Metabolic Engineering of Cofactor F420 Production in Mycobacterium smegmatis

Cofactor F420 is a unique electron carrier in a number of microorganisms including Archaea and Mycobacteria. It has been shown that F420 has a direct and important role in archaeal energy metabolism whereas the role of F420 in mycobacterial metabolism has only begun to be uncovered in the last few years. It has been suggested that cofactor F420 has a role in the pathogenesis of M. tuberculosis, the causative agent of tuberculosis. In the absence of a commercial source for F420, M. smegmatis has previously been used to provide this cofactor for studies of the F420-dependent proteins from mycobacterial species. Three proteins have been shown to be involved in the F420 biosynthesis in Mycobacteria and three other proteins have been demonstrated to be involved in F420 metabolism. Here we report the over-expression of all of these proteins in M. smegmatis and testing of their importance for F420 production. The results indicate that co–expression of the F420 biosynthetic proteins can give rise to a much higher F420 production level. This was achieved by designing and preparing a new T7 promoter–based co-expression shuttle vector. A combination of co–expression of the F420 biosynthetic proteins and fine-tuning of the culture media has enabled us to achieve F420 production levels of up to 10 times higher compared with the wild type M. smegmatis strain. The high levels of the F420 produced in this study provide a suitable source of this cofactor for studies of F420-dependent proteins from other microorganisms and for possible biotechnological applications.


Introduction
The cofactor F 420 was first identified chemically in methanogenic Archaea in 1972 [1], although a compound with similar characteristics was previously described in Mycobacteria in the early 1960s [2,3]. Since its discovery, F 420 and its precursor FO (so called 5-deazaflavins) have been found in a variety of (micro)organisms, including Archaea, bacteria and eukaryotic species (Table 1). F 420 is named on the basis of its intense absorbance/ fluorescence at 420 nm (emission 480 nm), which is redox dependent and is lost upon reduction of the cofactor. It also has unique chemical and biological characteristics; the isoalloxazine chromophore of F 420 is structurally very similar to that of the flavins (FMN and FAD), although it is functionally similar to NAD(P) + (Figure 1). Functionally, F 420 is a two-electron carrier involved in hydride transfer reactions. The redox potential of F 420 H 2 /F 420 +2e 2 (2360 mV) is lower than those of the classical hydrogen carriers NAD(P)H/NAD(P)+2e 2 (2320 mV) and FADH 2 /FAD+2e 2 (2219 mV) [4,5].
A key biosynthetic precursor of F 420 is FO (7,8-didemethyl-8hydroxy-5-deazariboflavin), comprising an isoalloxazine ring and ribitol moieties. Formation of F 420 follows a series of biochemical reactions and is completed by the addition of a phospholactate group, and finally a poly-glutamate tail in which L-glutamate residues are linked together via c-glutamyl bonds ( Figure 2) [6,7]. The length of the poly-glutamate tail constitutes the main difference between the F 420 cofactors from different microorganisms, the number of residues varying from 2-9. There are suggestions, however, that the type of aor c-glutamyl linkage in the terminal glutamate residue could also be different in some Archaeal species [6,8,9,10].
F 420 is not commercially available and researchers working on F 420 -dependent proteins have to prepare it as required. With the discovery of new F 420 -dependent enzymes and increasing interest in F 420 -dependent reactions, especially in the case of the pathogen Mycobacterium tuberculosis (Mtb), a resource with high yields of F 420 production is required. F 420 has been previously purified from various microorganisms, including Archaea (Methanobacterium, Methanococcus and Methanosarcina species) and Actinomycetes (Actinomadura, Actinoplanes, Streptomyces, Rhodococcus, Nocardia and Mycobacteria species), with differing yields [11]. F 420 purification in all cases, however, essentially follows the same principle; precipitation of cellular proteins using heat or an organic solvent, followed by separation of F 420 from remaining cellular components based on its acidic nature [4]. In order to purify F 420 , a number of different chromatographic steps have been used, including ion exchange, adsorption, HPLC and gel filtration chromatography [6,9,11]. Isabelle et al. have reported thorough analyses of F 420 -producing microorganisms, and based on ''ease of growth, fewer hazards, and lower costs'' concluded that M. smegmatis is the best source for F 420 production, providing there is no requirement for a particular number of glutamate residues in the F 420 polyglutamate tail [11].
Our initial F 420 purification trials indicated that M. smegmatis transformed to over-express the M. tuberculosis protein FGD1 (F 420 -dependent glucose-6-phosphate dehydrogenase 1) could produce higher levels of F 420 compared with the wild type strains. This observation prompted us to thoroughly investigate the effects on F 420 production of over-expression of other proteins known to be involved in F 420 biosynthesis and metabolism in Mycobacteria. These include three proteins in the F 420 biosynthetic pathway, viz. FbiA (Rv3261) [12], FbiB (Rv3262) [12] and FbiC (Rv1173) [13] and three other proteins which are shown to be involved in F 420 metabolism: FGD1 (Rv0407) [9,14], Ddn (Rv3547) [15] and Rv0132c (author's unpublished data).
Here we describe the development of vectors to co-express Mtb proteins in M. smegmatis. We further show that by co-expressing enzymes associated with F 420 production and manipulating growth conditions, greatly increased levels of F 420 can be obtained. With the growing recognition that F 420 plays a crucial role in Mycobacteria and other organisms, this readily available source of the cofactor will be useful for testing its physiological and biochemical roles, and for possible applications in biotechnology.

