118
Biochemical Society Transactions (2006) Volume 34, part 1
Developing structure-based models to predict substrate specificity of D-group (Type II) molybdenum enzymes: application to a molybdo-enzyme of unknown function from Archaeoglobus fulgidus E.J. Dridge*, D.J. Richardson†, R.J. Lewis* and C.S. Butler*1 *Institute for Cell and Molecular Biosciences, University of Newcastle, Newcastle upon Tyne NE2 4HH, U.K., and †School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, U.K.
Abstract The AF0174–AF0176 gene cluster in Archaeoglobus fulgidus encodes a putative oxyanion reductase of the D-type (Type II) family of molybdo-enzymes. Sequence analysis reveals that the catalytic subunit AF0176 shares low identity (31–32%) and similarity (41–42%) to both NarG and SerA, the catalytic components of the respiratory nitrate and selenate reductases respectively. Consequently, predicting the oxyanion substrate selectivity of AF0176 has proved difficult based solely on sequence alignments. In the present study, we have modelled both AF0176 and SerA on the recently determined X-ray structure of the NAR (nitrate reductase) from Escherichia coli and have identified a number of key amino acid residues, conserved in all known NAR sequences, including AF0176, that we speculate may enhance selectivity towards trigonal planar (NO3 − ) rather than tetrahedral (SeO4 2− and ClO4 − ) substrates.
Introduction
Substrate selectivity of AF0176
Archaeoglobus fulgidus is a strict anaerobe found only in hydrothermal vents and has been isolated near Italy, in the North Sea and from hot springs in Yellowstone National Park. Although its respiratory metabolism is considered to be that of a strict sulphate reducer, the genome sequence of A. fulgidus suggests that this may not be the case [1] and it has been suggested that several molybdopterin-binding oxidoreductases, with polysulphide, nitrate, DMSO and thiosulphate as potential substrates, might contribute to energizing the cell membrane [1]. Of particular interest is the gene cluster AF0174–AF0176 that has been postulated to encode an unusual respiratory nitrate reductase of the NAR (nitrate reductase) family [2]. However, further analysis of this gene cluster and the translated proteins raises questions regarding its substrate specificity. Notably, the distinct similarity between AF0175–AF0176 and SerAB (catalytic components of periplasmic selenate reductase) from Thauera selenatis [3], combined with the fact that deep hydrothermal vent fluids are rich in metalloid compounds including selenium and arsenic [4], suggests that AF0174–AF0176 could equally participate in a selenate reduction/respiration pathway (Figure 1A).
AF0176 is a putative molybdopterin-binding subunit and shows 32% identity with and 42% similarity to SerA and is of a similar size (99.4 kDa). It also shows 31% identity with and 41% similarity to NarG (catalytic subunit of membranebound nitrate reductase) from Escherichia coli, but is approx. 40 kDa smaller. AF0176 has a cysteine-rich cluster (motif: CX3 CX3 CX29 C) towards the N-terminal end of the polypeptide thought to be involved in the co-ordination of an iron–sulphur cluster, probably of the [4Fe-4S] type. This motif is also conserved in both SerA and NarG; however, in both cases, the first cysteine residue is replaced with a histidine. The involvement of this motif in the co-ordination of an iron–sulphur cluster in NarG has been shown recently in the crystal structure. The presence of this iron–sulphur cluster in SerA has yet to be confirmed. The aspartate residue that co-ordinates the Mo in the bis-MGD (bis-molybdopterin guanine dinucleotide) cofactor (Asp222 in E. coli NarG and Asp209 in T. selenatis SerA) is also conserved in AF0176 (Asp180 ) and places this putative nitrate reductase in the Dgroup (Type II) of molybdo-enzymes (Figure 1B). The Nterminus has a leader sequence that is predicted to be cleaved between residues 18 and 19 of the sequence [5]. The leader peptide has a twin-arginine motif (residues 5 and 6) and, consequently, AF0176 is predicted to be translocated across the membrane by an archaeal TAT (twin-arginine translocase) system [6]. This is in contrast with the cytoplasmic facing NAR-type nitrate reductase from the mesophilic bacteria. The presence of this TAT leader sequence places AF0176 in a
Key words: Archaeoglobus fulgidus, hyperthermophile, molecular modelling, molybdenum, selenate reductase, tungsten. Abbreviations used: NAR, nitrate reductase; bis-MGD, bis-molybdopterin guanine dinucleotide; NarG, catalytic subunit of membrane-bound nitrate reductase; SerA, catalytic subunit of periplasmic selenate reductase; TAT, twin-arginine translocase; TMAO, trimethylamine N-oxide. 1 To whom correspondence should be addressed (email
[email protected]).
