Proline isomerization in the C-terminal region of HSP27
- 1.2k Downloads
In mammals, small heat-shock proteins (sHSPs) typically assemble into interconverting, polydisperse oligomers. The dynamic exchange of sHSP oligomers is regulated, at least in part, by molecular interactions between the α-crystallin domain and the C-terminal region (CTR). Here we report solution-state nuclear magnetic resonance (NMR) spectroscopy investigations of the conformation and dynamics of the disordered and flexible CTR of human HSP27, a systemically expressed sHSP. We observed multiple NMR signals for residues in the vicinity of proline 194, and we determined that, while all observed forms are highly disordered, the extra resonances arise from cis-trans peptidyl-prolyl isomerization about the G193-P194 peptide bond. The cis-P194 state is populated to near 15% at physiological temperatures, and, although both cis- and trans-P194 forms of the CTR are flexible and dynamic, both states show a residual but differing tendency to adopt β-strand conformations. In NMR spectra of an isolated CTR peptide, we observed similar evidence for isomerization involving proline 182, found within the IPI/V motif. Collectively, these data indicate a potential role for cis-trans proline isomerization in regulating the oligomerization of sHSPs.
KeywordsMolecular chaperones Small heat-shock proteins cis-trans proline isomerization Nuclear magnetic resonance spectroscopy Intrinsically disordered proteins
Evidence for this flexibility of the CTR has come primarily from solution-state nuclear magnetic resonance (NMR) spectroscopy, which is well suited to studying unstructured regions within proteins. Disorder in the CTR has been observed in αA- and αB-crystallin (human HSPB4 and HSPB5, respectively) (Carver et al. 1992), HSP27 (HSPB1) (Carver and Lindner 1998), murine Hsp20 (van de Klundert et al. 1998), murine Hsp25 (Carver et al. 1995a), and Saccharomyces cerevisiae Hsp26 (Benesch et al. 2010). Detailed interrogation of the CTR in αB-crystallin has revealed that, in solution above 0 ºC, the IPI/V motif is transiently bound to the ACD at only a few percent of the total (Baldwin et al. 2011a; Baldwin and Kay 2012), and that the independent binding of two CTRs, either in an intra- or inter-molecular manner, triggers the dissociation of a monomer from an oligomer (Baldwin et al. 2011b; 2012). The strength and kinetics of the interaction are responsive to amino acid substitutions (Smulders et al. 1996; Hilton et al. 2013a), such that it confers some selectivity between sHSPs (Delbecq et al. 2012). These interactions appear to be finely tuned to regulate both sHSP assembly and hetero-oligomerization (Hochberg and Benesch 2014; Delbecq et al. 2015).
HSP27 is a systemically expressed sHSP that plays a central role in maintaining protein homeostasis (Arrigo 2001), but its elevated expression often results in poor prognosis in multiple cancers (Ciocca and Calderwood 2005). Numerous post-translational modifications (PTMs) have been reported in the CTR of HSP27 (Blakytny et al. 1997; Sakamoto et al. 2002; Nagaraj et al. 2012; Ylikallio et al. 2015), some of which directly influence biological activity. Notably, argipyrimidation of R188 in the CTR of HSP27 prevents caspase-3- and caspase-9-mediated apoptosis, prolonging the life of human carcinoma cells (Sakamoto et al. 2002). Moreover, mutations in the CTR of HSP27 are implicated in neuromuscular and neurodegenerative diseases including amyotrophic lateral sclerosis (Capponi et al. 2016) and Charcot-Marie-Tooth disease (Evgrafov et al. 2004; Kijima et al. 2005; D’Ydewalle et al. 2011; Chalova et al. 2014; Ylikallio et al. 2015). The high density of PTMs and disease-associated mutations in the CTR of HSP27 allude to a significant functional role for this region in health, and malfunction in disease.
