A S52P mutation in the ‘α-crystallin domain’ of Mycobacterium leprae HSP18 reduces its oligomeric size and chaperone function

Authors


Abstract

Mycobacterium leprae HSP18 is a small heat shock protein (sHSP). It is a major immunodominant antigen of M. leprae pathogen. Previously, we have reported the existence of two M. leprae HSP18 variants in various leprotic patients. One of the variants has serine at position 52, whereas the other one has proline at the same position. We have also reported that HSP18 having proline at position 52 (HSP18P52) is a nonameric protein and exhibits chaperone function. However, the structural and functional characterization of wild-type HSP18 having serine at position 52 (HSP18S52) is yet to be explored. Furthermore, the implications of the S52P mutation on the structure and chaperone function of HSP18 are not well understood. Therefore, we cloned and purified these two HSP18 variants. We found that HSP18S52 is also a molecular chaperone and an oligomeric protein. Intrinsic tryptophan fluorescence and far-UV CD measurements revealed that the S52P mutation altered the tertiary and secondary structure of HSP18. This point mutation also reduced the oligomeric assembly and decreased the surface hydrophobicity of HSP18, as revealed by HPLC and 4,4′-dianilino-1,1′-binaphthyl-5,5′-disulfonic acid binding studies, respectively. Mutant protein was less stable against thermal and chemical denaturation and was more susceptible towards tryptic cleavage than wild-type HSP18. HSP18P52 had lower chaperone function and was less effective in protecting thermal killing of Escherichia coli than HSP18S52. Taken together, our data suggest that serine 52 is important for the larger oligomerization and chaperone function of HSP18. Because both variants differ in stability and function, they may have different roles in the survival of M. leprae in infected hosts.

Structured digital abstract

Abbreviations
ACD

α-crystallin domain

bis-ANS

4,4′-dianilino-1,1′-binaphthyl-5,5′-disulfonic acid, dipotassium salt

CFU

colony-forming unit

FRET

Förster resonance energy transfer

Gu-HCl

guanidine hydrochloride

IPTG

isopropyl thio-β-d-galactoside

MDH

malate dehydrogenase

sHSP

small heat shock protein

Introduction

Heat shock proteins (HSPs) are a group of proteins found in all prokaryotic and eukaryotic cells. The term ‘heat shock protein’ evolved from the work carried out by Ritossa [1] in 1962. HSPs act as a molecular chaperone that assists the proper folding of newly synthesized proteins. Moreover, they also play a crucial role in preventing the aggregation of stressed proteins, degrading misfolded proteins and transporting proteins in all prokaryotic and eukaryotic cells [2]. HSPs are classified into two major families: large and small heat shock protein (sHSP). Generally, sHSPs are characterized by a conserved ‘α-crystallin domain’ (ACD) flanked by a highly flexible C-terminal extension and a poorly conserved N-terminal region. Mycobacterium tuberculosis HSP16.3 (Rv2031c), HSP27, wheat HSP16.9, α-crystallin, etc., are classical examples of sHSPs [3, 4]. sHSPs are generally considered to be ATP-independent molecular chaperones [5]. They often exhibit refolding of denatured proteins [6, 7], although they are less efficient than the large HSPs [3]. At elevated temperatures, their chaperone activity enhances significantly [8].

Despite extensive studies, a number of issues relating to the chaperone function of sHSPs still remain unclear. One such issue concerns the role of oligomeric structure on the chaperone function of different sHSPs. Generally, the ‘α-crystallin domain’, which is a conserved structural element within the sHSP family, is the building block of oligomeric structure of different sHSPs [9-11]. Whether the dissociation or association of oligomeric assembly is required for chaperone function of sHSPs is still not understood. For example, the entire N-terminal deletion from the HSP26 sequence produced a smaller oligomeric form (24-mer to dimer) that had less chaperone function than the 24-mer wild-type HSP26 [12]. Similarly, truncation of critical amino acid residues from the N-terminal region of two oligomeric α-HSPs (24-mer HSPH and HSPF) gave rise to a range of smaller oligomeric forms (dimer–octamer) that were devoid of chaperone activity [13]. Leroux et al. [14] also showed that monomeric Caenorhabditis elegans HSP12.6 did not exhibit molecular chaperone function. All of the above evidence revealed that the dissociation of oligomeric assembly resulted in the decrease in chaperone function of sHSPs. On the other hand, the formation of larger oligomeric assemblies in α-crystallin leads to a loss of chaperone function [15-17]. αA- and αB-crystallin usually exists as an 800-kDa complex comprising 32 subunits. Several cataract causing point mutations in the ACD of αA-crystallin (G98R and R116C) and in the ACD of αB-crystallin (R120G) generated a 1.5- to three-fold larger oligomeric assembly with reduced chaperone function [15-17]. In addition, there are several studies claiming that the chaperone function of some sHSPs depends on the dynamics of dissociation of their oligomeric assembly. Dissociation of 24 subunits of human HSP27 into the tetrameric form is essential for binding destabilized T4 lysozyme efficiently [18]. A similar kind of oligomeric dissociation (nonamer to trimer) is essential for the chaperone-like activity of M. tuberculosis HSP16.3 [19]. Ehrnsperger et al. [20] also showed a higher chaperone function for murine HSP25 when its oligomeric size was reduced from hexadecamer to tetramer. Thus, it is evident that the relationship between oligomeric structure/size and chaperone function is different for different sHSPs and that every sHSP has an optimal oligomeric size to exhibit proper chaperone function. Despite the totally mixed viewpoints existing in the literature, it is well established that most sHSPs with the conserved ACD have strong dependence on an oligomeric structure for their function.

One such sHSP that possesses an ‘α-crystallin domain’ belonging to the mycobacterial sHSP family is M. leprae HSP18. It is a class 3 (acr3) sHSP [21] and homologues of sHSPs of the same class are found in Mycobacterium smegmatis, Mycobacterium marinum and Mycobacterium avium [22]. In 1988, Nerland et al. [23] characterized this sHSP. M. leprae HSP18 is a major T-cell antigen with immunodominant epitopes. Previous studies revealed that the T-cell epitopes carrying amino acids 1–38 and 41–55 of M. leprae HSP18 were found to be cross-reactive with the Mtuberculosis complex, Mavium and other mycobacteria [24]. This 18-kDa gene is transcriptionally activated during intracellular growth in macrophages and may be involved in the survival of M. leprae within the macrophages [25].