C 2006
Biochemical Society
The 11th Nitrogen Cycle Meeting 2005
Figure 1 Diagram showing similarities and differences between nitrate reductase, selenate reductase and AF0174–AF0176 (A) Subunit composition of selenate reductase (SerABC) from T. selenatis, nitrate reductase (NarGHI) from E. coli and putative oxyanion reductase (AF0174–AF0176) from A. fulgidus. (B) Segment alignment showing selected residues that may confer substrate selectivity. Mo ligand Asp209/180/222 is highlighted by an open box. TS, T. selenatis SerA; AF, A. fulgidus AF0176; EC, E. coli NarG.
distinct subgroup of D-group molybdo-enzymes and in addition to SerA, sequence alignment predicts that other members include chlorate reductase (ClrA; Ideonella dechloratans [7]), dimethyl sulphide dehydrogenase (DdhA; Rhodovulum sulfidophilum [8]), ethylbenzene dehydrogenase (EbdA; Azoarcus sp.-like strain EbN1 [9]) and nitrate reductase (Nar1; Haloarcula marismortui [10]). AF0175 (27.8 kDa) is a putative molybdopterin oxidoreductase iron–sulphur-binding subunit and shows 29% identity with and 37% similarity to SerB and very low identity with NarH, the iron–sulphurcontaining component of the nitrate reductases. AF0175 contains four cysteine-rich clusters (each with four cysteine residues) that are highly conserved with SerB and are predicted to bind four [4Fe-4S] iron–sulphur clusters. AF0174 (45.1 kDa) is a membrane subunit and shares little similarity to other sequences in the genome database. Structural analysis using 3D-PSSM has revealed low structural similarity to NrfD,
the membrane component of the Nrf pathway that conducts electrons from the quinol pool to the terminal components in the periplasm. The final component AF0173 (17.6 kDa) is predicted to encode a homologue of SerD and NarJ, and act as a specific chaperone protein possibly involved in the assembly of AF0176 [11]. The structural similarity and low sequence identity of AF0176 with both SerA and NarG provides us with a unique opportunity to identify particular amino acid residues within the substrate entry channel and around the Mo cofactor that may control substrate selectivity of the D-group oxyanion reductases. Using molecular modelling based on the recent three-dimensional X-ray structure of NarG from E. coli [12], together with the amino acid sequence of selenate reductase from T. selenatis, we have identified a number of key residues absolutely conserved in all known NAR sequences, including AF0176, that we speculate may enhance selectivity towards C 2006
Biochemical Society
119
120
Biochemical Society Transactions (2006) Volume 34, part 1
Figure 2 Molecular structure of oxyanion reductase active site Homology modelling indicates that some residues around the Mo (grey sphere) cofactor are not conserved amongst nitrate and selenate reductases. For instance, although the Asp222 in NarG (blue) is conserved in SerA and AF0176 (Asp180 , green), Asn52 in NarG and AF0176 (Asn49 ) is replaced with the chemically inert Gly76 in SerA (green sphere). For clarity, one of the pterin cofactors has been excluded from the Figure.