Structural information about HSP27 remains relatively limited, as it assembles into a polydisperse ensemble of oligomers centered around ~600 kDa (Jovcevski et al. 2015). However, it comprises an ~80-residue ACD that folds into an immunoglobulin-like domain (Baranova et al. 2011; Hochberg et al. 2014; Rajagopal et al. 2015) (Fig. 1b). The CTR contains an abundance of charged, polar, and proline residues (Fig. 1a), which are known to promote disordered backbone conformations (Theillet et al. 2014), and has been crystallized bound to the ACD (Hochberg et al. 2014) (Fig. 1b). In crystal structures of αA-, αB-crystallin, and HSP27 ACD-CTR complexes, the CTR forms an extended conformation in which the I-P peptide bond in the IPI/V motif adopts the trans-conformation, enabling penetration of the adjacent residues into the hydrophobic groove of the ACD. Here, we have probed the conformation and dynamics of the HSP27 CTR by means of solution-state NMR spectroscopy. We report that the CTR is predominately disordered, and adopts distinct unfolded structural ensembles that interconvert on the seconds timescale or longer via cis-trans peptidyl-prolyl isomerization at its two proline residues. Notably, we show that the cis-conformers are populated 20- to 30-fold higher than expected for a fully disordered peptide of this sequence without any proline residues. Given the importance of the CTRs of mammalian sHSPs in assembly and recognition (Delbecq and Klevit 2013; Hochberg and Benesch 2014; Treweek et al. 2015), we speculate that proline isomerization may play a role in regulating the assembly of sHSPs.
The C-terminal region of HSP27 populates multiple conformations
We next interrogated the conformations and dynamics of the CTR using NMR spectroscopy. Using [U-13C,15N]-HSP27, we recorded a two-dimensional (2D) 1H-15N heteronuclear single-quantum coherence (HSQC) spectrum, which correlates the chemical shifts of 1H nuclei that are covalently attached to 15N nuclei (e.g., amide and amine bonds) (Fig. 2a). We observed sharp resonances with limited dispersion in the 1H dimension, indicating that these resonances arise from dynamically disordered residues lacking a fixed tertiary structure. In order to assign the observed resonances to specific HSP27 residues, we recorded three-dimensional (3D) NMR spectra from [U-13C, 15N]-labeled HSP27 that correlate 13C, 15N, and 1HN chemical shifts of neighboring amino acids whose nuclei are linked through a network of covalent bonds (Sattler et al. 1999). This allowed us to unambiguously assign the observed resonances to residues E178–K205 (Fig. 2b). These data therefore reveal that the final 28 residues of HSP27, including the IPV motif, are predominately disordered in solution.
Proline residues in the HSP27 CTR undergo cis-trans isomerization
In our assignments, we noticed that a subset of HSP27 CTR residues yielded two unique signals per residue, with appreciably different signal intensities. This reveals two CTR conformational ensembles that interconvert with each other relatively slowly, on a timescale of seconds or longer. Interestingly, the subset of residues with two NMR signals all clustered near a single proline residue, P194 (Fig. 2c). In a disordered chain, the majority of amino acids are expected to form peptide bonds that predominantly adopt the trans conformation (>99.5%) (Fig. 1c, right). By contrast, peptide bonds involving proline residues (X-Pro) have an increased population of the cis conformation (5-15%) (Weiss et al. 1998; Theillet et al. 2014) (Fig. 1c, left). The inter-conversion of the two forms involves the rearrangement of a covalent bond, which results in a substantially slower rate of exchange, typically on the seconds timescale or slower under physiological conditions (Grathwohl and Wüthrich 1981; Weiss et al. 1998; Wedemeyer et al. 2002). Our observation of multiple resonances for residues in the vicinity of P194 (Fig. 2c) hence alluded to potential cis-trans proline isomerization about the G193-P194 peptide bond.
We directly assessed the possibility of a cis-Pro conformation in the G193-P194 peptide bond by analyzing the 13C side-chain chemical shifts in P194 (Fig. 2d). The difference in 13C chemical shifts between Pro-Cβ and Pro-Cγ nuclei is highly diagnostic of cis (~10 ppm) or trans (~5 ppm) peptide bond conformations (Schubert et al. 2002; Shen and Bax 2010). Using a 3D C(CO)NH experiment (Grzesiek et al. 1993), we correlated adjacent side-chain 13C chemical shifts to show clear relations between resonances. Such an experiment reveals that the two resonances from E195 can be attributed to being directly adjacent to either the cis or trans form of P194 (Fig. 2d). This finding is independently supported by a calculation using PROMEGA (Shen and Bax 2010), which predicts cis- or trans-Pro conformations based on 13C chemical shifts. Similarly, additional minor peaks from other residues near P194 arise from the formation of the cis-P194 bond (Fig. S1). Together, these data clearly demonstrate that proline isomerization occurs about the G193-P194 peptide bond.
Signal intensity coming from the second proline residue (P182) in the CTR of HSP27 in the 3D C(CO)NH experiment was insufficient to determine any populations of cis-P182, likely owing to its proximity to the slowly tumbling oligomer bulk. To address the likelihood of isomerization about the I181-P182 peptide bond, we therefore prepared a non-isotopically enriched peptide comprising residues E178–E186. In the methyl region of the 1H-13C HSQC spectrum of this peptide, we would expect eight resonances in the absence of isomerization. However, we observed 15 peaks arising from methyl-bearing residues (Fig. S2). The appearance of these additional resonances demonstrates that the peptide adopts multiple structural ensembles, with the most probable cause being proline isomerization. Given that the IPV motif is largely unbound in the oligomers, it is highly likely that P182 exists in both cis and trans forms.
