Introduction

Tuberculosis (TB) caused by Mycobacterium tuberculosis (Mtb) remains one of the leading causes of death worldwide. According to statistical data from the World Health Organization (WHO), about 10.4 million people became ill with TB and 1.4 million died from the disease in 2015 (WHO 2016). Over the past 40 years, an effective four-drug combination regimen of isoniazid, rifampicin, pyrazinamide, and ethambutol has been developed and employed as the first choice for TB treatment. However, long-term use of a limited number of drugs has led to significant non-compliance, side effects, and drug resistance (Riccardi et al. 2009). The prevalence of multidrug-resistant (MDR) and extensively drug resistant (XDR) Mtb strains has become a new serious public health problem in many countries (Lauzardo and Peloquin 2012). Thus, there is an urgent need to find new anti-TB drugs with novel targets.

Ferredoxins are acidic, low molecular weight, soluble iron-sulfur proteins that exist in diverse organisms. Ferredoxins mainly act as electron transfers in redox systems that are involved in a range of important metabolic processes such as cytochrome P450-associated reactions (i.e., hydroxylation, epoxidation, peroxidation, sulfoxidation, dealkylation, or deamination of a substrate) (Ewen et al. 2011). In bacteria, a classic cytochrome P450 mono-oxygenase (CYP) system generally consists of a ferredoxin, a ferredoxin-NAD(P)H reductase (FNR), and the CYP. Electrons required for a CYP catalytic reaction are transferred from a reduced nicotinamide cofactor, such as NADH or NADPH, to the CYP via FNR and ferredoxin (Werck-Reichhart and Feyereisen 2000). The general electron transfer pathway for a bacterial CYP system follows as NAD(P)H → FNR → ferredoxin → CYP (Ewen et al. 2011). The genome sequence of Mtb revealed that the pathogen encodes 20 different CYPs with five ferredoxins, two FNRs (FdrA and FprA), and two FDRX (fusion protein of FNR and ferredoxin) as their putative electron transfer partners (Munro et al. 2003).

In recent studies, some of the Mtb CYPs and their redox partners were found to be essential for the pathogen’s viability or pathogenicity and have been suggested to be potential candidates for Mtb drug targeting (Fischer et al. 2002; Ouellet et al. 2010; Zhang et al. 2006). However, identification of the correct redox partners for CYPs in Mtb remains a big challenge since only a minority of CYPs and CYP-related redox proteins have been structurally and/or functionally characterized, and as redox protein-coding genes normally are distant from most of the CYP genes at their genomic loci. In addition, heterologous expression of Mtb CYPs or putative redox partner proteins in other host system, such as in Escherichia coli, has often remained unsuccessful and ineffective. This has further impeded characterization of the redox system for the pathogen (McLean et al. 2006a, b). Among the five putative ferredoxins, including FdX (Rv0763c), Rv1786 (Rv1786), FdxA (Rv2007c), FdxC (Rv1177), and FdxD (Rv3503c), in Mtb pathogenic strain H37Rv, to date, only FdX has been identified as a redox partner to support catalysis with CYP51, while all the rest ferredoxins have remained poorly characterized for their structures or functions (Ouellet et al. 2010).

Rv1786, a putative 3Fe-4S ferredoxin encoded by H37Rv gene Rv1786, is located downstream of CYP143A1 (Rv1785c) and has been considered a potential natural electron transfer partner protein for CYP143A1. Despite the fact that little is currently known about the structure or possible functions of CYP143A1, it appears to be highly conserved across the Mycobacterium genus including pathogenic and non-pathogenic species by phylogenetic analysis, which suggests the CYP enzyme may perform important general functions in Mtb. Hence, given the adjacent location of Rv1786 to CYP143A1, characterizing the structural and biochemical properties of Rv1786 and validating the members of the Rv1786-based redox couples would be very useful for further understanding of the potential catalytic activity of the CYP enzyme.

In this study, Rv1786 from the highly pathogenic Mtb strain H37Rv was heterologously expressed in E. coli, purified as a His-tagged construct and characterized by SDS-PAGE, western blot, MALDI-TOF/TOF-MS, UV/Visible (UV/Vis), and electron paramagnetic resonance (EPR) spectroscopic analysis. To reveal possible reductase partners of Rv1786 in redox coupling system, interactions between Rv1786 and each of FdrA, FprA, and CYP143A1 were evaluated by surface plasmon resonance (SPR) assays and steady state kinetic studies. Our results show that although both reductase proteins appeared to be involved in transferring electron from Rv1786, FprA had a stronger binding affinity for Rv1786, and FdrA had a better coupling activity with Rv1786. To the best of our knowledge, this is the first report on the preliminary structural and biochemical characterization of Mtb Rv1786 as a redox partner.

