Israel Hanukoglu and Tamar Gutfinger
Department of Hormone Research, Weizmann Institute of Science, Rehovot 76100, Israel
(This paper reports the discovery of the NADP binding site consensus motif. The findings of this study were further confirmed by structural analysis of the enzyme (see reprint).
The full reference for this article is: Hanukoglu I., and Gutfinger T. cDNA sequence of adrenodoxin reductase. Identification of NADP-binding sites in oxidoreductases. Eur J Biochem 180:479-484, 1989.
Enzyme Commission number: Adrenodoxin reductase (E. C. # 1.18.1.2)
Summary
Adrenodoxin reductase is an NADP dependent flavoenzyme which functions as the reductase of mitochondrial P450 systems. We sequenced two adrenodoxin reductase cDNAs isolated from a bovine adrenal cortex cDNA library. The deduced amino acid sequence shows no homology with the sequence of the microsomal P450 systems or other known protein sequences. Nonetheless, by sequence analysis and comparisons with known sequences of dinucleotide binding folds of two NADP binding flavoenzymes two regions of adrenodoxin reductase sequence were identified as the FAD and NADP binding sites. These analyses revealed a consensus sequence for the NADP binding dinucleotide fold (GxGxxAxxxAxxxxxxG - in one letter amino acid code) that differs from FAD and NAD binding dinucleotide fold sequences. In the data base of protein sequences the NADP binding site sequence appears solely in NADP-dependent enzymes, the binding sites of which were not known to date. Thus, this sequence may be used for identification of a certain type of NADP binding sites of enzymes that show no significant sequence similarity.
Introduction
Adrenodoxin reductase is the first enzyme in the mitochondrial P450 systems which catalyze several critical steps in the biosynthesis of steroid hormones in steroidogenic tissues [1-3], and the biosynthesis of bile acids [4, 5] and vitamin D [6] in the liver and kidney, respectively. These P450 systems are composed of three enzymes which constitute an electron transfer chain located on the matrix side of the inner mitochondrial membrane [7, 8]. The function of adrenodoxin reductase is to transfer electrons from NADPH to a specific ferredoxin (adrenodoxin), which in turn transfers them one at a time to the P450 [1, 9, 10]. Adrenodoxin reductase cDNA has been recently isolated from both bovine and human steroidogenic tissue cDNA libraries [11-14] and used to demonstrate that adrenodoxin reductase is encoded by a single gene in both genomes [11, 14]. Thus, the same adrenodoxin reductase serves the different mitochondrial P450 systems in steroidogenic tissues.
As an NADP dependent electron transfer protein, adrenodoxin reductase belongs to a group of NAD and NADP dependent oxidoreductases which constitute a large collection of enzymes that vary widely in their sizes, substrate specificities, and sequences. The structures of a large number of NAD binding proteins have been studied extensively. Comparisons of the NAD binding domains of these enzymes indicate that their dinucleotide (ADP) binding sites share a similar βαβ fold which forms a pocket to accommodate the coenzyme [15-18]. The sequences of these binding sites reveal a nearly universally conserved Gly-X-Gly-X-X-Gly sequence which appears at the border between the first β-sheet strand and the α-helix allowing the formation of a tight turn [17, 18]. The structures of only two NADP binding proteins have been well characterized to date [19, 20]. Glutathione reductase, like adrenodoxin reductase, is an FAD containing flavoenzyme. The ADP binding sites in both its FAD and NADP domains form a βαβ fold similar to that of NAD [16, 17, 19] (For convenience we shall use the terms FAD and NADP "binding sites" to refer to the binding site for the ADP portion of these molecules). In contrast, the NADP binding site of dihydrofolate reductase shows a different structure [20].
In the absence of sequence information on adrenodoxin reductase its structural relationship to other NAD(P) dependent enzymes remained unknown. In this study we report the cDNA sequence of adrenodoxin reductase and examine its similarity to other oxidoreductases. By sequence comparisons and analyses one region of adrenodoxin reductase is identified as the NADP binding site, and a modified dinucleotide binding site consensus sequence is used to uniquely identify additional NADP binding enzymes in the protein sequence databases.
