structural communications\(\def\hfill{\hskip 5em}\def\hfil{\hskip 3em}\def\eqno#1{\hfil {#1}}\)

Journal logoSTRUCTURAL BIOLOGY
COMMUNICATIONS
ISSN: 2053-230X
Volume 67| Part 9| September 2011| Pages 1129-1136

Solution structure of an arsenate reductase-related protein, YffB, from Brucella melitensis, the etiological agent responsible for brucellosis

aSeattle Structural Genomics Center for Infectious Disease, http://www.ssgcid.org , USA,bBiological Sciences Division, Pacific Northwest National Laboratory, Richland, Washington, USA,cDepartment of Medicine, University of Washington, Seattle, Washington, USA,dSeattle Biomedical Research Institute, Seattle, Washington, USA, and eDepartment of Medical Education and Biomedical Informatics and Department of Global Health, University of Washington, Seattle, Washington, USA
*Correspondence e-mail: garry.buchko@pnnl.gov

(Received 25 January 2011; accepted 19 February 2011; online 16 August 2011)

Brucella melitensis is the etiological agent responsible for brucellosis. Present in the B. melitensis genome is a 116-residue protein related to arsenate reductases (Bm-YffB; BR0369). Arsenate reductases (ArsC) convert arsenate ion (H2AsO4), a compound that is toxic to bacteria, to arsenite ion (AsO2), a product that may be efficiently exported out of the cell. Consequently, Bm-YffB is a potential drug target because if arsenate reduction is the protein's major biological function then disabling the cell's ability to reduce arsenate would make these cells more sensitive to the deleterious effects of arsenate. Size-exclusion chromatography and NMR spectroscopy indicate that Bm-YffB is a monomer in solution. The solution structure of Bm-YffB (PDB entry 2kok ) shows that the protein consists of two domains: a four-stranded mixed β-sheet flanked by two α-helices on one side and an α-helical bundle. The α/β domain is characteristic of the fold of thioredoxin-like proteins and the overall structure is generally similar to those of known arsenate reductases despite the marginal sequence similarity. Chemical shift perturbation studies with 15N-labeled Bm-YffB show that the protein binds reduced glutathione at a site adjacent to a region similar to the HX3CX3R catalytic sequence motif that is important for arsenic detoxification activity in the classical arsenate-reductase family of proteins. The latter observation supports the hypothesis that the ArsC-YffB family of proteins may function as glutathione-dependent thiol reductases. However, comparison of the structure of Bm-YffB with the structures of proteins from the classical ArsC family suggest that the mechanism and possibly the function of Bm-YffB and other related proteins (ArsC-YffB) may differ from those of the ArsC family of proteins.

1. Introduction

Brucella melitensis is a facultative intracellular bacterial pathogen that exhibits a host preference for goats and sheep. It is one of the Brucella species identified as being responsible for brucellosis, a zoonotic disease that causes abortions and stillbirths in animals and Malta fever in humans. The disease is transmitted to humans primarily via contact with infected animals (alive or dead) or the consumption of unpasteurized dairy products (Young, 1995[Young, E. J. (1995). Clin. Infect. Dis. 21, 283-289.]). In undeveloped regions of the Mediterranean, Asia, Africa and Latin America brucellosis infections are a major problem in livestock and human populations, causing severe economic hardship (Corbel, 1997[Corbel, M. J. (1997). Emerg. Infect. Dis. 3, 213-221.]). In humans, the disease is multi-systemic and the symptoms are nonspecific, including fever, chills, malaise, dementia, fatigue, headaches, nausea, vomiting and constipation. Infections can be successfully treated with antimicrobial agents that are able to penetrate well into the host cells, often a combination of doxycycline and streptomycin or doxycycline and rifampin for prolonged periods of time (Young, 1995[Young, E. J. (1995). Clin. Infect. Dis. 21, 283-289.]; Rubinstein et al., 1991[Rubinstein, E., Lang, R., Shasha, B., Hagar, B., Diamanstein, L., Joseph, G., Anderson, M. & Harrison, K. (1991). Antimicrob. Agents Chemother. 35, 1925-1927.]). There is much interest in the pathobiology of strains of Brucella because it is a potential agent of biological warfare and its pathogenicity is unique in that the organism does not display any obvious `classical' virulence factors (Moreno & Moriyon, 2002[Moreno, E. & Moriyon, I. (2002). Proc. Natl Acad. Sci. USA, 99, 1-3.]).

Arsenic, a naturally occurring metalloid element that is frequently abundant in the environment (Messens & Silver, 2006[Messens, J. & Silver, S. (2006). J. Mol. Biol. 362, 1-17.]), is a human carcinogen (Shi et al., 2004[Shi, H., Shi, X. & Liu, K. J. (2004). Mol. Cell. Biochem. 255, 67-78.]) that is also toxic to most forms of life. The toxic effects arise from the reactivity of arsenic ions with protein thiols. To counter the toxic effects, a family of arsenic detoxification enzymes has evolved that convert arsenate ion (H2AsO4), the highly reactive form of arsenic, to arsenite ion (AsO2), a compound that may be effectively transported outside the cell (Mukhopadhyay et al., 2002[Mukhopadhyay, R., Rosen, B. P., Phung, L. T. & Silver, S. (2002). FEMS Microbiol. Rev. 26, 311-325.]). In B. melitensis, YffB, a protein with marginal sequence similarity to the classical family of arsenate reductases (ArsC), is found that may play a role in reducing arsenate (DelVecchio et al., 2002[DelVecchio, V. G. et al. (2002). Proc. Natl Acad. Sci. USA, 99, 443-448.]). This 13.5 kDa protein (Bm-YffB) is a potential drug target because if arsenate reduction is this protein's major biological function, con­tributing to the organism's virulence, then disabling this protein and the cell's ability to reduce arsenate would make B. melitensis more sensitive to the deleterious effects of endogenous arsenate. Towards understanding the biological function of Bm-YffB and providing a blueprint for structure-based drug design (Myler et al., 2009[Myler, P. J., Stacy, R., Stewart, L. J., Staker, B. L., Van Voorhis, W. C. & Buchko, G. W. (2009). Infect. Disord. Drug Targets, 9, 493-506.]) based on this protein, the solution structure of Bm-YffB was determined. Its thermostability was measured by CD spectroscopy and its structure was compared with those of a similar protein, Pseudomonas aeruginosa YffB (Pa-YffB; PDB entry 1rw1 ; Teplyakov et al., 2004[Teplyakov, A., Pullalarevu, S., Obmolova, G., Doseeva, V., Galkin, A., Herzberg, O., Dauter, M., Dauter, Z. & Gilliland, G. L. (2004). BMC Struct. Biol. 4, 5.]), and a protein known to reduce arsenate, Escherichia coli ArsC (Ec-­ArsC; PDB entry 1id9 ; Martin et al., 2001[Martin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P. & Edwards, B. F. (2001). Structure, 9, 1071-1081.]).

