Interactions of the intact FsrC membrane histidine kinase with its pheromone ligand GBAP revealed through synchrotron radiation circular dichroism

Biochimica et Biophysica Acta 1818 (2012) 1595–1602

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbamem

Interactions of the intact FsrC membrane histidine kinase with its pheromone ligand GBAP revealed through synchrotron radiation circular dichroism Simon G. Patching a, Shalini Edara a, Pikyee Ma a, 1, Jiro Nakayama b, Rohanah Hussain c, Giuliano Siligardi c, Mary K. Phillips-Jones a,⁎ a b c

Astbury Centre for Structural Molecular Biology, Faculty of Biological Sciences, University of Leeds, Leeds, LS2 9JT, UK Department of Bioscience and Biotechnology, Faculty of Agriculture, Graduate School, Kyushu University, Fukuoka 812–8581, Japan Diamond Light Source Ltd, Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0DE, UK

a r t i c l e

i n f o

Article history: Received 18 December 2011 Received in revised form 8 February 2012 Accepted 10 February 2012 Available online 17 February 2012 Keywords: Membrane histidine kinase FsrC Gelatinase biosynthesis-activating pheromone (GBAP) Synchrotron radiation circular dichroism Ligand interactions kd

a b s t r a c t FsrC is the membrane-bound histidine kinase component of the Fsr two-component signal transduction system involved in quorum sensing in the hospital-acquired infection agent Enterococcus faecalis. Synchrotron radiation circular dichroism spectroscopy was used here to study the intact purified protein solubilised in detergent micelles. Conditions required for FsrC stability in detergent were firstly determined and tested by prolonged exposure of stabilised protein to far-ultraviolet radiation. Using stabilised purified protein, far-ultraviolet synchrotron radiation circular dichroism revealed that FsrC is 61% α-helical and that it is relatively thermostable, retaining at least 57% secondary structural integrity at 90 °C in the presence or absence of gelatinase biosynthesis-activating pheromone (GBAP). Whilst binding of the quorum pheromone ligand GBAP did not significantly affect FsrC secondary structure, near-ultraviolet spectra revealed that the tertiary structure in the regions of the Tyr and Trp residues was significantly affected. Titration experiments revealed a calculated kd value of 2 μM indicative of relatively loose binding of gelatinase biosynthesis-activating pheromone to FsrC. Although use of synchrotron radiation circular dichroism has been applied to membrane proteins previously, to our knowledge this is the first report of its use to determine a kd value for an intact membrane protein. Based on our findings, we suggest that synchrotron radiation circular dichroism will be a valuable technique for characterising ligand binding by other membrane sensor kinases and indeed other membrane proteins in general. It further provides a valuable screening tool for membrane protein stability under a range of detergent conditions prior to downstream structural methods such as crystallisation and NMR experiments particularly when lower detergent concentrations are used. Crown Copyright © 2012 Published by Elsevier B.V. All rights reserved.

1. Introduction One of the most important mechanisms by which bacteria sense and respond to changes in their environment is mediated through two-component signal transduction systems. These systems generally comprise a histidine protein kinase (or membrane sensor kinase) Abbreviations: DDM, n-dodecyl-β-D-maltoside; SRCD, synchrotron radiation circular dichroism; CD, circular dichroism; GBAP, gelatinase biosynthesis-activating pheromone; IPTG, iso-propylthiogalactoside; HT, high tension; UV, ultraviolet; MRW, mean residue weight; CMC, critical micelle concentration; ESD, extracellular sensory domain; TM, transmembrane domain; SDS-PAGE, sodium dodecyl sulphate-polyacrylamide gel electrophoresis; kd, dissociation constant ⁎ Corresponding author at: Astbury Centre for Structural Molecular Biology, Garstang South Building, Faculty of Biological Sciences, University of Leeds, Leeds, LS2 9JT, UK. Tel.: +44 113 343 5624; fax: +44 113 343 3167. E-mail addresses: [email protected] (S.G. Patching), [email protected] (S. Edara), [email protected] (P. Ma), [email protected] (J. Nakayama) , [email protected] (R. Hussain), [email protected] (G. Siligardi), [email protected] (M.K. Phillips-Jones). 1 Present address: Instituto de Tecnologia Quimica e Biologica, Universidade Nova de Lisboa.

which is usually membrane-bound and is responsible for sensing a specific environmental signal (or set of signals), and a partner response regulator which mediates a specific response, usually up- or down-regulating specific genes. Of the two components, much less is understood about the sensor kinase proteins, particularly with regard to ligand (signal) interactions. This is largely due to the hydrophobic nature of these membrane proteins which presents technical challenges for studying these proteins. To overcome this, one approach has been to devise in vivo reporters of sensor activation and inhibition in response to ligand; this has been particularly successful for identifying ligand specificity determinants and the rules for ligand-receptor recognition in the Agr quorum system of Staphylococcus aureus [1–3]. Another approach has been to express and purify the transmembrane sensory domains in isolation [4,5]. Despite this progress, little quantitative binding data have yet emerged. Furthermore, studies of protein fragments are likely to be unsuitable for observing the expected longer range structural impacts of ligand binding. Thus, studies using the intact versions of these membrane proteins is crucial. Previous work by our group overcame many of the challenges associated with working with these full-length membrane proteins

0005-2736/$ – see front matter. Crown Copyright © 2012 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2012.02.015

1596

S.G. Patching et al. / Biochimica et Biophysica Acta 1818 (2012) 1595–1602

when we successfully overexpressed and purified thirteen of the sixteen intact membrane sensor kinases from Enterococcus faecalis, an important cause of hospital-acquired infections in the UK [6]. One of these proteins was FsrC, a quorum (or cell density) sensor which regulates expression of gelatinase (GelE), serine protease (SprE) and other genes important for virulence (Fig. 1A (transmembrane region)) [7,8]. The quorum signalling ligand for FsrC is gelatinase biosynthesis-activating pheromone, (GBAP), a small autoinducing lactone-linked cyclic peptide of 11 amino acid residues (Fig. 1B) [9]. We showed that intact recombinant FsrC, which was purified within detergent micellar complexes, was active and responsive to GBAP: using in vitro autophosphorylation assays the addition of two-fold GBAP increased the levels of phosphorylated FsrC by ten-fold [6].

