Adherence of Mycoplasma pneumoniae to human alveolar macrophages

IMMUNOLOGY AND MEDICAL MICROBIOLOGY

ELSEVIER

FEMS Immunology

and Medical Microbiology

15 (1996) 135-141

Adherence of Mycoplasma pneumoniae to human alveolar macrophages Abed Athamna a, Mordechai R. Kramer b, Itzhak Kahane a3* a Department of Membrane and Ultrastructure Research, The Hebrew University. Hadassah Medical School, PO Box 12272, Jerusalem 91120, Israel b Pulmonary Institute. Hadassah University Hospital, Jerusalem 91120, Israel Received 20 March 1996; revised 11 June 1996; accepted

12 June 1996

Abstract The human pathogen Mycoplasma pneumoniae causes primary atypical-cold agglutinin-positive pneumonia. Since alveolar macrophages intl:rnalize mycoplasma as part of their immune defense, we studied characteristics of the human macrophage receptor for opsonized and nonopsonized M. pneumoniae. The glass-adhering subpopulation of h4. pneumoniae attached more than the non-adherent subpopulation. The attachment was dose-dependent and enhanced by opsonization in the presence of human serum. It is inhibited by sulfated compounds such as dextran-sulfate and polyanetholsulfonic acid, but not by dextran or several monosaccharides, suggesting that sulfated glycolipids on the macrophage surface may act as receptors for M. pneumoniae binding. In addition, sialylated compounds, such as fetuin and al-acid glycoprotein, were found to be potent inhibitors of the attachment, also indicating the role of sialic acid residue in recognition and attachment of M. pneumoniae

to human

Keywords: Adherence:

alveolar

macrophages.

Human alveolar macrophage;

Mycoplasma pneumoniae; Phagocytosis;

1. Introduction The alveolar macrophages are the first and the predominant phagocytic cells in the lung that come into contact with invading microorganisms. Their antibacterial functions requires attachment of the bacteria to the phagocytic cells. The interaction between the cells is mediated either by opsonins (immune globulins or complement factors) or by surface components’ recognition followed by phagocytosis and subsequent destruction of the bacteria (for a

* Corresponding (2) 346-198.

author.

Tel: +972

0928-8244/96/$15.00 Copyright PiI SO928-8244(96)00055-7

(2) 758-155;

0 1996 Federation

Fax:

+972

of European

Sialic acid; Sulfated lipid

review, see [l]). In vitro studies, conducted on the interaction between mycoplasmas and several species of phagocytic cells of various animals, strongly suggest the role of opsonins in mediating the attachment between the mycoplasmas and various types of macrophages [I-4]. The attachment of mycoplasmas to phagocytic cells in the absence of opsonins is a feature of several, but not all, Mycoplasma species [l-6]. It was also shown that in animals without specific antibodies to mycoplasma, the outcome of the initial interaction with phagocytic cells plays an important role in determining the development of infection [7]. The human pathogen M. pneumoniae is a major pathogen responsible for about 25% of hospitalization, primarily of children and young Microbiological

Societies. Published

by Elsevier Science B.V.

136

A. Athamna et al./ FEMS Immunology and Medical Microbiology 15 (1996) 135-141

adults with pneumonia [g-lo]. Data on the interaction between the human pathogen M. pneumoniae, which colonizes the epithelial lining of the human respiratory tract and the human alveolar macrophages (HA+& are described in this paper. We demonstrate that the glass-adhering phenotype of M. pneumoniae, strain M129, strongly adheres to HA$s compared to the glass nonadhering Ml29 phenotype. The binding was enhanced in the presence of human serum. Based on inhibition studies, we show that sialic acid residues and sulfated components contribute to the recognition and attachment of these bacteria to HA&.

2. Materials

and methods

2.1. Mycoplasma

and growth conditions

The study was conducted on the wild-type, pathogenic and cytoadherence-positive M. pneumoniae strain Ml29 in its early (20th) passage. Organisms were grown in Roux bottles containing 70 ml of Channock medium and the bottles were incubated horizontally at 37°C for 72 h until the phenol red indicator changed to light yellow. At any given time, part of the organisms grown in microcolonies adhering to the glass surface (glass-adhering ‘phenotype’), while others are grown in the medium (nonadhering ‘phenotype’) [ 111. The medium with the nonadhering bacteria was transferred to centrifuge tubes. The glass-adhering bacteria were scraped off with a rubber policeman into 5 ml of 0.25 M NaCl. Separated bacterial populations were harvested by centrifugation at 18 OOOxg for 20 min at 4°C. The pellet was resuspended in 10 ml of 0.25 M NaCl and then washed twice with 0.25 M NaCl as above. Protein content of the M. pneumoniae suspension was determined by the method of Bradford [ 121 using bovine serum albumin (BSA, Sigma Chemicals Co.) as a standard. The viable count assay showed that 1 pg of M. pneumoniae represents about lo7 cfu. 2.2. Isolation of human alveolar macrophages Human alveolar macrophages were obtained from bronchoalveolar lavage from patients undergoing routine bronchoscopy for evaluation of various dis-