2. Materials and methods

2.1. Cloning, expression and purification

The Bm-Yffb gene (BR0369; YP_4138591.1) was amplified using the genomic DNA of B. melitensis biovar Abortus 2308 and the oligonucleotide primers 5′-GGGTCCTGGTTCGATGAGTGTGA­CCATTTACGGCATC-3′ (forward) and 5′-CTTGTTCGTGCTG­TTTATTATAGCTTAAAATAAGCTTCATACTGCG-3′ (reverse) (Invitrogen, Carlsbad, California, USA). The amplified Bm-YffB gene was then gel-purified, treated with T4 DNA polymerase and annealed into the NruI/PmeI-digested expression vector AVA0421 at a site that provided a 21-residue tag containing six consecutive histidine residues (MAHHHHHHMGTLEAQTQGPGS-) at the N-­terminus of the expressed protein (Choi et al., 2011[Choi, R., Kelley, A., Leibly, D., Nakazawa Hewitt, S., Napuli, A. & Van Voorhis, W. (2011). Acta Cryst. F67, 998-1005.]). The recombinant plasmid was transformed into E. coli BL21(DE3)R3-pRARE2 cells (a gift from SGC Toronto, Toronto, Ontario, Canada) using a heat-shock method. Uniformly 15N- and 15N-,13C-labeled Bm-­YffB was obtained by growing the transformed cells (310 K) in minimal medium (Miller) containing 15NH4Cl (1 mg ml−1) and D-­[13C6]-glucose (2.0 mg ml−1) supplemented with FeCl3 (50 µg ml−1) and the antibiotics chloramphenicol (35 µg ml−1) and ampicillin (100 µg ml−1). Once the cells reached an OD600 of ∼0.8, the cells were cooled to 298 K and protein expression was induced with iso­propyl β-D-1-thiogalactopyranoside (0.026 µg ml−1). After approximately 5 h, the cells were harvested by mild centrifugation and frozen at 193 K. The frozen pellet was later thawed and resuspended in ∼35 ml lysis buffer (0.3 M NaCl, 50 mM sodium phosphate, 10 mM imidazole, pH 8.0) brought to 0.2 mM phenylmethylsulfonyl fluoride (PMSF) prior to three passes through a French press (SLM Instruments, Rochester, New York, USA). Following 60 s sonication (SLM Instruments, Rochester, New York, USA) the cell debris was pelleted by centrifugation at 25 000g for 1 h in a JA-20 rotor (Beckman Instruments, Fullerton, California, USA). The supernatant was then passed through a 0.45 µm syringe filter and applied onto an Ni–NTA affinity column (Qiagen, Valencia, California, USA) containing ∼25 ml resin. The column was washed stepwise by gravity with 40 ml buffer (0.3 M NaCl, 50 ml sodium phosphate pH 8.0) containing increasing con­centrations of imidazole (5, 10, 20, 50 and 250 mM). Bm-YffB eluted primarily in the 250 mM imidazole wash. Following exchange into 3C cleavage buffer by overnight dialysis in 4 l cleavage buffer (150 mM NaCl, 20 mM Tris–HCl, pH 8.4) the protein was concentrated to ∼2 ml (Amicon Centriprep-10) and the N-terminal polyhistidine tag was removed by overnight incubation with 3C protease (1 µg per 50 µg target protein) at 279 K (Bryan et al., 2011[Bryan, C. M., Bhandari, J., Napuli, A. J., Leibly, D. J., Choi, R., Kelley, A., Van Voorhis, W. C., Edwards, T. E. & Stewart, L. J. (2011). Acta Cryst. F67, 1010-1014.]). Using a flow rate of 1.0 ml min−1, the reaction solution was then loaded onto a Superdex 75 HiLoad 16/60 column (GE Healthcare, Piscataway, New Jersey, USA) to simultaneously purify the protein and exchange it into NMR buffer (100 mM NaCl, 20 mM Tris–HCl, 1.0 mM dithiothreitol, pH 7.1). The band containing Bm-YffB (retention time 78 min) was collected and the volume was reduced (Amicon Centriprep-10) to generate NMR samples in the 1–2 mM range (Lowry analysis). SDS–PAGE analysis of the final NMR samples showed the protein to be greater than ∼95% pure.

2.2. Circular dichroism spectroscopy

An Aviv Model 410 spectropolarimeter (Lakewood, New Jersey, USA) calibrated with an aqueous solution of ammonium D-(+)-camphorsulfonate was used to collect circular dichroism data from a 0.05 mM Bm-YffB sample in NMR buffer in a quartz cell of 0.1 cm path length. A thermal denaturation curve was obtained by recording the ellipticity at 216 nm in 2.0 K intervals from 283 to 353 K. A quantitative estimation of the melting temperature, Tm, was obtained by taking a first derivative of the thermal denaturation curve using the Aviv software (Greenfield, 2006[Greenfield, N. J. (2006). Nature Protoc. 6, 2527-2535.]). Steady-state wavelength spectra for Bm-YffB were recorded in 0.5 nm increments between 200 and 260 nm at 298 and 353 K. Each reported steady-state wavelength spectrum was the result of averaging two consecutive scans with a bandwidth of 1.0 nm and a time constant of 1.0 s. These spectra were processed by subtracting a blank spectrum from the protein spectrum and then automatically line-smoothing the data using the Aviv software.