M. leprae HSP18 isolated from different strains contain serine at position 52 [26]. In our earlier study, we found two M. leprae HSP18 variants in different leprotic patients [27]. In these two HSP18 variants, we observed a point mutation at position 52. Serine at position 52 mutates to proline. Subsequently, we characterized the HSP18 variant bearing proline at position 52 (HSP18P52) [28]. Nine subunits of HSP18P52 usually combine together to form a large oligomer of 173 kDa. We also found that HSP18P52 is a molecular chaperone. However, the functional behaviour of wild-type HSP18 (HSP18S52) is still not understood properly. Despite having less idea about the functionality of wild-type HSP18, this sHSP has been used for the development of second-generation vaccines for leprosy as a carrier protein [29]. Very few attempts have been made to understand the structure of this protein. Mitra et al. [30] identified tryptic digested fragments in M. leprae HSP18, which appeared to be a potent stimulator of CD4+ T-cell responses in normal and leprotic patients. Costa et al. [29] showed that the native conformation of wild-type HSP18 was preserved after hydrophobic modification by acylation. We also made an attempt to understand the structure of wild-type and mutant HSP18 and theoretically demonstrated that mutant HSP18 variant (HSP18P52) has a different secondary structure compared to that of the wild-type HSP18 (HSP18S52) [31]. However, the effect of this point mutation (S52P mutation) on the tertiary, quaternary structure and chaperone function of HSP18 remains to be investigated. In the present study, we report the complete characterization of wild-type HSP18 (HSP18S52) and the impact of this point mutation at position 52 of HSP18 on its structure and chaperone function using different biophysical and biochemical techniques.

Results and Discussion

Many antigens have crucial implications in the immune response of M. leprae pathogen [32-34]. Several of these antigens, with molecular weights in the range 10–70 kDa, belong to the HSP family [35]. Many of these HSPs contribute to virulence and persistence of M. leprae pathogen. For example, 10-kDa sHSP of M. leprae is a major T-cell antigen and may have ‘immunoprophylactic’ potential for the disease leprosy [36]. HSP65 is another such example of the major immune targets in leprosy and tuberculosis. Singh et al. [37] reported evidence for the existence of molecular mimicry between M. leprae HSP65 and cytokeratin-10 of keratin (host protein) and established the role of this HSP in the pathogenesis of leprosy by clinical manifestation. 18-kDa antigen (HSP18) of M. leprae also plays an important role in the survival of M. leprae pathogen in host cells. Pessolani et al. [38] demonstrated that the expression of this antigen might be regulated by iron. It was also shown that the 18-kDa protein formed a high iron content native complex of 380 kDa, which could act as an iron depository moiety during iron deficient conditions inside host cells and might enable the survival of M. leprae under iron-deficient conditions. A recent study by Maheshwari and Dharmalingam [39] revealed that the expression of M. leprae HSP18 could be induced under conditions of hypoxia, nutrient depletion and oxidative stress. The same study also showed that M. leprae HSP18 might facilitate the survival of M. leprae under various stress conditions and that autophosphorylation of this protein may be the probable mechanism for such survival.

The sequence of M. leprae HSP18 has been divided into three distinct regions: N-terminal region comprising residues 1–38, ‘α-crystallin domain’ comprising residues 39–121 and a flexible C-terminal tail with residues 122–148. Sequence homology among αA-crystallin, αB-crystallin, Methanococcus jannaschii HSP16.5, wheat HSP16.9, M. tuberculosis HSP16.3 (Rv2031c), M. tuberculosis acr2 (Rv0251c) and M. leprae HSP18 clearly revealed that M. leprae HSP18 contains a highly conserved ‘α-crystallin domain’ (Fig. 1). It is generally assumed that the ACD of various sHSPs binds the stressed prone client proteins, which helps them to exhibit proper chaperone function [40, 41]. The three-dimensional structure of this domain has been determined by X-ray crystallography for many sHSPs, clarifying the structural orientation of ACD and also helping to understand the role of ACD with respect to the oligomeric structure and chaperone function of these sHSPs [42-45]. Several point mutations in the ACD of well known sHSPs perturbed their oligomeric structure and chaperone function [16, 17, 46-48]. We also observed a point mutation (S52P) in the ACD of M. leprae HSP18 [27]. We hypothesized that the S52P mutation in ACD of HSP18 also has an impact on the structure and chaperone function of this sHSP.

Figure 1.

Sequence alignment of seven sHSPs. The amino acid sequence alignment between the seven sHSPs, αA-crystallin (AlphaA), αB-crystallin (AlphaB), M. jannaschii HSP16.5 (Hsp16.5), wheat HSP16.9 (Hsp16.9), M. tuberculosis Rv2031c, HSP16.3 (Hsp16.3), M. tuberculosis Rv0251c, acr2 (MTB-Acr2) and M. leprae HSP18 (Hsp18) was performed using multiple sequence alignment software (t-coffee; http://www.ebi.ac.uk/Tools/msa/tcoffee/). All seven sHSPs have a conserved ‘α-crystallin domain’ (highlighted in turquoise).

To confirm our hypothesis, we prepared wild-type (HSP18S52) and mutant (HSP18P52) HSP18 proteins. We cloned and expressed hexa-His-tagged HSP18S52 and HSP18P52 in E. coli M15 cells and then purified these proteins. After the proteins passed through a Ni2+ column, we checked the purity of these two proteins by SDS/PAGE (Fig. S1). Similar to our previous study [28], this protein purification procedure yielded HSP18P52 with a double band, whereas we observed only a single band for HSP18S52. We already identified the exact molecular weight of the major (19.3 kDa) and minor band (16.7 kDa) of HSP18P52 by MS [28]. In the present study, we determined the molecular weight of HSP18S52 by ESI-MS. The expected molecular weight of HSP18S52 is 19313.8 Da including the amino acids added during construction. ESI-MS revealed that the molecular weight for HSP18S52 is 19315.4 Da, which further confirms the presence of serine at position 52 in HSP18S52.