Preparation of New Mycobacterial Vectors
The pYUB1049 vector (5795 bp) is a product of ligation between the vectors pMS134 and pET28b-cmaA2 [16], resulting in a vector with a cloned gene between NdeI and BamHI restriction sites. The pYUB1049 vector was subjected to restriction digestion using NcoI (single site) and BlpI (two sites) restriction sites, in order to obtain a linear vector without the multiple cloning site. The plasmid was first digested to completion with NcoI (Roche Applied Science) and dephosphorylated using calf intestinal alkaline phosphatase (New England Biolabs) followed by ethanol precipitation. The NcoI-cut linear pYUB1049 vector was subjected to a partial digestion with BlpI (BpuI102I isoschizomer, Fermentas) for 20 minutes and the reactions stopped using 5 mL 0.5 M EDTA. The digested vector was run on a 0.5% agarose gel and a DNA fragment corresponding to 4705 bp was excised and gel-purified.

PCR Amplification and Cloning
The open reading frames (ORFs) encoding Rv3261 (FbiA), Rv3262 (FbiB), Rv1173 (FbiC), Rv0407 (FGD1), Rv3547 (Ddn) and Rv0132c were amplified from M. tuberculosis H37Rv genomic DNA using Pwo, Pfx or PrimeStar polymerases with the primers outlined in Table 2. All constructs were cloned with either N-or C-terminal His 6 -tags. The amplified products for the FGD1 [17] and Rv0132c constructs were cloned using restriction/ligation cloning into the pYUB1049/pYUB28b vectors. The constructs were transformed into E. coli Top10 cells and plated on low salt LB agar medium supplemented with 50 mg/mL hygromycin B to select for colonies harbouring the plasmid. Positive clones were verified using colony PCR, restriction digestion and sequencing.
All other ORFs were cloned using the GatewayH cloning system into the pDESTsmg vector [18]. The GatewayH cloning system uses a nested PCR method involving two rounds of amplification in which the second round uses the product of the first round as template. Gene-specific primers are used in the first round PCR to amplify the gene of interest and generic primers are used in the second round amplification to incorporate the required recombination sites for subsequent cloning. The PCR products were cloned by recombination into pDONR221 (Invitrogen) using BP Clonase TM (Invitrogen), to generate the entry clones. The constructs were transformed into E. coli Top10 cells and plated on LB agar medium containing 50 mg/mL of kanamycin. Positive clones were verified using BsrGI digestion and sequencing. These positive entry clones were recombined in vitro with pDESTsmg, in an LR reaction using LR Clonase TM (Invitrogen), to generate a M. smegmatis expression construct. Following transformation of recombinant pDESTsmg plasmids, positive clones were selected on low salt LB agar plates supplemented with 50 mg/mL hygromycin B and were verified using BsrGI digestion.
The pYUBDuet vector was used to clone the F 420 biosynthetic ORFs (FbiAB and FbiC) together using restriction/ligation cloning. Both FbiC and FbiAB ORFs were amplified using PfuUltra Fusion HS DNA polymerase (Stratagene) using the primers outlined in Table 2. FbiC was first cloned using NcoI/ HindIII restriction sites and the FbiAB operon was subsequently cloned using NdeI/EcoRV restriction sites.

