Oral Microbiol Immunol 1992: 7; 204-211

Survey of a receptor protein in human erythrocytes for hemagglutinin of Porphyromonas gingivaiis

H. Hayashi, A. Nagata, D. Hinode, M. Sato, R. Nakamura Department of Preventive Dentistry, School of Dentistry, University of Tokushima, Japan

Hayashi H, Nagata A, Hinode D, Sato M, Nakamura R. Survey of a receptor protein in human erythroeytes for hemagglutinin o/Porphyromonas gingivalis. Oral Miei-obiol Immunol 1992: 7: 204-211. The purpose of this study is to survey a receptor proteiti in human erythrocyte membrane for the hetnagglutinin (HA) of Porphyromonas gingivalis. Hutnan erythrocytes were modified by either chymotrypsin or P. gingivalis HA along with the disappearance of their hctnagglutinating ability and the rctnoval of the band 3 proteiti. By preparative electrophoresis, this protein was isolated atid purified frotn hutnan erythrocytes. The purified protein showed strong inhibitory activity for hetnagglutination and the binding to P. gingivalis cells, whose binding sites were calculated to be approximately 9000, suggesting its binding to the active site of HA. Hemagglutinin purified from P. gingivalis by affinity absorption to sheep erythrocyte ghosts possessed strong trypsin-like activity, and both the HA and the enzyme activities were inhibited by arginine. Specific tnodification of arginyl residues in human erythrocytes by phenylglyoxal diminished the hetnagglutinating ability. From the similarity of the inhibition profile and possible active sites between HA and the trypsin-like protease, it is suggested that hemaggiutination may occur as a result of the primary reaction of the enzyme (protease) and the substrate. These results suggest that band 3 may be a key protein in human erythrocyte metnbrane for HA frotn P. gingivalis and its binding sites may be arginyl residues of the protein.

Hemagglutinating activity (HA activity) of pathogenic bacteria has been noticed in relation to their adhesive capacity to host tissues, which is an early step in bacterial infections. Porphyromonas gingivalis. a possible pathogenic bacterium of adult periodontitis, can also agglutinate erythrocytes, and this property may correlate with its colonization in the periodontai lesion. The hetnagglutinating substance of P. gingivalis has been purified and characterized by Inoshita et al. (13) and Okuda et al. (26). Recently, Mouton et al. (20) have demonstrated that the hemagglutinating adhesin HA-Ag2 of P. gingivatis is itnmunochemically distinct from fimbrilin; that is, the antibody against HA-Ag2 reacts with antigens in outer metnbrancs prepared as extracellular vesicles and/or in shcared-cell outer tnetnbranes, but does not react with antigens in fimbriae. Sev-

eral studies have been performed on the active protein of P. gingivatis as a hetnagglutinin (HA) or an adhesin. However, little is known of the receptor protein of this substance on erythrocytes. From the inhibition results of the hetnagglutination by arginine, Mayrand & Holt (19) have suggested that arginine may function as a contact residue between the bacterial cell receptor and its counterpart on the erythrocyte during agglutination. More recently, Ni Eidhin & Mouton (23) have inferred that the erythrocyte membrane receptor for the HA-Ag2 of P. gingivalis is presutnably spectrin. In this study, we provide evidence that the receptor protein in hutnan erythrocyte metnbrane for the HA of P. gingivalis 381 tnay be a band 3 protein; moreover, arginyl residues appear to be important in the interaction be-

Key words: Porphyromonas gingivaiis: hem-

agglutination: human erythrocyte: band 3: receptor protein Ryo Nakamura, Department of Preventive Dentistry, University of Tokushima School of Dentistry, 18-15, Kuramotocho-3-chrome, Tokushima City, 770, Japan Accepted for publication November 27, 1991

tween this protein and the P. gingivatis HA. Material and methods Bacterial strain