To characterize the functionality of HSP18S52 and to assess whether the S52P mutation altered the chaperone function of M. leprae HSP18, the chaperone activity of two HSP18 variants was explored using two different client proteins. Dithiothreitol-induced insulin aggregation profiles in the absence or presence of HSP18 variants are shown in Fig. 2A. At a ratio of 1 : 1 (w/w) of HSP18S52 to insulin, 35% protection against aggregation was observed (Fig. 2A, trace 3, and Fig. 2C). By contrast, at the same chaperone to client protein ratio, the chaperone function of HSP18P52 was found to be reduced by 10% (Fig. 2A, trace 2, and Fig. 2C). When βL-crystallin was used as the client protein, a more dramatic effect was observed. We noted that HSP18P52 inhibited protein aggregation marginally (5%) at a chaperone : substrate ratio of 1 : 1 (w/w) (Fig. 2B, trace 2, and Fig. 2D). However, at the same ratio, HSP18S52 inhibited protein aggregation by 38% more than HSP18P52 (Fig. 2B, trace 3, and Fig. 2D). To test whether these two HSP18 variants aggregate at 60 °C, aggregation assays were also performed at this temperature with these sHSPs alone as a control experiment. It was observed that both protein variants showed no aggregation, which revealed their stability at higher temperature (Fig. 2B, traces 4 and 5). The chaperone function of two HSP18 variants was also compared by varying the ratio between the chaperone and client proteins. In each case, HSP18S52 was found to be superior to HSP18P52 in preventing the aggregation of client proteins (Fig. S2). Taken together, our findings indicate that wild-type HSP18 is a molecular chaperone. Similar to other sHSPs, it can prevent the aggregation of chemically-induced and heat-induced client proteins. In addition, our data indicate that S52P point mutation reduced the chaperone function of M. leprae HSP18.

Figure 2.

Chaperone function of HSP18 proteins. (A) dithiothreitol-induced aggregation of 0.35 mg·mL−1 insulin at 25 °C. Traces: 1, client protein (CP) alone; 2, CP + HSP18P52; 3, CP + HSP18S52. (B) Thermal aggregation of 0.15 mg·mL−1 βL-crystallin at 60 °C in the absence and presence of different HSP18 proteins. Traces: 1, client protein (CP) alone; 2, CP + HSP18P52; 3, CP + HSP18S52; 4, HSP18P52 alone; 5, HSP18S52 alone. Percentage protection by different HSP18 proteins against insulin aggregation (C) and βL-crystallin aggregation (D). The chaperone : client protein ratio (w/w) was 1 : 1 for both aggregation assays. Data are the mean ± SD from triplicate determinations.

GroEL and DnaK are the most well characterized large HSPs in terms of refolding of client proteins [7, 49]. sHSPs also exhibit this property [5-7], although the mechanism of refolding of unfolded client proteins by this class of proteins is still unclear. Information about the refolding ability of HSP18 has not been reported in the literature. Therefore, to obtain an idea about the refolding ability of HSP18, we studied the ability of two HSP18 variants (HSP18S52 and HSP18P52) to reactivate fully denatured malate dehydrogenase (MDH). When fully denatured MDH was allowed to refold by diluting in appropriate refolding buffer in absence of HSP18, maximum activity recovered was only approximately 5% (Fig. 3, trace 1). In the presence of 20 μm HSP18S52, approximately 26% activity of MDH could be regained (Fig. 3, trace 4). At the same protein concentration, the recovery of enzyme activity with HSP18P52 was reduced by approximately 10% (Fig. 3, trace 3). To test the specificity of two HSP18 variants, the renaturation of MDH was also performed in presence of BSA (20 μm). It was found that, unlike the two HSP18 variants, BSA failed to reactivate the chemically denatured substrate enzyme (Fig. 3, trace 2). Thus, the reactivation assay revealed that both variants of HSP18 (HSP18S52 and HSP18P52) can assist the refolding of denatured enzymes and also that the S52P mutation decreased the refolding ability of HSP18.

Figure 3.

Reactivation of MDH by HSP18. The enzyme activity of MDH was destroyed by incubation in 6 m Gu-HCl solution at 25 °C. Trace 1, 10 nm MDH alone; Trace 2, 10 nm MDH + 20 μm BSA; Trace 3, 10 nm MDH + 20 μm HSP18P52; Trace 4, 10 nm MDH + 20 μm HSP18S52; refolding was initiated by 100-fold dilution of MDH (1 μm) in 6 m Gu-HCl in refolding buffer (pH 7.5) containing 50 mm phosphate, 10 mm magnesium acetate, and 5 mm dithiothreitol. Each data point is the mean of triplicate measurements and error bars indicate the SD.

The M. leprae HSP18 gene is transcriptionally activated during intracellular growth in macrophages [25]. RT-PCR studies clearly revealed that in vivo over-expression of this protein plays a key role in M. leprae infection [50]. The results of several other studies also suggested that over-expression of sHSPs has a protective effect on the survival of E. coli against thermal stress [51]. To identify differences between the protective effect of over-expression of HSP18S52 and HSP18P52 in vivo, a similar assay was performed. The number of colony-forming units (CFUs) was counted in cultures of E. coli in the absence and presence of HSP18 variants after heat shock at 48 °C. The proportion of viable cells that survived after heat shock was plotted at different post heat shock time points (Fig. 4A). After 6 h of heat shock at 48 °C, the number of surviving cells that contained empty control pQE31 vector or uninduced cells that contained HSP18S52 or HSP18P52 was negligible. The viability of cells after 6 h of bacterial cell cultures containing over-expressed HSP18S52 decreased by three orders of magnitude only. On the other hand, the viability of the cells that over-expressed HSP18P52 was decreased further by five orders of magnitude. Protein degradation, if any, resulting from the heat shock was also analyzed by SDS/PAGE (Fig. 4B,C). The profiles revealed no protein degradation and, instead, showed the intense expression of both proteins at 0, 3 and 6 h after heat shock at 48 °C. The protective effect of HSP18S52 on cell survival was stronger than HSP18P52. Both an in vivo cell viability experiment and in vitro aggregation and refolding assays revealed that mutation of serine 52 in the ACD markedly reduces the protection ability/chaperone function of HSP18. Based on these findings, we propose that the presence of serine at position 52 is important for enabling the proper chaperone functionality of this sHSP.

