GMC oxidoreductases

GMC oxidoreductases

J. Mol. Biol. (1992) 223, 811-814 GMC Oxidoreductases A Newly Defined Family of Homologous Catalytic Activities Proteins with Diverse Douglas R. Ca...

425KB Sizes 1 Downloads 33 Views

J. Mol. Biol. (1992) 223, 811-814

GMC Oxidoreductases A Newly Defined Family of Homologous Catalytic Activities

Proteins with Diverse

Douglas R. Cavener Dkpartment

of Molecular Biology, Vanderbilt Nashville, TN 37235, U.S.A.

University

(Received 24 July 1991; accepted 22 October 1991) Sequence comparison of Drosophila melanogaster glucose dehydrogenase, Eseherichia coli choline dehydrogenase, Aspergillus niger glucose oxidase and Hansenula polymorpha methanol oxidase indicates that these four diverse flavoproteins are homologous, defining a new family of proteins named the GMC oxidoreductases. These enzymes contain a canonical ADP-binding B&fold close to their amino termini as found in other flavoenzymes. This domain is encoded by a single exon of the D. melanogaster glucose dehydrogenase gene. Keywords:

evolution;

oxidoreductase;

t Abbreviations used: GLD, glucose dehydrogenase; MOX, methanol oxidase; GOX, glucose oxidase; CHD, choline dehydrogenase; GMC, glucose-methanol-choline; GR, glutathione reductase; LpDH, lipoamide dehydrogenase; PND, pyridine nucleotide disulfide. $03.00/O

methanol;

choline

enzymes contain a variable number of unrelated residues at their amino termini before the wellconserved ADP-binding site. With the exception of a 60 residue insertion found in MOX, the sequences require only a few small insertion/deletions to optimize their alignment. Pairwise sequence identities range between 23 to (Table 1). 32% D. melanogaster GLD and E. coli CHD exhibit the highest degree of similarity despite the obvious expectation that GLD and GOX should since they at least share glucose as a substrate. The statistical significance of each pairwise alignment was evaluated by comparing the similarity of the primary amino acid sequence of one enzyme with 200 randomized sequences of the other enzyme using the RDF2 program of W. Pearson (University of Virginia, U.S.A.). In all cases the alignment score of the true sequences were greater than 20 standard deviations above the mean of the alignment scores of the randomized sequences of the same amino acid composition, indicating a highly significant match of the true sequences. These four proteins were also compared to four other flavoenzymes, human glutathione reductase (GR; Thieme et al., 1981) Axobacter vinelandii lipoamide dehydrogenase (LpDH; Westphal et al., 1988), Xtreptomyces cholesterol oxidase (Ishizaki et al., 1989) and Pseudomonas aeruginosa p-hydroxybenzoate hydroxylase (Entsch et al., 1988) using the RFD2 program. These comparisons were statistically insignificant; all alignment scores were less than one standard deviation above the mean of the scores of 200 randomized sequences for each pairwise comparison. For GR

Recently the sequences of a diverse group of flavoenzymes have been reported including glucose dehydrogenase (GLDT; EC 1.1.99.10) from Drosophila melanogaster (Krasney et al,, 1990)) methanol oxidase (MOX; EC 1.1.3.13) from the yeasts Hansenula polymorpha (Ledeboer et al., 1985) and Pichia pastoris (Koutz et al., 1989), glucose oxidase (GOX; EC 1.1.3.4) from Aspergillus niger (Frederick et al., 1990) and choline dehydrogenase (CHD; EC 1.1.99.1) from Escherichia coli (Lamark et al., 1991). With the exception of GLD and GOX, these enzymes use different substrates (Fig. 1). In addition, GOX and MOX yield hydrogen peroxide as a product whereas GLD and CHD do not. As expected, these four flavoenzymes contain a 30 amino acid region near their amino termini (Fig. 2) corresponding to the well-characterized ADP-binding pap-fold (Wierenga et al., 1983). Given the disparate reactions that these enzymes catalyze, the remaining primary sequence would be expected to be dissimilar. Surprisingly however, these four are similar throughout their primary enzymes sequence and are thus undoubtedly homologous proteins (Fig. 2). Accordingly I propose to name this group of homologous proteins the glucose-methanoloxidoreductase family. These choline (GMC)