P. gingivatis 381, which was originally isolated from subgingival plaque by Dr. S. S. Socransky, Forsyth Dental Center, Boston, MA, was obtained by courtesy of Dr. Y. Yamatnoto, Sunstar Inc., Osaka, Japan, and was maintaitied in our Iabotatory by weekly transfer on anaerobic basal tnedium with 5% sheep blood. The bacteriutn was grown atiaerobically at 37 "C as described previously (27). Hemaggiutination assay

HA activity was detertnined in roundbottomed microtiter plates (Limbro Di-

Receptor of hemaggtutinin of P. gingivalis vision Flow Laboratories, Hamden, CT). Fifty jid of the samples for HA activity were serially diluted in two-fold steps in isotoiiic Tris-buffer-sodium chloride, pH 7.5 (TBS; 10 mM Tris hydrochloride buffer containing 0.15 M sodiutn chloride), and an equal volutne of 1% washed hutnan erythrocytes (4x lO"" cells) was added. The mixtures were incubated for 90 tnin at 37°C, and then the hemaggiutination was assayed by visual inspection. The hemagglutinating titer was defined as the reciprocal of the highest dilution showing hetnagglutination, and one unit of the activity was expressed as the amount of HA sufficient to agglutinate a specified number of erythrocytes (per well) utider standard assay conditiotis. For the detertnination of hemaggiutination inhibition, 25 /NA inhibited the HA activity (Table 4). The strongest inhibitory activity was found in N-benzoyl-Pro-Phe-Arg-pNA, and the lowest inhibitory concentration . in a well was 2.9 pM. However, when arginine residue existed in a peptide as amino terminal such as Arg-Pro-/jNA, no inhibition for the HA activity was observed. Also, L-Arg-/iNA showed very weak inhibition. All of the synthetic substrates that showed inhibition for the HA activity were hydrolyzed very well by the purified HA, and other substrates that did not show any inhibition were not at all hydrolyzed by the HA (Table 4).

Table 4. Inhibitory activity of some synthetic substrates for aminopeptidases on the HA activity and their hydrolysis by HA Substrate N-/)Tosyl-Gly-Pro-Arg-/)NA N-/)Tosyl-Gly-Pro-Lys-/;NA N-CBZ-Gly-Pro-Arg-/;NA Arg-Pro-pNA Gly-Pro-/;NA Nt-Boc-Leu-Ser-Thr-Arg-/7NA L-Arg-/;NA Na-Benzoyl-DL-Arg-/)NA N-Benzoyl-Pro-Phe-Arg-/;NA

Lowest inhibitory concentration for hemaggiutination (/(M)

Relative hydrolysis*

13.5 >376 26.8 >538 >760 5.3 340 71.8 2.9

122 NG 84.6 NG NG 138 63.9 100 110

• Relative hydrolysis is expressed as that the hydrolysis of Na-Benzoyl-DL-Arg-/;NA is 100 under the specified conditions as described in the text. When optical density was less than 0.01 for the assay of the hydrolyzed /;-nitroaniline, the activity was considered negligible (NG).

Reeeptor of hemagglutinin of P. gingivalis of incubation time were applied to the experiment of saturation kinetics of '"Ilabelled band 3 protein. Fig. 5 shows the saturation kinetics for the binding of band 3 protein to P. gingivalis 381. When iticreasing amounts of '•'^I-labellcd band 3 protein were added to the reaction mixture, the binding to bacterial cells increased. The inset shows the analysis by the Scatchard tnethod (30), and it suggests the single binding site and an apparent Kj is calculated to be 77 nM. At saturation, 2.5 X 10' bacterial cells bound 37.8 ptnol of band 3 protein, that is, 9000 binding sites per bacterial cell were calculated.