The activity of intact FsrC was also shown to be specifically inhibited by siamycin I, a known inhibitor of the Fsr pathway, establishing the first identified target underlying Fsr inhibition by siamycin I [10]. Having established the suitability of purified intact FsrC for identifying ligands and inhibitors that modulate its activity, the present study was undertaken to initiate the acquisition of kinetic and binding data for their interactions. Here we use synchrotron radiation circular dichroism (SRCD) spectroscopy, a technique not used previously for the study of intact membrane sensor kinases, nor used previously in the near-UV region to obtain binding data for any intact membrane protein. We also define the conditions required to stabilise the membrane protein in detergent solution for reliable SRCD measurements, which were then employed here to provide the first quantitative binding data for any intact purified membrane sensor kinase with its ligand. 2. Materials and methods 2.1. Expression and purification of intact FsrC E. coli BL21 [DE3] cells harbouring pTTQHK15, the FsrC membrane protein expression plasmid which carries the intact fsrC gene of E. faecalis V583 fused at the 3’-end to a sequence encoding a hexahistidine motif [6], were cultured as described previously [6,10]. Briefly, cells were cultured at 37 °C in Luria-Bertani broth in 30- or 100-litre volumes in an aerated fermenter, and FsrC expression induced using 0.5 mM IPTG [6]. After 3 h further incubation at 37 °C, cells were harvested, lysed by explosive decompression and inner membranes isolated [6]. Intact FsrC was solubilised from inner membranes using 1% n-dodecyl-β-D-maltoside (DDM) detergent and purified via the engineered hexa-histidine tag using nickel affinity chromatography, all of which was performed as described in [6], except that 0.02% or 0.05% DDM was present in all buffers and throughout all purification steps, and a second purification step was undertaken in which eluted protein was re-applied to a freshly charged nickel column, washed with two column volumes of Wash buffer [6] containing 30 mM imidazole prior to elution [6]. Size exclusion chromatography using desalting BioRad Econo-Pac 10DG columns was undertaken to exchange preparations into 10 mM sodium phosphate buffer pH 7.5 containing 0.02% or 0.05% DDM. Exchanged protein was concentrated using Vivaspin Centrifugal Concentrators (Sartorius). 2.2. Western blotting FsrC preparations were subjected to SDS-polyacrylamide gel electrophoresis (5% stacking and 12.5% resolving acrylamide/bisacrylamide gels) [11] and transferred to Fluorotrans™ membrane (Pall BioSupport UK) for Western blot analysis using INDIA™ HisProbe-HRP for detection of histidine-tagged proteins, as described previously [6]. 2.3. SRCD measurements

Fig. 1. Production and confirmation of purified active his-tagged E. faecalis FsrC. (A) Predicted topology of the putative transmembrane and extracellular sensory regions of FsrC using the predictive tools outlined in Methods. Residues 1–217 of the total 464 are shown; those highlighted in black are identical amongst FsrC and all four specificity types of quorum sensor AgrC of Staphylococcus aureus [1–3], those in grey are unique to FsrC and those highlighted in white are residues that are similar to one or more AgrC sequences. B. Molecular structure of the GBAP pheromone [9]. (C) SDS-PAGE (5% stacking and 12.5% resolving gels) of 5 μg purified intact FsrC (52.9 kDa) visualised using Coomassie blue staining. M, molecular mass markers. (D) Western blot using 5 μg purified FsrC to detect the presence of the C-terminal hexahistidine tag; (E) In vitro autophosphorylation activity assays of: lane 1,purified FsrC (80 pmoles); and lane 2, purified FsrC (80 pmoles) in the presence of GBAP (160 pmoles Q-less version) used in the study. Assays were performed as described in Methods. Phosphorimager-based measurements of radiolabelled FsrC-P levels are shown below each lane. Arrows denote the positions of the monomeric (solid arrow) and minor levels of multimeric (dashed arrow) intact protein, which were verified by N-terminal sequencing and mass spectrometry [6].

SRCD experiments were performed using a nitrogen-flushed instrument on the B23 Synchotron Radiation CD Beamline at the Diamond Light Source, Oxfordshire, UK [12]. Samples were typically prepared in 10 mM sodium phosphate pH 7.5 containing 0.02 or 0.05% DDM (above the CMC value of 0.009%), and left to equilibrate for a period of time based on CD measurement showing no further change in the near UV region. Stabilisation/equilibration conditions included an incubation period of at least 2 h at 20 °C prior to obtaining spectra of samples prepared in 0.02% DDM. When ligand or other additions were required, a period of 20 min incubation was required prior to obtaining spectral data. Spectra were obtained routinely at 20 °C (unless otherwise stated). No data in which the HT of the detector (PMT) exceeded 600 V were included in the analyses [12,13].