Materials and methods

Template DNA, strains, and plasmid

Genomic DNA from Mtb strain H37Rv, which was kindly provided by Dr. J. Xu and Y. Lu, Department of Pharmacology, Beijing Tuberculosis and Thoracic Tumor Research Institute, Beijing Chest Hospital, was used as template DNA in this study. The Mtb strain H37Rv was deposited at the American Type Culture Collection as ATCC 25618. E. coli strains DH5α, BL21 (DE3), and Rosetta (DE3) as well as plasmid pET30a were used for cloning and expression. The E. coli strains DH5α and BL21 (DE3) were obtained from TaKaRa (Dalian, China), and the E. coli strain Rosetta (DE3) was obtained from TransGen Biotech (Beijing, China). The plasmid pET30a was purchased from Novagen (Madison, WI, USA). E. coli strains transformed with plasmids were cultured in LB (Luria-Bertani) medium (Oxoid, Basingstoke, UK) supplemented with appropriate antibiotics. Oligonucleotide primers for gene amplification, and containing appropriate restriction sites for cloning (Table 1), were synthesized by Shanghai Invitrogen (Shanghai, China).

Table 1 Oligonucleotide primers used in plasmid construction
Full size table

DNA manipulations

General DNA manipulations were carried out following standard protocols as described by Sambrook et al. (Sambrook and Russell 2006). Plasmid DNA was isolated with the EasyPure HiPure Plasmid MiniPrep Kit (TransGen Biotech, Beijing, China). PCR amplification products were purified from agarose gels with the EasyPure Quick Gel Extraction Kit (TransGen Biotech, Beijing, China).

Sequence alignment and phylogenetic analysis

All ferredoxin protein sequences were obtained from NCBI database. Since Rv1786 belongs to the putative ferredoxin with [3Fe-4S] cluster, all protein sequence alignments of ferredoxins with [3Fe-4S] cluster were analyzed. The resulting output sequences for various species were aligned and saved in ClustalX 2.0 format (Takagi et al. 2005). A phylogenetic tree was generated using MEGA 5.0 software (Oluwatoyin Japhet et al. 2012). The aligned ClustalX 2.0 format sequences were imported into the MEGA phylogeny tool, and a phylogenetic tree was generated using the “construct/test maximum likelihood” method.

Cloning of the genes coding for Rv1786, FdrA, FprA, and CYP143A1

Mtb gene Rv1786 (GeneID 885846) was amplified from the genomic DNA of Mtb strain H37Rv (NC_000962.2) by PCR using PrimeSTAR® HS DNA Polymerase with GC buffer (TaKaRa, Dalian, China). Primers used for amplification of Rv1786 gene are listed in Table 1. The forward primer contained a NdeI restriction site permitting cloning into the start codon (ATG) of pET30a, while the reverse primer was designed to produce an in-frame tetrahistidine C-terminal tag (underlined) and a stop codon (TGA) followed with a Hind III restriction site. The amplification conditions were 98 °C for 3 min; then 30 cycles of 98 °C for 15 s, 56 °C for 15 s, and 72 °C for 20 s; and finally, 72 °C for 10 min. The amplicon was ligated into a similarly restricted pET30a vector and introduced by transformation into E. coli Rosetta (DE3) (TransGen Biotech, Beijing, China), and resulting kanamycin- and chloramphenicol-resistant colonies containing the cloned Rv1786 were confirmed by PCR and sequencing. One of these clones, pET30a-Rv1786, was chosen for further study.

Mtb genes for FdrA (Rv0688), FprA (Rv3106), and CYP143A1(Rv1785c) were also amplified from the genomic DNA of Mtb H37Rv with primers in Table 1 by PCR using PrimeSTAR® HS DNA Polymerase with GC buffer (TaKaRa, Dalian, China). The amplification conditions of FdrA and FprA were 98 °C for 3 min; then 30 cycles of 98 °C for 15 s, 56 °C for 15 s, and 72 °C for 90 s; and finally, 72 °C for 10 min. The amplification conditions of CYP143A1 were different from the other genes, as an extra process called touchdown PCR was added before the normal process. The total conditions of CYP143A1 were 98 °C for 2 min, then 20 cycles of 98 °C for 10 s, 75 °C for 15 s, and 1 °C minus per cycle, then 68 °C for 2 min, followed by a normal PCR process. All the amplicons were ligated into corresponding restricted pET30a vector and transformed into E. coli BL21 (DE3) (TaKaRa, Dalian, China), and resulting kanamycin-resistant colonies containing the cloned FdrA, FprA, or CYP143A1 were confirmed by colony PCR and sequencing, respectively. Positive clones were chosen for further study.

