An enzyme controlled by light: the molecular mechanism of photoreactivity in nitrile hydratase

REVIEWS 75 Larese, F., De Zotti, R., Molinari, S. and Bovenzi, M. (1994) Occup. Environ. Med. 51, 548–552 76 Randez-Gil, F., Prieto, J. A., Murcia, A. and Sanz, P. (1995) J. Cereal Sci. 21, 185–193 77 Randez-Gil, F. and Sanz, P. (1993) FEMS Microbiol. Lett. 112, 119–124 78 Monfort, A., Blasco, A., Prieto, J. A. and Sanz, P. (1997) J. Cereal Sci. 26, 195–199 79 Monfort, A., Blasco, A., Prieto, J. A. and Sanz, P. (1996) Appl. Environ. Microbiol. 62, 3712–3715 80 Si, J. Q. (1997) Cereal Foods World 42, 802–807 81 De Winde, J. H., Thevelein, J. M. and Winderickx, J. (1997) in Yeast Stress Responses (Hohmann, S. and Mager, W. H., eds), pp. 7–52, Springer-Verlag 82 DeRisi, J. L., Iyer, V. R. and Brown, P. O. (1997) Science 278,

680–686 83 Spellman, P. T. et al. (1998) Mol. Biol. Cell 9, 3273–3297 84 Chu, S. et al. (1998) Science 282, 699–705 85 Schena, M., Heller, R. A., Theriault, T. P., Konrad, K., Lachenmeir, E. and Davis, R. W. (1998) Trends Biotechnol. 16, 301–306 86 Holstege, F. C. P. et al. (1998) Cell 95, 717–728 87 Crauwels, M., Winderickx, J., De Winde, J. H. and Thevelein, J. M. (1997) Yeast 13, 973–984 88 Van Rooijen, R. and Klaassen, P. (1998) in Genetic Modification in the Food Industry: A Strategy for Food Quality Improvement (1st edn) (Roller, S. and Harlander, S., eds), pp. 158–173, Blackie Academic & Professional 89 Roller, S. and Harlander, S. (1998) Genetic Modification in the Food Industry: A Strategy for Food Quality Improvement (1st edn), Blackie Academic & Professional

An enzyme controlled by light: the molecular mechanism of photoreactivity in nitrile hydratase Isao Endo, Masafumi Odaka and Masafumi Yohda Extensive studies have revealed the molecular mechanism of the photoreactivity of nitrile hydratase from Rhodococcus sp. N-771. In the inactive enzyme, nitric oxide is bound to the non-heme ferric iron at the catalytic center, stabilized by a claw-like structure formed by two post-translationally modified cysteines and a serine. The inactive nitrile hydratase is activated by the photoinduced release of the nitric oxide. This result might provide a means of designing novel photoreactive chemical compounds or proteins that would be applicable to biochips and light-controlled metabolic systems.

itrile hydratase (NHase) is a soluble metalloenzyme containing a non-heme iron1 or a noncorrinoid cobalt2 atom at the catalytic site. It consists of two subunits (a and b) in addition to the metal ion, M, with the basic stoichiometry a1b1M1 (Ref. 3). NHase is important in applied biotechnology, as it is used for the industrial production of acrylamide (more than 30 000 ton yr21)4–6. Both subunits have a molecular weight of approximately 23 kDa and their primary sequences are well conserved, although there is no apparent homology between the two subunits5. One of the best-characterized NHases comes from Rhodococcus sp. N-771 and belongs to the ironcontaining NHase family7. In vivo, the NHase activity of Rhodococcus sp. N-771 cells decreases during aerobic incubation in the dark and almost disappears after about 16 h in darkness (dark inactivation); the activity is almost completely recovered by light irradiation (Fig. 1)7,8. This photoreactivity of NHase is intrinsic to the enzyme: if the inactive form is purified from darkinactivated cells, it can be reactivated by light7, although

N

I. Endo ([email protected]), M. Odaka and M. Yohda are at the Biochemical Systems Laboratory, The Institute of Physical and Chemical Research (RIKEN), Hirosawa, Wako-shi, Saitama 351-0198, Japan.