The cis-P194 state is populated to approximately 15% in HSP27 oligomers
The intensities of NMR signals are related to the number of nuclei resonating at a given chemical shift and therefore reflect the population of a given conformation, provided the two forms undergo similar dynamics. We therefore extracted the peak intensities from both the cis- and trans-P194 conformations, and calculated the population of the cis state to be ~10% at 25 °C (Fig. 2e). In an independently prepared sample, we used two well-resolved and relatively intense resonances from adjacent resonances that we could unambiguously attribute to the cis or trans forms (A196, A197) to calculate the equilibrium constant K = [cis]/[trans] at 20, 30, and 40 °C in order to probe the enthalpic and entropic contributions to formation of the cis-P194 state (Fig. 2f). The population of the cis state increased to ~15% at 40 °C, our closest measure to body temperature. This variation in equilibrium position could be explained by the process being reversible and characterized by constant enthalpy and entropy values over the temperature range studied. A Van’t Hoff plot of the resultant equilibrium constant at each temperature indicates that the cis-P194 conformation is enthalpically disfavored (~10 kJ mol−1), but entropically favored (3 J K−1 mol−1) over the trans-P194 state (Fig. 2f). These results, derived from the full-length protein in heterogeneous oligomeric forms, are consistent with data obtained from small, proline-containing oligopeptides (Grathwohl and Wüthrich 1981; Eberhardt et al. 1993; Troganis et al. 2000).
Proline isomerization affects the propensity for residual β-strand structure in the CTR
Δδ values report on secondary structure, with significant deviations from zero over multiple residues reflecting β-strand or α-helical conformations, depending on the nucleus and the sign of Δδ. Of those available, the Cα and Cβ shifts are the most sensitive to residual secondary structure. Both the cis- and trans-P194 ensembles show a small tendency for increased and decreased Cα and Cβ chemical shifts, respectively, in the region 180–185, which indicates a low population of β-strands in both ensembles. For residues 193–199 in the vicinity of the conserved IXI/V motif, cis-P194 induced a small population of β-strand, whereas the trans-P194 ensemble is more disordered in this region. These findings were supported by calculation of the secondary structure propensity (SSP) (Marsh et al. 2006), where the propensity for β-strand formation was estimated to be on the order of 10% for residues 195–198 in the cis-P194 state (Fig. 3e). Similarly, using the software δ2D (Camilloni et al. 2012), we calculated an increased population of β-strand conformations in the region near cis-P194, albeit with a significantly higher propensity for β-strand formation than predicted by SSP (Fig. S3, compare to Fig. S4). The increase in β-strand conformation is accompanied by a concomitant loss of polyproline-II structure in the cis conformation. Polyproline-II conformations populate similar angles in the Ramachandran map as β-strands, but form a distinct hydrogen bonding pattern. Thus, since the CTR interacts with the ACD in a β-strand conformation, the formation of residual β-strand structure in the CTR would be expected to influence binding to the hydrophobic β4-β8 groove (Laganowsky et al. 2010; Laganowsky and Eisenberg 2010; Hochberg et al. 2014).
Both proline isomers are highly dynamic
Both 15N R 1 and 15N R 2 are less sensitive to rapid internal dynamics than the hetNOE, yet provide information regarding motions on the nanosecond timescale that are faster than the rotational correlation time. The values obtained for the cis- and trans-P194 forms are very similar (Fig. 4b, c), with the average values of R 1 and R 2 of (~1.2 and 5 s−1, respectively) and the low R 2/R 1 ratio (~4, Fig. S5) indicating that both cis- and trans-P194 conformational ensembles are also highly dynamic in the nanosecond regime (Fig. 4, S4, S5). In addition to these motions in the CTR of HSP27, it is possible, as in the case of αB-crystallin (Baldwin et al. 2011a), that there are contributions from conformational exchange to the observed R 2 values in the IPV motif. Taken together, our data have revealed that the CTR, including the IPV motif, is highly dynamic on the picosecond-nanosecond timescale, with the magnitude of these motions increasing upon moving away from the ACD toward the C-terminus.