S.G. Patching et al. / Biochimica et Biophysica Acta 1818 (2012) 1595–1602

For measurements in the far UV range (180–260 nm) the concentration of FsrC employed was 6 μM. Using sample volumes of 200 μl, four scans were acquired using an integration time of 1 second, a pathlength of 0.02 cm and a slit width of 0.5 mm equivalent to 1.2 nm bandwidth. To investigate the effects of GBAP on the FsrC spectrum in this region, the protein was incubated in the presence of either 12 μM GBAP (dissolved in 3.09% acetonitrile) or an equivalent volume of acetonitrile. The final concentration of acetonitrile was 0.4%. Secondary structural content was determined by analysis of the far-UV spectral data using the CONTINLL and CDSSTR algorithms [14–16] of B23 beamline OLIS Globalworks software. To determine if GBAP binding affects the thermal stability of FsrC, stabilised FsrC was incubated in the presence and absence of two-fold GBAP at 5 °C and spectra obtained over a range of temperatures starting at 5 °C and increasing incrementally to 90 °C, with 10 min equilibration time followed by a return to 20 °C in one step of 30 (min) equilibration time. For measurements in the near UV region (250–340 nm) experiments were performed in 10 mM sodium phosphate (pH 7.5) containing 0.05% DDM using 100 μl sample volumes at 20 °C. In general, a concentration of 20 μM FsrC was employed in cells of 1 cm pathlength. To investigate the effect of five-fold molar excess GBAP, averaged data from 10 scans (integration time 1 s) were obtained for samples containing FsrC in the presence of five-fold (100 μM) GBAP (dissolved in 0.36% acetonitrile), FsrC in the presence of 0.36% acetonitrile, GBAP (dissolved in 0.36% acetonitrile) (100 μM), and 0.36% acetonitrile. For titration experiments, GBAP was added in a stepwise manner as described previously [13], by adding small volumes of 2–3 μl of ligand stocks directly into the 100 μl protein sample (20 μM) and incubating for 20 min at 20 °C prior to obtaining spectra. GBAP was tested at molar ratios of 0.5:1, 1:1, 2:1, 3:1 and 5:1 GBAP: FsrC with a 20 minute incubation period following each addition before acquiring the spectra. At titration point 5:1 (the maximum GBAP:FsrC employed), sample volumes were 111 μl and contained 0.36% acetonitrile. Again, 10 scans were acquired, each with an integration time of 1 sec (75 min per titration point). That the accumulating acetonitrile itself exerted no significant effect on the FsrC spectrum was verified by subtraction of acetonitrile controls derived from the spectrum of the maximum acetonitrile concentration. The dissociation constant kd was determined using the experimental criteria described previously [13] and by analysis using the B23 inhouse program CD titration© v1.6 [Myatt, Siligardi et al., pers. commun.] based on non-linear regression analysis [17]. All spectra of relevant background buffers, solvents, GBAP alone etc were subtracted from the experimental data, which were then either converted into mean residue ellipticity using the equation: [θ]mrw,λ = MRW x θλ/10 x d x c, where θλ is the observed ellipticity (degrees) at wavelength λ, d is the pathlength (cm), and c is the concentration (g/ml), or analysed using CD titration© v1.6. The mean residue weight (MRW) for recombinant FsrC was taken to be 114.72. Data were converted to ΔA for kd calculation. The cyclic GBAP pheromone ligands used in this study were prepared synthetically. Peptides QN(SPNIFGQWM) and Ac-N(SPNIFGQWM) have been described previously, and have been shown to exhibit comparable levels of activation of FsrC autophosphorylation activity [18,19]. 2.4. Activity assays Autophosphorylation activity assays were performed as described previously [6,20], except that 360 pmoles of purified protein were used in the 15 μl assay reactions. In GBAP reactions, 160.5 pmoles of the peptide were also added, resulting in a FsrC:GBAP molar ratio of 1:0.45. In all reactions, equivalent volumes of acetonitrile (used to dissolve GBAP) were added and samples were preincubated for 20 min at room temperature prior to initiation of the assays through addition of radiolabelled ATP.

1597

2.5. Protein determination FsrC concentrations were determined spectrophotometrically in an Implen nanophotometer instrument using a molar extinction coefficient, ε (FsrC-His6), for recombinant FsrC of 38,390 M − 1 cm − 1. 2.6. Topology predictions Predicted topology of the putative transmembrane and extracellular sensory regions of the source FsrC protein used in the present study, originating from E. faecalis V583, were determined using TMHMM (http:// www.cbs.dtu.dk/services/TMHMM), SOSUI (http://bp.nuap.nagoya-u.ac. jp/sosui/), TMpred (http://www.ch.embnet.org/software/TMPRED_form. html) and TopPred (http://bioweb.pasteur.fr/seqanal/interfaces/toppred. html) programs. 3. Results 3.1. Preparation and confirmation of purified and active intact FsrC We have previously developed methods for the successful expression in E. coli of intact FsrC, for its solubilisation from E. coli inner membranes and for its purification via the hexa-histidine tag engineered at the C-terminus of the protein [6,10,20]. Fig. 1C shows the successful employment of these methods, which were further optimised here, for the purification of intact FsrC. In common with many membrane proteins, FsrC exhibits typically anomalous migration in SDS-PAGE; the apparent mass of ~43 kDa was reported previously [6] and contrasts with the predicted mass of 52.9 kDa [6]. The identity of the recombinant protein was confirmed by mass spectrometry [6], by routine detection of the C-terminal hexa-histidine tag in Western blotting experiments (Fig. 1D) and by verification of its retained activity and response to GBAP monitored through autophosphorylation activity assays; Fig. 1E shows that the presence of two-fold GBAP used in this study elicited a significant (twenty-fold) increase in FsrC-P levels in the activity assays, as expected [6]. Therefore, the intact FsrC protein was confirmed as a purified and active protein under the DDM detergent conditions employed and was therefore used for the following SRCD experiments. 3.2. Secondary structure, structural integrity and stabilisation of purified intact FsrC Although the activity assays undertaken both in the present study (Fig. 1E) and previously [6,10] confirm that intact FsrC is prepared as an active (and therefore presumably folded) protein, retention of secondary structural integrity by intact FsrC preparations was investigated using SRCD in the far-UV range, especially in light of the relatively harsh detergent-based purification procedure used to obtain the purified membrane protein. The spectra obtained were typical of an α-helical membrane protein (Fig. 2), thereby confirming the retention of FsrC structural integrity. The purification procedure required for producing the intact membrane protein was therefore confirmed as suitable for obtaining folded protein which retains structural integrity. Such retention of secondary structure post-purification was also found to be the case during our previous studies of other intact sensor kinases which employed similar methods (e.g. [20]). Our initial SRCD studies revealed that consecutive spectra of the same sample, whilst consistently typical of well folded protein, were variable and generally not superimposable in terms of intensity magnitude. Stabilisation/equilibration trials of FsrC were therefore undertaken to identify conditions under which the protein-micelle complex gave reliable, consistent and superimposable CD spectra. After trialling several different sets of conditions, FsrC was successfully stabilised through incubating freshly diluted protein stocks for

