Communication Vol. No. Issue of January 15, OF 1991 by The ... · Communication Vol. 266, No. 2, Issue of January 15, pp. 689-692,1991 THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by

Communication Vol. 266, No. 2, Issue of January 15, pp. 689-692,1991 THE JOURNAL OF BIOLOGICAL CHEMISTRY

0 1991 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U. S. A.

Specific Detection of Quinoproteins by Redox-cycling Staining*

(Received for publication, July 25, 1990) Mercedes A. PazS8, Rudolf FliickigerS, Andra Boakll, Herbert M. Kaganll, and Paul M. Gallop8 From the $Laboratory of Human Biochemistry, Children’s Hospital and the Harvard Schools of Medicine and Dental Medicine, Boston, Massachusetts 02146, and the llDepartment of Biochemistry, Boston University School of Medicine, Boston, Massachusetts 02115

Quinones and related quinonoid substances catalyze redox cycling at an alkaline pH in the presence of excess glycine as reductant. With nitroblue tetrazolium and oxygen present there is concomitant reduction of the tetrazolium to formazan. This property of quinon- oid compounds is used for the specific staining of quin- oproteins, separated by sodium dodecyl sulfate-poly- acrylamide gel electrophoresis and electroblotted onto nitrocellulose. The dopa-containing vitelline proteins and the 6-hydroxydopa-containing bovine serum amine oxidase are stained with the nitroblue tetrazo- lium/glycinate reagent. Also, the mammalian quino- proteins, diamine oxidase and lysyl oxidase, purported to contain pyrroloquinoline quinone, tested positive in this procedure. No quinonoid components were de- tected in three putative pyrroloquinoline quinone-con- taining quinoproteins, dopamine &hydroxylase, lipox- ygenase, and peptidylglycine-amidating monoxygen- ase. Redox-cycling staining therefore confirms the presence of covalently bound quinones in the copper- dependent amine oxidases, but not in two putative quinoprotein oxygenases. Clarification of the biolog- ical significance of quinolation should be facilitated by identification of quinoproteins using this approach.

Covalently bound quinonoid cofactors are purported to occur in various bacterial and mammalian enzymes, referred to as quinoproteins (1). While the structure of the dissociable quinone cofactor is pyrroloquinoline quinone (PQQ)’ in some copper-dependent bacterial dehydrogenases, the structure of covalently bound quinonoid components is uncertain in both

AGO4727 (to P. M. G.), AGO7723 (to P. M. G.), AR 18880 (to H. M. * This work was supported by National Institutes of Health Grants

K.), and HL 19717 (to H. M. K.) and by a Milton Fund award (to R. F.). Preliminary results for the study were obtained in part with quinoproteins generously provided by D. M. Dooley, Amherst College, Amherst, MA. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed Children’s Hos- pital, Enders 12, 320 Longwood Ave., Boston MA 02115. Tel.: 617- 738-6838; Fax: 617-730-0226.

’ The abbreviations used are: PQQ, pyrroloquinoline quinone; NBT, nitroblue tetrazolium; dopa, 3,4-dihydroxyphenylalanine; topa, 3,4,6-trihydroxyphenylalanine; SDS-PAGE, sodium dodecyl sulfate- polyacrylamide gel electrophoresis.

bacterial and mammalian quinoproteins. However, for bovine serum amine oxidase, the quinonoid component has recently been established as a 6-hydroxydopa or 6-hydroxydopaqui- none residue (2).

Earlier evidence suggested that PQQ is the orthoquinone cofactor in the mammalian quinoproteins, lysyl oxidase from bovine aorta (3), and human placenta (4), pig kidney diamine oxidase ( 5 ) , bovine plasma amine oxidase (6, 7), bovine adre- nal dopamine @-hydroxylase (8), dog liver choline dehydro- genase (9), pig kidney dopa decarboxylase (lo), and possible in peptidylglycine-amidating monooxygenase (1).

Identification of quinone cofactors with carbonyl reagents is difficult and may give ambiguous results (11). This is illustrated by the fact that plasma amine oxidase which was reported to contain PQQ (6) has now been unambiguously shown to contain a 6-hydroxydopa residue (2).

In view of the potential physiological (12) and pharmaco- logical (13-15) significance of quinone cofactors, definitive detection and identification of quinonoid compounds in pro- teins is essential.