trigonal planar (NO3 − ) rather than tetrahedral (SeO4 2− and ClO4 − ) substrates. Of particular interest are residues that surround the Mo cofactor (Asn49 , Tyr175 and Pro182 ) and those that dictate the conformation of the putative substrate entry channel [Thr51 , Gln192 and Thr193 (numbers refer to position in the AF0176 sequence); Figure 1B]. These residues are replaced by Gly76 , Ser204 , Tyr211 , Val78 , Ala221 and Arg222 in SerA and in future studies will represent excellent targets for site-directed mutagenesis analysis. The asparagine residue ˚ (1 A ˚ = 0.1 nm) away (Asn52 in NarG) is positioned at 3.9 A from the Mo atom, serves a structural role positioning the [4Fe-4S] cluster and could form a hydrogen bond to bound substrate, potentially dictating an active site suitable for binding nitrate specifically [12]. Our model shows that this asparagine residue is also conserved in AF0176 (Asn49 ), but is replaced with glycine (Gly76 ) in the active site of SerA (Figure 2) and clearly provides a more open structure around the Mo cofactor. Furthermore, a tyrosine residue (Tyr220 ) in NarG, which is replaced by asparagine (Asn178 ) in AF0176, is further substituted for threonine (Thr207 ) in SerA, again making the active site of SerA more open and hydrophilic. Interestingly, the SerA residues Gly76 , Thr207 , Tyr211 , Val78 and Ala221 are conserved also in the chlorate reductase from I. dechloratans, an enzyme that reduces selenate, bromate, iodate and nitrate in addition to chlorate [7]. Based on these structural predictions, it would seem highly likely that C 2006
Biochemical Society
AF0176 will display nitrate reductase activity and may not selectively reduce selenate.
Is AF0176 a molybdenum- or tungsten-containing enzyme? The growth of hyperthermophilic archaea appears to be obligately dependent on tungsten and some are incapable of utilizing molybdenum: consequently, they express a number of novel tungsto-enzymes [13]. These tungsten complexes have greater thermostability than their molybdenum counterparts due to stronger π -interactions, and this may explain why they are utilized at temperatures near boiling point. In the context of this work, this raises an interesting question. Is AF0176 a tungsto- or molybdo-enzyme? Given that the atomic and ionic radii and the chemical properties (oxidation states IV, V and VI) of tungsten are very similar to those of molybdenum, there has been much interest to see whether tungsten can substitute for molybdenum in a range of molybdo-enzymes and retain their activity [14]. A number of publications have reported that the replacement of molybdenum by tungsten results in inactive proteins, including nitrogenase from Azotobacter vinelandii [15], nitrate reductase from plants [16], hepatic sulphite oxidase [17] and formate dehydrogenase from Methanobacterium formicicum [18]. It has been suggested that molybdenum is more suitable for catalysing reactions with relatively
The 11th Nitrogen Cycle Meeting 2005
high redox potentials and the reason why tungstensubstituted enzymes are inactive is because of the lower reduction potential of the W(VI)/(V) and W(V)/(IV) couples. However, two bacterial periplasmic molybdo-enzymes, the DMSO reductase from Rhodobacter capsulatus [19] and the periplasmic TMAO (trimethylamine N-oxide) reductase from E. coli [20], can function with tungsten at the active site, resulting in an enzyme with increased catalytic efficiency, increased thermal stability and broader substrate selectivity [20]. Furthermore, the periplasmic nitrate reductase (NAP) from Paracoccus pantotrophus can also function with W at the active site, albeit with reduced activity and decreased substrate binding affinity [21]. These enzymes, although TATtranslocated, are classified as Type I (C-group) reductases, in which the MGD cofactor is covalently attached to the protein via a cysteine ligand rather than the aspartate ligand in the D-group reductases. These Type I (C-group) reductases have yet to be identified in the hyperthermophilic archaea. However, these studies do demonstrate that in some enzyme systems both molybdenum and tungsten can not only catalyse the same chemical reaction, but also catalyse reactions at relatively high redox potential [TMAO/TMA (trimethylamine), Em = +130 mV; DMSO/DMS, Em = +160 mV; nitrate/nitrite, Em = +400 mV]. Recent biochemical characterization of the TAT-dependent nitrate reductase from Pyrobaculum aerophilium has revealed that the enzyme displays very high catalytic activity (V max for nitrate ∼1162 s−1 ) and has adapted to function with either molybdenum or tungsten at the active site [22].