244

the activated NHase cannot be inactivated by darkness in vitro. NHases from Rhodococcus sp. N-7749 and R31210 have also been reported to be activated by light irradiation. A comparison of the nucleotide sequences shows that these NHases are identical to that from Rhodococcus sp. N-77111–13. Photoactivation induces UV–visible absorptional changes in NHase14 (Fig. 2). The inactive NHase shows two absorption peaks, at 280 and 370 nm; upon light irradiation, the intensity of absorbance at 280 nm decreases by 12–15% and the peak at 370 nm almost disappears. Instead, a shoulder at around 400 nm and a small, broad peak at 710 nm appear. Flash-photolysis studies show that these spectral changes occur within 50 nsec15. Spectral changes of electron-spin resonance (ESR)16 and Mössbauer17 were also accompanied by photoactivation. These results led to the view that the non-heme catalytic center functions as the photoreactive site. The role of nitric oxide Fourier-transform infrared (FT-IR) difference spectra of NHase before and after photoactivation show large negative bands at around 1850 cm21, where biochemical compounds do not usually produce a signal18. The suggestion that nitric oxide (NO) was responsible

0167-7799/99/$ – see front matter © 1999 Elsevier Science. All rights reserved. PII: S0167-7799(99)01303-7

TIBTECH JUNE 1999 (VOL 17)

1853

50

After irradiation

Before irradiation 0 0

10

20 30 Time (h)

40

50

Figure 1 The dark inactivation and photoactivation of nitrile hydratase (NHase) in Rhodococcus sp. N-771. Cells were cultured as described in Ref. 14. The washed cell suspension was irradiated with white light for 30 min just before incubation in the dark. Measurements were taken either before or after a 30-min exposure to white light on ice. A unit of NHase activity is defined as the amount of enzyme that produces 1 (mmole propioamide) min21.

Before irradiation After irradiation

2.0

1.8

1.5

1.2 0.6

1.0 0.5

0.0 300

400 500 600 Wavelength (nm)

700

1865

1880 1860 1840 1820 Wavenumber (cm21)

0.002

1800

Active 0.002

1900

1880 1860 1840 1820 Wavenumber (cm21)

Active + NO

1900

trends in Biotechnology

Absorbance

2.5

(31021) 2.4

1900

Absorbance

10

Absorbance

20

Absorbance

30

1844 Inactive

1880 1860 1840 1820 Wavenumber (cm21)

Active + NO light

1900

1880 1860 1840 1820 Wavenumber (cm21)

1800

0.002

1800

0.002

1800

trends in Biotechnology

Absorbance

40

trends in Biotechnology

NHase activity [units (mg dry cell–1)]

REVIEWS

Figure 2 UV–visible spectra of the photoreactive nitrile hydratase (NHase). The abosorption spectra were measured before and after light irradiation for 15 min on ice. The grey lines are plotted on the lefthand axis, the black lines on the right-hand axis.

Figure 3 The effects of exogenous nitric oxide and photoirradiation on Fouriertransform infrared spectra of nitrile hydratase (NHase). Modified with permission from Ref. 19. Copyright 1997 American Chemical Society.

for this result was confirmed by its isotope shift using labeling18; NHase purified from the darkinactivated cells has NO molecules bound to the nonheme iron center, which are released by light excitation. To prove this hypothesis, the effect of NO on the NHase activity was examined19. Introducing NO gas into a solution of active NHase results in the complete disappearance of activity. However, approximately 86% of the activity of the native active enzyme is restored following light irradiation, indicating that NHase inactivated by NO exposure can be photoactivated in the same way as the native enzyme. This suggests that the NHase inactivated by exogenous NO is identical to the inactive native form of the enzyme. The small decrease in the yield of photoactivation is probably due to partial denaturation of NHase by NO. The FT-IR spectra of the NHase samples in the typical NO stretching region are shown in Fig. 320. The in-

active NHase displays a band at 1853 cm21 with shoulders at 1865 cm21 and 1844 cm21. Upon photoactivation, the 1853 cm21 band and the 1844 cm21 shoulder disappear, while the small band at 1865 cm21 remains. When the active NHase is again inactivated by NO, the signals at 1853 cm21 and 1844 cm21 reappear with almost fully recovered intensity. As with the native enzyme, these NO peaks disappear upon subsequent light irradiation. The intensity of the 1865 cm21 band also increases after introducing NO but remains after light irradiation; this band may be due to NO molecules nonspecifically bound to or trapped within the enzyme. ESR and UV–visible spectral changes are also induced by NO binding and photoirradiation19. Furthermore, the interaction between NO and the non-heme iron has been detected directly by resonance Raman spectroscopy21. These results indicate that the NO molecule bound to the non-heme iron(III) center is released upon

