate in this process.
IbpB, or sHSPs in general, function as a robust molecular chaperone to act upon a large diversity of substrate proteins in living cells growing under fluctuating conditions [23,27], therefore making sHSP-substrate interactions complex, and involving multiple sites on the sHSP. It has been demonstrated that the substratebinding residues of IbpB are located predominantly in the N-terminal arm [23]. Here, we chose a total of five IbpB Bpa variants with Bpa incorporated in their N-terminal arm, to investigate the IbpB-substrate interaction. We found decrease in crosslinked cellular proteins in all tested IbpB Bpa variants when returned for further incubation at 30°C after 50°C heat stress. Fu et al. proposed according to their in vivo photocrosslinking data that there are three types of substrate-binding residues in IbpB, classified into types I and II residues activated at low and normal temperatures, respectively, and type III residue mediated oligomerization at low temperature but switched to substrate binding at heat shock temperature [23]. In the N-terminal arm of IbpB, all these three types of substrate-binding residues are distributed. Here, we did not choose the type I residues, as they are capable of mediating substrate binding at 30°C, that would make it not convenient to characterize the interaction dynamics of IbpB-substrate from 30°C to 50°C, and back to 30°C. In contrast, four type II residues (Phe-4, Leu-6, Trp-13, and Ala-20) and one type III residue (Phe-32) which primarily mediate self-oligomerization at 30°C and switch to substrate binding at 50°C were subjected to in vivo photo-crosslinking analysis. Our results showed that, unlike F4Bpa, four other Bpa variants (L6Bpa, W13Bpa, A20Bpa, and F32Bpa) displayed the similar pattern in the decrease of photo-crosslinked bound proteins during the recovery stage. That means the rate of substrate proteins release of these IbpB Bpa variants does not depend on the type of the residue with Bpa incorporated, suggesting that the interaction dynamics between IbpB and its substrates is complicated. We speculate that the positions of Bpa incorporation and the properties of bound substrates may be attributed to the difference in the substrate release rate. It seems that some substrates remain bound longer on certain residues of the IbpB N-terminal arm. Since all the Bpa variants have been checked for function (as shown in Fig. 2), the complexity of the interaction between IbpB and its substrates is not due to the effect of Bpa incorporation.
Furthermore, this is the first study to investigate the substrate release of sHSPs in living cells by using photo-crosslinking, which is an effective approach to obtain very dynamic and even short-lived interactions that happen in vivo. Our study is of interest in methodology for investigating the interaction dynamics between molecular chaperones and their substrates in living cells.
E. coli strains and Prof Peter Schultz for the aminoacyl-tRNA synthetase and tRNA expression vectors.