Conclusions The archaeal gene cluster AF0174–AF0176 from A. fulgidus is postulated to be either a nitrate and/or selenate reductase due to its sequence identity with both the selenate reductase SerA and the nitrate reductase NarG, with a conserved aspartate residue (Asp180 ) placing the putative reductase in the D-group (Type II) of molybdo-enzymes. Molecular modelling has identified a number of residues that may control substrate specificity of the enzyme, and based on this analysis, we predict that nitrate is a more likely substrate than selenate. Although predicted to be a molybdo-enzyme, the growth of the hyperthermophilic archaea may be obligately dependent on tungsten; we therefore suggest that tungsten may also function at the active site of AF0176. In order to test these hypotheses, we are currently developing expression systems for AF0174–AF0176 in E. coli with the aim of determining
substrate selectivity, probing the role of key active site residues using site-directed mutagenesis and assessing the effects of exchanging molybdenum/tungsten at the active site. This work was supported by BBSRC (Biotechnology and Biological Sciences Research Council) research grants 13/P17219 and BBS/B/ 10110 to C.S.B. and D.J.R. E.J.D. is in receipt of a BBSRC PMS Committee Ph.D. studentship.
References 1 Klenk, H.P., Clayton, R.A., Tomb, J.F., White, O., Nelson, K.E., Ketchum, K.A., Dodson, R.J., Gwinn, M., Hickey, E.K., Peterson, J.D. et al. (1997) Nature (London) 390, 364–370 2 Richardson, D.J., Berks, B.C., Russell, D.A., Spiro, S. and Taylor, C.J. (2001) Cell. Mol. Life Sci. 58, 165–178 3 Schroder, I., Rech, S., Krafft, T. and Macy, J.M. (1997) J. Biol. Chem. 272, 23765–23768 4 Butterfield, D.A., McDuff, R.E., Mottl, M., Lilley, M.D., Massoth, G.J. and Lupton, J.E. (1994) Geophys. Res. 99, 9561–9583 5 Nielsen, H., Engelbrecht, J., Brunak, S. and von Heijne, G. (1997) Protein Eng. 10, 1–6 6 Palmer, T. and Berks, B.C. (2003) Microbiology 149, 547–556 7 Danielsson, H., Stenklo, K., Karlsson, J. and Nilsson, T. (2003) Appl. Environ. Microbiol. 69, 5585–5592 8 McDevitt, C.A., Hugenholtz, P., Hanson, G.R. and McEwan, A.G. (2002) Mol. Microbiol. 44, 1575–1587 9 Johnson, H.A., Pelletier, D.A. and Spormann, A.M. (2001) J. Bacteriol. 183, 4536–4542 10 Yoshimatsu, K., Iwasaki, T. and Fujiwara, T. (2002) FEBS Lett. 516, 145–150 11 Krafft, T., Bowen, A., Theis, F. and Macy, J.M. (2000) DNA Sequence 10, 365–377 12 Jormakka, M., Richardson, D., Byrne, B. and Iwata, S. (2004) Structure 12, 95–104 13 Johnson, M.K., Rees, D.C. and Adams, M.W.W. (1996) Chem. Rev. 96, 2817–2840 14 Garner, C.D. and Stewart, L.J. (2002) Met. Ions Biol. Syst. 39, 699–726 15 Hales, B.J. and Case, E.E. (1987) J. Biol. Chem. 262, 16205–16211 16 Deng, M.D., Moureaux, T. and Caboche, M. (1989) Plant Physiol. 91, 304–309 17 Johnson, J.L., Cohen, H.J. and Rajagopalan, K.V. (1974) J. Biol. Chem. 249, 5046–5055 18 May, H.D., Patel, P.S. and Ferry, J.G. (1988) J. Bacteriol. 170, 3384–3389 19 Stewart, L.J., Bailey, S., Bennett, B., Charnock, J.M., Garner, C.D. and McAlpine, A.S. (2000) J. Mol. Biol. 299, 593–600 20 Buc, J., Santini, C.L., Giordani, R., Czjzek, M., Wu, L.F. and Giordano, G. (1999) Mol. Microbiol. 32, 159–168 21 Gates, A.J., Huges, R.O., Sharp, S.R., Millington, P.D., Nilavongse, A., Cole, J.A., Leach, E.R., Jepson, B., Richardson, D.J. and Butler, C.S. (2003) FEMS Microbiol. Lett. 220, 261–269 22 Afshar, S., Johnson, E., de Vries, S. and Schroder, I. (2001) J. Bacteriol. 183, 5491–5495 Received 22 September 2005
C 2006
Biochemical Society
121