2.3. Nuclear magnetic resonance spectroscopy

Varian 800-, 750- and 600-Inova spectrometers equipped with 1H/13C/15N triple-resonance probes and pulse-field gradients were used to collect the NMR data required for resonance assignments and structure determination. The NMR data, which were collected from 1–2 mM samples at 293 K, were processed with Felix2007 (Felix NMR Inc., San Diego, California, USA) and analyzed with Sparky (v.3.115; Goddard & Kneller, 2008[Goddard, T. D. & Kneller, D. G. (2008). Sparky 3. University of California, San Francisco, USA. http://www.cgl.ucsf.edu/home/sparky/ .]). Assignments of the 1H, 13C and 15N chemical shifts for the backbone and side-chain resonances were made from standard 2D 1H–15N HSQC, 1H–13C HSQC, HBCBC­GCDHD and HBCBCGCDCHE experiments and 3D HNCACB, CBCA(CO)NH, HNCO, HCC-TOCSY-NNH and CC-TOCSY-NNH experiments using Varian Protein Pack pulse programs. Chemical shifts were referenced to DSS (DSS = 0 p.p.m.) using indirect methods (Wishart et al., 1995[Wishart, D. S., Bigam, C. G., Yao, J., Abildgaard, F., Dyson, H. J., Oldfield, E., Markley, J. L. & Sykes, B. D. (1995). J. Biomol. NMR, 6, 135-140.]). Distance restraints were obtained from a suite of 3D 13C- and 15N-edited NOESY-HSQC experiments using a mixing time of 80 ms. Deuterium-exchange studies were performed by lyophilizing a 15N-­labeled NMR sample, re-dissolving it in 99.8% D2O and immediately collecting 1H–15N HSQC spectra 10, 20 and 60 min after exchange. An overall rotational correlation time, tc, was estimated from backbone-amide 15N T1/T1ρ ratios (Farrow et al., 1994[Farrow, N. A., Muhandiram, R., Singer, A. U., Pascal, S. M., Kay, C. M., Gish, G., Shoelson, S. E., Pawson, T., Forman-Kay, J. D. & Kay, L. E. (1994). Biochemistry, 33, 5984-6003.]; Buchko et al., 2008[Buchko, G. W., Tarasevich, B. J., Bekhazi, J., Snead, M. L. & Shaw, W. J. (2008). Biochemistry, 47, 13215-13222.]). A chemical shift perturbation experiment was performed by adding aliquots of reduced glutathione in NMR buffer (50 mM) to a 0.5 mM sample of 15N-­labeled Bm-YffB. Following gentle agitation, 1H–15N HSQC spectra were collected at glutathione:Bm-YffB molar ratios of 0.3:1, 0.6:1, 1:1 and 2:1.

2.4. Structure calculations

The chemical shifts for Bm-YffB were assigned using conventional methods (Cavanagh et al., 1996[Cavanagh, J., Palmer, A. G., Fairbrother, W. & Skelton, N. (1996). NMR Spectroscopy: Principles and Practice. San Diego: Academic Press.]) and were deposited in the Biological Magnetic Resonance Data Bank (BMRB) under accession No. 16517. Using these 1H, 13C and 15N chemical shift assignments and the peak-picked data from 13C- and 15N-edited NOESY-HSQC experiments as initial inputs, structure calculations were performed iteratively using CYANA (v.2.1; Güntert, 2004[Güntert, P. (2004). Methods Mol. Biol. 278, 353-378.]). 184 dihedral angle restraints for both φ and ψ were introduced on the basis of the elements of secondary structure identified in the early structural ensembles and TALOS calculations (Cornilescu et al., 1999[Cornilescu, G., Delaglio, F. & Bax, A. (1999). J. Biomol. NMR, 13, 289-302.]). Near the end of the iterative process, 84 hydrogen-bond restraints (1.8–2.0 and 2.7–3.0 Å for the NH—O and N—O distances, respectively) were introduced into the structure calculations on the basis of proximity in early structure calculations and the observation of slowly exchanging amides in the deuterium-exchange experiment. On the basis of the chemical shift difference between the proline 13Cβ and 13Cγ atoms (Schubert et al., 2002[Schubert, M., Labudde, D., Oschkinat, H. & Schmieder, P. (2002). J. Biomol. NMR, 24, 149-154.]), Pro93 was placed in the cis conformation. From the final set of 100 calculated structures, the 20 with the lowest target function were selected and this ensemble was deposited in the Protein Data Bank (PDB) under PDB code 2kok . Structural quality was assessed using the Protein Structure Validation Suite (PSVS; v.1.3; Bhattacharya et al., 2007[Bhattacharya, A., Tejero, R. & Montelione, G. T. (2007). Proteins, 66, 778-795.]). Note that the deposited structures were not refined with explicit water (Linge et al., 2003[Linge, J. P., Williams, M. A., Spronk, C. A., Bonvin, A. M. & Nilges, M. (2003). Proteins, 50, 496-506.]) because these calculations con­tinuously introduced unfavorable steric clashes into the structures despite numerous attempts to adjust the parameters. A summary of the structure statistics is provided in Table 1[link].

Table 1
Summary of the structural statistics for Bm-YffB

All statistics are for the 20-structure ensemble deposited in the PDB (2kok ).