Figure 4.

Cell viability of two HSP18 variants. pQE31-HSP18S52, pQE31-HSP18P52 and pQE31 empty vector containing no inserted gene (empty control vector) were expressed at 37 °C and induced with 0.4 mm IPTG when A600 of 0.8 was reached for cell cultures. After induction for 2 h and heat shock to 48 °C, the cells were incubated for a further 6 h. Samples were taken at different time points beginning at the time of heat shock, plated and the CFU was recorded. The proportion of viable cells containing pQE31-HSP18S52, pQE31-HSP18P52 and pQE31 empty control vector were plotted at different post heat shock time points (A). Protein expression of HSP18S52 (B) and HSP18P52 (C) was analyzed by 12% SDS/PAGE. Lanes 1–3 are the empty control pQE31 vector at 0-, 3- and 6-h time points after 2 h of induction and post heat shock at 48 °C. Lane 4 is the protein expression of cells containing the pQE31 vector alone, not induced with IPTG. Lane 5 is the protein marker. Lane 6 is the protein expression of cells containing the pQE31-HSP18S52/HSP18P52 vector, not induced with IPTG. Lanes 7–9 are the protein expression of pQE31-HSP18S52/HSP18P52 vector at 0-, 3- and 6-h time points after 2 h of induction and post heat shock. The experiment was repeated at least five times using triplicates of each cell culture.

To understand the molecular basis behind the decrease in the chaperone function of HSP18P52, the structural changes in the protein were determined. sHSPs have many hydrophobic patches at the surface and these hydrophobic patches generally bind aggregation prone client proteins through hydrophobic interactions. Numerous independent studies have quantified the hydrophobic patches at the surface of these proteins with the aid of different hydrophobic probes such as 4,4′-dianilino-1,1′-binaphthyl-5,5′-disulfonic acid, dipotassium salt (bis-ANS), 2-(p-toluidinyl)naphthalene-6-sulfonic acid, sodium salt, etc. [6, 40, 52, 53]. Several studies suggested a strong correlation between chaperone function and surface-exposed hydrophobicity in sHSPs [6, 53], although a others have failed to find such a correlation [54, 55]. Notwithstanding this controversy, it is generally considered that surface hydrophobicity governs the chaperone function of sHSPs. We explored the relationship between decreased chaperone activity in HSP18P52 and hydrophobic patches on its surface using a hydrophobic probe, bis-ANS. The refolding and survival of E. coli against thermal stress and aggregation assays were carried out at 25, 48 and 60 °C, respectively. Therefore, we measured the surface hydrophobicity at these temperatures. As shown in Fig. 5, the fluorescence intensity of bis-ANS bound to the mutant protein was 18–33% lower compared to that of wild-type protein. These results suggest that the S52P mutation led to a decrease in the hydrophobicity of HSP18, which reduced the chaperone activity of this protein.

Figure 5.

Effect of S52P mutation on the surface hydrophobicity of HSP18. The surface hydrophobicity of HSP18S52 and HSP18P52 was estimated using a hydrophobic probe: bis-ANS at (A) 25 °C; (B) 48 °C and (C) 60 °C. The protein concentration was 0.05 mg·mL−1 and the bis-ANS concentration was 10 m. The fluorescence spectrum of bis-ANS bound to different samples was recorded in the range 450–550 nm at respective temperatures (25, 48 and 60 °C). The excitation wavelength was 390 nm.

Numerous studies have reported that N-terminal as well as early part of ‘α-crystallin domain’ of sHSPs is responsible for binding of client proteins and bis-ANS [40, 41]. If any tryptophan moiety is in close proximity to bis-ANS binding sites in a protein, Förster resonance energy transfer (FRET) occurs between the tryptophan residue (donor molecule) and the bound bis-ANS (acceptor molecule) [56]. The location of bis-ANS binding for HSP18 has not been reported in the literature. Therefore, FRET studies were performed to obtain knowledge of the bis-ANS binding domain for HSP18. When HSP18S52 and HSP18P52 were excited at 295 nm, a considerable amount of FRET was observed between the single tryptophan residue (W33) and bound bis-ANS in both proteins (Fig. 6), which suggests that bis-ANS binding sites are located in close proximity to the tryptophan containing domain (i.e. N-terminal domain of HSP18). At any bis-ANS concentration, the ratio of tryptophan fluorescence to bound bis-ANS fluorescence at the respective emission maxima was always lower for the wild-type protein than for its S52P mutant, which suggests that the probability of FRET is higher for wild-type protein than that of its mutant. We can also conclude that the S52P mutation altered the mutual orientation, overall flexibility and/or distance between W33 and bis-ANS bound to HSP18.

Figure 6.

FRET between tryptophan (W) residue and bound bis-ANS. FRET between the single tryptophan residue (W33) of HSP18S52 (A) or HSP18P52 (B) and bound bis-ANS was observed with excitation of tryptophan at 295 nm and fluorescence emission in the range 310–550 nm was measured. The protein concentration was 3 μm and the added bis-ANS concentrations were 0 (trace 1), 2 (trace 2), 5 (trace 3) and 10 μm (trace 4).

Tryptophan (W) fluorescence along with near- and far-UV CD techniques were used to determine whether the altered chaperone function and surface hyrophobicity were accompanied by any change in tertiary and secondary structure. Perturbation in the secondary structure of HSP18 as a result of the S52P mutation was indicated by the far-UV CD spectra (Fig. 7). Quantitative analysis of the far-UV CD data using continll software showed that HSP18S52 is a major β-sheet protein (Table 1). The content of β-sheet decreased slightly (5.8%) with a concomitant increase in random coil structure (8.6%) in HSP18P52. These data further confirm that the S52P mutation perturbed the secondary structure of HSP18. Previously, our group theoretically predicted such an alteration in the secondary structure of HSP18 as a result of the S52P mutation [31].

Table 1. Percentage level of different secondary structural elements in HSP18S52 and HSP18P52 using continll software
Proteinsα-helix (%)β-sheet (%)β-turn (%)Random coil (%)
HSP18S523.635.719.640.7
HSP18P522.629.917.849.3
Figure 7.