15N

TIBTECH JUNE 1999 (VOL 17)

245

REVIEWS

The photoreactive iron center The Fe-type NHases have been shown to be unique enzymes, with a mononuclear low-spin ferric iron1. Combined with this observation, extended studies of the photoactivated enzyme have suggested that the iron-site has a distorted six-coordinate structure with two cysteine thiolate and three histidine immidazoles,

and a solvent ligand10,29,30. From comparison with the amino acid sequence of other iron–sulfur proteins, a cysteine cluster in the a subunit [(109)CSLCSC(114)] was suggested to be an iron-binding site31. However, the location of the iron center in NHase has not been investigated directly. The UV–visible absorption spectrum of inactive NHase is not affected by unfolding in 6 M urea and the photoresponsive peaks at 280 and 370 nm diminish following light irradiation14. This implies that the nonheme iron center is stable and able to respond to light even in an unfolded state, at least in the nitrosylated state. Studies on isolated a and b subunits of the inactive NHase demonstrated that, before light irradiation, the absorption spectrum of the isolated a subunit was similar to that of the inactive NHase denatured with 6 M urea. Upon light irradiation, the peak responsible for the nitrosyl iron center disappears, but the absorbance peaks characteristic of the active enzyme do not appear. Conversely, the isolated b subunit only shows a peak at 280 nm and light irradiation does not induce any spectral change. This would indicate that the non-heme iron center is solely located on the a subunit in the inactive state. When equimolar amounts of the a and b subunits are incubated together, approximately 30% of the activity of native NHase is recovered after light irradiation14. This suggests that significant amounts of the recovered a and b subunits are correctly refolded and able to assemble together to reconstitute NHase, and that the addition of cofactors is not required for reconstitution. This suggests that the non-heme iron center is very tightly bound to the subunits. To identify the iron-binding site, the a subunit isolated from the inactive NHase was cleaved with trypsin and the digest was then analysed by reversed-phase HPLC32. The chromatogram at 370 nm showed only one peak, at 20.6 min, and protein sequence analysis revealed that this peak contained a peptide fragment of 24 amino acids (NK24). This region is highly conserved in all known NHases and contains the Cys cluster. The proportion of Fe atoms in NK24 was found by inductively coupled plasma mass spectrometry to be

RCN hn Fe3+

NO

RCN Fe3+

NO

a

RCONH2

H2O

H2O

a

b

b

NO

Inactive

Active

trends in Biotechnology

photoactivation, and that the activated NHase can be reconverted to the original inactive form by NO binding. The amount of NO release from the NHase upon photoactivation has been estimated using a NO scavenger to be about one NO molecule per enzyme molecule19,22. The mechanism of photoactivation and dark inactivation of NHase that has been deduced from these experiments is illustrated in Fig. 4. In vivo and in the dark, NHase is inactivated by the direct binding of NO to the ferric iron in the non-heme iron center. This inactivated NHase is reactivated by light-induced breakage of the Fe–NO bond and subsequent NO release. The quantum yield of the photoactivation has been estimated to be 0.48 by laser flash-photolysis studies19. Recently, Bonnet et al. demonstrated that the Fe-type NHase from Comamonas testosteroni NI1, which does not show any photoreactivity in vivo, is also inactivated by exogenous NO and reactivated by photoirradiation23. This suggests that binding of NO and its photodissociation may be common in Fe-type NHases. Probably, the photoreactivity in vivo depends on NO production by the bacterium. It is known that NO plays several important roles in biological systems; in higher animals such as mammals, these include blood-pressure control, neurotransmission and the immune response24,25, and occur by NO binding to various heme and non-heme iron proteins with high affinity and regulating their biological function26–28. For instance, the activity of guanylyl cyclase is regulated by NO binding to its heme iron26,27. Thus, we believe that finding NHase activity to be regulated by NO in combination with light in bacteria provides a new aspect of biological function for NO.

Figure 4 Putative NO-regulation scheme for photoreactive nitrile hydratase (NHase) from Rhodococcus sp. N-771. Modified with permission from Ref. 19. Copyright 1997 American Chemical Society.