Materials and methods
The pAR cDNA was isolated from a lambda gt11 cDNA expression library constructed from bovine adrenal cortex mRNA [11]. We sequenced an additional cloned cDNA which was identified by rescreening of the cDNA library using the pAR cDNA as a probe.
The DNA sequencing was done as previously described [21-23]. BamHI, BanI, BsshII, EcoRI, SalI, and XmaI sites, and the first AvaI site from the 5' end of clone pAR were used for end labelling (Fig. 1). Secondary digestions were done with BglI, HphI, KpnI, PstI or one of the enzymes mentioned above.
The NBRF and GenBank sequence databases were searched with the programs Seqnce (Delaney Software, Vancouver, Canada) and MicroGenie [24] on an IBM Personal Computer, and University of Wisconsin-Madison Sequence Analysis Program on a Micro Vax II computer. Homology matrix comparisons and hydrophobicity and protein secondary structure analyses, were done as described [23, 25].
Results
The characteristics of the sequenced cDNAs
The correct reading frame of the pAR cDNA insert was identified using the sequences of partial tryptic fragments of adrenodoxin reductase which we had determined previously [11] (Fig. 1). Our sequence is in nearly perfect agreement with the sequence of Sagara et al. [12]. However, the sequences of these four independent clones show many differences from the cDNA sequence of adrenodoxin reductase recently reported by Nonaka et al. [13]. The differences include many insertions and deletions that alter the reading frame of translation for tens of residues. These discrepancies are probably a result of sequencing errors as their sequence was not cross-checked by sequencing an independent clone.
Adrenodoxin reductase sequence shows no homology with other oxidoreductase sequences
The comparison of the sequence of adrenodoxin reductase with the NBRF protein sequence database revealed no significant matches with any sequence using the programs noted in Materials and Methods. To detect low but significant similarity between adrenodoxin reductase and other oxidoreductases we carried out homology matrix analyses at low identity cutoff values (>25%) for segments of 30 residues (the FAD and NADP binding sites are 30 residues in other oxidoreductases). These analyses also failed to reveal similarity between adrenodoxin reductase and other oxidoreductase sequences including microsomal P450 reductase [26-28] and spinach ferredoxin reductase [29].
Identification of the FAD binding site of adrenodoxin reductase
In adrenodoxin reductase the dinucleotide binding site consensus sequence [30] is found only at the amino terminus region (Fig. 2). Secondary structure prediction analysis of this region provided further support for the identification of this segment as an ADP binding βαβ-fold. Sagara et al. [12] also noted that this region may be an FAD or NADP binding site. But, as the FAD binding domains of many flavoenzymes appear to be located close to the amino terminus of the protein, this region is most likely the FAD binding site of adrenodoxin reductase (Fig. 2).
Identification of the NADP binding site of adrenodoxin reductase
The adrenodoxin reductase sequence lacks a second Gly-X-Gly-X-X-Gly sequence that might indicate an NADP binding site. Hence we sought a putative NADP binding site by two criteria: 1) a sequence similarity with the NADP binding domain of the human glutathione reductase the crystal structure of which is known [16, 19]; and 2) a secondary structure that is similar to a βαβ fold. As noted below only one region was found to fulfil both of these criteria.
The sequences of only two NADP binding enzymes that are homologous to the human glutathione reductase are known [37, 42]. These three sequences contain a consensus sequence which distinguishes the NADP binding site from the FAD binding site by the first Ala residue (Fig. 2). This sequence is also found in adrenodoxin reductase with the substitution of a proline for the last glycine (Pro and Gly are both helix breakers and show tendency to occur in turns) (Fig. 2).