Expression in M. smegmatis
All expression constructs were electroporated individually into the M. smegmatis strain mc 2 4517. Preparation of electrocompetent cells and electroporation procedures were performed following published protocols [19]. Briefly, M. smegmatis mc 2 4517 cells were grown at 37uC in 7H9/ADC/Tween80 or LB/Tween80 containing 50 mg/mL kanamycin until an OD600 ,0.7. Cells were harvested and washed three times in 10% ice-cold glycerol and finally resuspended in 10% ice-cold glycerol. Single aliquots of the resulting competent cells (40 mL) were transformed with 1 mL of DNA and a further 260 mL of 10% glycerol in 0.2 cm cuvettes. Electroporation was performed using a Bio-Rad Gene Pulser set to the following parameters: R = 1000 V, Q = 25 mF and V = 2.5 kV. Cells were immediately harvested with 1 mL 7H9/ ADC/Tween80 (Difco TM and BBL TM Middlebrook) or LB/ Tween80 and incubated for 3 h at 37uC with shaking. Positive transformants were selected by plating on 7H10/ADC (Difco TM and BBL TM Middlebrook) or LBT agar plates containing 50 mg/ mL each of kanamycin and hygromycin B.

Western Blot Analyses
M. smegmatis cells expressing different constructs were lysed twice using a cell disruptor (Constant Systems Ltd.) and centrifuged at 16,0006g to pellet non-lysed cells and other insoluble material. Protein samples were separated on a 15% SDS-PAGE gel and transferred to polyvinylidene difluoride (PVDF) membranes using a wet transfer protocol (200mA, 3 hours) [21]. His-tagged recombinant proteins were detected using a mouse monoclonal anti-His antibody and horseradish peroxidase-conjugated anti-mouse antibody (GE Healthcare). The Luminol (ECL plus kit, GE Healthcare) chemiluminescence was detected using an LAS4000 imaging system (Fujifilm).