Expression, purification, and spectra analysis of Rv1786, FdrA, FprA, and CYP143A1

Transformant E. coli Rosetta (DE3) cells containing pET30a-Rv1786 were incubated in LB medium supplemented with 30 μg/mL kanamycin and 34 μg/mL chloramphenicol at 37 °C until the OD600 reading was up to 0.6. Then, 0.5 mM isopropyl 1-thio-β-D-galactopyranoside (IPTG) was added to induce target gene expression. The transformant cells were then incubated for a further 4 h. The cells were harvested by centrifugation at 6000×g for 10 min at 4 °C, washed with 0.85% NaCl twice, and resuspended in PBS (pH 7.4). The cells were lysed on ice by ultrasonication for 20 min with cycles of 2-s sonication followed by 6-s rest. After centrifugation at 12,000×g, 4 °C for 20 min, the supernatants were loaded onto Ni-IDA-Sefinose™ Resin (Sangon Biotech, Shanghai, China). The brown-colored fractions were eluted with a linear gradient of 20–500 mM imidazole. Each eluted fraction was analyzed by SDS-PAGE. Fractions containing relatively pure protein were pooled and concentrated to a volume of approximately 2 mL using ultrafiltration with Amicon concentrators (Merck Millipore, MA, USA) at 4 °C. After dialysis against 50 mM Tris-HCl (pH 7.4) overnight to remove excess imidazole, the fractions were loaded onto Q Sefinose™ FF agarose (Sangon Biotech, Shanghai, China) and eluted with a gradient NaCl/Tris-HCl buffer (pH 6.0). Samples were pooled and concentrated to 100 μL by ultrafiltration using Amicon concentrators at 1500×g. The concentrated samples were further dialyzed into 50 mM Tris-HCl (pH 7.4). A final purification step was performed for the concentrated protein via a Superdex 75 10/300GL using an AKTA purification system (GE Healthcare, Pittsburgh, PA). Eluted fractions were again analyzed by SDS-PAGE. The fractions were pooled and concentrated as before, and dialyzed into 50 mM Tris-HCl, pH 7.4, and stored at −80 °C until use.

Procedures used for the expression and purification of FdrA, FprA, and CYP143A1 were the same as described for Rv1786 above, except that the induction of the expression for FdrA and FprA was carried out with 1 mM IPTG at 22 and 25 °C, respectively. The expression of CYP143A1 (Rv1785c) was induced with 0.75 mM IPTG at 22 °C in the presence of the heme precursor delta-aminolevulinic acid (ΔALA; 100 μM).

SDS-PAGE and UV/Vis spectrophotometry analyses were performed to identify the purity, correct molecular weight, and recognizable absorption spectrum of the purified proteins containing iron-sulfur cluster. The UV/Vis spectroscopic analysis of the samples was carried out in 96-well plate (Corning, MA, USA). All the proteins were examined by PerkinElmer EnSpire® 2300 Multimode Plate Readerz (PerkinElmer, MA, USA) in potassium phosphate buffer (pH 7.4). The wavelength range used for spectral scanning was 260 to 700 nm.

Western blot analysis and EPR spectroscopic analysis of Rv1786

Western blot analysis was performed to confirm the correct expression of His4-tagged purified Rv1786 protein. One microgram of purified protein was loaded in 12% gel and separated by SDS-PAGE. The sponge and filter paper were presoaked in transfer buffer (25 mM Tris base, 190 mM glycine, 20% methanol). A polyvinylidene fluoride (PVDF; pore size 0.22 μm; Millipore, MA, USA) membrane was cut to the same size of the gel and submerged into 10 mL of 100% methanol for 5 min. The protein was transferred onto the PVDF membrane in transfer buffer at 250 mA for 90 min at room temperature (RT). The blotted PVDF membrane was incubated in 5% skim milk in Tris-buffered saline (TBS) with Tween-20 (TBST) at RT for 1 h with shaking to block non-specific binding. The membrane was then incubated overnight at 4 °C with rabbit polyclonal anti-His4 tag antibody (Sangon Biotech, Shanghai, China) diluted 1:500 v/v with TBST containing 5% skim milk. After a washing step, the membrane was incubated with horse radish peroxidase (HPR)-conjugated goat anti-rabbit IgG antibody (Sigma, St. Louis, MO, USA) diluted 1:5000 v/v in 5% skim milk TBST. Subsequently, the membrane was stained using 3,3′,5,5′-tetramethylbenzidine (TMB; Sigma, St. Louis, MO, USA) as the electrochemical substrate of HPR.

Test samples of the purified Mtb Rv1786 protein for EPR spectroscopy were prepared at a concentration of 1.25 mg/mL in 50 mM Tris buffer. EPR spectra of Rv1786 protein were recorded at 10 K using a Bruker E580 spectrometer operated at X-band with a modulation frequency of 9.371 GHz and an incident microwave power of 2.377 mW. The sweep time was set as 120 s and modulation amplitude (g) was 1.000 G. Rv1786 spectra in both the oxidized and sodium dithionite-reduced forms were recorded.