To check the statistical significance of this match (Fig. 2) we searched the NBRF database to determine the frequency of the following consensus sequence in the complete population of sequences. (in one letter amino acid code, wherein x represents any amino acid):
Sequence #1: xxxxxGxGxxAxxxAxxxxxxGxxxxxxx
Structure : ββββββ αααααααααααααα ββββββ
This search yielded only five matches (from nearly 5000 sequences). Two of these were for human glutathione reductase, and mercuric reductase (E. coli glutathione reductase was not included in the database) (Fig. 2). The third matching sequence included only Gly and Ala and was thus eliminated. The fourth match was an NADP specific glutamate dehydrogenase from N. crassa [34]. The homologous enzyme from E. coli contains one conservative change in the consensus sequence (Fig. 2). These two sequences show no homology with other NADP binding enzymes listed in Fig. 2, and there was no previous identification of their NADP binding sites. The fifth match was a bacteriophage protein [47]. Since the function of this protein is not defined the significance of the match is not known.
The five matches noted above were a subset of a total of 42 different sequences found when the search was carried by substituting "X" instead of the last "G" in the consensus sequence #1. All but one of the remainder (37 sequences), were incompatible with the more detailed consensus sequence shown in Fig. 2. The majority of these were eliminated by a simple rule: if a proline or a stretch of glycines appears in a position that corresponds to the β-sheet or α-helix portion of the sequence #1 it would break the secondary structure [48] and thus would not be compatible with a βαβ-fold forming sequence.
The one exception was the NADP specific octopine synthase sequence which showed perfect compatibility with the consensus sequence in Fig. 2 except that the third Gly was moved by one position. This sequence showed no significant similarity with the full sequence of adrenodoxin reductase or any of the other enzymes.
A search for the consensus sequence in recently published NADP binding enzyme sequences which were not yet included in the NBRF database revealed that NADP specific malic enzyme also possesses the consensus sequence with the exception of the position of the last Gly (Fig. 2).
Negative control: NAD and FAD binding site consensus sequences do not recognize any NADP binding enzymes
If the consensus sequence #1 indicates a distinction between the NADP vs. NAD and FAD binding sites then a search of the sequence database using NAD and FAD consensus sequences should not identify any NADP binding enzymes. A search of the NBRF library using FAD and NAD consensus sequence (substituting glycine instead of the first alanine in the consensus sequence #1) indeed identified FAD and NAD binding proteins but it did not identify a single NADP binding protein. Two of the proteins identified in this search are hypothetical proteins derived from open reading frames [49, 50]. Since their sequences matched perfectly the detailed consensus sequence (Fig. 2) these two "hypothetical proteins" probably function as oxidoreductases. Yet, whether they bind a dinucleotide remains to be determined.
Predicted secondary structure of the putative NADP binding site
The conformational propensities [48] of the NADP binding sites of all but two of the enzymes listed in Fig. 2 showed a profile that is perfectly compatible with a βαβ-fold structure. However, the profiles for adrenodoxin reductase and malic enzyme showed higher helix potential at the expected position of the second β-sheet strand. The sequence of adrenodoxin reductase shows no other region that fits a βαβ-fold structure better than this region despite the difference noted. In an ADP binding βαβ-fold the two β-sheet strands form the same β-pleated sheet [19]. It is possible that the second strand was substituted by a different β-sheet strand forming region in the course of evolution.
Analysis of the hydrophobicity of adrenodoxin reductase
Adrenodoxin reductase associates with the inner mitochondrial membrane wherein the P450 system is located. The profile of hydrophobicity using two scales of hydrophobicity [51, 52] did not show any segment that satisfies the criteria for prediction of a membrane spanning region in adrenodoxin reductase sequence.
Discussion
Is the region identified by the consensus sequence indeed the NADP binding site?
The hypothesis that the regions identified by the consensus sequence #1 in adrenodoxin reductase and other NADP binding enzymes are indeed NADP binding sites is supported by the following lines of evidence: 1. Among nearly 5000 different protein sequences known to date, this sequence appears solely in specific NADP binding enzymes. 2. The secondary structure of the region wherein this sequence is located is unequivocally predicted to be an ADP binding βαβ-fold for the newly identified NADP binding sites (see above for reservation for adrenodoxin reductase). 3. If the NADP binding flavoenzymes listed in Fig. 2 shared a common ancestor then the FAD and NADP binding sites of these enzymes would be expected to be at similar distances from one another, and this is indeed the case for adrenodoxin reductase (Fig. 2).