0022%2836/92/030811~4

glucose;

811

0 1992 Academic Press Limited

D. R. Cawner

812 Choline choline

Glucose glucose

dehydrogenase + unknown

(CHD)

acceptor

dehydrogenase + unknown

---

9

betaine

aPdehyde

+

reduced

acceptor

(GLD)

acceptor

9

6. g 1uconolactone

9

6-gluconolactone

f reduced

acceptor

Glucose oxidase (GOX) glucose + 02 Methanol

oxidase

methanol

Figure

(MOX)

9

+ 02

1. Catalytic

activities

and LPDH (and probably for p-hydroxybenzoate hydroxylase) this result is not surprising since a large fraction of their sequence is involved in binding NADPH (Theime et al., 1981), a cofactor not utilized by the GMC family of enzymes. The GMC oxidoreductase family demonstrates the flexibility of evolution to produce enzymes with diverse catalytic activities while retaining the same overall archit,ecture. Perhaps the three-dimensional structure of the ancestral protein of this family was compatible with binding to a number of substrates yielding a low level of catalysis for each. Subtle mutational refinements could have led to the high specificity displayed by the extant enzymes. Unfortunately the catalytic centers have not been mapped for any of the four enzymes. Moreover, two of the enzymes, GLD and CHD, have unknown acceptors. The evolutionary origins of the GMC oxidoreduc-

Table 1 Protein

sequence similarity among oxidoreductases GLD

GLD CHD GOX iMoxt

32.0 24.6 234

CHD (74.8) 27.3 25.4

GOX (33.4) (282)

+ II202

the GMC

MOX (37.3) (356) (21.3)

23.8

The “/b similarities for each pairwise comparisons are given in the lower left sector. The significance of these similarities was evaluated using the RDF2 program of W. Pearson. For each comparison 200 randomized protein sequences were generated based upon the amino acid composition of 1 of the 2 proteins. Alignment scores were calculated for each random sequence as compared to the other protein sequence. Then the alignment score of the true sequence was compared to the distribution of the 200 alignment scores for the randomized sequences. Shown in the upper right sector in parentheses is the number of standard deviations above the mean for each of the alignment scores of the true sequences. More than 3 standard deviations above the mean is highly significant statistically (p << 0.01). t For these calculations the large MOX insertion relative to the other 3 enzymes was deleted.

acetaldehyde

+ Hz02

of the G:MC oxidoreductases.

tase and pyridine nucieotide disulfide (PND) oxidoreductase (e.g. GR, LpDH and mercuric reductase) families both predate the divergence of prokaryotes and eukaryotes. These two enzyme groups appear to have evolved at a simila,r rate since pairwise comparisons between family members within each group share approximately 25”/b identity (Table 1; Greer & Perham, 1986). The striking feature of both groups is the presence of the ADP-binding flap-fold always found near the amino terminus. The high degree of similarity of the ADP binding site and its invariant position near the amino Derminus, strongly argues that it is an homologous domain. Since the GMC and PND oxidoreductase families are otherwise completely divergent, I speculate that these two families share a single primordial exon encoding the ADP binding domain. Indeed, the ADP-binding domain of Drosophila GLD is contained entirely within a 40 codon exon (Krasney et aZ.; 1990; Cavener & Krasney, 1991). Presumably this exon was shuffled or alternatively spliced into the ancestral genes of the GMC and PND families. This is an important paradigm relative to the debate over the antiquity of introns (Gilbert, 1978; Doolitt,le, 1985; Dorit et aZ.: 1990; Patthy, 1991) since the examples of protein domains corresponding to exons of ancient proteins are hotly debated and few in number. With the exception of the yeast methanol oxidase gene, the exon-intron structures of the other PND and GMC genes are unknown. The methanol oxidase gene is not particularly informative since it contains no introns (Ledeboer et al., 1985; Koutz et al.; 19891, typical of most yeast genes. It would be particularly useful to determine the exon-intron structure of the human glutathione reductase and hpoamide dehydrogena,se genes since mammalian genes are t’ypicaliy riddled w&h introns. The author is indebted to Dr Steven HenikofF for running the PAThIAT matrix analysis and to Beth Cavener for technical assistance. This work was support,ed by PiTH GM34170.