246

TIBTECH JUNE 1999 (VOL 17)

REVIEWS

0.79 molecule21 (Ref. 32). To obtain the minimum peptide segment required for the nitrosylated iron center, NK24 was further treated by thermolysin, carboxypeptidase Y and leucine-amino-peptidase M32. The isolated peptide was an 11 amino acid segment from Ile107 to Trp117, and this fragment exhibited an absorbance peak at 370 nm, which diminished following light irradiation. Interestingly, mass-spectrometric analysis of NK24 prepared from photoactivated NHase showed that the molecular mass of the peptide is 30.3–31.6 kDa, which is larger than the calculated value32. Carboxymethylation with reduction and protein sequence analysis of the peptide revealed that Cys112 is post-translationally modified by 32 Da. By amino acid analysis of the intact NHase from Rhodococcus sp. N-771 after derivatization with a fluorescent-labeling reagent, the 112th residue of the a subunit (Cys) was found to be posttranslationally oxidized to a sulfinic-acid derivative. Thus, the photoreactive nitrosylated non-heme iron center appears to be composed of the short peptide segment containing the Cys cluster with a Cys–SO2H derivative32. Surprisingly, further research revealed that another cysteine residue in the short peptide segment is also post-translationally modified (see below). X-ray crystallography reveals a unique structure The crystal structure of photoactivated NHase from Rhodococcus sp. R312 has recently been described at a resolution of 2.65 Å33. The a subunit consists of a long Nterminal arm and a C-terminal domain that forms a novel fold. This fold has a four-layered structure, abba, with unusual connections between the b strands. The b subunit also contains a long N-terminal extension, a helical domain and a C-terminal domain that folds into a b roll. The two subunits form a tight heterodimer that appears to be the functional unit of the enzyme, with the iron center located in a cavity at the subunit– subunit interface. This iron center is composed of four residues from the a subunit (aCys109, aCys112, aSer113, aCys114). The ligands to the non-heme iron atom are sulfur atoms from the three cysteine residues and two main-chain amide-nitrogen atoms (aSer113, aCys114). These five ligands are positioned at five vertices of an octahedron, the sixth site of which is unoccupied in the crystal structure. Thus, the iron site is highly unusual and only the second reported example of peptide–amide coordination in a metalloprotein34. The structure resembles a hybrid of the iron centers of heme and Fe–S proteins, but the crystal structure determined by Huang et al.33 did not show any posttranslational modifications of Cys112, despite the fact that NHase from Rhodococcus sp. R312 is identical to that from Rhodococcus sp. N-771. The crystal structure of the nitrosylated photoreactive NHase has since been resolved at a resolution of 1.7 Å35. The folding pattern was almost identical to that determined by Huang et al.33, showing that the conformation is conserved between the inactive and photoactivated enzymes. The atoms coordinating to the iron in the nitrosylated state were also identical to those in the photoactivated state, except for the nitrogen of NO. The NO occupied the sixth coordination site, which is accessible from the solvent and is thought TIBTECH JUNE 1999 (VOL 17)

to be the catalytic site. This result supports the previous findings that direct binding of NO to the nonheme iron center causes the inactivation of NHase19,21. The structural similarities of the active site and overall structure between the inactive and active forms of NHase strongly suggest that the inactivation is caused by occupation of the catalytic site by a NO molecule, not by exchange of the ligands provided by the a subunit. The electron-density map showed three areas of extra electron density around the Sg atoms of aCys112 and aCys114. Two of these correspond to the two oxygen atoms of the sulfinyl group; that is, the posttranslational modification of aCys112 was confirmed by the X-ray crystal structure. The other area of extra electron density was close to the Sg atom of aCys114, indicating that aCys114 may also be post-translationally modified. To confirm this, we examined the tryptic digest of the inactive enzyme by Fourier-transform iron-cyclotron-resonance mass spectrometry at neutral pH35. The result showed that the molecular mass of the catalytic iron-center–peptide complex (NK24) was 16.02 Da larger than the expected value, which is equivalent to one oxygen atom. It was proposed that aCys114 is post-translationally modified to a cysteinesulfenic acid (Cys–SOH). In acidic conditions, the bound NO and iron are both released, resulting in disulfide-bond formation between aCys–SOH114 and aCys10935, which is why this modification was not detected in the previous analysis by mass spectrometry at acidic pH32. Three oxygen atoms, Od1 of aCys-SO2H112, Od of aCys-SOH114 and Og of aSer113, protrude like claws from the plane containing the iron atom (and thus have been termed the ‘claw setting’) and hold the NO molecule in their center (Fig. 5). The three oxygen atoms are located close enough to the NO molecule to have a strong interaction with it. The association constants of NO for ferric irons are generally much lower than for the ferrous ions found in heme proteins36, but the nitrosyl ferric-iron complex of the NHase is stable under aerobic condition for over a year as long as it is kept in the dark. Moreover, the nitrosylated iron center is stable even when present as a proteolytic fragment from aIle107 to aTrp11732. The extraordinary stability of the nitrosyl ferric iron center in Fe-type NHase could be attributed to the interaction between NO and the oxygen atoms; the FT-IR difference spectrum shows that local structural changes occur around the iron center upon light irradiation18, suggesting that light irradiation not only breaks the Fe–N bond but also induces a structural change in the claw setting that weakens the interaction between the oxygen atoms and NO. Other similar proteins Cys–SO2H and Cys–SOH have been found in several other proteins. One example of a protein containing Cys–SO2H is SP-22, a substrate protein for mitochondrial ATP-dependent protease found in bovine adrenal cortex37. The 47th amino acid of SP-22 was found to be a sulfinic acid by fast-atom-bombardment mass spectrometry. The fact that the vicinal sequence is highly conserved between several homologous proteins suggests that this residue is important to the biological function of SP-22.