eases, mostly patients undergoing evaluation for lung malignancy or interstitial lung disease. All patients signed informed consent for the procedure, which was not different from a standard procedure. Patients with purulent or bloody secretion were excluded. Bronchoalveolar lavage was obtained by a method described elsewhere [ 131. Briefly, a fiber-optic bronchoscope (Pentax 15) was introduced through the nostril and passed through the vocal cord under local anesthesia. After complete inspection of the tracheobronchial tree, the tip of the bronchoscope was wedged into the orifice of the right middle lobe or lingua bronchus. Three aliquots of 50 ml normal saline were instilled through the working channel of the bronchoscope and aspirated into a collecting tube. Samples were sent to microbiology and cytology laboratories and 30-40 ml were used for this study. The samples were centrifuged at 1000 rpm for 5 min at 4°C to sediment the macrophages. Cells were washed twice with phosphate-buffered saline (PBS; 0.15 M NaCl, 0.02 M phosphate, pH 7.2) and their number was evaluated microscopically using a hemacytometer. Over 95% of the cells were mononuclear phagocytes. 2.3. Determination by ELBA of M. pneumoniae tachment to human alveolar macrophages

at-

The attachment of the bacteria to phagocytic cells was determined by a modification of an ELISA method described previously [14,15]. Monolayers of phagocytes (2-5 X lo4 cells/well) were prepared by allowing the cells to adhere to the bottom of the wells of 96-well, flat bottom microtiter plate (Dynatech M129B) for 1 h at 37’C. The wells were washed twice with PBS and 5 X 107-8 X lo8 cfu of M. pneumoniae in PBS + 1% BSA were added. After 30 min at 4°C the monolayers were washed three times with PBS to remove unattached bacteria. To evaluate the extent of the bacterial attachment to the phagocytes, 100 p,l of specific antibacterial antiserum diluted 1:lOO in PBS + 10% horse serum (HS) was added to each well. After incubation for 1 h at 37°C the monolayers were washed five times with PBS and 100 p,l of alkaline-phosphatase labeled goat-anti-rabbit immunoglobulin G diluted 1: 1000 in PBS + 10% HS were added to each well and incubated for 1 h at 37°C. After five washes with PBS, 100 pl of substrate solution (p-nitrophenyl phos-

A. Athamnael al,/ FEMS Immunologyand MedicalMicrobiology15 (1996) 135-141

phate, Sigma Chemical Co.> was added to each well. The color was allowed to develop for 15-30 min and the absorbance at 405 nm was determined by ELISA plate reader (SLT-Labinstruments, Austria). Controls consisted of wells without phagocytes or wells with macrophages but lacking bacteria to insure that the macrophages did not cross-react with the first antibody and to determine the nonspecific binding of antibodies. For inhibition studies, 50 p.1 of the indicated concentrations of inhibitors in PBS was added to wells coated with HA&s followed by addition of 50 ~1 of 2 X lo* cfu of M. pneumoniae. These were incubated for 30 min at 37°C. The extent of inhibition was determined as percent attachment of the bacteria in the presence of the inhibitor relative to the control. 2.4. Opsonization

of M. pneumoniae

A suspension of bacteria (100 ~1 of lo9 cfu/ml> in PBS containing O.Ol% MgCl, and 0.01% CaCl, (PBS-MgCa) was incubated at 37°C for 30 min with 100 kl/ml of undiluted normal human serum. The opsonized bacteria were washed by centrifugation as above. Bacteria were suspended in 200 ~1 of PBSMgCa and the opsonized bacteria were immediately tested for their ability to attach to HA$s as described above [ 161. 2.5. Electron microscopy Samples for scanning electron microscopy (SEM) were prepared by incubating 1 ml of 5 X lo4 HA$s with the glass-adhering grown M. pneumoniae (lo* cfu) for 30 min at 4°C. Unattached bacteria were removed by washing with PBS by centrifugation at 500 rpm for 5 min at 4.“C. The specimens of HA+s with adhering mycoplasmas were prepared for the SEM as described prevllously for red blood cells [ 171 using a JOEL 35 SEM equipped with a LaB, gun at an accelerating voltage of 35 kV.