FO and F 420 Characterization
M. smegmatis cells expressing different M. tuberculosis proteins were grown in identical conditions to late log phase or stationary phase. In all expression cultures the ZYP-5052 autoinduction media was used for F 420 production experiments and the media to flask volume ratio was kept constant at 20%. In order to optimize the media for F 420 production, the ZY component of ZYM-5052 media was replaced by commonly used media bases including 26 ZY, YT (0.8% tryptone, 0.5% yeast extract and 42.77 mM NaCl), TB (1.2% tryptone, 2.4% yeast extract and 0.4% glycerol), SOB (2% tryptone, 0.5% yeast extract, 8.56 mM NaCl, 2.5 mM KCl and 10 mM MgCl 2 ) and SOC (SOB with 20 mM glucose). Iron and sulphur supplements (ferric ammonium citrate, ferric citrate and ferrous sulphate all at 0.1 mg/mL and L-cysteine at 1 mM) were also added to the expression media as a possible requirement for the FbiC enzyme. L-glutamate and manganese chloride (1 mM final concentration) were also added to the expression media to evaluate their necessity for FbiB-mediated F 420 production [22].
To ascertain the optimum growth period for F 420 production, eight identical cultures of M. smegmatis cells expressing the recombinant FbiABC construct were set up. Each culture had a wild type M. smegmatis culture as a control. At 24 h intervals, one culture each of control and recombinant FbiABC-expressing M. smegmatis cells were harvested and processed to monitor the F 420  production level. The procedure was carried out for eight days and the F 420 production ratio for each day was calculated by dividing the F 420 fluorescence from FbiABC-expressing cells by fluorescence of the wild type control. M. smegmatis cells were centrifuged for 15 min at 160006g and the resulting media were used for FO characterization. The cell pellets were washed with 25 mM sodium phosphate buffer, pH 7.0 and were subsequently resuspended in 1 mL of the same buffer per 100 mg of cells (wet weight). The cell suspensions were autoclaved at 121uC for 15 min to break the cells open and were then centrifuged for 15 min at 160006g. Fluorescence of the media and the extract were monitored using excitation wavelength of 420 nm (405610 nm filter) and emission wavelength of 480 nm (485615 nm filter). All fluorescence experiments were performed using an EnVision Multilabel plate reader (Perkin Elmer) in a 96well plate format and were carried out in triplicate.
The autoclaved cell extracts were further purified using a HiTrap QFF ion exchange column (GE Healthcare) to separate the intracellular FO from the F 420 . The extract was run on the column pre-equilibrated with 25 mM sodium phosphate buffer, pH 7.0 and was subsequently washed with five column volumes of buffer. Two yellow fractions were eluted at 200 and 500 mM NaCl, respectively. The purified fractions were used for mass spectrometry analysis, together with the media from the previous step. The media (1 mL) was treated with an equal volume of cold acetone to precipitate the protein and the solution was then evaporated down to ,0.5 mL to drive off the acetone. A mix of water and 5% aqueous methanol with 0.1% formic acid was added to bring the final concentration of methanol to less than 1% (total volume 4 mL). All samples were then applied to a preequilibrated Alltech Maxi-Clean 300 mg large pore 100Å C-18 SPE cartridge and washed with 4 mL 5% methanol containing 0.1% formic acid followed by 4 mL 10% methanol. Compounds were eluted with 4 mL 80% methanol containing 5 mM ammonium bicarbonate pH 8.5. Eluates were evaporated under nitrogen and redissolved in 80% methanol and 20 mM ammonium acetate ready for mass spectrometry. Samples were infused at 3 mL/min under negative electrospray conditions into an LTQ-FT mass spectrometer (Thermo Scientific). The ion intensity data were obtained using a source voltage of 2.5 kV and capillary temperature of 225uC. Ions were examined in both the ion trap and ion cyclotron resonance cells, the latter to obtain high resolution (100,000 at m/z 400) accurate mass data. This was necessary to confirm the atomic composition of the ions and help deconvolute the contribution of metal ion adducts (Na + /K + ) to the levels of individual poly-glutamate species. Up to four sodium ions were adducted to produce some double charged negative ions.

New Mycobacterial Expression Vectors
The pYUB1049 vector does not provide an intact multiple cloning site and does not support C-terminal His-tag expression. In order to overcome these obstacles, the pYUB1049 vector was subjected to a restriction digestion using NcoI and BlpI enzymes and a linear fragment lacking the multiple cloning site was obtained. The resulting fragment was used as a backbone that could be ligated to the intact multiple cloning site from the pET28b or pETDuet-1 vectors to produce the pYUB28b and pYUBDuet vectors, respectively. Figure 3 provides a schematic representation of the vectors with the list of unique restriction sites that can be used for cloning.