MALDI-TOF/TOF-MS of Rv1786

The molecular mass of the purified Mtb Rv1786 protein was determined using a Bruker Ultraflex III MALDI-TOF/TOF-MS spectrometer equipped with Smart-beam laser (200 Hz). The matrix-assisted laser desorption/ionization MALDI-TOF/TOF MS result was obtained in the liner positive mode with an accelerating voltage of 20 kV. The protein samples were dissolved in Tris-HCl to a concentration of 1.25 mg/mL (pH 7.4). Theoretical molecular mass of the protein was calculated by the Compute pI/Mw tool.

SPR analysis

The protein-protein interactions between Rv1786 and each of FprA, FdrA, and CYP143A1 were analyzed by PlexArray HT (V3 version) label-free biomolecular interaction detection system (Plexera Bioscience, Seattle, WA, USA), which is based on SPR biosensing technology as described previously (Hu et al. 2014; Wu et al. 2015; Zhu et al. 2014). Rv1786 was dissolved in 25 mM Tris-HCl (pH 8.0) to achieve a concentration of 0.5 mg/mL, and immobilized onto a Nanocapture sensor chip (Plexera Bioscience, Seattle, WA, USA) by QArray mini spotter (Genetix Ltd., New Milton, Hampshire, UK) at 25 °C with humidity at 65%. Each sample was spotted in replicate on the chip, and each spot contained 0.1 μL of sample solution. During each measurement cycle, PBS running buffer (10 mM, pH 6.4) was first flowed through the chip surface at a rate of 2 μL/s to obtain a stable baseline, and the purified FprA, FdrA, or CYP143A1 was then injected as analyte at various concentrations in the running buffer for detecting their binding affinity to Rv1786 fixed on the chip surface. For binding measurements, four different concentrations of each analyte protein (500, 1000, 2000, and 4000 nM) were used to flow through the chip at a rate of 2 μL/s with an association time of 300 s and a dissociation time of 300 s. After each measurement cycle, the chip surface was regenerated with a solution of glycine-HCl (pH 2.0) followed by a PBS running buffer washing to remove bound proteins which enables the sensor chip to be reused for additional analyte injections. Real-time binding signals were recorded and analyzed by Data Analysis Module (Plexera Bioscience, Seattle, WA, USA). All possible binding curves were fitted by Langmuir equation, and the constants characterizing the association rate (or the on-rate, k a or k on), the dissociation rate (or the off-rate, k d or k off), and the equilibrium binding (K D ) were generated. In the SPR measurements, parallelly spotted mouse IgG (mIgG; Plexera Bioscience, Seattle, WA, USA) and bovine serum albumin (BSA; Plexera Bioscience, Seattle, WA, USA) on the chip with anti-mIgG (Plexera Bioscience, Seattle, WA, USA) flowing past were used as positive and negative controls, respectively.

Catalytic activity assay for FdrA or FprA in the presence of Rv1786

The FdrA- or FprA-mediated reduction of cytochrome c in the presence of Rv1786 was measured in 50 mM potassium phosphate buffer (pH 7.5). The reaction mixture contained cytochrome c (Sangon Biotech, Shanghai, China), different molar ratio of FdrA:Rv1786 or FprA:Rv1786 (1:0, 1:1, 1:2, 1:4), and various concentrations of NAD(P)H (0–150 μM). The catalytic activity of Rv1786-dependent electron transferring towards cytochrome c was detected in a 96-well plate and monitored by PerkinElmer EnSpire® 2300 Multimode Plate Reader at 550 nm as described previously (Fischer et al. 2002; Zanno et al. 2005). All reactions were initiated by adding the substrate, NAD(P)H, and recorded in triplicate over a period of 1 min. Steady-state kinetic parameters for each cytochrome c assay were determined with a constant concentration of cytochrome c (500 μM) and varying the concentrations of NADH or NADPH. Unless otherwise stated, NADH and NADPH were substrates for FdrA-Rv1786 and FprA-Rv1786 reaction systems, respectively (Fischer et al. 2002; Zanno et al. 2005). All data manipulation and analysis were performed using SigmaPlot 13.0.