1598

S.G. Patching et al. / Biochimica et Biophysica Acta 1818 (2012) 1595–1602

Fig. 2. Determination of FsrC stability during SRCD. (A) Far UV spectra of purified FsrC (6 μM) at 0 h (solid black line, unsymboled), 1.5 h (dashed line, unfilled square), and 2.5 h (dashed line, unfilled circle) following sample preparation in 10 mM sodium phosphate pH 7.5 containing 0.02% DDM at 20 °C. Repeat spectra after 1.5 h involving sample removal and reloading were also included (dashed line, unfilled triangle). Data shown are the average of four spectra. Spectra of stabilised FsrC before and after exposure to far UV radiation, in: (B) the absence and (C) the presence of two-fold GBAP. Spectrum 1, immediately after stabilisation (solid line); and spectrum 40, following 100 min exposure to light radiation (190–260 nm) during 39 consecutive scans (dashed line). The buffer conditions are described in Methods and in the presence of 1.2% acetonitrile. Data from each individual scan (unsmoothed) are shown.

a suitable period prior to collecting spectral data, as described in Methods. For example, Fig. 2A shows that superimposable spectra for FsrC in 0.02% DDM are obtained following incubation for at least 1.5 hr prior to data collection, and that a high level of reproducibility is possible thereafter. Reproducibility and stability were further confirmed in comparisons of an initial spectrum of freshly stabilised FsrC with that obtained following 100 min exposure to 180–260 nm radiation during 39 repeated scans in the instrument. The measured repeated spectra obtained were again almost indistinguishable, both in the presence or absence of the GBAP ligand (Fig. 2B and C). When detergent concentrations were increased to 0.05%, stability was achieved more rapidly, removing the requirement for the stabilisation/equilibration step, as examplified by measurements in the near-UV region (250–340 nm) (Appendix A: Fig A.1.). We have therefore identified conditions for reliable stabilisation of intact FsrC for

SRCD studies. Whilst previous studies have shown that choice of detergent type can affect activity, conformation and determinations of secondary structure composition of membrane proteins [21,22 and references therein], this is the first report of time- and concentration-dependent stabilisation required under detergent conditions greater than the CMC. The reasons for the required stabilisation period are unclear. It is possible that it simply reflects the time necessary for the protein-micelle complex to stabilise following the initial dilution of the protein stock to prepare the CD sample. If so, the dilution effects appear to be minimised through use of higher concentrations of detergent, as demonstrated here. Indeed, the higher concentration of 0.05% is more safely above the CMC for DDM of 0.009%, and therefore the system may reach stability more quickly. All subsequent experiments employed these stabilisation conditions.

S.G. Patching et al. / Biochimica et Biophysica Acta 1818 (2012) 1595–1602

3.3. Characterisation of FsrC-GBAP ligand interactions The effect of GBAP on the secondary structural conformation of FsrC was investigated using SRCD in the far-UV (180–260 nm) region. Fig. 3A shows that the addition of a two-fold molar excess of GBAP elicited no significant change in the FsrC spectrum, indicating no detectable effect of the ligand on the secondary structural composition of FsrC (Fig. 3A). When the far-UV spectra were compared with those of reference proteins of known secondary structural composition using two algorithms [14–16], a predominantly α-helical content of 62.5% (averaged data) for FsrC in the absence and presence of GBAP was revealed (Table 1; full data in Appendix B: Table B.1). This is in good agreement to that of the theoretical predicted α-helical content of 66% (obtained as the mean value derived using PREDATOR, PredictProtein, PSIPRED, SABLE, SAM and YAPSIN predictive tools — Appendix B: Table B.2). Importantly, Table 1 confirms that there is no significant effect of GBAP on FsrC secondary structure composition. Thermal studies showed very similar data in the presence and absence of the GBAP ligand with no change in the melting temperature (Tm) of FsrC (Fig. 3B, Appendix C: Fig. C.1). Comparisons of the peak/ trough maxima at each temperature reveal FsrC to be relatively thermostable, retaining 57–73% of secondary structure at 90 °C (Fig. 3C). Temperature-dependent loss of secondary structural integrity appears to be mainly irreversible however, since a return to 20 °C resulted in an insignificant recovery of structural integrity. In summary, there was no discernible effect of GBAP on the thermal stability of the protein. To determine whether GBAP binding affects the tertiary conformation of FsrC, SRCD measurements in the near-UV region encompassing the environments of the aromatic residues Tyr, Trp and Phe and of any disulphide bonds, were undertaken. A higher detergent concentration of 0.05% DDM (rather than 0.02%) was employed here, since stability/equilibration conditions are attained more rapidly (see above), and there are no significant changes in the near- or farUV spectra using this higher detergent level (Appendix C: Fig. C.2). Fig. 4A shows the near-UV spectra (250–340 nm) for intact FsrC in the presence and absence of a five-fold concentration of added GBAP.