PCR Amplification and Cloning
Six different ORFs which are believed to be involved in F 420 biosynthesis (FbiA, FbiB and FbiC) or F 420 metabolism (FGD1, Ddn and Rv0132c) were amplified and cloned for expression in M. smegmatis as His-tagged proteins. Assuming that FbiA and FbiB ORFs are transcribed as a single operon, we investigated the possibility of cloning and co-expression of the whole F 420 biosynthetic pathway (FbiAB and FbiC) in order to boost F 420 production yield. The pYUBDuet co-expression vector was designed, prepared and subsequently used to clone FbiC and FbiAB ORFs, making it possible to express three different proteins from a single vector. All three proteins were expressed in their native form without His-tags. The six F 420 biosynthetic or metabolic ORFs cloned into pYUB1049/pYUB28b/pDESTsmg vectors were expressed in M. smegmatis as individual proteins. Each of these proteins were cloned with either N-or C-terminal His-tags, making it possible to detect the protein expression using monoclonal anti-His antibodies. The western blotting experiments indicated that all proteins were expressed in M. smegmatis cells, as shown by appearance of correctsized bands for the appropriate proteins (data not shown).
The expression of proteins from the pYUBDuet vector could not be detected using western blotting, as they did not contain any tags; however, their successful expression could be inferred from FO and F 420 production as discussed later.

Cofactor F 420 Production
Individual M. smegmatis cultures harbouring six different constructs (FbiA, FbiB, FbiC, FGD1, Ddn and Rv0132c) were grown in order to find out the over-expression effect of these targets on F 420 production. Three different media were initially used to express the proteins; LBT with IPTG induction, MDG with no or low induction using IPTG, and ZYM-5052 autoinduction media. Based on growth rate and cell mass, ZYM-5052 media was selected as the best media and was used to continue F 420 production experiments. The fluorescence signals of the expression media and the cell extracts were monitored at 420 nm, enabling the detection of both FO and F 420 . It has been reported that FO comprises 1-7% of the total intracellular deazaflavin in Mycobacteria [8]; we used fluorescence at 420 nm to evaluate the F 420 contents of the cellular extracts without taking into account the small portion of the fluorescence signal coming from FO.
The experimental results indicate that FGD1 over-expression increases F 420 production by almost two-fold compared to the wild type strain (Figure 4, A). Cells expressing other Mtb proteins did not show a significant increase in F 420 yield, however. Cells expressing the FbiC construct (pDEST-FbiC) showed a strong blue-green colour in the media. This is presumably due to the presence of fluorescent FO in the media which diffuses out of the cells as FO does not have any charge on the molecule to cause retention inside the cell (Figure 4, A) [23]. Mass spectrometry confirmed that FO was indeed responsible for the distinct fluorescence of the media (m/ z 362.09870 [M-H] 2 ; C 16 H 16 N 3 O 7 requires 362.09882). This observation could be explained by over-expression of the FbiC protein leading to higher FO synthesis. Because the cells could not convert the over-produced FO to F 420 , the excess was presumably lost from the cells, either by diffusion or by active export.
This observation provided the motivation for us to co-express the FbiAB operon together with FbiC, hoping that over-expressed FbiA and FbiB proteins would be able to convert the synthesised FO into F 420 inside the cells. The pYUBDuet vector was used to clone FbiABC ORFs together; FbiC was first cloned, resulting in the pYUBDuet-FbiC construct, after which FbiAB was introduced to obtain pYUBDuet-FbiABC. Both these constructs were used to investigate the effect on FO/F 420 production (Figure 4, B). Cells expressing FbiC alone (pYUBDuet-FbiC) consistently showed more than 10-fold higher FO levels in the expression media compared to the wild type strains. It is an interesting observation that FO production by the pYUBDuet-FbiC construct is much higher (.50%) than by the pDESTsmg-FbiC construct, with the former expressing FbiC as the native protein whereas the latter has an Nterminal His-tag. In contrast, F 420 production from pYUBDuet-FbiC was not significantly elevated compared to wild type. By expressing the FbiAB operon together with FbiC (pYUBDuet-FbiABC), however, F 420 production was consistently more than five times higher inside the cells (Figure 4, B). These results clearly indicate that the cells express functional recombinant proteins resulting in much higher intracellular F 420 levels.
M. smegmatis cells expressing the pYUBDuet-FbiABC construct were then used to find out the optimum time period for F 420 production. The F 420 production was monitored for eight days using ZYM-5052 media and the F 420 production ratio was calculated and plotted versus the day of culture. The results indicated that the F 420 levels were the highest on day four of the culture, after which the levels gradually decreased. Based on this result, the best time to harvest the cells for F 420 purification is 4-5 days after setting up the expression culture (Figure 4, C). Subsequently, a set of experiments was performed to find out the best media formulation to grow the cells for F 420 production using an autoinduction protocol. ZY produced the highest F 420 yield among ZY, YT, TB, SOB and SOC media. Bioinformatic analysis has indicated that FbiC is a protein with possible Fe-S clusters. In addition the reaction catalyzed by an archaeal homologue of FbiB requires L-glutamate and manganese chloride [22]. The expression media were therefore also supplemented with iron/sulphur and L-glutamate/manganese additives. The results indicated that supplementation of the expression media with either of these additives does indeed increase the F 420 production yield (Figure 4, D). Surprisingly, cultures with an L-glutamate/ manganese supplement did not have extra FO in the media, implying that the cells could convert all the produced FO into F 420 inside the cells (Figure 4, D). It seems that the limiting factor in producing F 420 from over-produced FO was the supply of the required L-glutamate/manganese.
The FO/F 420 produced by the cells expressing the FbiABC construct was purified and analysed using mass spectrometry. The results show two predominant fractions; a 200 mM NaCl fraction mainly composed of FO and a 500 mM NaCl fraction of exclusively F 420 with more than 95% being F 420 -6 and F 420 -7 species ( Figure 5). This result is in line with the previously published results of F 420 extracted from the wild type M. smegmatis cells having the major species of F 420 -5 to F 420 -7 [9,11], implying that the over-expression of the FbiABC construct does not change the F 420 production profile.