. CHD Gm [email protected]?TaIt mx

IRSQCD~PCGPASS KB’LWSSLWs~HYIFSNG~LLl’DP~V

60 37

3

I!AI

. axI l$JFDYIII~~S--~DYFFD&~--~ (;ID PDYEXDFM-m ax ScXrVDYIIs-VL~-~RGPIIPD~Y~ KXpD-mem

54 1U 92 6.t

[email protected]

cm GID Gss

e--c’--r-SiPv -1GSDlD~USZNKXYWPRQWItZT~~ IFGssvD~TvE-

110 172

xx

IDSKTATF-YSSRPS-KA-LNG-RRAIVPCANILM;GSS

LJXBWQTmGNGI%GsTL

ml P~YYR%&‘-----NDY~‘P--P;I GtDP

V

147

WDYDLXR?

115 157 220

aXPVG-KFPYN-PPL

GX-&~~YSPr

207

XX

160

-ZNSTDELLPLMW”““‘m-PT

cand&adR-m&'G2l.3 GKDSIAIVS--~W~mSWS~aYxVKAIMSAVEDRGVPm-~~-aPNYCJRP---N263 KRragFLmAm~~FKTs-~ . CZID Grn ax IKR

LT-HII IHI~LIWH-~-vSLQFGs-TFKIL llfJvLl~~RA~~~~IL IFLITSTKCDKVIIE-mm-mWB

cm GID ax

QFWW&BDII?L~ IJsGvam--ADTAPw EYSGIWKSILE!?IG~W-~~-

=QRsGI-

274 S

219

VGVEN&STIP-I&T&I;

267 329 320

VfMEVVLSA(14VNSPHIL

278

MSCGTISSPLVL &-2-r;Lyp&

215 319 378 333

WHYm-mB .

.

.

IQYI-Fm?

mD imqp&-bB GD ~I~~L~QL= mx -TFm-Ysaw KXAAEDQWY~I-vem

367 427 422

Kr-awmA-m-m

39

caDv,@'-b'AIIFG

414

QD ~-~;rS.RSIQIZPAWNPRB3?I~+lYL Q3XDWIWNVAYSELFWI-~~~YF FXX DKPIMHYSVI~~

480 477 447

ST(IFVRITS?%W+%PDFDm .

cslDs*d&w&

l

QD TH-ElmvmLmm-~~w axIJT-EID~~sN~ rmxmmEaxAP~'""~HHe~sP~m~ . . m Gm ax &fnclTGExKL~YH~~KPTPKND~LH.sD~ cGl&uMm& (31DETAFW G;u) GP~SHDP.-,‘IMPK-VSSWIWAPA axRP-M”“mmS k&X EWEGSKIAPKamm-S% a3DIMIW

GmlnaiEK-~ CXDt -TIG%Wl!LVAl3XGYSGSDI.D

KIXAMA-LKISDAIL~~ Figure 2. Sequence alignment of the alignments and the PATMAT (Henikoff regions. The results of these 2 methods domain is underlined.