Restraints for structure calculations
 Total NOEs 1490
 Intraresidue NOEs 436
 Sequential (ii + 1) NOEs 449
 Medium-range (ii + j; 1 < j ≤ 4) NOEs 282
 Long-range (ii + jj > 4) NOEs 323
φ-angle restraints 94
ψ-angle restraints 94
 Hydrogen-bond restraints 84
Structure calculations
 No. of structures calculated 100
 No. of structures used in ensemble 20
Structures with restraint violations
 Distance restraint violations > 0.05 Å 0
 Dihedral restraint violations > 2° 0
R.m.s.d. from mean (Å)
 Backbone N—Cα—C=O atoms 0.69 ± 0.16
 Heavy atoms 1.19 ± 0.11
Ramachandran plot summary for selected residues (from PROCHECK)
 Most favored regions (%) 89.8
 Additionally allowed regions (%) 10.2
 Generously favored regions (%) 0.0
 Disallowed regions (%) 0.0
Global quality scores for ordered residues§
PROCHECK (all) −2.84 (−0.48)
PROCHECK (φ, ψ) −0.28 (−0.15)
MolProbity clash score −0.86 (13.91)
†Calculated for the ordered residues Val3–Ile8, Cys11–Ala59, Thr61–Asp98 and Lys100–Lys115.
‡Calculated for the central ordered core, Ser2–Ile8 and Cys11–Lys115, using PSVS.
§Z scores; values in parentheses are raw values.

The amino-acid sequence of Bm-YffB deposited in the PDB and BMRB is numbered sequentially, Gly1–Leu120, starting with the four non-native residues (GPGS-) that remained after 3C protease treatment. However, here the first four non-native residues are numbered sequentially with asterisks (Gly1*–Ser4*) and the first native residue, Met5 in the PDB and BMRB depositions, is labeled Met1.

3. Results and discussion

3.1. Solution structure of Bm-ArsC

The elution time of Bm-YffB on a Superdex 75 HiLoad 16/60 column was within a range consistent with a monomeric ∼14 kDa protein (data not shown). Such a conclusion was corroborated by the experimentally estimated rotational correlation time determined for Bm-YffB, 9.1 ± 0.2 ns (293 K), a value that is more consistent with a monomeric ∼14 kDa protein than an ∼28 kDa dimer (Bhattacharjya et al., 2004[Bhattacharjya, S., Xu, P., Gingras, R., Shaykhutdinov, R., Wu, C., Whiteway, M. & Ni, F. (2004). J. Mol. Biol. 344, 1071-1087.]). The line widths and chemical shift dispersion of the 1H–15N HSQC spectrum for Bm-YffB, shown in Fig. 1[link], were also characteristic of a folded monomeric protein with a molecular weight in the 15 kDa range. As illustrated in Fig. 1[link], 121 of the expected 123 amide resonances for Bm-YffB [120 − (6 prolines + Gly1*)] were assigned in the 1H–15N HSQC spectrum. Amide cross-peaks for Asp16 and Phe109 were not unambiguously identified. On the basis of these amide assignments and extensive assignment of the 13Cα and side-chain proton and carbon chemical shifts (BMRB ID 16517), an ensemble of structures was calculated (Fig. 2[link]a) that satisfied all of the available experimental NMR data (NOEs, chemical shifts, deuterium-exchange experiments and TALOS calculations).

[Figure 1]
Figure 1
Assigned 1H–15N HSQC spectrum of double-labeled Bm-YffB collected at 293 K in NMR buffer (100 mM NaCl, 20 mM Tris–HCl, 1.0 mM DTT, pH 7.1) at a 1H resonance frequency of 750 MHz. Side-chain –NH2 resonances are indicated by dashed horizontal lines (red) and the exchangeable ring resonances for Trp20 and Trp53 are identified with an `r'.
[Figure 2]
Figure 2
(a) Superposition of the cartoon representations of the ensemble of structures calculated for Bm-YffB (PDB entry 2kok ), with α-helices colored blue and β-strands colored gold. (b) Cartoon representation of the structure most similar to the average structure of the ensemble, with the four β-strands and seven α-helices labeled.

As summarized in Table 1[link], a total of 1490 interproton distance restraints, 84 hydrogen-bond restraints and 188 dihedral angle restraints were used in the final structure calculations. Each member of the final ensemble of 20 calculated structures agreed well with the experimental data, with no upper limit violation of greater than 0.05 Å and no torsion-angle violation of greater than 2°. Analysis of the ensemble with the PSVS validation-software package (Bhattacharya et al., 2007[Bhattacharya, A., Tejero, R. & Montelione, G. T. (2007). Proteins, 66, 778-795.]) also showed that the quality of the structures in the final ensemble was good. The Ramachandran statistics for all residues in the ensemble were overwhelmingly in acceptable space [90% of the (φ, ψ) pairs for Bm-YffB were found in the most favored regions and 10% were within additionally allowed regions] and all the structure-quality Z scores were acceptable (>−5).

The final set of 20 calculated structures in the ensemble converge well, as shown mathematically by the statistics in Table 1[link] and visually by the superposition in Fig. 2[link](a). The r.m.s.d.s of the structured core regions (Val3–Ile8, Cys11–Ala59, Thr61–Asp98 and Lys100–Lys115) in the ensemble from the mean structure are 0.7 Å for the backbone atoms (N—Cα—C=O) and 1.2 Å for all heavy atoms. Fig. 2[link](b) illustrates the single structure in the ensemble that is nearest to the mean structure and Fig. 3[link] shows a stereoview of this single structure. The protein consists of two domains: a four-stranded mixed β-sheet flanked by two α-helices on one face and an α-helical bundle. The α/β domain is composed of a β4–β3–β1–β2 β-sheet, with β4–β3–β1 aligned antiparallel and β1–β2 aligned parallel. α-Helix 1 (Asp12–His24) is tucked in behind β1–β2, and α7 (Pro107–Phe114) is tucked in behind β4–β3. The α/β domain is characteristic of the fold of thioredoxin-like proteins and usually contains a cis-proline on the N-­terminal side of β3 that plays a role in the integrity of the active site (Martin, 1995[Martin, J. L. (1995). Structure, 3, 245-250.]). Bm-YffB also contains a proline, Pro93, at this position and analysis of the 13Cβ and 13Cγ chemical shifts for this residue indicates that it is also in the cis conformation (Table 2[link]; Schubert et al., 2002[Schubert, M., Labudde, D., Oschkinat, H. & Schmieder, P. (2002). J. Biomol. NMR, 24, 149-154.]). The other domain in Bm-YffB is an α-helical bundle dominated by α3 (Ala40–Thr49) and α6 (Ala76–Ala85), with three short helices [α2 (Tyr33–Glu36), α4 (Thr61–Lys65) and α5 (Glu68–Ser72)] around them. Hydrophobic interactions between the side chains of residues on the interface of these two domains, including Phe46 (α3), Thr49 (α3), Tyr6 (β1) and Leu101 (β4), assist in holding the domains together. As shown in Fig. 4[link], the protein con­tains a polarized distribution of charges on its surface, with a dominance of positive surfaces. Such a distribution of charges has been observed in the ArsC protein family, as well as for Pa-YffB, and would favor the binding of anions such as arsenate (Teplyakov et al., 2004[Teplyakov, A., Pullalarevu, S., Obmolova, G., Doseeva, V., Galkin, A., Herzberg, O., Dauter, M., Dauter, Z. & Gilliland, G. L. (2004). BMC Struct. Biol. 4, 5.]).