At present we do not know the structural significance of the observed sequence differences between NAD and NADP binding sites (Fig. 2). Some of the conserved differences may play a role in determining NAD vs. NADP specificity of the binding site. With the availability of the cDNAs the structural role of the sequences identified in this paper can be examined by production of altered molecules from cDNAs modified using site directed mutagenesis techniques.
What are the different types of NADP binding sites?
The sequences of several NADP dependent enzymes do not include the NADP consensus sequence (Fig. 2): cytochrome P450 reductase [26-28], spinach ferredoxin reductase [29], quinone (menadione) reductase [53], 6-phospho-gluconate dehydrogenase [54], and glucose-6-phosphate dehydrogenase [55, 56]. Yet, the presence of this consensus sequence in certain NADP binding enzymes which share no significant overall sequence similarity indicates that its conservation may be mandated by the structural requirements of coenzyme and protein association. Hence, we expect that most if not all NADP binding sites that are structurally similar to a βαβ-fold will have the conserved consensus sequence. Following the nomenclature of Wierenga et al. [17], we suggest that these be referred to as Type I NADP binding sites. The enzymes that do not have the NADP consensus sequence may represent additional structurally distinct classes.
Acknowledgements
I. H. is the incumbent of the Delta Research Career Development Chair. This research was supported in part by the U. S. National Institutes of Health grant # AM33830 to I. H.
References
1. Lambeth, J. D., Seybert, D. W., Lancaster, J. R., Salerno, J. C. and Kamin, H. (1982) Mol. Cell. Biochem. 45, 13-31.
2. Hanukoglu, I. and Hanukoglu, Z. (1986) Eur. J. Biochem. 157, 27-31.
3. Jefcoate C. R., McNamara, B. C. and Dibartolomeis, M. J. (1986) Endocrine Res. 12, 315-350.
4. Wikvall, K. (1984) J. Biol. Chem. 259, 3800-3804.
5. Waki, N., Hiwatashi, A. and Ichikawa, Y. (1986) FEBS Lett. 195, 87-91.
6. Hiwatashi, A., Nishii, Y. and Ichikawa, Y. (1982) Biochem. Biophys. Res. Comm. 105, 320-327.
7. Churchill, P. F., deAlvare, L. R. and Kimura, T. (1978) J. Biol. Chem. 253, 4924-4929.
8. Farkash, Y., Timberg, R. and Orly, J. (1986) Endocrinology 118, 1353-1365.
9. Omura, T., Sanders, E., Estabrook, R. W., Cooper, D. Y. and Rosenthal, O. (1966) Arch. Biochem. Biophys. 117, 660-673.
10. Hanukoglu, I. and Jefcoate, C. R. (1980) J. Biol. Chem. 255, 3057-3061.
11. Hanukoglu, I., Gutfinger, T., Haniu, M. and Shively, J. E. (1987) Eur. J. Biochem. 169, 449-455.
12. Sagara, Y., Takata, Y., Miyata, T., Hara, T., and Horiuchi, T. (1987) J. Biochem. 102, 1333-1336.
13. Nonaka, Y., Murakami, H., Yabusaki, Y., Kuramitsu, S., Kagamiyama, H., Yamano, T. and Okamoto, M. (1987) Biochem. Biophys. Res. Comm. 145, 1239-1247.
14. Solish, S. B., Picado-Leonard, J., Morel, Y., Kuhn, R. W., Mohandas, T. K., Hanukoglu, I., and Miller, W. L. (1988). Proc. Natl. Acad. Sci. USA. 85, 7104-7108.