Results

Sequence similarity analysis of [3Fe-4S] ferredoxins across Mycobacterium genus

As indicated in Fig. 1a, b, the sequence similarity and a phylogenetic tree of the amino acid sequences of Rv1786 were analyzed in different mycobacterial species. The highlighted regions in Fig. 1 display a CXXHXXC(X)nCP motif, which is a typical iron-sulfur cluster-binding domain found in [3Fe-4S] cluster-containing ferredoxins. An alignment of the Rv1786 orthologs revealed that Rv1786 is highly conserved in different Mtb strains (100% identity) and among closely related M. tuberculosis complex (MTBC) species, which includes Mycobacterium canettii (99%), Mycobacterium africanum (99%), and Mycobacterium bovis (96%). Although sequence identities of Rv1786 orthologs remain considerable high in other pathogenic mycobacterial species such as Mycobacterium kansasii (86%) and Mycobacterium marinum (80%), lower similarities were obtained in a non-pathogenic mycobacterial member, Mycobacterium smegmatis (37%). When aligning with Rv1786 orthologs from bacteria outside the genus Mycobacterium including Rhodopseudomonas palustris (36%), Bacillus thermoproteolyticus (31%), and Terrabacter species (23%), sequence identities matched low (Shirakawa et al. 2005; Takagi et al. 2005; Zhang et al. 2014). In addition, Rv1786 shares low similarities with FdxD (32%) and FdX (31%), the other two [3Fe-4S] cluster-containing ferredoxins expressed in Mtb H37Rv (Fig. 1a, b). Rv1786 also shows low sequence similarity with Fdmor, a [3Fe-4S] cluster-containing ferredoxin encoded by part of morABC operon region from Mycobacterium sp. strain HE5 (29%) and an intracellular redox sensor WhiB3 (31%) from Mtb H37Rv (Sielaff et al. 2001; Singh et al. 2007).

Fig. 1
figure 1

Sequence alignment and phylogenetic analysis for Rv1786. a Alignment of the amino acid sequences of [3Fe-4S] ferredoxins in different species. b Phylogenetic tree illustrating the evolutionary relationship between Rv1786 and other [3Fe-4S] ferredoxins in different species. Bacterial species and the DDBJ/EMBL/GenBank reference sequences of the [3Fe-4S] ferredoxins included M. tuberculosis H37Rv (Rv1786 NP_216302, Fdx AFN48639, FdxD NP_218020, WhiB3 NP_217933), M. tuberculosis CTRI-2 (ferredoxin AEN00239), M. bovis (ferredoxin WP_019284145), M. africanum (ferredoxin WP_031669679), M. canettii (ferredoxin WP_014000948), M. kansasii (ferredoxin WP_023363981), M. marinum (ferredoxin WP_011740946), M. parascrofulaceum ATCC BAA-614 (ferredoxin EFG79933), M. vanbaalenii PYR-1 (ferredoxin AAS49916), M. sp. HE5 (Fdmor AAV54065), M. abscessus (ferredoxin CPY35196), M. smegmatis (ferredoxin WP_003897302), R. palustris Haa2 (HaPuxC 4ID8_A). Baci. thermoproteolyticus (ferredoxin 1WTF-A), Ter. sp. DBF63 (DbfA3 BAJ39939), Sulf. tokodaii str. 7 (ferredoxin BAB67388)

Full size image

Heterologous expression and spectroscopic properties of Rv1786, FdrA, FprA, and CYP143A1

Mtb Rv1786, FdrA, FprA, and CYP143A1 were expressed in the T7-based vector pET30a with C-terminal His-tag in E. coli as described in the “Materials and methods” section. All the His-tagged proteins could be obtained in soluble forms from E. coli lysates, and all the affinity-purified fractions from the nickel-charged resin displayed a brown color, which indicates the presence of iron-sulfur clusters, even after the dialysis and purification procedures. The SDS-PAGE analysis of the purified Mtb proteins is shown in Fig. 2a–d. FdrA, FprA, and CYP143A1 were all electrophoretically pure and shown to have approximate apparent masses of 48, 53, and 45.9 kDa, respectively, consistent with the expected masses from the amino acid sequences of the respective forms of the His-tagged proteins as previously reported (Fischer et al. 2002; Swami 2015; Zanno et al. 2005). However, the SDS-PAGE of Rv1786 provided an estimated molecular mass of about 17 kDa (Fig. 2d), that is about 9 kDa larger than its predicted molecular weight (7.93 kDa). To validate that the protein band corresponded to the expected Rv1786-His protein, we verified by western blot analysis that the His-tagged fusion protein could be detected using anti-His antibodies (Fig. 2e). To further confirm the identity of the protein, we used MALDI-TOF/TOF-MS where a major peak was obtained at 7935.8 Da, which is consistent with the theoretical average molecular mass of 7933.8 Da (Fig. 3). In order to confirm the iron-sulfur clusters ligated in the Mtb proteins, UV/Vis absorption spectra were collected using the scanning mode of PerkinElmer EnSpire® 2300 Multimode Plate Reader. As shown in Fig. 4a, a typical ferredoxin spectrum was obtained with a pliable peak at 406 nm for Rv1786. The outline of the UV/Vis absorption spectrum of Rv1786 is similar to FdX’s except for missing a shoulder of ~450 nm (Imai et al. 1983; Ricagno et al. 2007; McLean et al. 2006c). The UV/Vis spectrum of CYP143A1 is shown in Fig. 4b. Pure ferric CYP143A1 has a typical P450 UV/Visible spectrum, with a Soret band at 416 nm, and Q-bands at ~572 and 535 nm, which is consistent with results reported previously (Swami 2015). The UV/Vis spectra of FdrA and FprA (Fig. 4c, d) are exactly the same as was reported previously (Fig. 4c, d) (Fischer et al. 2002; Zanno et al. 2005).