Table 2
Chemical shift difference between the proline 13Cβ and 13Cγ atoms in Bm-YffB

Residue Δβγ (p.p.m.)
Pro2* 5.0
Pro51 5.0
Pro67 4.6
Pro87 4.2
Pro93 8.7
Pro107 4.2
Average for trans-Pro 4.51 ± 1.17
Average for cis-Pro 9.64 ± 1.27
†Average values obtained from Schubert et al. (2002[Schubert, M., Labudde, D., Oschkinat, H. & Schmieder, P. (2002). J. Biomol. NMR, 24, 149-154.]).
[Figure 3]
Figure 3
Stereoview showing a cartoon representation of the structure most similar to the average structure in the ensemble, with the protein rainbow-colored (ROYGBIV) from the N-terminus to the C-terminus.
[Figure 4]
Figure 4
Maps generated using PyMOL (DeLano, 2002[DeLano, W. L. (2002). PyMOL. http://www.pymol.org .]) of the electrostatic potential at the solvent-accessible surface of Bm-YffB. The long axis of the protein is illustrated and is sequentially rotated 90° about the horizontal axis four times.

3.2. Circular dichroism profile and thermal stability of Bm-ArsC

Circular dichroism (CD) spectroscopy is very sensitive to changes in a protein's backbone and, consequently, is a powerful tool to rapidly probe the conformation of proteins in solution and to assess the effect of variables such as pH, salt content and temperature on the structure of a protein (Woody, 1974[Woody, R. W. (1974). In Peptides, Polypeptides and Proteins, edited by E. R. Blout, F. A. Bovey, M. Goodman & N. Lotan. New York: John Wiley & Sons.]). Fig. 5[link](a) shows the steady-state CD spectrum of Bm-YffB collected at 298 K. The dominant feature of the spectrum is characteristic of α-helical secondary structure: a double minimum at approximately 220 and 208 nm and an extrapolated maximum around 195 nm (Holzwarth & Doty, 1965[Holzwarth, G. M. & Doty, P. (1965). J. Am. Chem. Soc. 87, 218-228.]; Greenfield, 2006[Greenfield, N. J. (2006). Nature Protoc. 6, 2527-2535.]). Such an observation is expected given the amount of helical structure (46%) observed in the solution structure of the protein (Figs. 2[link] and 3[link]). Note that the double minimum is skewed and is more intense around 220 nm, which is likely to be due to the contributions of other elements of secondary structure to the CD steady-state spectrum.

[Figure 5]
Figure 5
(a) Circular dichroism steady-state wavelength spectrum for Bm-YffB (0.05 mM) in NMR buffer collected at 298 K. (b) The CD thermal melt for Bm-YffB obtained by measuring the ellipticity at 216 nm in 2.0 K intervals between 283 and 353 K. (c) The first derivative of the thermal melt curve shows that the protein has a melting temperature of 326.6 K.

By monitoring the increase in the ellipticity at a specific wavelength with increasing temperature, the thermal stability of a protein may be measured and a melting temperature (Tm) estimated for the transition between a structured and an unstructured state (Karantzeni et al., 2003[Karantzeni, I., Ruiz, C., Liu, C.-C. & Licata, V. J. (2003). Biochem. J. 374, 785-792.]). As shown in Fig. 5[link](b), a gradual increase in ellipticity at 220 nm is observed for Bm-YffB up to ∼323 K, followed by a more rapid increase in ellipticity that tails off and plateaus at ∼333 K. Visual inspection of the sample after heating to 353 K showed evidence of precipitation, indicating that the unfolding was irreversible and that the CD data may not be analyzed thermodynamically (Karantzeni et al., 2003[Karantzeni, I., Ruiz, C., Liu, C.-C. & Licata, V. J. (2003). Biochem. J. 374, 785-792.]). However, a quantitative estimation of the Tm for this transition may still be obtained by assuming a two-state model and taking a first derivative of the curve shown in Fig. 5[link](b) (Greenfield, 2006[Greenfield, N. J. (2006). Nature Protoc. 6, 2527-2535.]). The maximum of this first derivative, shown in Fig. 5[link](c), is 326.6 K.