247

REVIEWS

Acknowledgments We would like to thank our collaborators: M. Nakasako, N. Kamiya, K. Takio, T. Noguchi, M. Hoshino, T. Nagamune, N. Dohmae, M. Chijimatsu, Y. Kobayashi, H. Kandori, J. Honda, M. Tsujimura, S. Nagashima and many students.

NO aCys1142SOH

aSer113

aCys1122SO2H trends in Biotechnology

FeIII

Figure 5 The ‘claw setting’ of nitrosylated nitrile hydratase. Key: carbon, grey; nitrogen, blue; oxygen, red; sulfur, green; iron, yellow. Reproduced with permission from Ref. 35.

It has been shown that two flavoproteins, NADH oxidase and NADH peroxidase from Enterococcus faecalis 10C1, possess Cys–SOH residues at non-flavin redox centers in the native enzymes38,39. These modified Cys residues play essential roles in the catalysis, being switched between Cys–SOH and Cys–S2 during turnover38. Furthermore, the Cys–SOH/Cys–SH redox cycle may be important for the DNA-binding activity of transcription factors including E2 protein40, OxyR41,42, Jun and Fos43, and nuclear-factor I44. However, none of these proteins are metalloproteins, and no other proteins contain both post-translational modifications together. Therefore, the post-translational modifications in the Fe-type NHase is unique in both structure and function. The mechanism by which aCys112 and aCys114 are post-translationally modified is still unknown. NO molecules might oxidize these cysteine residues in Rhodococcus sp. N-771; it has been reported that NO and ONOO2 (which is produced from O22 and NO) do oxidize protein sulfhydryls45,46. Alternatively, these thiols might be oxidized by a self-catalytic mechanism, as in the case of copper amine oxidase, in which a specific Tyr residue is oxidized to a topaquinone in the presence of Cu(II) and O247. Conclusions The structure of NHase from Rhodococcus sp. N-771 has now been fully elucidated, and this revealed that the unprecedented ‘claw setting’ gives the Fe-type NHases their photoreactivity. The cysteine modifications and the ‘claw setting’ structure that have been revealed are, however, probably not limited to Fe-type NHases. The development of an NO sensor and a photosensor based on the ‘claw setting’ of the NHase are under way. In general, biological reactions are controlled by temperature, pH and substrate concentration. If we can control biological reactions with light as well, we will be able to control the cell’s metabolic activity more specifically. Then, the photoregulation of metabolism will be an excellent tool in fermentation technology, and the Fe-type NHase would be a good model for designing novel photocontrollable enzymes.