Fig. 2
figure 2

SDS-PAGE of purified recombinant proteins. a CYP143A1. b FdrA. c FprA. d Rv1786. Proteins were stained with Coomassie blue in 12% gel. Lane M indicates size-marker proteins. e Expression of His-tagged Rv1786 detected by Western blot

Full size image
Fig. 3
figure 3

Identification of Rv1786 molecular weight by MALDI-TOF/TOF MS analysis. The main ion peak obtained was 7935.8 Da, which was consistent with the theoretical average molecular mass 7933.81 Da

Full size image
Fig. 4
figure 4

UV/Vis absorption spectra of purified recombinant proteins. a Rv1786. b CYP143A1. c FdrA. d FprA. Samples were dissolved in potassium phosphate buffer (pH 7.4). The spectra were recorded at room temperature between 260 and 660 nm. The Soret spectral maxima are at 406, 416, 370, and 450 nm, respectively

Full size image

EPR spectroscopic analysis of Rv1786

EPR spectroscopy has long been the method of choice for revealing structural information of iron-sulfur cluster-containing proteins, such as ferredoxins. EPR X-band spectra were monitored for Rv1786 to provide further insights into the heme ligation environment and spin-state equilibrium in the external status. As shown in Fig. 5, the EPR signal for Rv1786 is typical for a [3Fe-4S] cluster-containing protein, with a strong g = 2.02 resonance, characteristic of [3Fe-4S]+S = 1/2 clusters, which is in agreement with the EPR result of FdX, the only Mtb [3Fe-4S] ferredoxin member being well characterized in detail to date (Cheng et al. 2013; Kyritsis et al. 1999; Liu et al. 2013; McLean et al. 2006c).

Fig. 5
figure 5

EPR spectra of Rv1786. EPR spectra of oxidized Mtb Rv1786 were recorded at a concentration of 1.25 mg/mL in 50 mM Tris buffer at 10 K. Instrument settings were as follows: microwave frequency = 9.371 GHz, microwave power = 2.377 mW, sweep time = 120 s, modulation amplitude = 1.000 G. No EPR signal was detectable for the dithionite-reduced Rv1786

Full size image

Detection of interactions between Rv1786 and putative redox partners by SPR

SPR is an optical technique for detection of interactions between two different molecules, in which one is mobile and one is fixed on a thin gold film (Drescher et al. 2009). To discover the interaction modes between Rv1786 and putative redox partners including ferredoxin-NAD(P)H reductase FprA, FdrA, and CYP143A1, SPR analyses were used to explore the affinity between the proteins (Fig. 6 and Fig. S1). As shown in Fig. 6 and Fig. S1b–d, the interaction bindings of FprA, FdrA, and CYP143A1 with Rv1786 revealed that each of the three proteins displayed binding capacity to Rv1786, yet at different levels, and with signal intensities increasing with increasing concentration of the proteins. The SPR response signals of the controls are presented in Fig. S1a. The binding signal of the negative control (BSA) appeared flat-lined as the baseline due to a very low non-specific binding. The FdrA and FprA showed a significant binding affinity to Rv1786 with K D values of 2.59 × 10−6 and 1.14 × 10−6 M, respectively, while CYP143A1 showed an affinity to Rv1786 with a K D value of 3.37 × 10−7 M, which is about 10 times lower than that of the two FNRs (Table 2). Since the lower the value of K D corresponds to the more tight binding between proteins, it was indicated that CYP143A1 has a significantly higher affinity for Rv1786 than FdrA or FprA. Based on the ratio of the experimentally off- and on-rate (k d /k a ) as well as rank order distributions of the equilibrium binding constants (K D , equals to k d /k a ) for each Rv1786-based coupling on an on-off quadrant plot as shown in Table 2 and Fig. 6, affinity reactions between Rv1786 and each of FdrA, FprA, and CYP143A1 can be described as a slower on-rate/faster off-rate binding, a faster on-rate/faster off-rate binding, and a faster on-rate/slower off-rate binding (Fig. 6), respectively.