Far-UV CD spectra of two HSP18 variants. Spectra were recorded for 0.2 mg·mL−1 protein (in 10 mm phosphate buffer, pH 7.5) using a path length cell of 1 mm. The data interval was 1 nm.

The tryptophan fluorescence of the mutant protein was 20% higher than that of the wild-type protein (Fig. 8), which implied that the S52P mutation perturbed the microenvironment of the tryptophan (W33) residue. An alteration in λmax of the intrinsic tryptophan fluorescence spectra was also observed. A red shift of 6.5 nm further confirmed that HSP18P52 is in a more unfolded state than HSP18S52. We speculate that the N-terminus is more relaxed in the mutant protein, which results in the exposure of W33 to a more polar environment. Thus, HSP18P52 exhibits greater tryptophan fluorescence. A previous study showed similar changes in tryptophan fluorescence by subtle modifications in the ACD of one sHSP [57]. The near-UV CD spectra of these two proteins agreed with our intrinsic tryptophan fluorescence data (data not shown).

Figure 8.

Intrinsic tryptophan fluorescence spectra of HSP18S52 and HSP18P52. Tryptophan fluorescence spectra of different proteins (0.05 mg·mL−1) were recorded in the range 310–400 nm at 25 °C. The excitation wavelength was 295 nm. Excitation and emission slit widths were 5 nm each. Data were recorded at a wavelength resolution of 0.5 nm.

To determine whether the S52P mutation affected the oligomeric structure (i.e. quaternary structure) of the protein, gel filtration chromatography was performed using a TSK-GEL G4000SWXL column (Tosoh Bioscience LLC, King of Prussia, PA, USA). A single peak at an elution volume of 8.74 mL (Fig. 9A, peak 1) was observed for HSP18S52, whereas, for HSP18P52, this shifted slightly and appeared at an elution volume of 9.93 mL (Fig. 9B, peak 2). To calculate the oligomeric mass of these two proteins, a standard curve [log (molecular weight) versus elution volume] was generated with molecular weight standards: thyroglobulin (669 kDa), apoferritin (443 kDa), β-amylase (200 kDa), alcohol dehydrogenase (150 kDa) and ovalbumin (45 kDa) (Fig. 9A, inset). With the aid of this standard curve, the oligomeric mass of HSP18S52 was found to be 565 kDa. By contrast, the oligomeric mass of HSP18P52 was 177 kDa. This is in agreement with our previous study where we reported that the oligomeric mass of HSP18P52 was 173 kDa [28]. Taken together, these data suggest that subtle changes in the secondary and tertiary structure caused by the S52P mutation in the ACD of HSP18 led to a decrease in the mean oligomeric size, consisting of approximately 29 subunits in the wild-type protein to approximately nine subunits in the mutant protein. This reduction in the oligomeric assembly of HSP18 as a result of the S52P mutation may be the basis for its decreased chaperone function. Previous studies suggested that point mutations in the ACD of sHSPs are often associated with a decrease in oligomeric size and chaperone function [48, 58-60]. Overall, our results demonstrate a strong correlation between oligomeric size and chaperone function of HSP18.

Figure 9.

Estimation of oligomeric mass of recombinant HSP18 proteins. Gel-filtration profile of HSP18S52 (A) and HSP18P52 (B). TSK-GEL G4000SWXL column (7.8 mm × 30 cm; 5 μm) was first equilibrated with 50 mm phosphate buffer (pH 7.5). Subsequently, 20 μL HSP18S52 or HSP18P52 (0.5 mg·mL−1) were injected into the column. The flow rate of the column was maintained at 0.5 ml·min−1. The oligomeric mass of HSP18 proteins was estimated with the aid of a standard curve (A, inset).

Apart from oligomerization, structural stability is also required for the chaperone function of sHSPs [6, 52]. To determine the alteration in structural stability caused by the S52P mutation, the thermodynamic stability of HSP18S52 and HSP18P52 protein was compared. Equilibrium urea unfolding was measured by monitoring tryptophan fluorescence of the proteins at various urea concentrations. λmax values for native and unfolded HSP18S52 were recorded at 340 and 355 nm, whereas, for HSP18P52, λmax values for those two states were recorded at 346.5 and 355 nm, respectively. We plotted the ratio of intensities at native and unfolded states as a function of urea concentration (Fig. 10). An approximate estimate of the transition midpoint (C1/2) from the sigmoidal analysis of the denaturation profiles indicated that the C1/2 value decreased drastically from 2.61 m of urea for HSP18S52 to 0.88 m of urea for HSP18P52 (Fig. 10 and Table 2). The decrease in transition midpoint (C1/2) indicated that the presence of proline at position 52 destabilized the overall structural integrity of HSP18. To quantify the stability against chemical denaturation, all of the profiles were analyzed with the aid of a global three state fitting procedure, in according with the equation:

display math(1)
Table 2. C1/2 and ΔG0 values of HSP18S52 and HSP18P52 at 25 °C
ProteinsC1/2 (m)ΔG0 (kJ·mol−1)
HSP18S522.61 ± 0.1220.91 ± 0.48
HSP18P520.88 ± 0.0612.20 ± 1.24
Figure 10.

Thermodynamic stability of HSP18S52 and HSP18P52. Equilibrium urea unfolding profile for 0.05 mg·mL−1 of HSP18S52 (●) and HSP18P52 (○) at 25 °C. The profile has been normalized to a scale of 0–1. Symbols represent the experimental data points and the solid lines represent the best fit according to the three state model described by Eqn. (1).

where FN, FI and FU are the fluorescence intensities for 100% native, 100% intermediate and 100% unfolded form, respectively. ∆G10 refers to the standard free energy change between the native and intermediate form and ∆G20 refers to the standard free energy change between the intermediate and unfolded form. ∆G0, being the sum of ∆G10 and ∆G20, refers to the standard free energy change of unfolding (between native and unfolded form) at zero urea concentration. The fitted parameters are listed in Table 2. The standard free energy change of HSP18S52 unfolding at 25 °C is 20.91 kJ·mol−1. This value of ΔG0 compares well with a value of 21–22 kJ·mol−1 reported by Nagaraj et al. [52] for the unfolding of three sHSPs (HSP27, αA- and αB-crystallin) at 25 °C at an infinite dilution of urea. Sun et al. [61] also reported that the ΔG0 value of guanidine hydrochloride (Gu-HCl) induced unfolding of αA-crystallin at 25 °C is 24 kJ·mol−1, which is also quite comparable with the value shown in Table 2 for HSP18S52. The ΔG0 value for HSP18P52 reduced to 12.20 kJ·mol−1, indicating a decrease in thermodynamic stability (ΔΔG0) by approximately 8.71 kJ·mol−1. Several studies found that the increased chaperone function of α-crystallin is often associated with the greater structural stability of this protein [6, 53]. Therefore, we can conclude that the reduction in the overall structural stability causes a decrease in chaperone function of HSP18P52.