15. Rossman, M. G., Liljas, A., Branden, C.-I. and Banaszak, L. J. (1975) Enzymes 9, 61-102.
16. Rice, D. W., Schulz, G. E. and Guest, J. R. (1984) J. Mol. Biol. 174, 483-496.
17. Wierenga, R. K., DeMaeyer, M. C. H. and Hol, W. G. J. (1985) Biochemistry 24, 1346-1357.
18. Wierenga, R. K., Terpstra, P. and Hol, W. G. J. (1986) J. Mol. Biol. 187, 101-107.
19. Karplus, P. A. and Schulz, G. E. (1987) J. Mol. Biol. 195, 701-729.
20. Matthews, D. A., Alden, R. A., Bolin, J. T., Freer, S. T., Hamlin, R., Xuong, N., Kraut, J., Poe, M., Williams, M. and Hoogsteen, K. (1977) Science 197, 452-455.
21. Maxam, A. M. and Gilbert, W. (1980) Methods Enzymol. 65, 499-560.
22. Chuvpilo, S. A. and Kravchenko, V. V. (1984) FEBS Lett. 179, 34-36.
23. Hanukoglu, I. and Fuchs, E. (1982) Cell 31, 243-252.
24. Queen, C., and Korn, L. J. (1984) Nucl. Acids Res. 12, 581-599.
25. Hanukoglu, I. and Fuchs, E. (1983) Cell 33, 915-924.
26. Porter, T. D. and Kasper, C. B. (1986) Biochemistry 25, 1682-1687.
27. Haniu, M., Iyanagi, T., Miller, P., Lee, T. D. and Shively, J. E. (1986) Biochemistry 25, 7906-7911.
28. Vogel, F. and Lumper, L. (1986) Biochem. J. 236, 871-878.
29. Karplus, P. A., Walsh, K. E. and Herriott, J. R. (1984) Biochemistry 23, 6576-6583.
30. Schulz, G. E., Schirmer, R. H., and Pai, E. F. (1982) J. Mol. Biol. 160, 287-308.
31. Ronchi, S., Minchiotti, L., Curti, B., Zapponi, M. C. and Bridgen, J. (1976) Biochim. Biophys. Acta 427, 634-643.
32. Cole, S. T. (1982) Eur. J. Biochem. 122, 479-484.
33. McPherson, M. J. and Wootton, J. C. (1983) Nucl. Acids Res. 11, 5257-5266.
34. Kinnaird, J. H. and Fincham J. R. S. (1983) Gene 26, 253-260.
35. Moye, W. S., Amuro, N., Rao, J. K. M. and Zalkin, H. (1985) J. Biol. Chem. 260, 8502-8508.
36. Nagasu, T. and Hall, B.D. (1985) Gene 37, 247-253.
37. Greer, S. and Perham, R. N. (1986) Biochemistry 25, 2736-2742.
38. Krauth-Siegel, R. L., Blatterspiel, R., Saleh, M., Schiltz, E., Schirmer, R. H. and Untucht-Grau, R. (1982) Eur. J. Biochem. 121, 259-267.
39. Stephens, P. E., Lewis, H. M., Darlison, M. G. and Guest, J. R. (1983) Eur. J. Biochem. 135, 519-527.
40. Otulakowski, G. and Robinson, B. H. (1987) J. Biol. Chem. 262, 17313-17318.
41. Magnuson, M. A., Morioka, H., Tecce, M. F. and Nikodem, V. M. (1986) J. Biol. Chem. 261, 1183-1186.
42. Brown, N. L., Ford, S. J., Pridmore, R. D. and Fritzinger, D. C. (1983) Biochemistry 22, 4089-4095.
43. Young, I. G., Rogers, B. L., Campbell, H. D., Jaworowski, A. and Shaw, D. C. (1981) Eur. J. Biochem. 116, 165-170.
44. De Greve, H., Dhaese, P., Seurinck, J., Lemmers, M., Van Montagu, M. and Schell, J. (1982) J. Mol. Appl. Genet. 1, 499-511.
45. Weijer, W.J., Hofsteenge, J., Vereijken, J.M., Jekel, P.A. and Beintema, J. J. (1982) Biochim. Biophys. Acta 704, 385-388.