Fig. 6
figure 6

Distribution of kinetic parameters of protein-protein interactions determined by SPR on an on-off quadrant plot. Rv1786 was dissolved in 25 mM Tris-HCl (pH 8.0) and immobilized onto a Nanocapture sensor chip (Plexera Bioscience, Seattle, WA, USA) by QArray mini spotter (Genetix Ltd., New Milton, Hampshire, UK). The purified FdrA, FprA, or CYP143A1 was then injected as analyte at various concentrations in the running buffer for detecting their binding affinity to Rv1786 spotted on the chip surface. Four different concentrations of each analyte protein (500, 1000, 2000, and 4000 nM) were used to flow through the chip. All possible binding curves were fitted by Langmuir equation, and the constants characterizing the on-rate (k on), the off-rate (k off), and the equilibrium binding (K D ) were generated

Full size image
Table 2 Binding parameters of SPR measurements
Full size table

Steady state kinetic assay

Since NADH and NADPH were characterized as efficient reductants for FdrA and FprA, respectively, the capacity of Rv1786-mediated electron transferring towards the final electron acceptor, cytochrome c, was determined in the presence of NADH-FdrA or NADPH-FprA in steady-state kinetic assays (Fischer et al. 2002; Zanno et al. 2005). The kinetic constants for two FNRs (FdrA or FprA) coupled with increasing quantities of Rv1786 in the presence of the substrate are listed in Tables 3 and 4. A weak cytochrome c reductase activity was monitored for both FdrA and FprA in the absence of the ferredoxin Rv1786 as electron transfer protein (FdrA k cat = 2.396 ± 0.160 min−1, K m  = 2.762 ± 0.758 μM, k cat/K m  = 0.87 × 106 M−1 min−1; FprA k cat = 2.988 ± 0.476 min−1, K m  = 10.65 ± 2.75 μM, k cat/K m  = 0.28 × 106 M−1 min−1). In comparison with FdrA alone, the reduction rate and efficiency of cytochrome c showed approximately a 22- to 45-fold higher (k cat = 68.16 ± 4.16 to 111.2 ± 1.8 min−1, k cat/K m  = 1.91 × 107 to 3.91 × 107 M−1 min−1) for FdrA in the presence of varying levels of Rv1786. The overall catalytic efficiency of the reaction is also faster for FprA with supplement of Rv1786 (k cat = 4.909 ± 0.273 to 7.785 ± 1.210 min−1 and k cat/K m  = 0.33 × 106 to 0.73 × 106 M−1 min−1 for FprA:Rv1786 compared with k cat = 2.988 ± 0.476 min−1, k cat/K m  = 0.28 × 106 M−1 min−1 for FprA alone), though the parameters were not altered as significantly as what observed in NAPDH-FdrA-Rv1786-cytochrome c catalytic system. However, the K m values for FdrA or FprA remain relatively constant in the absence or presence of varying levels of Rv1786, which suggests that Rv1786 has little effect on the affinity of the two FNRs for the cytochrome c.

Table 3 Kinetic parameters for reductase activity of FdrA coupled by series concentration of Rv1786
Full size table
Table 4 Kinetic parameters for reductase activity of FprA coupled by series concentration of Rv1786
Full size table

Discussion

Revealing the biochemical features of uncharacterized proteins in Mtb will extend our understanding of key life processes required for the pathogen in infection, latency, and proliferation, which will help us to identify essential structural or functional proteins as potential target for the development of new drugs against the pathogen. The Mtb genome encodes a limited number of ferredoxins, FNRs or fusions of ferredoxin, and FNR-like domains that function as putative electron transfers to support the catalytic activities of up to 20 different CYP enzymes (Munro et al. 2003; Ouellet et al. 2010). The Mtb CYPs are putative drug target proteins due to their essential roles in the complex lipid metabolism of Mtb and their high-affinity binding to anti-fungal azole drugs (Zhang et al. 2006). The number of ferredoxin and FNR genes is thus significantly fewer than the CYP genes in the Mtb genome. Thus, the CYP enzymes and their putative redox partners cannot have a one-to-one relationship, which suggests that these redox proteins could play a “hub-like” role in the CYP-associated electron transfer pathways. Despite the importance of the redox proteins in CYP-related catalytic processes in Mtb, the structural or functional properties for the majority of the putative Mtb ferredoxins remain unclear. Among the uncharacterized Mtb ferredoxins, Rv1786 is one of the most obvious candidates for CYP-related redox partners, as it is adjacent to the CYP143A1 (Rv1785c) in the chromosome. Considering that correct ferredoxin-FNR coupling is a prerequisite for efficiency of electron transferring in the redox system, it is crucial to identify the effective combination of Rv1786 and putative endogenous FNRs or fusions containing FNR-like domains from Mtb (Lewis and Hlavica 2000).