We have found that PQQ oxidizes glycine at an alkaline pH but not valine. The superoxide released in the cycling reaction of PQQ with sodium glycinate reduces nitroblue tetrazolium to its formazan, allowing the detection of pico- moles of PQQ (16, 17). This oxidation of glycine is not restricted to PQQ but appears to be a general property of quinones. We have therefore used this reaction for the iden- tification of PQQ-like substances and quinonoid compounds in biological fluids, quinoproteins, and quinopeptides (18). We now report how this redox-cycling property of quinones can be used for the detection of quinoproteins after separation and electroblotting onto nitrocellulose. Results of a survey on several putative mammalian quinoproteins are presented.

EXPERIMENTAL PROCEDURES

Materiak-Glycine was either electrophoresis grade from Bio-Rad or AnalR from BDH; PQQ was obtained from Fluka, Switzerland, or as a gift from the Mitsubishi Gas Chemical Co., Japan; NBT, dopa, and 6-hydroxydopa (topa) from Sigma; phenanthrenequinone and 1,2,4-trihydroxybenzene were from Aldrich. Pig kidney diamine oxi- dase from Sigma was partially purified by passage over hydroxyapa- tite. The following purified proteins were kindly provided to us: dopamine P-hydroxylase (T. Kokubo, IRL, Ciba-Geigy, Japan and J . Klinman, University of California, Berkeley); bovine serum amine oxidase (J. Klinman); methylamine dehydrogenase from bacterium W3A1 and methylamine oxidase from Arthrobacter P1 (B. McIntire, University of California, San Francisco); vitelline B and C from the liver fluke, Fasciola hepatica (J. H. Waite, University of Delaware); frog peptidylglycine-amidating monooxygenase expressed with a bac- ulovirus expression vector system (Y. Nishikawa, IRL, Ciba-Geigy, Japan).

NBTIGlycinate Assay-To 100-pl sample aliquots were added 300 pl of potassium glycinate (2 M, pH lo), 100 pl of a 10 mg/ml sodium borohydride-reduced and dialyzed serum albumin solution (17), and 1 ml of nitroblue tetrazolium (0.24 mM NBT) in 2 M potassium glycinate, pH 10. Reagents were pipetted into tubes immersed in an ice-water bath. The reaction was started by incubation of the tubes in a water bath at 25 “C in the dark. After 1 h the absorbance at 530 nm was determined.

SDS-PAGE-Polyacrylamide (7.5-12%) gel electrophoresis was performed under reducing conditions according to Laemmli (19). A Mini-Protean I1 apparatus (Bio-Rad) with a 0.75-mm spacer was used. The samples were diluted with sample buffer and heated for 5 min at 95” C. Electrophoresis was stopped when the bromphenol blue dye reached the bottom of the gel.

689

690 Detection of Quinoproteins Protein Blotting-Electrophoretic transfer of proteins from the

polyacrylamide gel onto nitrocellulose paper (0.45-pm membrane, Micron Separations Inc.) was performed a t 100 V for 1 h and 10 "C using a pH 8.3 blotting buffer containing 25 mM Tris, 192 mM glycine, and 20% methanol.

Staining for Quinoproteins-Quinoproteins were detected by stain- ing with NBT (0.24 mM in 2 M potassium glycinate, pH 10). The nitrocellulose paper was immersed in the glycinate/NBT solution for 45 min in the dark resulting in a blue-purple stain of quinoprotein bands and no staining of other proteins. This time was optimal for the detection of quinoproteins. If valinate replaces glycinate, there is no staining of quinoproteins with NBT. The more reactive quinopro- teins, bovine serum amine oxidase, methylamine dehydrogenase, methylamine oxidase, vitelline B and C, stain in a shorter time. The intensity of the stain did not increase after 45 min with diamine oxidase or lysyl oxidase. The nitrocellulose was washed and/or stored in 0.1 M sodium borate, pH 10, a t 4 "C. Protein was stained with Ponceau S (0.1% in 5% acetic acid) resulting in a red stain, while the already stained quinoproteins remained blue-purple.

RESULTS AND DISCUSSION

The reactivity of various quinones and quinonoid com- pounds in the NBT/glycinate assay is shown in Fig. 1. PQQ is the most reactive quinonoid substance followed by phen- anthrenequinone, and the quinonoid compounds, topa, 1,2,4- trihydroxybenzene, menadione, and dopa. These data show that redox cycling with NBT/glycinate detects quinonoid compounds with considerable sensitivity. An important fea- ture of this approach is that it detects quinonoid compounds irrespective of their initial oxidation state. Detection of qui- nonoid compounds using carbonyl reagents requires that they be present in the quinone form.