ported previously by Inoshita et al. (13) and Nishikata & Yoshimura (24), in terms of the SDS-PAGE pattern and the molecular size of HA (Fig. 6). The contact residue between HA and the membrane protein or erythrocytes may be arginine and/or lysitie residues, because these atnino acids were reported to inhibit hemaggiutination of HA frotn P. gingivatis (13, 26). The fact that argininc was used to release HA from erythrocyte ghosts in this study supported the concept of binding of HA to arginyl residue in the erythrocytes. Furthermore, the loss of hemagglutinating activity by tnodification with PGO proved this concept because this reagent is generally used to modify selectively Discussion arginyl residues in the peptide specifiThe hemagglutinin used for this study cally (31). was isolated and purified from P. ginNi Eidhin ct al. (23) suggested that givalis by procedures described above the erythrocyte membrane receptor for using affinity absorption to the glutar- HA-Ag2 is one of the extrinsic proteins aldehyde-fixed ghosts prepared from and presutned it to be spectrin. Howsheep erythrocytes. This isolation and ever, spectrin is one of the extrinsic propurification procedure was followed be- teins that exist on the phospholipid bicause it was simple and reasotiable, and layers, but no part of thetn is exposed the idea comes frotn utilizing its biologi- to the outside of the erythrocyte memcal ability of direct binding to erythro- bratie (2, 18, 22). On the other hand, cyte membrane. The purified HA may band 3, glycophorin A, band 4.5 and have some contamination, but we ob- band 7 are called intrinsic proteins or tained very similar results to those re- trans-membrane proteins, which exist in

209

200K1

A

B

Fig. 6. SDS-PAGE pattern of the purified HA. Lanes; A. Marker proteins. B. Purified HA. Staining was carried out with silver stain 11.

o

20

c '3 "o

0.7 0.6

CO

•a

0.5

cts

0.4

m 10

0.3

_0J

0.2

"3

0.1 0

20

30

40

Bound (nM)

0

10

20

30

40

50

Band 3 p r o t e i n free

60 (pmol)

Fig. 5. Dose-response curves for binding of '"I-labelled band 3 protein to P. gingivatis 381. Bacteria (2.5x10' cells) and the increasing amount of the labelled band 3 protein were incubated at 4''C for 1 h as described in the text. Each value shows the average of quadruplicate determinations. • — • ; total binding; • — • : nonspecific binding; • — • : specific binding. The inset shows the Scatchard analysis of the specific binding data.

the phospholipid bilayers, but these proteins transfixed the tnembrane and part of them are exposed to the outside of the crythrocyte membrane (7, 15, 29). By the treattnent of the surface of the erythrocyte membrane with HA and chymotrypsin, sotne proteins were removed, with the consequent loss of their hemagglutinating activity. One of these proteins is supposed to be a receptor protein. Among the proteins that disappeared by the treatment, some extrinsic proteins such as ankylin, band 4.1 and hetn, which should not be affected, are included. The reason has not been clarified yet, but it is conceivable that these 4 protein bands were eliminated because they detached from the tnembrane, which was caused by the disintegration of band 3. Besides, these proteins are considered to bind to band 3 at the inner surface of the metnbrane, and to be indirectly fixed to the tnembrane by combining with band 3 (7, 28, 32). PGO selectively modified arginyl residues of

210

Hayashi et al.