1599

Table 1 Secondary structure analysis of purified FsrC in DDM micelles using the mean values derived from the CONTINLL and CDSSTR algorithms⁎. Secondary structure

H1: α-helix H2: distorted α-helix S1: β-strand S2: distorted β-strand T: turn Unordered

Composition FsrC

FsrC + GBAP⁎⁎

0.410 ± 0.017 0.203 ± 0.012 0.043 ± 0.006 0.037 ± 0.006 0.113 ± 0.012 0.190 ± 0.036

0.430 ± 0.017 0.21 ± 0.010 0.037 ± 0.006 0.033 ± 0.006 0.11 ± 0.010 0.183 ± 0.025

⁎ Using the SP43 and SP43 plus 5 denatured protein datasets (Appendix Table B.1). ⁎⁎ 2-fold molar excess of GBAP.

Signals from Phe occur within 255–270 nm, signals from Tyr predominate at 275–282 nm and those from Trp mainly occur within 285–305 nm [23,24], though there are overlapping contributions from the Tyr and Trp signals in these regions. The spectra shown in Fig. 4A are corrected for any signals arising from other sample components (buffer/0.05% DDM/acetonitrile) or from GBAP itself, as appropriate, and we therefore attribute the change in the spectrum to an interaction arising from the binding of GBAP to FsrC. The largest changes in the spectrum are in a wavelength region that corresponds with Trp and Tyr residues, so these residues in FsrC and/or in GBAP are likely to be involved in the binding interaction. To quantify the binding between FsrC and GBAP observed, a method was developed for titrating increasing concentrations of GBAP with FsrC into the sample cell and from such data to derive a kd value for the interaction. Fig. 4B shows the FsrC spectra obtained in response to increasing concentrations of GBAP up to a molar excess of 5:1. The program CD titration© v1.6 was used to calculate from these data the kd value, taking into account (i) any contributions of buffer, GBAP and acetonitrile by subtracting the relevant controls shown in Fig. 4C; and (ii) for each titration point, the dilution of the sample. The change in absorbance at 277 nm (Fig. 4B) following each ligand addition was calculated, and Appendix D: Fig. D.1 shows

Fig. 3. The effect of GBAP on FsrC and its thermal stability revealed through SRCD in the far UV region of the spectrum. (A) FsrC spectra in the presence and absence of two molar equivalents of GBAP at 20 °C. Solid line, FsrC + acetonitrile; dashed line, FsrC + GBAP. (B) Thermal stability of FsrC alone with no GBAP present. Spectra were obtained at each of the temperatures described in Methods. Spectra for FsrC with GBAP present (showing similar data) are shown in Appendix C.1.; (C) thermal stability of FsrC in the presence and absence of two-fold GBAP and expressed as % initial mean residue ellipticity obtained at 5 °C; solid line, FsrC (plus equivalent acetonitrile (1%)); dashed line, FsrC in the presence of two-fold GBAP (dissolved in acetonitrile). Spectra were obtained as described in Methods and are the average of four scans (data unsmoothed).

1600

S.G. Patching et al. / Biochimica et Biophysica Acta 1818 (2012) 1595–1602

Fig. 5. Estimation of the binding affinity (kd value) for GBAP binding to FsrC from SRCD titration measurements in the near-UV region at 20 °C. The titration and control spectra shown in Fig. 4 were used in the program CD titration© v.1.6 to calculate the change in absorbance at 277 nm following each GBAP addition. The experimental data were compared to binding curves matching a range of possible kd values; the experimental data fitted best to a model binding curve for a kd of 2 μM (shown). Comparisons of the experimental data with other model binding curves (kd values from 1.0 to 10.0 μM) are shown in Appendix D: Fig. D.1.

the full range of kd fits tested. The best fit was to a model binding curve with a kd value of 2 μM (Fig. 5). This binding affinity and the concentration range at which the curve reaches saturation is consistent with our GBAP-activated assays of autophosphorylation activity performed with purified FsrC [6,10]. 4. Discussion

Fig. 4. Detection and titration of GBAP binding to FsrC using SRCD measurements in the near-UV region at 20 °C. (A) Effect of GBAP binding; spectra for FsrC (20 μM) in the absence (solid line) and presence (dashed line) of five-fold molar excess of GBAP. The spectra shown are fully corrected for buffer/acetonitrile/GBAP signals by subtraction of the control spectra shown in (C). The reactions contained 20 μM FsrC, 0.05% DDM and either 0.36% acetonitrile or 100 μM GBAP with 0.36% acetonitrile. The spectral regions predominantly contributed by aromatic residues (Phe, Tyr, Trp) are shown in grey. (B) Titration spectra for GBAP binding. The grey dashed line indicates the 277 nm wavelength used for estimating the kd value. (C) Control spectra used for detection and titration of GBAP binding. Spectra are shown for the following separate samples: buffer alone; buffer + total acetonitrile at titration point 5:1 (buffer+ 0.36% acetonitrile); buffer +total GBAP at titration point 5:1 (buffer+ 100 μM GBAP with 0.36% acetonitrile); FsrC alone (buffer +20 μM FsrC); FsrC +total acetonitrile at titration point 5:1 (buffer + 20 μM FsrC+ 0.36% acetonitrile); FsrC +total GBAP at titration point 5:1 (buffer+ 20 μM FsrC+ 100 μM GBAP with 0.36% acetonitrile). All samples were stabilised as described in Methods. Unsmoothed data shown.