3.3. Comparison with related structures

The PDB was searched for structures similar to Bm-YffB using the DALI search engine (Holm & Rosenström, 2010[Holm, L. & Rosenström, P. (2010). Nucleic Acids Res. 38, W545-W549.]). The search indicated that the structure of Bm-YffB was most similar (Z score = 13.4) to that of the conserved hypothetical protein YffB from P. aeruginosa (PDB entry 1rw1 ; Teplyakov et al., 2004[Teplyakov, A., Pullalarevu, S., Obmolova, G., Doseeva, V., Galkin, A., Herzberg, O., Dauter, M., Dauter, Z. & Gilliland, G. L. (2004). BMC Struct. Biol. 4, 5.]). This similarity is evident in Fig. 6[link], which shows a superposition of the two structures with the program SuperPose (Maiti et al., 2004[Maiti, R., van Domselaar, G. H., Zhang, H. & Wishart, D. S. (2004). Nucleic Acids Res. 23, W590-W594.]). It is evident that both proteins have the same number of β-strands and α-helices organized into a similar two-domain structure. Indeed, using Bm-YffB residues Ser2–Lys115 the backbone r.m.s.d. is 3.78 Å between Bm-YffB and the equivalent region in Pa-YffB. Such similarity between the structures of the two proteins is reasonable given that the amino-acid sequences of the two proteins are 56% identical and 70% similar.

[Figure 6]
Figure 6
Superposition of the structure closest to the mean for Bm-YffB (PDB entry 2kok , blue) on the crystal structure determined for Pa-YffB (PDB entry 1rw1 , magenta; Teplyakov et al., 2004[Teplyakov, A., Pullalarevu, S., Obmolova, G., Doseeva, V., Galkin, A., Herzberg, O., Dauter, M., Dauter, Z. & Gilliland, G. L. (2004). BMC Struct. Biol. 4, 5.]) using the program SuperPose (Maiti et al., 2004[Maiti, R., van Domselaar, G. H., Zhang, H. & Wishart, D. S. (2004). Nucleic Acids Res. 23, W590-W594.]).

After Pa-YffB, the DALI search identified a number of proteins with Z scores between 11.3 and 10.0 that were annotated as arsenate reductases (ArsC) or Spx regulatory proteins. While the structure of Bm-YffB produced high DALI Z scores with these two types of proteins, the sequence identity between Bm-YffB and these proteins was only 14–27%. The Spx protein is a global transcription regulatory protein that does not bind DNA as part of its regulatory role but instead binds to the α-subunit of RNA polymerase to control gene expression in response to disulfide-stress conditions (Nakano et al., 2005[Nakano, S., Erwin, K. N., Ralle, M. & Zuber, P. (2005). Mol. Microbiol. 55, 498-510.]). In Spx it is hypothesized that disulfide-bond formation in a CXXC motif triggers events that result in an increase in the cellular levels of thioredoxin and thioredoxin reductase (Nakano et al., 2003[Nakano, S., Küster-Schöck, E., Grossman, A. D. & Zuber, P. (2003). Proc. Natl Acad. Sci. USA, 100, 13603-13608.]). Because Bm-YffB is a monomer and contains only a single cysteine residue, it is unlikely that it functions as a regulatory protein using a mechanism similar to that which may be employed by Spx. On the other hand, Bm-YffB contains a sequence that is somewhat similar to the HX3CX3R catalytic sequence motif that is important for arsenic detoxification activity in the arsenate-reductase family of proteins (ArsC; Martin et al., 2001[Martin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P. & Edwards, B. F. (2001). Structure, 9, 1071-1081.]); hence, it is more likely that Bm-YffB has a function related to arsenate reduction.

Fig. 7[link] shows a superposition of a representative structure of Bm-YffB (PDB entry 2kok ) with the crystal structure of E. coli ArsC (Ec-ArsC) in the native form (PDB entry 1id9 ; Martin et al., 2001[Martin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P. & Edwards, B. F. (2001). Structure, 9, 1071-1081.]) using the program SuperPose (Maiti et al., 2004[Maiti, R., van Domselaar, G. H., Zhang, H. & Wishart, D. S. (2004). Nucleic Acids Res. 23, W590-W594.]). Apart from an extra two-stranded β-sheet at the C-terminus of Ec-ArsC, the two structures are generally similar to each other: a four-stranded mixed β-sheet is flanked by a number of α-helices that generate a similar overall fold. One major difference between the two structures, highlighted in Fig. 7[link], is the presence of a triad of arginine residues, Arg60, Arg94 and Arg106 (red), in Ec-ArsC that is absent in Bm-YffB. These three arginines have been reported to be essential for enzymatic function, forming a thiasahydroxyl adduct when bound to arsenate (Martin et al., 2001[Martin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P. & Edwards, B. F. (2001). Structure, 9, 1071-1081.]). In the Bm-YffB structure only one arginine, Arg96 (black), is found in this area. Such a major structural difference suggests a difference in substrate specificity between Bm-YffB and Ec-ArsC and it has been postulated that the ArsC-YffB family of proteins might be able to bind glutathione in the absence of arsenate (Teplyakov et al., 2004[Teplyakov, A., Pullalarevu, S., Obmolova, G., Doseeva, V., Galkin, A., Herzberg, O., Dauter, M., Dauter, Z. & Gilliland, G. L. (2004). BMC Struct. Biol. 4, 5.]).

[Figure 7]
Figure 7
Superposition of the structure closest to the mean for Bm-YffB (PDB entry 2kok , blue) on the crystal structure determined for E. coli ArsC (PDB entry 1id9 , magenta; Martin et al., 2001[Martin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P. & Edwards, B. F. (2001). Structure, 9, 1071-1081.]) using the program SuperPose (Maiti et al., 2004[Maiti, R., van Domselaar, G. H., Zhang, H. & Wishart, D. S. (2004). Nucleic Acids Res. 23, W590-W594.]). The side chains of the three arginine residues in Ec-ArsC (Arg60, Arg94 and Arg107) that form a thiasahydroxyl adduct that is essential for enzymatic function are labeled and colored red. The only residue in Bm-YffB that is equivalent to those in the arginine triad of Ec-ArsC is Arg96; this side chain is colored black.