The yield of formazan produced in the redox cycling which varies with the different quinones tested depends on the extent of dimerization and polymerization to redox-inert end products. With PQQ polymerization is likely to be limited at pH 10 because of charge repulsion between PQQ molecules caused by its three carboxylate groups. This explains in part why PQQ redox cycles more efficiently than phenanthrene quinone. Such a dimerization-polymerization process cannot occur with protein-bound quinone cofactors. As a conse- quence, redox cycling is likely to be more pronounced with quinones bound to proteins than with quinones in solution.

Fig. 2 demonstrates that NBT/glycinate can be used as a specific stain for quinoproteins, separated, and electroblotted onto nitrocellulose. Panels A and B show the staining of the mammalian quinoproteins, bovine serum amine oxidase (lane 1 ) and pig kidney diamine oxidase (lune 2). Panel B shows

1 t

0.09 ' ' " " " ' 1 I , ' " ' ' ' I ' " . ' ' ' . I ' ' ' " ' ' ~ 1 1 , , . . . . .1 1 Ob 1 0 ' 10' 109 10 ' 10 '

pmoles of Compound FIG. 1. Reactivity of quinones and quinonoid compounds in

the NBT/glycinate assay. 0, PQQ; 0, PAQ; W, topa; 0, 1,2,4- trihydroxybenzene; A, dopa; 'I, menadione.

the NBT/glycinate staining pattern and panel A the same electroblot counterstained red for protein with Ponceau S. The partially purified diamine oxidase is specifically stained with NBT/glycinate (B , lane 2) , while the contaminating proteins stain only with Ponceau S (A, lune 2) . Purified bovine serum amine oxidase, containing the quinonoid cofactor 6- hydroxydopa also stains strongly with NBT/glycinate (B , lane 1) . The electrophoretic heterogeneity is likely to be caused by the glycosylation state of this protein (20).

The specificity of NBT/glycinate staining is also apparent with the bacterial quinoproteins methylamine oxidase (21) and methylamine dehydrogenase (22) (D, lunes 1 and 2 ) . The quinone-containing a-subunits (80-90 kDa) of the methyla- mine oxidase stain strongly (D, lane l ) with NBT/glycinate; few protein bands are seen with the Ponceau S (C, lune 1 ). In the methylamine dehydrogenase (a&), only the 15-kDa sub- unit containing the covalently bound quinone cofactor stains with NBT/glycinate (D, lune 2 uersus C, lune 2).

Several well characterized proteins which do not contain quinonoid cofactors do not stain with NBT/glycinate, includ- ing the flavoprotein xanthine oxidase, the heme-containing cytochrome c, and the glycoprotein fetuin.

The ketoamine structure which is present in nonenzymat- ically glycosylated (glycated) proteins does not support redox cycling. Unlike the quinones ketoamines react weakly in both NBT/glycinate and NBT/valinate. Furthermore, reduction with sodium borohydride eliminates reactivity from ketoam- ines, but not from quinones, since redox cycling is initiated from both quinones and hydroquinones. A weak positive reaction observed with certain preparations of albumin is reduced by prior reduction with sodium borohydride, indicat- ing that glycation was responsible for part of this weak reactivity.

NBT/glycinate staining was strongly positive with dopa- containing vitelline B (31 kDa) and vitelline C (16 kDa) (not shown). In these proteins, multiple dopa residues are present from which dopaquinone can be formed (23). As the quinone cofactor of methylamine dehydrogenase appears to be neither PQQ nor 6-hydroxydopa? NBT/glycinate detects a broad spectrum of quinonoid cofactors.

Apart from the 6-hydroxydopa-containing serum amine oxidase, to date only two other mammalian quinoproteins, diamine oxidase and lysyl oxidase, tested positive. Appreciable color was apparent with about 10 pg of protein. When com- pared with the other quinoproteins the reactivity of these two putative PQQ-containing oxidases is lower. This may be due to the following: (i) the quinone cofactor in these quinopro- teins is not PQQ and redox cycles less efficiently; (ii) the quinone cofactor may have undergone an irreversible modifi- cation during purification and may no longer be detectable by redox cycling, but this appears unlikely; (iii) the quinone cofactor(s) are present in less than stoichiometric amounts and this seems to be the case with lysyl oxidase. Cofactor titration of purified bovine lysyl oxidase with [14C]phenylhy- drazine yields fractional values ranging from 0.2 to 0.35 mol/ mol protein (3), values that quantitatively agree with the degree of covalent labeling of the active site with the mecha- nism-based inactivator, 14C-labeled P-aminopropionitrile (24).