band 3 protein, with the greater part of arginyl residues present on the outer surface of the crythrocyte membrane being localized to this protein (1). Thus, the band 3 represented a likely candidate as an crythrocyte receptor for HA. Accordingly, the band 3 protein was isolated and purified, and strong inhibition of hemaggiutination by this purified band 3 protein was confirmed. The binding study on '"'I-labelled band 3 protein proved its strong affinity to P. gingivalis cells. Murakami et al. (21) have reported the binding of a histatin 5 to P. gingivatis that shows the inhibitory activity for hemaggiutination. They calculated the binding site on the P. gingivatis cells to be 3600. In comparison with this data, P. gingivalis cells have approximately 2.5 times binding sites for band 3 protein. In defining the mechanism of hemaggiutination by P. gingivalis HA, the proteolytic activity of this substance is of special interest because there is some evidence that the hemaggiutination may occur as a result of the production of enzyme-substrate complex by the primary enzyme reaction between the protease and a protein containing arginyl residues. This evidence iticludes the following; 1) HA possesses trypsin-like activity as described above and reported recently by Nishikata et al. (24). A similar kind of evidence was observed on Vibrio cholerae (11). 2) Inhibition profiles of HA activity arc very similar to those of trypsin-like protease (25). 3) Many synthetic substrates for arginine aminopeptidase possess inhibitory activity for hemaggiutination by the purified HA. 4) Some membrane proteins of the erythrocytes were digested as described above. 5) Arginyl residues in the peptide are the binding sites for both HA and the trypsin-like protease. Hence, the receptor protein on the erythrocyte membrane could serve as a substrate for the trypsin-like protease, possibly via the carboxyl site of arginine, which is recognized by trypsin-like .protease. Among proteins consisting of the crythrocyte membrane, band 3 is the major component that contains arginyl residue abundantly (1), and could be a substrate of the trypsin-like protease. The results of this study suggest that band 3 may represent a key receptor protein in the erythrocyte membrane for HA frotn P. gingivalis, and that arginyl residues in this protein are its binding sites. Hemaggiutination may occur as a result of enzytne-substrate reaction be-

tween a trypsin-like protease and the band 3 protein. However, further study will be needed to confirm this mechanism.

Acknowledgements

We gratefully acknowledge the kind advice of Professor H. Ishida, Department of Pharmacology for the binding assay of band 3 protein. This study was supported in part by a Grant-in-Aid for Scientific Research (A-02404078) from the Ministry of Education, Science and Culture of Japan (1990-93).

References 1. Bjerrutn PJ, Wieth JO, Borders CL Jr. Selective phenylgiyoxalation of functionally essential arginyl residues in the erythrocyte anion transport protein. J Gen Physiol 1983; 81; 453-484. 2. Branton D, Cohen CM, Tyler J. Interaction of cytoskeletal proteins on the human erythrocytes membrane. Cell 1981; 24; 24-32. 3. Cabantchik ZI, Rothstein A. Membrane protein related to anion permeability of human blood cells. J Membrane Biol 1974; 15; 227-248. 4. Erianger BF, Kokowsky N, Cohen W. The preparation and properties of two new chromogenic substrates of trypsin. Arch Biochem Biophys 1961; 95; 271-278. 5. Fukuda M, Eshdat Y, Tarone G, Marehesi VT. Isolation and characterization of peptides derived from the cytoplasmic segment of band 3, the predominant intrinsic membrane protein of the human erythrocytes. J Biol Chem 1978; 253; 2419-2428. 6. Giga Y, Sutoh K, Ikai A. A new multimeric hetnagglutinin from coelomic fiuid of the sea urchin Anihocidaris cra.ssi.spina. Biochem 1985; 24; 4461-4467. 7. Haest CWM. Interactions between metnbrane skeleton proteins and the intrinsic domain of the erythrocyte membrane. Biochim Biophys Acta 1982; 694; 331-352. 8. Hartree EF. Determination of protein; a modification of the Lowry tnethod that gives a litiear photomeric response. Anal Biochem 1972; 48; 422-427. 9. Hayashi H, Morioka M, Sato M, Nakamura R. Isolation of hemagglutinin from Porphyromonas gingivatis using erythrocyte ghost as an affinity adsorbent and its morphological observation. Proc Third World Congr Prev Dentistry. Fukuoka; Matsukuma Printing (in press). 10. Hinode D, Hayashi H, Nakamura R.