3.4. Chemical shift perturbation studies with reduced glutathione

In support of the hypothesis that the ArsC-YffB family of proteins may bind glutathione in the absence of arsenate ions, it has been shown by mass spectrometry that P. aeruginosa YffB can bind glutathione (Teplyakov et al., 2004[Teplyakov, A., Pullalarevu, S., Obmolova, G., Doseeva, V., Galkin, A., Herzberg, O., Dauter, M., Dauter, Z. & Gilliland, G. L. (2004). BMC Struct. Biol. 4, 5.]). One of the advantages of structure determination by NMR-based methods is that following the chemical shift assignment of the 1H–15N HSQC spectrum, not only is it possible to quickly detect ligand binding to the protein by chemical shift perturbation experiments, but, it is also possible to map the location of the binding interaction on the three-dimensional structure of the protein (Buchko et al., 1999[Buchko, G. W., Daughdrill, G. W., de Lorimier, R., Rao, B. K., Isern, N. G., Lingbeck, J. M., Taylor, J. S., Wold, M. S., Gochin, M., Spicer, L. D., Lowry, D. F. & Kennedy, M. A. (1999). Biochemistry, 38, 15116-15128.]; Zuiderweg, 2002[Zuiderweg, E. R. (2002). Biochemistry, 41, 1-7.]). This is possible because protein–ligand interactions often manifest as changes in the chemical environment of the nuclei at the interface of ligand binding which are accompanied by perturbations in the measurable chemical shifts of the backbone 1HN and 15N resonances. Upon identifying the resonances that experience a binding-dependent chemical shift or intensity perturbation, it is possible to map the location of the binding site onto the structure of the protein. Fig. 8[link] is the result of such an experiment following the titration of reduced glutathione into a 15N-­labeled sample of Bm-YffB. At a 2:1 molar ratio of reduced glutathione:Bm-YffB, a subset of seven resonances are observed to shift. In Fig. 9[link] these perturbed resonances are mapped onto the structure of Bm-YffB and it is observed that they all cluster into one area (red) adjacent to the potential GX3CX3K catalytic sequence motif (yellow). The latter observation adds support to the hypothesis that Bm-YffB and other proteins in the YffB family may function as glutathione-dependent thiol reductases.

[Figure 8]
Figure 8
Overlay of the 1H–15N HSQC spectrum of 15N-labeled Bm-YffB (black) on the spectrum collected in the presence of a 2:1 molar ratio of glutathione:Bm-YffB (magenta). Residues that were significantly perturbed are labeled. The spectrum at an ∼1:1 molar ratio of glutathione:Bm-YffB was similar to that shown here at a 2:1 molar ratio. Data were collected at a proton resonance frequency of 600 MHz at 293 K.
[Figure 9]
Figure 9
Surface representation of the structure of Bm-YffB shown with a labeled transparent cartoon representation. Regions that may potentially be important for the function of the protein are labeled. The lone equivalent residue to those present in the arginine triad of Ec-ArsC, Arg96, is colored black, residues in the potential GX3CX3K catalytic sequence motif are colored yellow and the residues perturbed by the addition of reduced glutathione (labeled in Fig. 8[link]) are colored red.

4. Conclusions

The solution structure determined for Bm-YffB is similar to the crystal structure reported for the YffB protein from P. aeruginosa: a four-stranded mixed β-sheet flanked by two α-helices on one side and an α-helical bundle. While the protein sequences of Bm-YffB and Pa-­YffB are marginally similar to the amino-acid sequences of classical arsenate reductases (ArsC), comparison of the structures of YffB proteins and ArsC proteins suggest that the substrate specificities and mechanisms of these two families of proteins may differ. Chemical shift perturbation studies with reduced glutathione corroborate the hypothesis that YffB proteins may function as glutathione-dependent thiol reductases. Further biochemical and biophysical studies will be necessary in order to identify the substrate and work out the details of the mechanism before Bm-YffB may be exploited for structure-based drug design. The solution structure and bio­physical data presented here for Bm-YffB will facilitate these efforts.

Supporting information


Acknowledgements

This research was funded by the National Institute of Allergy and Infectious Diseases, National Institute of Health, Department of Health and Human Services under Federal Contract No. HHSN272200700057C. The SSGCID internal ID for Bm-YffB is BrabA.00007.a. The majority of the research presented here was conducted at the W. R. Wiley Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the US Department of Energy's Office of Biological and Environmental Research (BER) program located at Pacific Northwest National Laboratory (PNNL). Battelle operates PNNL for the US Department of Energy.