Discussion
The cofactor F 420 has an important role in the metabolism of Archaea and has been the subject of numerous studies over the years since its identification. It is now clear that this importance applies also to Mycobacteria, for which there is growing evidence that F 420 plays a key role in defence against oxidative and nitrosative stress [5,24]. Consistent with this, the number of identified F 420dependent enzymes from Mycobacteria is growing, with nine new examples recently described [25]. A recent partial phylogenetic profiling study has proposed that there are at least 28 separate F 420dependent enzymes in M. tuberculosis, suggesting that F 420 has a pivotal role in redox reactions of this pathogenic mycobacterium [26]. Few of these enzymes have been characterised, however, and research into their functions, and the role of F 420 , are handicapped by the fact that there is no commercial source for this cofactor, which can only be obtained in relatively low yield from the wild type M. smegmatis strain. A major aim of this study was to increase the F 420 production yield in M. smegmatis by cloning and expression of the genes involved in F 420 production and metabolism.

Mycobacterial Expression Vectors
The pYUB1049 plasmid is a T7 promoter-based vector for which expression can be induced by IPTG or autoinduction. This vector has previously been used as a shuttle vector for cloning of Mycobacterial genes into E. coli and subsequent expression of proteins in M. smegmatis [17,27]. The pYUB1049 vector has been also converted to a GatewayH cloning system compatible vector, pDESTsmg [18]. In this study, two different vectors were designed and prepared from the parental pYUB1049 vector; the pYUB28b vector is used for restriction/ligation cloning of single genes with the capability of expressing N-and C-terminal His-tags, whereas the pYUBDuet vector is a co-expression vector for simultaneous expression of two genes in a Mycobacterial host. Our experimental results demonstrate the application of T7-promoter based coexpression vectors in M. smegmatis that could also be useful in other contexts. Although there have been previous reports of coexpression systems for Mycobacteria [28,29,30,31,32,33], to the best of our knowledge, this is the first T7-promoter based coexpression vector for a Mycobacterial host. The pYUB28b and pYUBDuet vectors, together with the pDESTsmg vector which has been previously developed in the authors' lab for the GatewayH cloning system, represent a repertoire of T7 promoter-based vectors which can be routinely used for expression of a wide range of ORFs in a Mycobacterial host.