Dynamic fluctuations of the CTR
Here we have shown that the final 28 residues in the CTR of the human sHSP HSP27 can be readily discerned by NMR spectroscopy using non-deuterated samples. The observation of their chemical shifts in the context of the 100–700-kDa oligomeric ensemble, combined with detailed analysis of chemical shifts and relaxation properties, reveals this region of HSP27 to be highly dynamic. These conclusions are in agreement with previous NMR studies of the flexible CTRs of mammalian sHSPs αA- and αB-crystallin (Carver et al. 1992; 1995b; Smulders et al. 1996; Carver and Lindner 1998; Treweek et al. 2010; Baldwin et al. 2012; Mainz et al. 2015; Delbecq et al. 2015), HSP27 (Carver and Lindner 1998), Hsp20 (van de Klundert et al. 1998), and Hsp25 (Carver et al. 1995a; Lindner et al. 2000) in which the final 10–15 residues were amenable to NMR investigation under non-heat-shock conditions.
In this study, we were able to observe and assign an additional 10 residues within the disordered CTR of HSP27 over those described previously (Carver and Lindner 1998), likely due to our usage of modern hardware, a higher magnetic field strength, and [U-13C,15N]-labeled samples, and we could thus probe the conformations and dynamics of the final 28 residues. Similarly, a recent NMR investigation of the CTR in human αB-crystallin was able to extend the number of observed residues to ~25 (Mainz et al. 2015). The extensions of these two sHSPs are therefore both predominantly disordered at physiological solution conditions. Notably, the IPI/V motifs in both cases are primarily detached from the oligomers, and therefore only make transient interactions with their own ACD or adjacent ACDs. Studies on αΑ- and αB-crystallin CTRs have demonstrated that rapid motions take place on the pico-nanosecond timescale (Treweek et al. 2010), and quantitative analysis of the relaxation properties of HSP25 Pro-194 reveals that its CTR has high motional flexibility (Esposito et al. 1998). We have found by 15N-relaxation that the disordered CTR of HSP27 also exhibits similar motions on the same timescale. These combined observations show that the motions at the end of several mammalian CTRs are effectively uncoupled from the slowly tumbling oligomers, such that they behave comparably to an unstructured peptide.
Cis-trans proline isomerization in the CTR
We identified that slow inter-conversion between cis- and trans-proline conformations within the G193-P194 peptide bond takes place in HSP27. Proline isomerization has also been reported in the extension of bovine αB-crystallin, where the authors were able to attribute the two peaks for V169 to cis and trans conformers of the K166-P167 peptide bond (Carver and Lindner 1998). The second proline residue in the CTR of HSP27 is situated within the highly conserved IPV motif. We cannot conclude for certain that cis-trans isomerization of P182 takes place since the chemical shifts near this residue are weak, rendering a lowly populated cis conformation unobservable. However, using methyl-based experiments on a peptide encompassing the IPV motif, we observed more NMR signals than would be expected from the amino acid sequence, which likely indicates cis-trans isomerization about the I181-P182 peptide bond. In support of this, we note that, when taking advantage of the enhanced signal-to-noise afforded by the use of perdeuterated protein, additional weak resonances were detected for residues near P160 in the IPI motif of human αB-crystallin (Mainz et al. 2015). This combined evidence suggests that both HSP27 and αB-crystallin undergo cis-trans proline isomerization in the IPI/V motif.
Cis-trans proline isomerization about the I181-P182 bond could play a significant role in HSP27 oligomerization. The conformation of the CTR when bound in the hydrophobic groove of the ACD is expected to be in the extended trans-P182 form, as has been observed in structures of other sHSP oligomers (Hilton et al. 2013b) and for the CTR peptide co-crystallized with the HSP27 ACD (Fig. 5a) (Hochberg et al. 2014). In HSP27, the I181-P182 peptide bond in the CTR adopts a trans-conformation, which allows the side-chains of I181 and V183 to dock into the hydrophobic groove created by the β4 and β8 strands (Fig. 5a). Rotation about the I181-P182 peptide bond to create a cis conformation would create a less energetically favorable state in which the hydrogen bonds with the β4 strand are broken and V183 and F185 are exposed to solvent (Fig. 5b). Therefore, we would expect that the trans-P182 form has a higher affinity for the ACD than the cis form. For the trans-I181-P182 state that contains residual β-strand conformation (~10%), this minor population could comprise a binding-compatible state. Additionally, any PTMs and mutations that affect the ratio and inter-conversion rate of these two forms could potentially affect interactions with the ACD. Notably, three inheritable mutations (T180I, P182S, and P182L) in this region lead to the onset of the motor neuropathy Charcot-Marie-Tooth disease (D’Ydewalle et al. 2011; Chalova et al. 2014).