References

First citationBhattacharjya, S., Xu, P., Gingras, R., Shaykhutdinov, R., Wu, C., Whiteway, M. & Ni, F. (2004). J. Mol. Biol. 344, 1071–1087.  Web of Science CrossRef PubMed CAS Google Scholar
First citationBhattacharya, A., Tejero, R. & Montelione, G. T. (2007). Proteins, 66, 778–795.  Web of Science CrossRef PubMed CAS Google Scholar
First citationBryan, C. M., Bhandari, J., Napuli, A. J., Leibly, D. J., Choi, R., Kelley, A., Van Voorhis, W. C., Edwards, T. E. & Stewart, L. J. (2011). Acta Cryst. F67, 1010–1014.  Web of Science CrossRef IUCr Journals Google Scholar
First citationBuchko, G. W., Daughdrill, G. W., de Lorimier, R., Rao, B. K., Isern, N. G., Lingbeck, J. M., Taylor, J. S., Wold, M. S., Gochin, M., Spicer, L. D., Lowry, D. F. & Kennedy, M. A. (1999). Biochemistry, 38, 15116–15128.  Web of Science CrossRef PubMed CAS Google Scholar
First citationBuchko, G. W., Tarasevich, B. J., Bekhazi, J., Snead, M. L. & Shaw, W. J. (2008). Biochemistry, 47, 13215–13222.  Web of Science CrossRef PubMed CAS Google Scholar
First citationCavanagh, J., Palmer, A. G., Fairbrother, W. & Skelton, N. (1996). NMR Spectroscopy: Principles and Practice. San Diego: Academic Press.  Google Scholar
First citationChoi, R., Kelley, A., Leibly, D., Nakazawa Hewitt, S., Napuli, A. & Van Voorhis, W. (2011). Acta Cryst. F67, 998–1005.  Web of Science CrossRef IUCr Journals Google Scholar
First citationCorbel, M. J. (1997). Emerg. Infect. Dis. 3, 213–221.  CrossRef CAS PubMed Google Scholar
First citationCornilescu, G., Delaglio, F. & Bax, A. (1999). J. Biomol. NMR, 13, 289–302.  Web of Science CrossRef PubMed CAS Google Scholar
First citationDeLano, W. L. (2002). PyMOL. http://www.pymol.orgGoogle Scholar
First citationDelVecchio, V. G. et al. (2002). Proc. Natl Acad. Sci. USA, 99, 443–448.  Web of Science CrossRef PubMed CAS Google Scholar
First citationFarrow, N. A., Muhandiram, R., Singer, A. U., Pascal, S. M., Kay, C. M., Gish, G., Shoelson, S. E., Pawson, T., Forman-Kay, J. D. & Kay, L. E. (1994). Biochemistry, 33, 5984–6003.  CrossRef CAS PubMed Web of Science Google Scholar
First citationGoddard, T. D. & Kneller, D. G. (2008). Sparky 3. University of California, San Francisco, USA. http://www.cgl.ucsf.edu/home/sparky/Google Scholar
First citationGreenfield, N. J. (2006). Nature Protoc. 6, 2527–2535.  Google Scholar
First citationGüntert, P. (2004). Methods Mol. Biol. 278, 353–378.  PubMed Google Scholar
First citationHolm, L. & Rosenström, P. (2010). Nucleic Acids Res. 38, W545–W549.  Web of Science CrossRef CAS PubMed Google Scholar
First citationHolzwarth, G. M. & Doty, P. (1965). J. Am. Chem. Soc. 87, 218–228.  CrossRef PubMed CAS Web of Science Google Scholar
First citationKarantzeni, I., Ruiz, C., Liu, C.-C. & Licata, V. J. (2003). Biochem. J. 374, 785–792.  CrossRef PubMed CAS Google Scholar
First citationLinge, J. P., Williams, M. A., Spronk, C. A., Bonvin, A. M. & Nilges, M. (2003). Proteins, 50, 496–506.  Web of Science CrossRef PubMed CAS Google Scholar
First citationMaiti, R., van Domselaar, G. H., Zhang, H. & Wishart, D. S. (2004). Nucleic Acids Res. 23, W590–W594.  Web of Science CrossRef Google Scholar
First citationMartin, J. L. (1995). Structure, 3, 245–250.  CrossRef CAS PubMed Web of Science Google Scholar
First citationMartin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P. & Edwards, B. F. (2001). Structure, 9, 1071–1081.  Web of Science CrossRef PubMed CAS Google Scholar
First citationMessens, J. & Silver, S. (2006). J. Mol. Biol. 362, 1–17.  Web of Science CrossRef PubMed CAS Google Scholar
First citationMoreno, E. & Moriyon, I. (2002). Proc. Natl Acad. Sci. USA, 99, 1–3.  Web of Science CrossRef PubMed CAS Google Scholar
First citationMukhopadhyay, R., Rosen, B. P., Phung, L. T. & Silver, S. (2002). FEMS Microbiol. Rev. 26, 311–325.  Web of Science CrossRef PubMed CAS Google Scholar
First citationMyler, P. J., Stacy, R., Stewart, L. J., Staker, B. L., Van Voorhis, W. C. & Buchko, G. W. (2009). Infect. Disord. Drug Targets, 9, 493–506.  CrossRef PubMed CAS Google Scholar
First citationNakano, S., Erwin, K. N., Ralle, M. & Zuber, P. (2005). Mol. Microbiol. 55, 498–510.  Web of Science CrossRef PubMed CAS Google Scholar
First citationNakano, S., Küster-Schöck, E., Grossman, A. D. & Zuber, P. (2003). Proc. Natl Acad. Sci. USA, 100, 13603–13608.  Web of Science CrossRef PubMed CAS Google Scholar
First citationRubinstein, E., Lang, R., Shasha, B., Hagar, B., Diamanstein, L., Joseph, G., Anderson, M. & Harrison, K. (1991). Antimicrob. Agents Chemother. 35, 1925–1927.  CrossRef PubMed CAS Google Scholar
First citationSchubert, M., Labudde, D., Oschkinat, H. & Schmieder, P. (2002). J. Biomol. NMR, 24, 149–154.  Web of Science CrossRef PubMed CAS Google Scholar
First citationShi, H., Shi, X. & Liu, K. J. (2004). Mol. Cell. Biochem. 255, 67–78.  Web of Science CrossRef PubMed CAS Google Scholar
First citationTeplyakov, A., Pullalarevu, S., Obmolova, G., Doseeva, V., Galkin, A., Herzberg, O., Dauter, M., Dauter, Z. & Gilliland, G. L. (2004). BMC Struct. Biol. 4, 5.  Google Scholar
First citationWishart, D. S., Bigam, C. G., Yao, J., Abildgaard, F., Dyson, H. J., Oldfield, E., Markley, J. L. & Sykes, B. D. (1995). J. Biomol. NMR, 6, 135–140.  CrossRef CAS PubMed Google Scholar
First citationWoody, R. W. (1974). In Peptides, Polypeptides and Proteins, edited by E. R. Blout, F. A. Bovey, M. Goodman & N. Lotan. New York: John Wiley & Sons.  Google Scholar
First citationYoung, E. J. (1995). Clin. Infect. Dis. 21, 283–289.  CrossRef CAS PubMed Google Scholar
First citationZuiderweg, E. R. (2002). Biochemistry, 41, 1–7.  Web of Science CrossRef PubMed CAS Google Scholar

This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.

Journal logoSTRUCTURAL BIOLOGY
COMMUNICATIONS
ISSN: 2053-230X
Volume 67| Part 9| September 2011| Pages 1129-1136
Follow Acta Cryst. F
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds