Crystal structure of Rv3899c^184–410, a hypothetical protein from Mycobacterium tuberculosis

The crystal structure of Rv3899c^184–410 was found to contain a compact helix bundle and an α/β/α sandwich folding domain.

Abstract

Rv3899c is a hypothetical protein from Mycobacterium tuberculosis which is conserved across mycobacteria. It is predicted to be secreted and has been found in culture filtrates. It has been proposed as a potential vaccine candidate; however, its biological function is unknown. Here, the global structure of Rv3899c^184–410, a fragment of Rv3899c, is reported. The structure resembles the shell of a sea snail, and its N- and C-termini form two relatively independent compact domains: an α/β/α sandwich folding domain and an α-helix bundle domain. There are no reported protein structures for any Rv3899c homologues; this structure provides the first structural glimpse of a new protein family consisting of Rv3899c and its homologues.

1. Introduction  

Tuberculosis (TB), caused by infection with the pathogen Mycobacterium tuberculosis, remains a threat to global public health. It is estimated that there were 9.6 million cases of tuberculosis worldwide and 1.5 million deaths from TB in 2014 (World Health Organization, 2015 ▸). Effective TB prevention and control is thus extremely important and urgent. The Bacillus Calmette–Guérin (BCG) vaccine is the only licensed TB vaccine, but it provides insufficient protection (Moliva et al., 2015 ▸). Improving this vaccine, or developing more effective vaccines, is thus an important strategy for improved TB prevention and control. The secretome of M. tuberculosis is likely to be an ideal source of vaccine candidates (Zheng et al., 2013 ▸).

Rv3899c, annotated as a conserved hypothetical protein, is composed of 410 amino acids and has a predicted molecular mass of 40.8 kDa (Lew et al., 2011 ▸). Its function is unknown and it only has clear homologues in the genus Mycobacterium. Little else is known about this protein except that it is secreted; it has a twin-arginine transport (TAT) signal and has been identified in culture filtrates using two-dimensional PAGE combined with liquid chromatography-coupled MS/MS (Målen et al., 2007 ▸). In addition, it has been found to be present in M. tuberculosis H37Rv-infected guinea pig lungs at 90 d, but not 30 d, post-infection (Kruh et al., 2010 ▸). Deng and coworkers identified protein Rv3899c as a potential immunogen that can induce a strong immunological reaction, implying its potential as a vaccine candidate (Deng et al., 2014 ▸). Of interest, the Rv3899c gene is located in the vicinity of the ESX-2 operon, suggesting a possible functional relationship between Rv3899c and the ESX-2 system, one of five type VII or ESAT-6 secretion systems (ESX-1 to ESX-5) that are essential for mycobacterial virulence and/or viability (Houben et al., 2014 ▸; Newton-Foot et al., 2016 ▸).

As a first step to studying Rv3899c at the molecular level, we have purified Rv3899c and solved the structure of a fragment of this protein (amino-acid residues 184–410). Its structure resembles the shell of a sea snail in shape and consists of an α/β/α sandwich folding domain and a α-helix bundle domain.

2. Materials and methods  

2.1. Macromolecule production  

The expression, purification and crystallization of Rv3899c have been reported previously (Song et al., 2015 ▸). In brief, Rv3899c protein was expressed in Escherichia coli BL21(DE3) cells and purified first with an Ni²⁺-chelating column (GE Healthcare) followed by a Superdex G200 size-exclusion chromatography column (GE Healthcare).

2.2. Crystallization  

The protein was crystallized using a solution consisting of 0.2 M ammonium acetate, 0.1 M bis-tris pH 5.5, 25% PEG 3350. The crystallized protein was previously shown to be a fragment of Rv3899c comprising residues 184–410 (Song et al., 2015 ▸).

2.3. Data collection and processing  

Diffraction data for native Rv3899c^184–410 were collected on a Rigaku R-AXIS IV⁺⁺ image plate using Cu Kα radiation at 100 K, processed with iMosflm (Battye et al., 2011 ▸) and scaled with SCALA (Evans, 2006 ▸) in the CCP4 program suite (Winn et al., 2011 ▸), as reported previously (Song et al., 2015 ▸). An Hg²⁺-derivative crystal was obtained by soaking the native crystal with 1 mM HgCl₂ for 24 h. Data for the Hg²⁺-derivative crystal were collected on a Rigaku Saturn 944 HG CCD using Cu Kα radiation at 100 K with an oscillation step of 1°, 180 s exposure time per image and a crystal-to-detector distance of 60 mm. The data were then processed with XDS (Kabsch, 2010 ▸) and scaled with XSCALE (Kabsch, 2010 ▸). Diffraction data-collection and processing statistics are summarized in Table 1 ▸.

Table 1. Data collection and processing.

	Rv3899c184–410	Hg2+ derivative
Diffraction source	Rigaku MicroMax-007 HF	Rigaku MicroMax-007 HF
Wavelength (Å)	1.5418	1.5418
Temperature (K)	100	100
Detector	R-AXIS IV++	Saturn CCD
Crystal-to-detector distance (mm)	150	60
Rotation range per image (°)	1	1
Total rotation range (°)	180	360
Exposure time per image (s)	240	180
Space group	P212121	P212121
a, b, c (Å)	50.15, 54.97, 75.80	47.54, 54.01, 74.41
Mosaicity (°)	0.87	0.92
Resolution range (Å)	44.51–1.90 (2.00–1.90)	45.54–2.08 (2.14–2.08)
Total No. of reflections	109835 (10690)	132756 (2190)
No. of unique reflections	16928 (2179)	19928 (757)
Completeness (%)	100 (93.0)	90.0 (46.3)
Multiplicity	6.6 (4.8)	6.6 (2.9)
〈I/σ(I)〉	13.1 (3.5)	15.8 (2.0)
R r.i.m. (%)	10.0 (42.5)	10.8 (81.7)
Overall B factor from Wilson plot (Å2)	17.4	12.2