The CTR of αB-crystallin also has a central proline residue within its IXI/V motif that could be prone to cis-trans isomerization. We previously found that this motif was also predominately disordered, and quantified its kinetics and thermodynamics of binding to the ACD of αB-crystallin using a combination of methyl-TROSY CPMG NMR experiments and native MS (Baldwin et al. 2011a, b). The lifetime of the bound ACD-CTR αB-crystallin complex varied strongly with temperature and pH, but was typically on the order of milliseconds at near-physiological conditions, and populated to approximately 1% (Baldwin et al. 2011a). However, the exchange of subunits between oligomers is mediated by monomers with lifetimes on the order of seconds under similar conditions (Baldwin et al. 2011b), much longer than CTR binding. Despite the differences in lifetime of 3 orders of magnitude, CTR binding and subunit exchange are kinetically coupled (Baldwin et al. 2011a), with the independent but concomitant binding of two adjacent CTRs occurring on the seconds timescale, and being necessary for the dissociation of a monomer from an oligomer (Baldwin et al. 2012; Hochberg and Benesch 2014).
The rate for cis-trans proline isomerization would be expected to be an additional 3 orders of magnitude slower, with lifetimes on the tens of minutes timescale (Wedemeyer et al. 2002). As the cis-trans isomerization rates are 6 orders of magnitude slower than the CTR binding process, it is unlikely that these events are directly correlated. It is interesting, however, to speculate that the two different proline conformers may have different affinities for the ACD. The trans form would perhaps be expected to bind the ACD transiently a large number of times during its lifetime before conversion to the relatively inert cis conformation. Consistent with this, mutation of the central proline has been shown to modulate both the kinetics and thermodynamics of oligomerization of αB-crystallin, resulting in slightly faster subunit exchange rates and an altered equilibrium oligomeric distribution (Hilton et al. 2013a). These results reveal that, while the conserved IXI/V motif within the CTR is predominantly disordered, it can nevertheless exert significant regulatory properties on the underlying oligomers by affecting the transiently populated bound states. As such, effects such as cis-trans proline isomerization may, in principle, play an important regulatory role.
Possible roles for cis-trans proline isomerization
Inside cells, cis-proline bonds play fundamental roles with widespread biological implications (Sarkar et al. 2007). The slow spontaneous rate of cis-trans proline isomerization often regulates the rates with which proteins fold (Wedemeyer et al. 2002) and cis-proline bonds can induce structural changes in membrane-embedded receptors (Lummis et al. 2005). A specific class of ubiquitous molecular chaperones, peptidyl-prolyl isomerases, such as trigger factor, function to catalyze the otherwise slow inherent inter-conversion of cis- and trans-Pro states (Gothel and Marahiel 1999). In HSP70, it has been proposed that a cis-Pro bond involving a universally conserved proline residue in the ATPase domain regulates the allosteric transition between nucleotide- and substrate-bound states (Vogel et al. 2006; Alderson et al. 2016). Considering the evidence of allosteric communication between the contact the CTR transiently makes with the ACD, and the dimer interface (Delbecq et al. 2012; Hochberg and Benesch 2014), we speculate that the cis-P182 and cis-P194 bonds could affect the oligomerization of HSP27, or even its interactions with cellular targets. In addition, the post-translational modification of residues in the CTR, such as the non-enzymatic argipyrimidation or R188 (Sakamoto et al. 2002), could be influenced by isomerization regulating solvent exposure.
We have employed solution-state NMR spectroscopy to probe the conformation and dynamics of the disordered CTR of HSP27. We found that the CTR contains residual β-strand structure in the highly conserved IPV motif, which could impact its subsequent binding to the ACD. In addition, our results demonstrate that both proline residues (P182, P194) within the CTR undergo cis-trans isomerization, with the cis-P194 conformation retaining its rapid internal dynamics, but leading to an increase in residual β-strand structure in residues 193–199. We speculate that cis-trans proline isomerization is a feature of those mammalian sHSP CTRs that are flexible, in particular HSP27 and αB-crystallin. Given that these two are the human sHSPs with the most widespread expression, are heavily up-regulated during stress, and have the largest number of interactors (Kampinga et al. 2015; Mymrikov et al. 2017), it is likely that the dynamic fluctuations of their CTRs play an important role in their canonical sHSP chaperone activity.