During the present study we have developed a robust method for obtaining reproducible SRCD spectra for intact purified FsrC solubilised in detergent micelles and for detecting and quantifying the binding of its pheromone ligand GBAP to the protein. We identified a number of important technical considerations that should be useful in general to investigate ligand binding by membrane proteins reconstituted in detergent micelles by SRCD spectroscopy. In particular, a sufficient incubation period following initial dilution of the protein in detergent to prepare the SRCD sample and following the addition of ligands is essential for achieving a stable sample and reproducible results. Stability was monitored by acquiring repeated scans until consecutive scans overlay satisfactorily. When we used FsrC in a buffer containing 0.02% DDM, an incubation period of at least 1.5 h was necessary for the sample to stabilise before acquiring spectra (Fig. 2), and 2 h was routinely adopted in this study. Such incubations were not required using 0.05% DDM for which 30 min were sufficient. In the titration experiments to obtain quantitative information on ligand binding an incubation period of 20 min following each addition of ligand was used. Ten repeated scans and an integration time of 1 s for acquiring each spectrum, which in total gives a time of approximately 75 min per titration point were carried out (Fig. 4). A sufficient number of scans to obtain a spectrum with a good signal-tonoise ratio is especially important for the titrations, particularly when small changes in the spectral intensity are observed. Small sample volume measurements are particularly advantageous for membrane protein studies in which protein yields are typically lower compared to those of most soluble proteins. Such low volume measurements in the near UV region for ligand titrations using a 2 mm × 2 mm window of a 1 cm pathlength cell were made possible by the B23 highly collimated beam [12]; these measurements would not have been feasible using a conventional benchtop instrument. Indeed, SRCD proved to be a useful screening method for determining secondary structure stability of FsrC in detergent micelles under different conditions, and will thus provide a valuable tool for identifying stabilised membrane protein preparations in low detergent conditions required for some downstream applications such as protein crystallisation or NMR structural measurements, where stability is

S.G. Patching et al. / Biochimica et Biophysica Acta 1818 (2012) 1595–1602

paramount (e.g. [25,26]). Our study highlights the importance of optimising detergent usage; indeed, previous SRCD measurements in the far-UV range have also identified the importance of detergent conditions for the stability of other membrane proteins, for reliable measurements of α-helical content and even for providing protein preparations in their bound or free states [21,27]. The present study is the first report of quantitative binding data derived by SRCD for an intact membrane sensor kinase with its ligand. SRCD and CD spectroscopy of membrane proteins has generally been used previously for confirmation of secondary and tertiary structural integrity of these proteins and protein domains (e.g. [4,5,20,28,29]) or their ligands including GBAP [18], and for investigations of protein stability (thermal- and urea-based denaturation) (e.g. [30,31]). For example, a near-UV CD- and fluorescence-based study of two small histidine kinases of Mycobacterium tuberculosis revealed their relatively high thermostability (which was also found to be the case here for FsrC (Fig. 3B and C)), and their characteristics of unfolding, including quantification of the energies involved [31]. However, to our knowledge there has been only one previous CD study of ligand binding by an intact membrane protein, which characterised interactions of the small multidrug resistance transporter protein SugE with its quaternary cation ligand, though again no titrations or binding data were reported [32]. Near-UV SRCD measurements revealed that GBAP binding affected the tertiary structure of purified intact FsrC resulting in changes in the environments of the aromatic residues (particularly Tyr and Trp) (Fig. 4). Indeed, the thermal stability data presented in Fig. 3C may suggest a small but discernible stabilising effect of the ligand (Fig. 3C). Native FsrC (51,507.4 Da) possesses three putative extracellular sensory domains that are likely to be involved in ligand binding (ESD1-3), separated by six putative transmembrane regions (TMs) and three relatively short intracellular regions (Fig. 1A) [33]. The protein possesses four Trp and eleven Tyr residues [7,9]; Trp56 and 118 are both predicted to reside within TM 2 and 4, respectively, though they are also in close proximity of (and could even therefore reside within) ESDs 1 and 2. The majority of the hydrophobic Tyr residues have predicted locations within the TMs, though Tyr98 and Tyr112 are predicted to reside in ESD2 (Fig. 1A). The SRCD signals of Trp and Tyr overlap and therefore it is difficult to identify which aromatic environments or indeed residues are affected by GBAP binding (Fig. 4). However, there are clearly several candidate Tyr and Trp residues affected, some of which have predicted locations within or close to ESDs 1 and 2 (Fig. 1A). The quorum sensor AgrC has a similar topology to that predicted for FsrC and reporter-based studies of AgrC activation and inhibition have identified ESDs 1 and 2 as specifically important in AIP pheromone recognition [3] and ESD 2 important for AIP specificity for each AgrC type [2]. However, the primary sequences of the staphylococcal AIPs and AgrC sensing domains exhibit little identity to FsrC and GBAP, though the possibility of general conserved similarities at the three-dimensional structural level for these small peptide pheromones is feasible. The Phe residues of FsrC have predicted locations mainly within the TMs but also within putative ESD1 (Phe47), ESD2 (Phe107 and Phe108) and ESD3 (Phe180 and Phe184) (Fig. 1A). However, we found little discernible effect of ligand binding on the total Phe signal (Fig. 4), suggesting that none of the Phe environments are significantly affected and that there is little conformational change in these environments. FsrC is sensitive to very low concentrations of GBAP in vivo and is activated in the presence of 1–5 nM GBAP in the environment [6,33]. By contrast, the kd for FsrC-GBAP interactions of 2 μM calculated in the present study demonstrates relatively loose binding of the activating ligand (Fig. 5). This result is consistent with the concentration range of GBAP required to activate purified FsrC reconstituted in DDM micelles in in vitro autophosphorylation assays [6]. Indeed, the micromolar kd value obtained here is also consistent with that reported for