We then compared the stability of HSP18S52 and HSP18P52 protein against thermal stress by far-UV CD measurement. CD measurement was performed at 217 nm (characteristic of the β-sheet secondary structure of the protein) over the temperature range 25–80 °C and then the fraction unfolded (αU) for both HSP18 proteins was calculated using the equation:

display math(2)

where θF is the ellipticity value at 25 °C for completely folded or native protein, θt is the observed ellipticity value at any temperature and θU is the ellipticity value at 80 °C for the completely denatured or unfolded state. We plotted fraction unfolded (αU) as a function of temperature (Fig. 11A). The thermal denaturation profiles of both HSP18 proteins show a sigmoidal shape. Sigmoidal analysis of these profiles (Fig. 11A, solid lines) and the plot of dαU/dt versus temperature (t) (Fig. 11B) demonstrated that HSP18S52 undergoes thermal unfolding with a mid-point transition or melting temperature (Tm) of 61 °C (Table 3). Interestingly, the Tm value obtained from the thermal denaturation profiles of other sHSPs (HSP22 and M. tuberculosis HSP16.3) also lies between 57 and 60 °C [62, 63]. FT-IR spectroscopy, CD and differential scanning calorimetry studies revealed that α-crystallin had a phase change transition temperature between 60 and 62 °C [64]. Maulucci et al. [65] also determined the phase transition temperature of the quaternary structure of this sHSP using static and dynamic light scattering measurements and found that it had a quaternary structural transition at 45 °C [65]. The S52P mutation shifted the Tm value as obtained from the temperature dependence curve of fraction unfolded (αU) from 61 to 54 °C (Fig. 11A and Table 3). The decrease in the mid-point transition or Tm value of HSP18P52 under thermal stress by approximately 8 °C clearly suggests that the S52P mutation significantly reduced the thermal stability of M. leprae HSP18. Because HSP18S52 and HSP18P52 are oligomeric proteins (Fig. 9), we calculated the change in enthalpy (H) of this thermal transition using the van't Hoff equation:

display math(3)
Table 3. Mid-point transition or Tm and van't Hoff enthalpy (ΔHvH) values associated with thermal denaturation of HSP18S52 and HSP18P52
ProteinsTm (°C)ΔHvH (kJ·mol−1)
HSP18S5261.4 ± 0.2126.37 ± 1.92
HSP18P5253.8 ± 0.3106.42 ± 1.39
Figure 11.

Structural stability of HSP18S52 and HSP18P52 against thermal stress. (A) Temperature induced changes in the fraction of unfolded state (αU) for HSP18S52 and HSP18P52 proteins. The profile has been normalized to a scale of 0–1. Symbols represent the experimental data points and the solid lines represent the best fit according to the sigmoidal analysis. (B) The first derivatives of the curves shown in (A). The maxima correspond to the mid-point transition or Tm for the thermal denaturation of both HSP18 variants. (C) van't Hoff plots for the data shown in (A). The values of van't Hoff enthalpy (ΔHvH) estimated for these transitions of HSP18S52 and HSP18P52 were equal to 126 and 106 kJ·mol−1, respectively.

where Keq(folded→unfolded) = αU/(1 − αU) and temperature in Kelvin. The value of van't Hoff enthalpy (ΔHvH) associated with the thermal transition of HSP18S52 was approximately 126 kJ·mol−1, which decreased to 106 kJ·mol−1 for HSP18P52 (Fig. 11C and Table 3). Both chemical and thermal denaturation experiments revealed that the S52P mutation in the ACD significantly reduces the structural stability of HSP18. We can also state that the presence of serine at position 52 is important for maintaining the structural stability of M. leprae HSP18 under various stress conditions.

We have also checked whether the S52P mutation changes the digestibility of HSP18 by trypsin. Figure 12 shows the SDS/PAGE profile of the tryptic digestion products of wild-type and mutant HSP18 at different digestion times using a 1 : 50 (w/w) ratio of trypsin to chaperone. After digestion for 45 min, it is clear that the S52P mutation has considerably aggravated the cleavage. After tryptic digestion for 60 min, trypsin completely digested HSP18P52, although several bands were distinctly visible in HSP18S52, indicating that the S52P mutation makes HSP18 more prone to tryptic digestion. We consider that decreased structural stability made the mutant protein more susceptible towards tryptic cleavage.

Figure 12.

SDS/PAGE profile of the trypsin digest of HSP18S52 and HSP18P52. Wild-type and mutant HSP18 (0.5 mg·mL−1 in 50 mm phosphate buffer, pH 7.5) was digested with trypsin for different time intervals at 37 °C. The trypsin to chaperone ratio was 1 : 50 (w/w).

In summary, the findings of the present study revealed that wild-type M. leprae HSP18 (HSP18S52) is a molecular chaperone and exists in larger oligomeric form. This protein exhibits several properties, such as refolding ability, aggregation prevention ability, etc., that are associated with other sHSPs. Direct implications of chaperone function of acr1 mycobacterial sHSP (M. tuberculosis HSP16.3) on the survival and virulence of M. tuberculosis pathogen are well documented in the literature [66, 67]. Because HSP18S52 prevents the aggregation of various stressed protein and helps E. coli to survive under thermal stress, the chaperone ability of HSP18S52 may play an important role in the survival of M. leprae pathogen in the host. We also found that the S52P mutation made HSP18 extremely unstable and more prone to tryptic cleavage. Whether this mutation has any impact on the host–pathogen interaction remains to be explored. Because both variants of HSP18 are over-expressed in vivo during infection and the presence of proline residue at position 52 perturbed the functionality of M. leprae HSP18, the utilization of HSP18 in drug targeting strategies remains a fundamental challenge. However, this point mutation in the ‘α-crystallin domain’ of HSP18 reduced the oligomeric assembly and decreased the chaperone activity of this protein. Because HSP18P52 is an extremely unstable protein, this protein may have different roles with respect to the survival of M. leprae pathogen at different stages of leprosy.