Materials and methods
HSP27 expression and purification from Escherichia coli
An ampicillin-resistant (AmpR) HSP27 plasmid was kindly donated by Prof. Heath Ecroyd (University of Wollongong). The HSP27 plasmid was transformed into BL21-Gold(DE3) competent cells and plated on an LB-agar plate containing 100 μg mL−1 of Amp. A single colony from the LB-agar plate was used to inoculate a 5-mL culture of LB medium with 100 μg mL−1 of Amp for expression of full-length HSP27. Following growth at 37 °C for 6 h, the 5-mL culture was transferred to 100 mL (LB medium) or 50 mL (M9 minimal medium) cultures and grown overnight at 37 °C, which were then used to inoculate 1 L (LB medium) or 500 mL (M9 minimal medium) cultures. The M9 minimal medium consisted of 6 g L−1 Na2HPO4, 3 g L−1 KH2PO4, 1 g L−1 NaCl, 1 g L−1 15NH4Cl, 6 g L−1 natural abundance glucose, 1 mM MgSO4, and 100 μΜ CaCl2. For preparation of uniformly-13C, 15N-labeled ([U-13C,15N]-) HSP27, the M9 minimal medium contained 2 g L−1 of [U-13C]-glucose and 1 g L−1 of 15NH4Cl. When the absorbance at 600 nm of these cultures reached between 0.6 and 0.8 units, IPTG was added to a final concentration of 100 μg mL−1 and protein expression ensued for 3 h at 37 °C. Cells were pelleted and frozen at −80 °C until use.
HSP27 was purified in a similar manner as described previously (Muranova et al. 2015) using anion exchange chromatography (AEX) with HiTrapQ HP columns (GE Healthcare) followed by size exclusion chromatography (SEC) on a Superdex S200 column (GE Healthcare). The final SEC buffer contained 20 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA at pH 7.0. After pooling HSP27-containing fractions that eluted from the Superdex S200 column, the samples were further purified with a second AEX step involving the usage of a HiTrap Capto Q ImpRes column (GE Healthcare). HSP27 was then buffer-exchanged using Amicon spin filters (Millipore) into 30 mM NaH2PO4, 100 mM NaCl, 2 mM EDTA, 2 mM NaN3, pH 7.0 with 6% D2O (NMR buffer with 100 mM NaCl).
Native mass spectrometry of HSP27
A 25-μM sample of natural abundance HSP27 was buffer exchanged into 200 mM ammonium acetate, pH 6.9. Data were collected on an ion-mobility Synapt G1 mass spectrometer (Waters) under ionization conditions that maintain non-covalent interactions, according to a protocol described previously (Kondrat et al. 2015).
Backbone NMR assignments for [U-13C,15N]-HSP27
All NMR spectroscopy experiments for resonance assignments of HSP27 were recorded on a 600-MHz Varian-Inova spectrometer equipped with a 5-mm z-axis gradient triple resonance, room temperature probe. [U-13C,15N]-HSP27 was prepared at a final concentration of 2 mM in NMR buffer with 100 mM NaCl. Standard assignment experiments were collected, namely 3D HNCO, HN(CO)CA, HN(CA)CO, HNCA, HNCACB, and C(CO)NH spectra (Sattler et al. 1999), and these experiments contained either uniform sampling or non-uniform sampling (NUS) schemes in the indirect dimensions. When NUS was employed, data were recorded with ~25% sparsity in the indirect dimensions and time-domain data were reconstructed with MddNMR (Kazimierczuk and Orekhov 2011). All 3D NMR spectra were acquired at 25 °C, processed with NMRPipe (Delaglio et al. 1995), and visualized with Sparky (Lee et al. 2015).
Natural abundance 13C NMR spectra of the HSP27 C-terminal tail peptide
A peptide encompassing the IPI/V motif of HSP27 was ordered from BioMatik. The sequence of the peptide was EITIPVTFE, with an Nα acetyl group at the N-terminus. Approximately 5 mg of peptide was dissolved in 600 μL of 30 mM NaH2PO4 at pH 12 to ensure solubility of the hydrophobic peptide. pH values below 12 led to precipitation of the peptide at the high concentrations that were required to detect 13C at natural abundance. A 2D 1H-13C HSQC spectrum was recorded on a 500-MHz Bruker Avance III spectrometer equipped with a cryogenically cooled probe. Since 13C is present at natural abundance to only 1% 13C, a minor cis population (~10%) would yield an effective NMR-observable concentration of only 0.1% of the total peptide. Thus from the 5 mg that were dissolved in NMR buffer at a concentration of ~8 mM, the effective total concentration of NMR-observable peptide is 80 μM and with an effective concentration of signals from a minor cis-P182 state near 8 μM.