Purification and characterization of three types of proteases from culture supernatant of Porphyromonas gingivatis. Infect Immun 1991; 59; 30603068. 11. Honda T, Hata-Naka A, Lertpocasombat K, Miwatani T. Production of monoclonal antibodies against a hemagglutinin/protease of Vit>rio ehoterae nonOl. FEMS Microbiol 1991; 78; 227-230. 12. Hunter WM, Greenwood FC. Pteparation of iodine-131 labelled human growth hormone of high specific activity. Nature 1962; 194; 495-496. 13. Inoshita E, Atnano A, Hanioka T, Tatnagawa H, Shizukuishi S, Tsunemitsu A. Isolation and some properties of exohemagglutinin from the culture medium of Bacteroides gingivalis 381. Infect Itnmun 1986; 52; 421-427. 14. Jennings ML. Kinetics and mechanism of anion transport in red blood cells. Ann Rev Physiol 1985; 47; 519-533. 15. Kopito RR, Lodish HF. Primary structure and transmembrane orientation of the murine anion exchange protein. Nature 1985; 3T6; 234-238. 16. Laetnmli UK. Cleavage of structural proteins during the assetnbly of the head of bacteriophage T^. Nature 1970; 227; 680-685. 17. Lukacovic MF, Feinstein MB, Sha'afi RI, Perrie S. Purification of stabilized band 3 protein of the human erythrocyte membtane and its reconstitution into liposomes. Biochem 1981; 20; 31453151. 18. Lux SE. Dissecting the red cell tnembrane skeleton. Nature 1979; 281; 426429. 19. Mayrand D, Holt SC. Biology of asaccharolytic black-pigmented Baeteroides species. Microbiol Rev 1988; 52: 134152. 20. Mouton C, Ni Eidhin D, Deslauriets M, Lamyl L. The hemagglutinating adhesin HA-Ag2 of Baeteroide.s gingivati.s is distinct from fimbrilin. Otal Microbiol Immunol 1991; 6; 6-11. 21. Murakami Y. Ishimoto T, Amano A, Tatnagawa H, Shizukuishi S, Tsunemitsu A. [Binding of a histatin to Porpliyromona.s gingivatis.] J Dent Health 1991: 41; 472-473 (in Japanese). 22. Nicolson GJ, Marchesi VT, Singer SJ. Localization of spectrin on the inner surface of human red blood cell membranes by ferritin-conjugated antibodies. J Cell Biol 1971; 51; 265-272. 23. Ni Eidhin D, Mouton C. Pteliminary identification of the erythrocyte membrane receptor for the hetnagglutinating adhesin HA-Ag2 of Baeieroides gingivatis. J Dent Res 1989; 68; 879. 24. Nishikata M, Yoshimura F. Characterization of Porphyromonas (Baeieroides) gingivalis hemagglutinin as a protease. Biochem Biophys Res Commun 1991; 178; 336-342. 25. Nishikata M, Yoshimura F, Nodasaka

Reeeptor of hemagglutinin of P. gingivalis Y. Possibility of Baeteroides gingivalis hemagglutinin possessing protease activity revealed by inhibition studies. Microbiol Immunol 1989; 33; 75-80. 26. Okuda K. Yamamoto A, Naito Y, Takazoe 1, Slots J, Genco RJ. Purification and properties of hetnagglutinin frotn culture supernatant of Bacteroides gingivalis. Infect Imtnun 1986: 54: 659-665. 27. Otsuka M, Endo J, Hinode D et al. Isolation and characterization of protease from cultute supernatant of Bacteroides

gingivalis. Infect Itntnun 1986: 22: 491498. 28. Pasternack GR. Anderson RA, Leto TL, Marchesi VT. Interactions between protein 4.1 and band 3. J Biol Chetii 1985: 260: 3676-3683. 29. Ross AH. Radhakrishnan R, Robson RJ, Khorana HG. The transtnembrane domain of glycophorin A as studied by crosslinking using photoactivatable phospholipid. J Biol Chetn 1982: 257: 4152^161.

211

30. Scatchard G. The attractions of proteins for small molecules and ions. Ann NY Acad Sci 1949: 51: 660-672. 31. Takahashi K. The reaction of phenylglyoxal with arginine residues in proteins. J Biol Chem 1968: 243: 6171-6179. 32. Walder JA, Catterjee R. Steck TL et al. The inteiaction of hetnoglobin with the cytoplastnic dotnain of band 3 of the human erythrocyte metnbrane. J Biol Chem 1984: 259: 10238-10246.