248

References 1 Sugiura, Y., Kuwahara, J., Nagasawa, T. and Yamada, H. (1987) J. Am. Chem. Soc. 109, 5848–5850 2 Brenann, B. A., Alms, G., Nelson, M. J., Durney, L. T. and Scarrow, R. C. (1996) J. Am. Chem. Soc. 118, 9194–9195 3 Odaka, M. et al. (1996) Biochem. Biophys. Res. Commun. 221, 146–150 4 Kobayashi, M. and Shimizu, S. (1998) Nat. Biotechnol. 16, 733–736 5 Kobayashi, M., Nagasawa, T. and Yamada, H. (1992) Trends Biotechnol. 10, 402–408 6 Yamada, H. and Kobayashi, M. (1996) Biosci. Biotechnol. Biochem. 60, 1391–1400 7 Ngamune, T. et al. (1990) Biochem. Biophys. Res. Commun. 168, 437–442 8 Nagamune, T. et al. (1990) Photochem. Photobiol. 51, 87–90 9 Nakajima, T., Takeuchi, K. and Yamada, H. (1987) Chem. Lett. 9, 1767–1770 10 Brennan, B. A., Cummings, J. G., Chase, D. B., Turner, I. M., Jr and Nelson, M. J. (1996) Biochemistry 35, 10068–10077 11 Ikehata, O., Nishiyama, M., Horinouchi, S. and Beppu, T. (1989) Eur. J. Biochem. 181, 563–570 12 Mayaux, J. F., Cerebelaud, E., Soubrier, F., Faucher, D. and Petre, D. (1990) J. Bacteriol. 172, 6764–6773 13 Yohda, M., Honda, J., Nagamune, T., Endo, I., Yoshida, T. and Miura, K. (1994) Ann. New York Acad. Sci. 721, 158–159 14 Tsujimura, M., Odaka, M., Nagashima, S., Yohda, M. and Endo, I. (1996) J. Biochem. (Tokyo) 119, 407–413 15 Honda, J. et al. (1994) Biochemistry 33, 3577–3583 16 Honda, J. et al. (1992) Ann. New York Acad. Sci. 672, 29–36 17 Honda, J. et al. (1992) FEBS Lett. 301, 177–180 18 Noguchi, T., Honda, J., Nagamune, T., Sasabe, H., Inoue, Y. and Endo, I. (1995) FEBS Lett. 358, 9–12 19 Odaka, M. et al. (1997) J. Am. Chem. Soc. 119, 3785–3791 20 Nakamoto, K. (1986) in Infrared and Raman Spectra of Inorganic and Coordination Compounds, pp. 309–310, Wiley 21 Noguchi, T., Hoshino, M., Tsujimura, M., Odaka, M., Inoue, Y. and Endo, I. (1996) Biochemistry 35, 16777–16781 22 Akaike, T. and Maeda, H. (1996) Methods Enzymol. 268, 211–221 23 Bonnet, D., Artaud, I., Moali, C., Pétré, D. and Mansuy, D. (1997) FEBS Lett. 409, 216–220 24 Nathan, C. (1992) FASEB J. 6, 3051–3064 25 Schmidt, H. H. H. W. and Walter, U. (1994) Cell 78, 919–925 26 Stamler, J. S. (1994) Cell 78, 931–936 27 Traylor, T. G. and Sharma, V. S. (1992) Biochemistry 31, 2847–2849 28 Tsai, A. (1994) FEBS Lett. 341, 141–145 29 Doan, P. E., Nelson, M. J., Jin, H. and Hoffman, B. M. (1996) J. Am. Chem. Soc. 118, 7014–7015 30 Scarrow, R. C. et al. (1996) Biochemistry 35, 10078–10088 31 Nelson, M. J. et al. (1991) J. Am. Chem. Soc. 113, 7072–7073 32 Tsujimura, M. et al. (1997) J. Biol. Chem. 272, 29454–29459 33 Huang, W., Jia, J., Cummings, J., Nelson, M., Schneider, G. and Lindqvist, Y. (1997) Structure 5, 691–699 34 Peters, J. W., Stowell, M. H. B., Soltis, S. M., Finnegan, M. G., Johnson, M. K. and Rees, D. C. (1997) Biochemistry 36, 1181–1187 35 Nagashima, S. et al. (1998) Nat. Struct. Biol. 5, 347–351 36 Hoshino, M., Ozawa, K., Seki, H. and Ford, P. C. (1993) J. Am. Chem. Soc. 115, 9568–9575 37 Watabe, S., Kohno, H., Kouyama, H., Hiroi, T., Yago, N. and Nakazawa, T. (1994) J. Biochem. (Tokyo) 115, 648–654 38 Claiborne, A., Miller, H., Parsonage, D. and Ross, R. P. (1993) FASEB J. 7, 1483–1490 39 Yeh, J. I., Claiborne, A. and Hol, W. G. J. (1996) Biochemistry 35, 9951–9957 40 McBride, A. A., Klausner, R. D. and Howley, P. M. (1992) Proc.

TIBTECH JUNE 1999 (VOL 17)