1601

other sensory proteins involved in signal transduction. For example, it is comparable to the kd value of 5.5 μM determined for the isolated periplasmic binding domain of another sensor kinase, CitA, using isothermal titration calorimetry measurements [34]. Furthermore, it differs as expected with the kd values obtained in vitro for other membrane protein families such as the lower-affinity transport proteins in which millilmolar affinities are typical [35]. We cannot rule out the possibility that the conformation of the protein in detergent is not as in vivo, as the detergent conditions are different to those in the native membrane. This could account for the lower binding affinity compared with that reported for in vivo activation. However, there are also inherent limiting features of most in vivo binding experiments to be taken into account, such as ligand activities in the presence of different cell densities used in the binding assays. We note relevant similarities between our detergent system and native membrane systems reported previously, including association period for binding and equilibration, which provide additional confidence that detergent systems are suitable for providing data of in vivo and physiological relevance [36]. The ability to determine binding affinity data for intact sensor kinases using SRCD, as demonstrated for the first time in the present study, will be highly informative in future comparative studies such as mutagenesis to identify residues involved in ligand binding and in inhibitor screening. Acknowledgements We thank Lianne M. Davis (University of Leeds) for provision of early CD data on FsrC using a benchtop instrument at the University of Leeds. We gratefully acknowledge the BBSRC (BB/D001641/1) (to MKPJ) for funding and the Diamond Light Source Ltd for beamtime allocation. SGP acknowledges the EPSRC (EP/G035695/1) for financial support. Appendix A. Supplementary data Supplementary data to this article can be found online at doi:10. 1016/j.bbamem.2012.02.015. References [1] J.S. Wright III, G.J. Lyon, E.A. George, T.W. Muir, R.P. Novick, Hydrophobic interactions drive ligand-receptor recognition for activation and inhibition of staphylococcal quorum sensing, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 16168–16173. [2] E. Geisinger, E.A. George, T.W. Muir, R.P. Novick, Identification of ligand specificity determinants in AgrC, the Staphylococcus aureus quorum-sensing receptor, J. Biol. Chem. 283 (2008) 8930–8938. [3] R.O. Jensen, K. Winzer, S.R. Clarke, W.C. Chan, P. Williams, Differential recognition of Staphylococcus aureus quorum-sensing signals depends on both extracellular loops 1 and 2 of the transmembrane sensor AgrC, J. Mol. Biol. 381 (2008) 300–309. [4] Y.P. Kim, K.J. Yeo, M.H. Kim, Y.-C. Kim, Y.H. Jeon, Structural characterisation of the intra-membrane histidine kinase YbdK from Bacillus subtilis in DPC micelles, Biochem. Biophys. Res. Commun. 391 (2010) 1506–1511. [5] K.J. Yeo, S.-N. Kwak, H.J. Kim, C. Cheong, M.H. Kim, Y.H. Jeon, Expression and characterisation of the intergral membrane domain of bacterial histidine kinase SCO3062 for structural studies, Biochem. Biophys. Res. Commun. 376 (2008) 409–413. [6] P. Ma, H.M. Yuille, V. Blessie, N. Göhring, Z. Iglói, K. Nishiguchi, J. Nakayama, P.J.F. Henderson, M.K. Phillips-Jones, Expression, purification and activities of the entire family of intact membrane sensor kinases from Enterococcus faecalis, Mol. Membr. Biol. 25 (2008) 449–473. [7] I.T. Paulsen, L. Banerjei, G.S.A. Myers, K.E. Nelson, R. Seshadri, T.D. Read, D.E. Fouts, J.A. Eisen, S.R. Gill, J.F. Heidelberg, H. Tettelin, R.J. Dodson, L. Umayam, L. Brinkac, M. Beanan, S. Daugherty, R.T. DeBoy, S. Durkin, J. Kolonay, R. Madupu, W. Nelson, J. Vamathevan, B. Tran, J. Upton, T. Hansen, J. Shetty, H. Khouri, T. Utterback, D. Radune, K.A. Ketchum, B.A. Dougherty, C.M. Fraser, Role of mobile DNA in the evolution of vancomycin-resistant Enterococcus faecalis, Science 299 (2003) 2071–2074. [8] L.E. Hancock, M. Perego, Two-component signal transduction in Enterococcus faecalis, J. Bacteriol. 184 (2002) 5819–5825. [9] J. Nakayama, Y. Cao, T. Horii, S. Sakuda, A.D.L. Akkermans, W. de Vos, H. Nagasawa, Gelatinase biosynthesis-activating pheromone: a peptide lactone

1602

[10]

[11]

[12]

[13]