'

447 514 510 503

*

467 539

. W& T I533

-VFmlA

563 Sw&IvH~~~II~

519

593 587 623

555 612 605 664 GMC oxidoreductases. The FASTA program was used to perform et al.; 1990) programs to perform multiple alignment of several were combined to arrive at the final alignment. The ADP-binding

pairwise conserved fiafi-fold

814

D. R. Cavener Murooka, Y. (1989). Nucleotide sequence of Lhe gene for cholesterol oxidase from a Streptomyces sp. J. Bacterial. 171, 5966601. Koutz, P., Davis; G. R., Stillman, C., Barringer, K., Gregg; J. & Thill, G. (1989). Structural comparison of the Pichia pastoris alcohol oxidase genes. Yeast, 5. 167-177. Krasney, P. A., Carr, C. M. & Cavener, D. R. (1990). Evolution of the ghmose dehydrogenase gene in

References Cavener, D. R. & Krasney, P. A. (1991). Drosophila glucose dehydrogenase and yeast alcohol oxidase are homologous and share N-terminal homology with other flavoenzymes. Mol. Biol. Evol. 8, 144-150. Doolittle, R. F. (1985). The genealogy of some r cently evolved vert,ebrate proteins. Trends Bioche w/ Sci. 10, 233-237. Dorit, R. L., Schoenbach, L. & Gilbert, W. (1990). How big is the universe of exons? Science, 250, 1377-1382. Entsch, B., Nan, Y., Weaich, K. & Scott, K. F. (1988). Sequence and organization of pobA, the gene coding for p-hydroxybenzoate hydroxylase, an inducible enzyme from Pseudomonas aeruginosa. Gene, 71, 279-291. Frederick, K. R., Tung, J., Emerick, R. S., Masiarz, F. R., Chamberlain, S. H., Vasavada, A., Rosenberg, S., Chakraborty, S., Schopfer, L. M. & Massey, M. (1990). Glucose oxidase from Aspergillus niger. Cloning, gene sequence, secretion from Saccharomyces cerevisiae and kinetic analysis of a yeast-derived enzyme. J. Biol. Chem. 265, 3793-3802. Gilbert, W. W. (1978). Why genes in pieces? Nature (London), 271, 501. Greer, S. & Perham, R. N. (1986). Glutathione reductase from Escherichia coli: cloning and sequence analysis of the gene and relationship to other flavoprotein disulfide oxidoreductases. Biochemistry, 25, 2736-2742. Henikoff, S., Wallace, J. C. & Brown, J. P. (1990). Finding protein similarities with nucleotide sequence databases. Methods Enzymol. 183, 111-132. Ishizaki, T., Hirayama, N., Shinkawa, H., Nimi, 0. C

Edited

by

Drosophila.

Mol.

Biol.

Evol.

7, 155-177.

Lamark, T., Kaasen; I., Eshoo, M. W., Falkenberg. I”., McDougall, 3. & Strom, A. R. (1991). DNA sequence and analysis of the bet genes encoding the osmoregulatory choline-glycine betaine pathway of Eaeherichia coli. Mol. Microbial. 5, 1049-1064. Ledeboer, A. M., Edens, L., Maat, J., Visser, C., Bos, J. W. 6 Verrips, C. T, (19%). Molecular cloning and characterization of a gene coding for methanol oxidase in Mansenula polymorpha. Nucl. Acids Res. 13, 306333082. Patthy, L. (1991). Exons-Original building blocks of proteins. BioEssays, 4, 187-192. Thieme, R., Pai, E. F., Schirmer, R. H. & Schulz, G. E. (1981). Three-dimensional structure of glutathione reductase at 2 A resolution. J. Mol. Biol. 152, 763-782. Westphal, A. H. & de Kok, A. (1988). Lipoamide dehydrogenase from Azotobacter vinelandii. Molecular cloning, organization and sequence analysis of the gene. Eur. J. Biochem. 172, 299-305. Wierenga, R. K.; Drenth, J. & Schulz, G. E. (1986). Prediction of the occurrence of the ADP-binding /[email protected] in proteins, using an amino acid sequence fingerprint. J. Mol. Biol. 167, 725-739.

F.

Cohen