In the present study, we identified by phylogenetic analysis that Rv1786 has a typical CXXHXXC(X)nCP iron-sulfur cluster-binding motif and is highly conserved among pathogenic members of Mycobacterium genus. These data suggests that the ferredoxin may be relevant for full virulence or infectivity of Mtb towards the human host. The genome sequence of Mtb H37Rv revealed two FNRs, FdrA (Rv0688) and FprA (Rv0886), as putative electron transfer partner proteins for ferredoxins (Ouellet et al. 2010). Since we can express both Mtb FNR proteins in a soluble and a stable form in E. coli, it is possible to reconstitute potential combinations between Rv1786 and each of the two FNRs in vitro (Fischer et al. 2002; Zanno et al. 2005). Based on the mobility in SDS-PAGE of the purified recombinant Mtb Rv1786, the molecular weight of the protein was estimated to be 17 kDa, which is obviously bigger than the prediction by the sequence of the His-tagged Rv1786. However, our MS analysis confirmed that the mass of the protein correlated to the predicted molecular mass, a result which also indicated that the reason to the protein retardation of the SDS-PAGE could not be due to a post-translational modification. However, the unpredicted migration behavior of the protein could instead be related to its acidic nature. It has been observed in other studies that acidic proteins (i.e., glutamate- and/or aspartate-rich proteins) often display a larger molecular mass in SDS-PAGE analysis than their predicted size (Guan et al. 2014; Niu and Guiltinan 1994; Sallantin et al. 1990). Considering that Rv1786 is a typical acidic protein with a high percentage (23.9%) of glutamate and aspartate residues and with a pI of 4.11, this could also be the explanation behind our SDS-PAGE result. The UV/Vis and EPR spectroscopic properties of the purified Rv1786 confirmed that the ferredoxin contained a [3Fe-4S] cluster and was folded correctly as proposed upon recombinant expression in E. coli.

The different affinity interactions between Rv1786 and each of FdrA, FprA, and CYP143A1 were analyzed with SPR measurements. According to the commonly accepted quantification standard of interaction features in SPR analyses, a significant protein-protein binding affinity can be demonstrated with the equilibrium binding constants (K D ) at 10−6 M level and an experimentally off- and on-rate can be ranked with 10−3 s−1 and 103 M−1 s−1 level as fast/slow cutoffs (Hu et al. 2014; Wu et al. 2015; Zhu et al. 2014). In the present study, a significant affinity coupling of Rv1786 to each of FdrA and FprA has been identified with K D values at 10−6 level. When comparing k d and k a values between Rv1786-FdrA and Rv1786-FprA couplings, a slightly stronger binding (a faster on-rate and a slower off-rate) is observed for Rv1786-FprA. In agreement with the prediction in terms of their genetic context, a significant binding affinity between Rv1786 and CYP143A1 was also demonstrated with k d , k a , and K D values at 10−4 s−1, 103 M−1 s−1, and 10−7 M level.

Since Rv1786 lies adjacent to the gene encoding CYP143A1 (Rv1785c), the protein encoded by the gene has been thought as a natural electron transferring partner for CYP143A1 in Mtb (McLean and Munro 2008; Ouellet et al. 2010). However, as the substrate for the CYP143A1 enzyme still remains unknown, the electron transfer function of Rv1786 is difficult to evaluate in an authentic CYP143A1-based catalytic reaction system. Considering that cytochrome c has been used for evaluation of the electron transfer activity of Mtb FdrA or FprA coupling with other Mtb ferredoxins in previous studies (Fischer et al. 2002; Zanno et al. 2005), we decided to use the enzyme as final electron acceptor to characterize the preference for the two FNRs in the absence or presence of Rv1786. The cytochrome c reduction assay demonstrates that Rv1786 can significantly enhance the catalytic rate and efficiency of the reduction reaction for both FdrA and FprA. The steady-state kinetic parameters for FdrA-Rv1786 suggest a better coupling activity of Rv1786 to FdrA than that of Rv1786 to FprA.

In summary, the observations from the present study provide a preliminary but rational evidence for characterization of Rv1786 as an active electron transfer in CYP-associated redox system in Mtb. Our results indicate that Rv1786 is able to couple with FdrA or FprA as redox partner. A better coupling activity was observed with FdrA while a stronger binding was found with FprA. Further studies involving Rv1786-deficient systems, crystal structure characterization, and exact function in CYP143A1, catalytic reactions would be required to fully estimate the extent of involvement of this important ferredoxin in virulence or infectivity of Mtb. The present study improves our understanding of the biochemical properties and function of the iron-sulfur cluster containing ferredoxin proteins in Mtb.