Analysis of residual secondary structure in the HSP27 C-terminal region
To analyze residual secondary structure in the CTR of HSP27, we utilized the software programs SSP (Marsh et al. 2006) and δ2D (Camilloni et al. 2012). For both SSP and δ2D, calculations were performed with 13Cα, 13Cβ, and 13CO chemical shifts or with 13Cα, 13Cβ, 13CO, 15N, and 1HN chemical shifts for both the cis-P194 and trans-P194 conformations.
15N spin relaxation experiments (R1, R2, NOE) on HSP27
Standard pulse sequences to measure hetNOE, longitudinal (T 1), and transverse (T 2) relaxation times were employed (Palmer 1993). T 1 and T 2 values are reported as their respective inverses, i.e., rates (R 1 and R 2), for convenience. The spectrometer temperature was calibrated with d4-methanol. We recorded 15N R 1, 15N R 2, and hetNOE datasets at 25 °C on a 2-mM sample of [U-13C,15N]-HSP27 in NMR buffer with 100 mM NaCl. All data sets mentioned here were processed with NMRPipe (Delaglio et al. 1995), visualized with Sparky (Lee et al. 2015), and further analyzed with FuDA (http://www.biochem.ucl.ac.uk/hansen/fuda/) to fit peak shapes. To obtain the 15N R 2 estimated values in Fig. S6, we used the average 15N R 2 value that was measured for a 20-kDa protein at a similar concentration (1 mM) on our spectrometer to then extrapolate to proteins that are 40 and 60 kDa in molecular mass. The 15N R 2 is proportional to the rotational correlation time of a molecule, which is itself roughly proportional to the molecular mass, assuming isotropic rotational diffusion.
Thermodynamic analysis of cis-trans G193-P194 isomerization
1H-15N HSQC spectra on [U-15 N]-HSP27 were recorded at 20, 30, and 40 °C. FuDA was used to fit peakshapes with uniform linewidths to extract reliable peak intensities. The equilibrium constant K was calculated as the population of cis-P194 divided by the population of trans-P194. The resultant values of K were fit to the van’t Hoff equation in order to extract the change in enthalpy (ΔH) and the change in entropy (ΔS) that accompany the trans- to cis-P194 reaction.
We thank Prof. Heath Ecroyd (University of Wollongong, Australia) for generously providing the HSP27 expression plasmid and Dr. Ad Bax (National Institutes of Health, USA) for helpful suggestions. TRA acknowledges funding from the NIDDK, the NIH Oxford-Cambridge Scholars Program, and Pembroke College, Oxford. JLPB acknowledges support from the Engineering and Physical Sciences Research Council (EP/J01835X/1), and AJB holds a David Phillips Fellowship from the Biotechnology and Biosciences Research Council (BB/J014346/1).
- Benesch JLP, Aquilina JA, Baldwin AJ, Rekas A, Stengel F, Lindner RA, Basha E, Devlin GL, Horwitz J, Vierling E, Carver JA, Robinson CV (2010) The quaternary organization and dynamics of the molecular chaperone HSP26 are thermally regulated. Chem Biol 17:1008–1017. doi: 10.1016/j.chembiol.2010.06.016 CrossRefPubMedPubMedCentralGoogle Scholar
- Capponi S, Geuens T, Geroldi A, Origone P, Verdiani S, Cichero E, Adriaenssens E, De Winter V, Bandettini di Poggio M, Barberis M, Chiò A, Fossa P, Mandich P, Bellone E, Timmerman V (2016) Molecular chaperones in the pathogenesis of amyotrophic lateral sclerosis: the role of HSPB1. Hum Mutat 37:1202–1208. doi: 10.1002/humu.23062 CrossRefPubMedPubMedCentralGoogle Scholar
- Chalova AS, Sudnitsyna MV, Strelkov SV, Gusev NB (2014) Characterization of human small heat shock protein HspB1 that carries C-terminal domain mutations associated with hereditary motor neuron diseases. Biochim Biophys Acta 1844:2116–2126. doi: 10.1016/j.bbapap.2014.09.005 CrossRefPubMedGoogle Scholar