NBT/glycinate staining was used to assess the quinoprotein nature of the putative quinoprotein oxygenases, dopamine P- hydroxylase, peptidylglycine-amidating monooxygenase (11, and soybean lipoxygenase (25). These three proteins did not react as quinoproteins with glycinate-NBT. The results ob- tained with two highly purified dopamine P-hydroxylase prep- arations are shown in Fig. 3. It is apparent that the commer-

W. S. McIntire, personal communication.

Detection of Quinoproteins 69 1

FIG. 2. NBT/glycinate staining of Mr quinoproteins on an SDS-PAGE 200 - electroblot. A: lane I , purified bovine 116 - serum amine oxidase; lane 2, partially 97 - purified pig kidney diamine oxidase. The 66 - electroblot was first stained with NBT/ glycinate (as in R ) and then counter- 63 - stained with Ponceau S. The molecular mass markers (kDa) are on the left side. H , same proteins as in A stained with NBT/glycinate; only quinoproteins are stained. C: lane 1, purified methylamine oxidase; lane 2, purified methylamine de- hydrogenase. The electroblot was stained first with NBT/glycinate (as in IN and then counterstained with Pon- ceau S. D, same proteins as in C stained with NBT/glycinate only; quinoproteins are stained specifically. A 1 2

.-~..~.-. . -.

Mr 200 -

116 - 97-

66-

43-

A 1 2 3 B 1 2 3 FIG. 3. NBT/glycinate staining of dopamine &hydroxylase

preparations on an SDS-PAGE electroblot. A and R, crude dopamine @-hydroxylase (lane 2) and purified dopamine @-hydroxyl- ase preparations (lanes I and 3 ) . A, NBT/glycinate stain which shows an unknown quinoprotein of about 97 kDa in the crude dopamine @- hydroxylase (lane 2) and in one purified dopamine P-hydroxylase preparation (arrow, lane I ) . R, same gel counterstained for protein with Ponceau S.

cia1 crude dopamine P-hydroxylase preparation (panel A , lane 2) also contains a 97-kDa quinoprotein band which is not dopamine P-hydroxylase. This band was removed completely during purification in one preparation (A , lane 3 ) but not completely in the other (A, lane 1 ). The Ponceau counterstain shown in B, lanes 1-3, detects several protein bands in the commercial dopamine P-hydroxylase. Copurification of this contaminant quinoprotein may have led to the erroneous identification by the hexanol procedure of dopamine P-hy- droxylase as a quinoprotein (7) which could not be confirmed (26). I t also explains why quinopeptides from a Pronase digest of crude dopamine P-hydroxylase separated by high pressure liquid chromatography could be detected with glycinate/NBT (18).

The frog peptidylglycine-amidating monooxygenase en- zyme and the soybean lipoxygenase (25) did not stain with NBT/glycinate (not shown). As with dopamine P-hydroxyl- ase, a purported phenylhydrazine product of PQQ has been isolated from a soybean lipoxygenase (27). This discrepant finding remains unexplained. However, based on our data it appears that these oxygenases are not quinoproteins.

The presence of covalently attached quinones is confirmed for the mammalian copper-dependent amine oxidases but not for other mammalian putative quinoproteins. The mamma- lian quinoprotein amine oxidases are important metabolic

B 1 2 C 1 2 D 1 2

enzymes. Diamine oxidase is the major enzyme which cata- bolizes putrescine, the substrate for the synthesis of the other polyamines, hence, its importance in the regulation of polya- mine levels in tissues and in the control of cell proliferation. Lysyl oxidase catalyzes the cross-linking of collagen and elastin and is the key enzyme regulating connective tissue maturation. Plasma amine oxidase controls the removal of biogenic amines from the circulation.

Quinolation may be a mechanism for the regulation of enzyme activity as it is suggested by the less than stoichio- metric quinone content in lysyl oxidase and serum amine oxidase. The redox-cycling approach presented here provides a tool to reliably determine the quinone content of proteins for the study of the biological significance of this process and the factors that affect protein quinolation.