46. Romeo, C., Moriwaki, N., Yasunobu, K. T., Gunsalus, I. C. and Koga, H. (1987) J. Protein Chem. 6, 253-261.
47. Garcia, J.A., Mendez, E. and Salas, M. (1984) Gene 30, 87-98.
48. Chou, P. Y. and Fasman, G. D. (1978) Adv. Enzymol. 47, 45-148.
49. Buhk, H.-J. and Messer, W. (1983) Gene 24, 265-279.
50. Gielen, J., De Beuckeleer, M., Seurinck, J., Deboeck, F., De Greve, H., Lemmers, M., Van Montagu, M. and Schell, J. (1984) EMBO J. 3, 835-846.
51. Eisenberg, D., Schwarz, E., Komaromy, M. and Wall, R. (1984) J. Mol. Biol. 179, 125-142.
52. Engelman, D. M., Steitz, T. A. and Goldman, A. (1986) Ann. Rev. Biophys. Biophys. Chem. 15, 321-353.
53. Robertson, J. A., Chen, H.-C. and Nebert, D. W. (1986) J. Biol. Chem. 261, 15794-15799.
54. Carne, A. and Walker, J. E. (1986) J. Biol. Chem. 258, 12895-12906.
55. Takizawa, T., Huang, I-Y., Ikuta, T. and Yoshida, A. (1983) Proc. Natl. Acad. Sci. USA 83, 4157-4161.
56. Persico, M. G., Viglietto, G., Martini, G., Toniolo, D., Paonessa, G., Moscateeli, C., Dono, R., Vulliamy, T., Luzzato, L. and D'Urso, M. (1986) Nucl. Acids Res. 14, 2511-2522.
Figure 1. The sequence of the pAR cDNA insert and the deduced amino acid sequence of adrenodoxin reductase. The numbering of the DNA starts with the first base of the codon for the first NH2 terminal residue of the mature adrenodoxin reductase (starting with Ser-Thr-Gln... ) [11, 13], based on the cDNA sequence [12]. The last six bases of the sequence and the poly(A) sequence are from clone pAR1. The sequence of pAR1 starts at about 1000 and shows three silent bp differences (1044:C, 1387:C, 1692:A), and one (1114:T) which changes residue 372 from Pro to Ser. The sequence of clone 16 of Sagara et al. [12] shows only two silent bp changes (1445:T, 1507:G) from the sequence of pAR, and the sequence of clone 12 shows two additional silent bp changes (1044:C, 1047:T). The sequence shown here does not include the first 150 bp of the pAR insert as this is an addition of unknown origin. The polyadenylation sequence AATAAA appears 21 bp before the poly(A).
Figure 2. Alignment of the putative FAD and NADP binding sites of adrenodoxin reductase with those of other oxidoreductases. The NADP binding sites for the enzymes marked with an asterisk are identified here for the first time. The secondary structure indicated is based on the known crystal structures of FAD and NADP binding sites of glutathione reductase [16, 19] (α, α-helix; β, β-sheet; T, turn). The sequence shown includes 10 residues on both sides of this βαβ fold. The number of the first residue shown is indicated before the sequence (1=the first residue of the amino terminus of the mature protein). In the consensus sequence () indicates a hydrophobic residue, (+, -) a charged residue, and () a hydrophilic residue. The sequences listed are: adrenodoxin reductase [Bovine: this paper and 12, human: 14]; D-Amino acid oxidase [31], fumarate reductase [32], glutamate dehydrogenase [Escherichia coli: 33, Neurospora crassa: 34, yeast: 35, 36], glutathione reductase [E. coli: 37, human: 38], lipoamide dehydrogenase [E. coli: 39, human: 40, porcine: 16, 40], malic enzyme [41], mercuric reductase [42], NADH dehydrogenase [43], octopine synthase [44], p-hydroxy-benzoate hydroxylase [45] and putidaredoxin reductase [46]. Abbreviations: DH, dehydrogenase; H, hydroxylase; R, reductase.