[14] [15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

S.G. Patching et al. / Biochimica et Biophysica Acta 1818 (2012) 1595–1602 that mediates a quorum sensing in Enterococcus faecalis, Mol. Microbiol. 41 (2001) 145–154. P. Ma, K. Nishiguchi, H.M. Yuille, L.M. Davis, J. Nakayama, M.K. Phillips-Jones, Anti-HIV siamycin I directly inhibits autophosphorylation activity of the bacterial FsrC quorum sensor and other ATP-dependent enzyme activities, FEBS Lett. 585 (2011) 2660–2664. J. Sambrook, E.F. Fritsch, T. Maniatis, in: C. Nolan (Ed.), Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1989. T. Javorfi, R. Hussain, D. Myatt, G. Siligardi, Measuring circular dichroism in a capillary cell using the B23 synchotron radiation CD beamline at Diamond Light Source, Chirality 22 (S1) (2010) E149–E153. G. Siligardi, B. Panaretou, P. Meyer, S. Singh, D.N. Woolfson, P.W. Piper, L.H. Pearl, C. Prodromou, Regulation of Hsp90 ATPase activity by the co-chaperone Cdc37p/p50, J. Biol. Chem. 277 (2002) 20151–20159. S.W. Provencher, J. Glockner, Estimation of globular protein secondary structure from circular dichroism, Biochemistry 20 (1981) 33–37. L.H.M. Van Stokkum, H.J.W. Spoelder, M. Bloemendal, R. Van Grondelle, F.C.A. Groen, Estimation of protein secondary structure and error analysis from CD spectra, Anal. Biochem. 191 (1990) 110–118. N. Sreerama, R.W. Woody, Estimation of protein secondary structure from circular dichroism spectra: comparison of CONTIN, SELCON and CDSSTR methods with an expanded reference set, Anal. Biochem. 287 (2000) 252–260. D.J. Freeman, G. Pattenden, A.F. Drake, G. Siligardi, Marine metabolites and metal chelation. CD studies of metal binding to Lissoclinum cyclopeptides, J. Chem. Soc. Perkin Trans. 2 (1998) 129–135. K. Nishiguchi, K. Nagata, M. Tanokura, K. Sonomoto, J. Nakayama, Structure-activity relationship of gelatinase biosynthesis-activating pheromone of Enterococcus faecalis, J. Bacteriol. 191 (2009) 641–650. J. Nakayama, Y. Cao, T. Horii, S. Sakuda, H. Nagasawa, Chemical synthesis and biological activity of the gelatinase biosynthesis activating pheromone of Enterococcus faecalis and its analogs, Biosci. Biotechnol. Biochem. 65 (2001) 2322–2325. C.A. Potter, A. Ward, C. Laguri, M.P. Williamson, P.J.F. Henderson, M.K. Phillips-Jones, Expression, purification and characterisation of the full-length histidine protein kinase RegB from Rhodobacter sphaeroides, J. Mol. Biol. 320 (2002) 201–213. O.V. Nolandt, T.H. Walther, S. Roth, J. Bürck, A.S. Ulrich, Structure analysis of the membrane protein TatCd from the Tat system of B. subtilis by circular dichroism, Biochim. Biophys. Acta 1788 (2009) 2238–2244. M.A. O'Malley, M.E. Helgeson, N.J. Wagner, A.S. Robinson, Toward rational design of protein detergent complexes: determinants of mixed micelles that are critical for the in vitro stabilisation of a G-protein coupled receptor, Biophys. J. 101 (2011) 1938–1948.

[23] S.M. Kelly, T.J. Jess, N.C. Price, How to study proteins by circular dichroism, Biochim. Biophys. Acta 1751 (2005) 119–139. [24] G. Siligardi, B. Samori, S. Melandri, M. Visconti, A.F. Drake, Correlations between biological activities and conformational properties of human, salmon, eel, porcine calcitonin and elcatonin elucidated by Circular Dichroism spectroscopy, Eur. J. Biochem. 221 (1994) 1117–1125. [25] C.G. Tate, Practical considerations of membrane protein instability during purification and crystallisation, Methods Mol. Biol. 601 (2010) 187–203. [26] S.G. Patching, NMR structures of polytopic integral membrane proteins, Mol. Membr. Biol. 28 (2011) 370–397. [27] A.J. Miles, A. Drechsler, K. Kristan, G. Anderluh, R.S. Norton, B.A. Wallace, F. Separovic, The effects of lipids on the structure of the eukaryotic cytolysin equinatoxin II: a synchotron radiation circular dichroism spectrscopic study, Biochim. Biophys. Acta 1778 (2008) 2091–2096. [28] B.A. Wallace, J.G. Lees, A.J.W. Orry, A. Lobley, R.W. Janes, Analyses of circular dichroism spectra of membrane proteins, Protein Sci. 12 (2003) 875–884. [29] G. Psakis, M. Saidijam, K. Shibayama, J. Polaczek, K.E. Bettaney, J.M. Baldwin, S.A. Baldwin, R. Hope, L.-O. Essen, R.C. Essenberg, P.J.F. Henderson, The sodiumdependent D-glucose transport protein of Helicobacter pylori, Mol. Microbiol. 71 (2009) 391–403. [30] N. Keegan, H. Ridley, J.H. Lakey, Discovery of biphasic thermal unfolding of OmpC with implications for surface loop stability, Biochemistry 49 (2010) 9715–9721. [31] R. Shrivastava, A.K. Das, Temperature and urea induced conformational changes of the histidine kinases from Mycobacterium tuberculosis, Int. J. Biol. Macromol. 41 (2007) 154–161. [32] D.C. Bay, R.J. Turner, Spectroscopic analysis of the intrinsic chromophores within small multidrug resistance protein SugE, Biochim. Biophys. Acta 1808 (2011) 2233–2244. [33] J. Nakayama, E. Tanaka, R. Kariyama, K. Nagata, K. Nishiguchi, R. Mitsuhata, Y. Uemura, M. Tanokura, H. Kumon, K. Sonomoto, Siamycin attenuates fsr quorum sensing mediated by a gelatinase biosynthesis-activating pheromone in Enterococcus faecalis, J. Bacteriol. 189 (2007) 1358–1365. [34] T. Gerharz, S. Reinelt, S. Kaspar, L. Scapozza, M. Bott, Identification of basic amino acids important for citrate binding by the periplasmic receptor domain of the sensor kinase CitA, Biochemistry 42 (2003) 5917–5924. [35] S.G. Patching, A.R. Brough, R.B. Herbert, A.R. Rajakarier, P.J.F. Henderson, D.A. Middleton, Substrate affinities for membranetransport proteins determined by 13C cross-polarisation magic-anglespinning NMR, J. Am. Chem. Soc. 126 (2004) 3072–3080. [36] D.S. Chapple, R. Hussain, C.L. Joannou, R.E.W. Hancock, E. Odell, R.W. Evans, G. Siligardi, Structure and association of human lactoferrin peptides with Escherichia coli lipopolysaccharides, Antimicrob. Agents Chemother. 48 (2004) 2190–2198.