F 420 Production
FbiC is annotated as FO synthase [34], catalysing the transfer of the hydroxybenzyl group from 4-hydroxyphenylpyruvate (a tyrosine precursor) to 5-amino-6-ribitylamino-2,4(1H,3H)-pyrimidinedione (an intermediate in flavin biosynthesis) to form FO ( Figure 2) [23]. FO is the first intermediate with a complete deazaflavin chromophore in the F 420 biosynthesis pathway [13,23] providing the rationale for believing that this reaction might be the rate limiting step in F 420 biosynthesis. FbiA and FbiB are believed to be involved in production of F 420 from the precursor FO molecule; FbiA in generating F 420 -0 from FO and FbiB in adding glutamate residues to F 420 -0 to produce F 420 with a polyglutamate tail of variable length (Figure 2) [12]. In all Mycobacterial species with the genome sequences completed to date (21 in total, as of October 2010) FbiA is located immediately upstream of FbiB (www.TBdb.org). A detailed analysis indicates that the start site for the FbiB ORF overlaps with the last four base pairs of the FbiA ORF, though in a different reading frame, implying that they might be transcribed as a single operon for expression. In fact, it has been shown in M. bovis that these two ORFs are transcribed together as a single mRNA species [12]. This genetic arrangement made it possible to co-express the FbiAB operon and FbiC gene together, using the pYUBDuet vector, in order to increase F 420 yield. Based on our results, the optimum condition to produce F 420 by M. smegmatis cells expressing recombinant FbiABC is a culture with autoinduction media using ZY base over 4-5 days supplemented with iron, sulphur, L-glutamate and manganese. Using these optimal conditions, the F 420 production yield was up to 10times higher compared with the wild type strains.
The main limiting factor in F 420 production, based on our results on over-expression of the three enzymes FbiA, FbiB and FbiC from the F 420 biosynthetic pathway, appears to be the availability of the FbiB reaction substrate/cofactor. It does not seem that the FbiC reaction is the limiting step of the pathway even when the media are not supplemented with L-glutamate/ manganese; excess FO was always present in high quantities in the media, indicating that the over-expressed FbiA and FbiB proteins are still not capable of converting all FO to F 420 . An alternative possibility is that FbiA and FbiB need other accessory protein(s) in order to perform the conversion more efficiently; in fact another ORF in M. smegmatis (MSMEG_2392) has been shown, by transposon mutagenesis studies, to be involved in F 420 biosynthesis from FO [35]. Biochemical studies need to be performed using purified enzymes in order to study the kinetics in detail and determine the rate limiting step of the pathway.
Our previous crystal structures of M. tuberculosis FGD1 [9], together with other F 420 -containing crystal structures from different Archaeal species [36,37,38,39], have indicated that the F 420 poly-glutamate tail is not required for reaction catalysis; the poly-glutamate tail is extended into the solvent and it seems that this is a conserved feature of the enzymes that use F 420 in oxidoreduction reactions. We propose that, therefore, the high yields of F 420 from M. smegmatis strains expressing the recombinant FbiABC proteins, regardless of the number of glutamate residues in the poly-glutamate tail, identify this as a valuable source of F 420 that might be used with enzymes purified from other microorganisms. Furthermore, the high yield of FO/F 420 also opens a door for possible biotechnological applications.

Depositions
The nucleotide sequences for pYUB28b and pYUBDuet vectors have been deposited in the National Centre for Biotechnology Information (NCBI) under GenBank HQ247814 and HQ247815 accession numbers, respectively. The vectors are available upon request. Figure 5. The F 420 production profile from M. smegmatis cells over-expressing the FbiABC construct. The F 420 -6 and F 420 -7 species constitute the main species as deduced using mass spectrometry. doi:10.1371/journal.pone.0015803.g005