Materials and methods

Materials

Dithiothreitol, bovine insulin, bis-ANS, Gu-HCl, urea and isopropyl thio-β-d-galactoside (IPTG) were obtained from Sigma Chemical Co. (St Louis, MO, USA). Amicon Ultra-15 centrifugal filters (10-kDa cut-off) were obtained from Millipore (Bedford, MA, USA). MDH, oxaloacetic acid, trypsin, NADH and all buffer salts (Tris, phosphate, etc.) were obtained from Sisco Research Laboratories (Mumbai, India). Bovine βL-crystallin was purified from bovine eye lenses, as described previously [68]. All other chemicals were of analytical grade.

Cloning, expression and purification of two HSP18 variants

We cloned and expressed wild-type and mutant HSP18 (HSP18S52 and HSP18P52) in E. coli M15 cells as described previously [27, 28]. Wild-type and mutant cDNA sequences were confirmed by restriction digestion and DNA sequence analysis. Proteins were over-expressed with 0.4 mm IPTG and purified. The bacterial pellet obtained after 2 h of induction was centrifuged at 10 000 g and suspended in lysis buffer (50 mm Tris, 500 mm NaCl, 10 mm imidazole, pH 8.0). Afterwards, sonication was carried out on ice via six duty cycles at 38% amplitude. The debris was removed from the solution by centrifugation at 10 000 g and 4 °C for 60 min. The supernatant was mixed with an appropriate amount of nickel-nitrilotriacetic acid resin in accordance with the manufacturer's instructions (Qiagen,Valencia, CA, USA) and gently mixed for 60 min at room temperature before loading into an empty column. The column was then washed with buffer (pH 8.0) containing 50 mm Tris, 500 mm NaCl, 10 mm imidazole and eluted with 50 mm Tris, 500 mm NaCl, 100–500 mm imidazole (pH 8.0). The purity of proteins in each fraction was assessed by 12% SDS/PAGE. The fractions that contain > 95% pure proteins were pooled and dialyzed extensively at 4 °C against 50 mm phosphate buffer (pH 7.5). The pooled protein was then concentrated using Amicon Ultra-15 centrifugal filters (Millipore) and stored at −20 °C. Concentration of two HSP18 variants was determined spectrophotometrically by measuring A278 using an extinction coefficient of 0.4 mg−1·mL·cm−1 for both proteins. The concentration of these two proteins was also determined using the Bradford assay. Because we reported the molecular weight of HSP18P52 earlier [28], in the present study, we only determined the molecular weight of HSP18S52 by ESI-MS. All of the biophysical assays were performed with at least three independent protein preparations.

Intrinsic tryptophan fluorescence measurements

The intrinsic tryptophan fluorescence spectra of proteins (0.05 mg·mL−1) in 50 mm phosphate buffer (pH 7.5) at 25 °C were recorded using a Fluoromax 4P Spectrofluorometer (Horiba Jobin Mayer, Edison, NJ, USA). The excitation wavelength was set to 295 nm and the emission spectra were recorded between 310 and 400 nm. Data were recorded at a wavelength resolution of 0.5 nm.

CD measurements

Far-UV CD spectra were measured at 25 °C using a Jasco 810 spectropolarimeter (Jasco, Inc., Tokyo, Japan). Spectra were recorded in the range 195–250 nm using a cylindrical quartz cell with a path length of 1 mm. Proteins (0.2 mg·mL−1) were dissolved in 10 mm phosphate buffer (pH 7.5). The reported spectra were the mean of five scans. Spectra were analyzed for secondary structure content using the curve-fitting software continll [52].

The near-UV CD spectra were measured at 25 °C using the same spectropolarimeter. The spectra were measured with a 1.0 mg·mL−1 protein solution in 50 mm phosphate buffer (pH 7.5). The reported spectra were the mean of five scans.

Determination of oligomeric mass by gel filtration chromatography

The oligomeric mass of HSP18S52 and HSP18P52 was determined by gel filtration chromatography in a HPLC (Waters Corp, Milford, MA, USA) instrument using a TSK-GEL G4000SWXL analytical gel filtration column (7.8 mm × 30 cm; 5 μm). After the column was equilibrated with 50 mm phosphate buffer (pH 7.5), 20 μL of HSP18S52 or HSP18P52 (0.5 mg·mL−1 in the equilibrium buffer) was injected into the column. The chromatogram was monitored by measuring A280. The column was calibrated with molecular weight standards: thyroglobulin (669 kDa), apoferritin (443 kDa), β-amylase (200 kDa), alcohol dehydrogenase (150 kDa) and ovalbumin (45 kDa) (Sigma Chemical Co.). The flow rate was maintained at 0.5 mL·min−1 for all measurements.

Determination of surface hydrophobicity of two HSP18 variants

The surface hydrophobicity of the HSP18S52 and HSP18P52 was measured with bis-ANS, a specific hydrophobic probe [6]. bis-ANS (10 μm) was added to HSP18S52 or HSP18P52 [0.05 mg·mL−1 in 50 mm phosphate buffer (pH 7.5)] and the mixture was incubated for 60 min at a variety of temperatures (25, 48 and 60 °C). Fluorescence emission spectra were recorded in the range 450–550 nm at the respective temperatures using an excitation wavelength of 390 nm. The excitation and emission band-passes were 2.5 and 5 nm, respectively. Data were recorded at a wavelength resolution of 0.5 nm.

Determination of structural perturbation by FRET studies

Perturbation in tertiary structure and available hydrophobic patches at the surface of HSP18 protein was assessed further by FRET studies. The FRET was measured between the single tryptophan residue (W33) of HSP18 proteins as an energy donor and bound bis-ANS as an acceptor. A wavelength of 295 nm was used to excite the tryptophan residue and emission spectra were recorded in the range 310–550 nm. The excitation and emission band-passes were 5 nm each. Data were recorded at a wavelength resolution of 0.5 nm. The concentration of HSP18S52 or HSP18P52 was 3 μm and the bis-ANS concentration was varied in the range 0–10 μm.