- D’Ydewalle C, Krishnan J, Chiheb DM, Van Damme P, Irobi J, Kozikowski AP, Berghe PV, Timmerman V, Robberecht W, Van Den Bosch L (2011) HDAC6 inhibitors reverse axonal loss in a mouse model of mutant HSPB1-induced Charcot-Marie-Tooth disease. Nat Med 17:968–974. doi: 10.1038/nm.2396 CrossRefPubMedGoogle Scholar
- Evgrafov OV, Mersiyanova I, Irobi J, Van Den Bosch L, Dierick I, Leung CL, Schagina O, Verpoorten N, Van Impe K, Fedotov V, Dadali E, Auer-Grumbach M, Windpassinger C, Wagner K, Mitrovic Z, Hilton-Jones D, Talbot K, Martin J-J, Vasserman N, Tverskaya S, Polyakov A, Liem RKH, Gettemans J, Robberecht W, De Jonghe P, Timmerman V (2004) Mutant small heat-shock protein 27 causes axonal Charcot-Marie-Tooth disease and distal hereditary motor neuropathy. Nat Genet 36:602–606. doi: 10.1038/ng1354 CrossRefPubMedGoogle Scholar
- Hochberg GKA, Ecroyd H, Liu C, Cox D, Cascio D, Sawaya MR, Collier MP, Stroud J, Carver JA, Baldwin AJ, Robinson CV, Eisenberg DS, Benesch JLP, Laganowsky A (2014) The structured core domain of αB-crystallin can prevent amyloid fibrillation and associated toxicity. Proc Natl Acad Sci 111:E1562–E1570. doi: 10.1073/pnas.1322673111 CrossRefPubMedPubMedCentralGoogle Scholar
- Kampinga HH, de Boer R, Beerstra N (2015) The multicolored world of the human HSPB family. Springer International Publishing. The big book on small heat shock proteins, Chapter 1, 3–26. doi: 10.1007/978-3-319-16077-1_1
- Kondrat FD, Struwe WB, Benesch JLP (2015) Native mass spectrometry: towards high-throughput structural proteomics. Methods Mol Biol 1261:349–371. doi: 10.1007/978-1-4939-2230-7_18
- Laganowsky A, Benesch JLP, Landau M, Ding L, Sawaya MR, Cascio D, Huang Q, Robinson CV, Horwitz J, Eisenberg D (2010) Crystal structures of truncated alphaA and alphaB crystallins reveal structural mechanisms of polydispersity important for eye lens function. Protein Sci 19:1031–1043. doi: 10.1002/pro.380 CrossRefPubMedPubMedCentralGoogle Scholar
- Lindner RA, Carver JA, Ehrnsperger M, Buchner J, Esposito G, Behlke J, Lutsch G, Kotlyarov A, Gaestel M (2000) Mouse Hsp25, a small heat shock protein. The role of its C-terminal extension in oligomerization and chaperone action. Eur J Biochem 267:1923–1932. doi: 10.1046/j.1432-1327.2000.01188.x CrossRefPubMedGoogle Scholar
- Muranova LK, Weeks SD, Strelkov SV, Gusev NB (2015) Characterization of mutants of human small heat shock protein HspB1 carrying replacements in the N-terminal domain and associated with hereditary motor neuron diseases. PLoS One 10:e0126248. doi: 10.1371/journal.pone.0126248 CrossRefPubMedPubMedCentralGoogle Scholar
- Nagaraj RH, Nahomi RB, Shanthakumar S, Linetsky M, Padmanabha S, Pasupuleti N, Wang B, Santhoshkumar P, Panda AK, Biswas A (2012) Acetylation of αA-crystallin in the human lens: effects on structure and chaperone function. Biochim Biophys Acta 1822:120–129. doi: 10.1016/j.bbadis.2011.11.011 CrossRefPubMedGoogle Scholar
- Sakamoto H, Mashima T, Yamamoto K, Tsuruo T (2002) Modulation of heat-shock protein 27 (Hsp27) antiactivity by methylgyloxal modification. J Biol Chem 277:45770–45775. doi: 10.1074/jbc.M207485200
- Troganis A, Gerothanassis IP, Athanassiou Z, Mavromoustakos T, Hawkes GE, Sakarellos C (2000) Thermodynamic origin of cis/trans isomers of a proline-containing β-turn model dipeptide in aqueous solution: a combined variable temperature 1H-NMR, two-dimensional 1H, 1H gradient enhanced nuclear Overhauser effect spectroscopy (NOESY), one-dimensional. Biopolymers 53:72–83. doi: 10.1002/(SICI)1097-0282(200001)53:1 CrossRefPubMedGoogle Scholar
- Ylikallio E, Konovalova S, Dhungana Y, Hilander T, Junna N, Partanen JV, Toppila JP, Auranen M, Tyynismaa H (2015) Truncated HSPB1 causes axonal neuropathy and impairs tolerance to unfolded protein stress. BBA Clin 3:233–242. doi: 10.1016/j.bbacli.2015.03.002 CrossRefPubMedPubMedCentralGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.