2.4. Structure solution and refinement  

The structure was solved by the single isomorphous replacement with anomalous scattering (SIRAS) method with AutoSol (Adams et al., 2010 ▸) in the PHENIX program suite (Adams et al., 2010 ▸) using both native and Hg²⁺-derivative data. Most of the model was built using AutoBuild in the PHENIX program suite and it was completed by manual modelling using Coot (Emsley et al., 2010 ▸). The model was refined with PHENIX (Adams et al., 2010 ▸), with the value of the parameter ‘X-ray data.remove outliers’ being TRUE (the default value) to exclude reflections with abnormal values that tend to reduce the performance of the refinement engine. Data-collection and model-refinement statistics are summarized in Table 2 ▸. The coordinate and structure-factor files have been deposited in the PDB under accession number 5imu. Structure figures were prepared using PyMOL v.1.8 (http://www.pymol.com).

Table 2. Structure refinement.

Resolution range (Å)	44.51–1.90
Completeness (%)	99.1
σ Cutoff	0
No. of reflections, working set	16078
No. of reflections, test set	850
Final R cryst	0.160
Final R free	0.190
No. of non-H atoms
Protein	1646
Ion	1
Water	348
Total	1995
R.m.s. deviations
Bonds (Å)	0.003
Angles (°)	0.616
Average B factors (Å2)
Protein	17.4
Ion	38.3
Water	28.2
Total	19.3
Ramachandran plot
Favoured regions (%)	98.60
Additionally allowed (%)	1.40

3. Results and discussion  

The Rv3899c protein is comprised of 410 amino acids and has a predicted molecular mass of 40.8 kDa. However, the stable crystallized protein studied here only comprised residues 184–410 and is referred to as Rv3899c^184–410 (Song et al., 2015 ▸). The crystal structure of Rv3899c^184–410 was determined here to 1.9 Å resolution, with R _cryst and R _free values of 0.160 and 0.190, respectively. The electron-density map was of good quality and residues 187–403, one potassium ion and 348 water molecules were built in the final model.

The structure resembles the shell of a sea snail (Fig. 1 ▸). It consists of eight α-helices and three β-strands arranged in the order α1–β1–β2–α2–β3–α3–α4–α5–α6–α7–α8 in the primary structure (Fig. 2 ▸). It can be divided into two relatively independent domains (Fig. 1 ▸ a): (i) an α/β/α sandwich folding domain located in the bottom half of the shell which consists of an antiparallel β-sheet formed by β-strands β1, β2 and β3, together with the peripheral helices α1, α2 and α3, and (ii) a compact helix bundle located in the upper half of the shell formed from helices α4–α8. In addition, a pocket, surrounded by helices α3, α7 and α8 and the loop connecting strands β2 and α2, was found at the interface of the two domains. The pocket has a surface area of 342 Ų and a volume of 359 ų as calculated using CASTp (Dundas et al., 2006 ▸). Negatively and positively charged regions are distributed at the top and bottom of this pocket, respectively. We propose that this pocket is a potential ligand-binding site and, based on the electrostatic potential surface surrounding the pocket, the ligand may be a protein with a complementary electrostatic potential surface and a ‘finger’ that inserts into the pocket.

Figure 1.

Overall structure of Rv3899c^184–410. Cartoon representation (a), topological diagram of the secondary structure (b) and the electrostatic potential surface (c) of Rv3899c^184–410. β-Strands and α-helices are coloured green and cyan, respectively. The proposed ligand-binding pocket is highlighted by a black ellipse.

Figure 2.

Sequence alignment and secondary-structure distribution of Rv3899c^184–410. Homologous proteins to Rv3899c from different species of Mycobacterium were identified by BLAST and seven were chosen for sequence-alignment analysis: M. avium (ABK67464), M. rhodesiae JS60 (EHB55683), M. kansasii 824 (ETZ97236), M. kyorinense (WP_045378944), M. marinum (WP_020731216), M. arupense (WP_046190023) and M. chimaera (WP_054585524). The multiple sequence alignment was performed with ClustalW (http://www.ebi.ac.uk/Tools/msa/clustalo/) and was displayed with secondary structures using ESPript (Robert & Gouet, 2014 ▸).

A search for homologous sequences in the Protein Data Bank using BLASTp (http://blast.ncbi.nlm.nih.gov/Blast.cgi) did not reveal any proteins with structures similar to Rv3899c^184–410 or to its N-terminal α/β/α sandwich lobe or C-terminal helix bundle. A subsequent search of the database of nonredundant protein sequences (nr) showed that Rv3899c (including its N-terminal α/β/α sandwich domain and C-terminal helix bundle) only has clear homologues in the genus Mycobacterium. Sequence alignment revealed that residues in the α/β/α sandwich fold are more conserved than those in the helix bundle, implying that the sandwich domain may play a more important functional role than the helix-bundle domain (Fig. 2 ▸).

In conclusion, we have solved the structure of the Rv3899c^184–410 protein, a secreted protein that has previously been identified as an antigen in culture filtrates. The structure reported here gives a first structural glimpse of a new protein family consisting of Rv3899c and its homologues. Functional studies are required to investigate its biochemical characteristics, its roles in host and bacterial cells, and its potential use as a vaccine candidate.

Supplementary Material

PDB reference: Rv3899c^184–410, 5imu