Estimation of the structural stability of two HSP18 variants

Trypsin digestion experiment

The effect of point mutation on the structural compactness of HSP18 was measured by comparing the trypsin digestibility of HSP18S52 and HSP18P52 [6, 53]. Wild-type and mutant HSP18 [0.5 mg·mL−1 in 50 mm phosphate buffer (pH 7.5)] were incubated with trypsin (at a ratio of 50 : 1, w/w) at 37 °C. Aliquots were withdrawn after different periods of digestion and the reaction was stopped immediately by adding soybean trypsin inhibitor. SDS/PAGE of the digested proteins was performed under reducing conditions in a Mini-PROTEAN 3 electrophoresis set-up (Bio-Rad, Hercules, CA, USA) using a linear 8–16% gradient polyacrylamide gel. Gels were scanned in a densitometer for quantitative analysis.

Thermal denaturation experiment

The structural stability of HSP18S52 and HSP18P52 was also determined using thermal induced unfolding experiments in Chirascan CD Spectropolarimeter (Applied Photophysics, Leatherhead, UK) equipped with a peltier system. The change in ellipticity at 217 nm was recorded stepwise between 25 and 80 °C in a quartz cell with a path length of 1 mm, allowing the samples to equilibrate at each temperature. Heating rate was set to 0.5 °C·min−1. Data were recorded at an interval of 2 °C. The protein concentration was 0.25 mg·mL−1 in 50 mm phosphate buffer (pH 7.5). Mid-point transition or Tm was calculated using sigmoidal analysis and a first-derivative plot [d(fraction unfolded)/dt versus temperature] as described previously [69] and the change in enthalpy associated with the thermal transition was calculated using the van't Hoff equation.

Chemical denaturation experiment

The structural stability of HSP18S52 and HSP18P52 was determined by an equilibrium chemical denaturation experiment. Both proteins (0.05 mg·mL−1 in 50 mm phosphate buffer, pH 7.5) were incubated separately with various urea concentrations (0–7 m) for 18 h at 25 °C. The intrinsic tryptophan fluorescence spectra of all samples were recorded in the range 310–400 nm using an excitation wavelength of 295 nm. The equilibrium unfolding profile was fit according to the three-state model as described previously [6, 52].

In vitro aggregation assays

The chaperone activity was determined with two client proteins: insulin and βL-crystallin as described previously [6, 68]. Both assays were performed with the aid of UV spectrophotometer (Perkin Elmer, Boston, MA, USA).

Insulin aggregation assay

Insulin (0.35 mg·mL−1) in 50 mm phosphate buffer (pH 7.5) was incubated in the absence and presence of 0.35 mg·mL−1 HSP18S52 and HSP18P52. Insulin aggregation was initiated by adding freshly prepared dithiothreitol to a final concentration of 20 mm and light scattering at 400 nm was monitored for 1 h in the kinetic mode at 25 °C.

βL-crystallin aggregation assay

βL-crystallin (0.15 mg·mL−1) in 50 mm phosphate buffer (pH 7.5) was incubated at 60 °C with or without two variants of HSP18 (0.15 mg·mL−1). Light scattering at 400 nm was monitored for 1 h in the kinetic mode. Both assays were carried out at different ratios of chaperone (HSP18 proteins) and client proteins.

In vivo cell viability experiment with two variants of HSP18 at 48 °C

The cell viability experiment was performed as described previously [51]. Briefly, an equal number of cells from an overnight culture of E. coli M15 strain carrying pQE31 (empty vector), pQE31-HSP18S52 and pQE31- HSP18P52 were inoculated into 25 mL of LB broth containing 100 μg·mL−1 ampicillin and 25 μg·mL−1 kanamycin and grown at 37 °C until A600 of 0.8 was reached. Protein expression was then induced with 0.4 mm IPTG. After 2 h of induction, samples were transferred to a shaking water bath at 48 °C. Samples were taken out at different time points (0, 3 and 6 h) post induction and scored for cell viability by plating on LB-agar plates containing 100 μg·mL−1 ampicillin and 25 μg·mL−1 kanamycin. Cell viability was determined by counting the CFUs on each plate after heat shock at 48 °C relative to the number of CFUs formed in each culture before heat shock. Protein expression in control and HSP18 expressing cultures were analyzed via 12% SDS/PAGE.

In vitro refolding assay

Chaperone activity was also determined by measuring HSP18-mediated refolding of MDH from its fully unfolded state as previously described [70]. Briefly, MDH (1 μm) was denatured in 6 m Gu-HCl for 8 h at 25 °C. Refolding of the enzyme was initiated by 100-fold dilution of the denatured MDH in a refolding buffer (pH 7.5), consisting of 50 mm phosphate, 10 mm magnesium acetate and 5 mm dithiothreitol, in absence and presence of 20 μm HSP18S52 and HSP18P52. The enzyme concentration was 10 nm during refolding. The activity of refolded enzyme was assayed by adding 20 μL of refolding mixture to 580 μL of refolding buffer descibed above containing 0.1 mm NADH and 0.4 mm oxaloacetic acid (pre-incubated at 25 °C) and by measuring the decrease in absorbance at 340 nm over time. To check whether the effect is specific to HSP18 variants, we performed the same experiment in the presence of 20 μm BSA.

Acknowledgements

We thank Professor K. P. Das (Bose Institute, Kolkata, India) for his useful suggestions during the HPLC experiments. S.K.N. acknowledges the receipt of an institute fellowship from the Institute of Technology Bhubaneswar to carry out this work. A.K.P. acknow-ledges CSIR, India grant 37(1535)/12/EMR-II for providing fellowship. K.D. thanks the Department of Biotechnology, New Delhi, India for grant no. BT/HRD/35/03/2010 (Distinguished Professorship Award). This work was supported from the Council of Scientific and Industrial Research, India grant 37(1535)/12/EMR-II (A.B.) and Department of Biotechnology, New Delhi, India grant CGESM/BT/03/002/87-Vol V (K.D.).

Ancillary

Advertisement