Acknowledgments-We wish to acknowledge the skillful technical assistance of A. Skalicki and C. Garey.

REFERENCES

1. Duine, J. A., and Jongejan, J . A. (1989) Annu. H F V . H ~ C J C ~ W I . 58,

2. Janes S. M., Mu, D., Wemmer, D., Smith, A. J . , Kaur, S., Maltby, D., Alma, L., Burlingame, A. L., and Klinman, d. P. (1990) Science 248,981-987

3. Williamson, P. R., Moog R. S., Dooley D. M., and Kagan, H. M. (1986) J. Riol. Chem. 261, 16302-16305

4. van der Meer, R. A., and Duine, J . A. (1986) Riochern. J. 239,

5. van der Meer, R. A., Jongejan, J. A., Frank, J., and Duine, d. A.

6. Lobenstein-Verbeek, C. L., Jonjegan, J. A., Frank, d . , and Duine,

7. Knowles, P. F., Pandeya, K. R., Rius, F. X., Spencer, C. M., Moog, R. S., McGuirl, M. A., andDooley, D. M. (1987) Biochern.

8. van der Meer, R. A,, Jongejan, J. A., and Duine, J. A. (1988)

9. Ameyama, M., Shinagawa, E., Matsushita, K., Takimoto, K., Nakashima, K., and Adachi, 0. (1985) Agric. Rid. C h m . 49,

10. Groen, B. W., van der Meer, R. A., and Duine, d. A. (1988) FfiL4S

403-426

789-791

(1986) FERS Lett. 206, 111-114

J. A. (1984) FEBS Lett. 170, 305-309

J . 24 1,603-608

FERS Lett. 231,303-307

3623-3626

Lett. 227,98-102 1. Gallop, P. M., Paz, M. A., Fluckiger, R., and Kagan, H. M. (1989)

Trends Riochem. Sci. 14, 343-346 2. Killgore, J., Smidt, C., Duich, L., Romero-Chapman, N., Tinker,

D., Reiser, K., Melko, M., Hyde, D., and Rucker, R. R. (1989) Science 245,850-852

3. Watanabe, A., Hobara, N., and Tsuji, T. (1988) Curr. Thw. Hcs.

4. Nishigori, H., Yasunaga, M., Mizumura, M., Lee, d. W., and 44,896-901

Iwatsuru, M. (1989) Life Sci. 45,593-598

692 Detection of Quinoproteins 15. Hobara, N., Watanabe, A., Kobayashi, M., Tsuji, T., Gomita, Y.,

and Araki, Y. (1988) Pharmacology 37 , 264-267 16. Fluckiger, R., Woodtli, T., and Gallop, P. M. (1988) Biochem.

Biophys. Res. Commun. 153, 353-358 17. Paz, M. A., Gallop, P. M., Torrelio, B. M., and Fluckiger, R.

(1988) Biochem. Biophys. Res. Commun. 154, 1330-1337 18. Paz, M. A., Fluckiger, R., Henson, E., and Gallop, P. M. (1989)

in PQQ and Quinoproteins (Jongejan, J. A., and Duine, J. A., eds) pp. 131-143, Kluwer Academic Publishers, Norwell, MA

19. Laemmli, U. K. (1970) Nature 227,680-685 20. Watanabe, K., and Yasunobu, K. T. (1970) J. Biol. Chem. 245 ,

21. McIntire, W. S. (1990) Methods Enzymol. 188, 227-235 4612-4617

22. Kenney, W. C., and McIntire, W. S. (1983) Biochemistry 22 ,

23. Waite, J. H., and Rice-Ficht, A. C. (1989) Biochemistry 28,6104- 6110

24. Tang, S. S., Trackman, P. C., and Kagan, H. M. (1983) J. Bid. Chem. 235,4331-4338

25. Veldink, G. A., Boelens, H., Maccarrone, M., van der Lecq, F., Vliegenthart, J. F. G., Paz, M. A., Fluckiger, R., and Gallop, P. M. (1990) FEBS Lett. 270,135-138

26. Robertson, J. G., Kumar, A., Mancewicz, J. A., and Villafranca, J. J. (1989) J. Biol. Chem. 264, 19916-19921

27. van der Meer, R. A,, and Duine, J. A. (1988) FEBS Lett. 236,

3858-3868

194-200