:

137

/t--f

:: 5 kl 5

0.4

0.2

o.oI/

0

20

Bacteria

40 added

60

80

(x107/well)

Fig. 1. Attachment of glass-adhering M. pneumoniae to human alveolar macrophages. Glass-adhering (0) and nonadhering (W) phenotypes of M. pneumoniae in various concentrations in a serum-free system were added to human alveolar macrophage monolayers onto microtiter plate wells for 30 min at 4°C. Non-attached bacteria were removed by washing the wells three times with PBS. The bacterial attachment values were determined by ELISA as described in Section 2, Materials and methods. Data are means of triplicates; vertical lines show the standard deviation.

moniae could be recognized and attached to HA+s monolayer without mediators such as opsonins. Attachment of M. pneumoniae to HA@ occurred in the absence of opsonins. Data revealed a linear increase in M. pneumoniae binding in the range of 5 X 107-4 X 10’ cfu per well. Attachment was saturated at a bacterial concentration above 4 X 10’ cfu per well (Fig. 1). 3.2. Scanning electron microscopy Attachment of the glass-adhering grown M. pneumoniae strain Ml29 to HA@ was also studied by SEM. Samples lacking opsonins revealed mycoplasma-like structures adhering to the HA$s (Fig. 2). Qualitative observations indicated that individual HA& exhibited varying number of mycoplasmas per cell. These results confirm those obtained by ELISA showing that Ml29 attached to HA$s in the absence of opsonins.

3. Results

3.1. Attachment of glass-adhering moniae Ml29

grown M. pneu-

to HA+s

We added the bacte:ria to HA+s monolayer in a serum free-medium to investigate whether M. pneu-

3.3. Attachment Ml29 to HA&

of nonadhering

M. pneumoniae

Attachment of the nonadhering M. pneumoniae Ml29 phenotype to HA+ was examined to compare

138

A. Athamna et al./ FEMS Immunology

and Medical Microbiology 15 (1996) 135-141

its ability to attach to phagocytes with the glass-adhering grown Ml29 phenotype. The attachment of the nonadhering M. pneumoniae Ml29 phenotype to HA@ was very low compared to the other phenotype (Fig. 3). Moreover, the attachment of the nonadhering grown bacteria to HA$s did not increase in response to increasing of the bacterial concentration added to the macrophage monolayers (Fig. 1). Bacterial treatment with human serum was done to compare the attachment of the bacteria to the macrophages in the presence or absence of opsonins. The treatment of the bacteria with normal human Fig. 3. Effect of human serum on the attachment of M. pneumoniae to human alveolar macrophages. Human alveolar macrophage monolayers were exposed to 2 x lo8 cfu of nonglass-adhering (A) or glass-adhering (C) phenotypes of M. pneumoniae in the absence (A, C) or presence (B, D) of human serum. After 30 min at 4”C, the non-attached bacteria were removed by washing the wells three times with PBS. The bacterial attachment values were determined by ELISA as described in Section 2, Materials and methods. Data are means of triplicates; vertical lines show the standard deviation.

serum enhances the attachment of both phenotypes, although the effect is more pronounced with the non-glass adhering phenotype (Fig. 3). 3.4. Inhibition HA&

Fig. 2. Scanning electron microscopy of human alveolar macrophages interacted with M. pneumoniae. A human alveolar macrophage%. pneumoniae suspension was incubated for 30 min at 4”C, followed by differential centrifugation at 500 rpm for 5 min at 4°C to remove nonattached bacteria. (a) Scanning electron micrographs showing a cluster of macrophages incubated with M. pneumoniae for 30 min at 4°C. (b) High magnification revealed the macrophage to be associated with M. pneumoniae-like structures (arrows). Note the lamillipoda extend from the cell surface (arrowheads). These ceils exhibit characteristic features followed by M. pneumoniae interaction. The SEM preparation was done as described in Section 2, Materials and methods.

of M. pneumoniae

attachment

to

It is known that the cells in the lung and tracheae are rich in surface sulfated glycolipids and sialylated glycoconjugates [ 181 and that M. pneumoniae recognize and attach to these ligands [ 181. We examined the effect of dextran sulfate and polyanetholsulfonic acid on the attachment of the bacteria to HA$s to see whether such ligands are present on HA+s. Dextran sulfate and polyanetholsulfonic acid inhibited the attachment of M. pneumoniae efficiently, but not completely (Fig. 4A). The inhibition was more effective with the increase of the dextran sulfate and polyanetholsulfonic acid concentration in the reaction mixture. This result also indicates that HA$s possess sulfated residues that are recognized by these bacteria and may act as receptor for M. pneumoniae attachment. It was shown that sialic acid residues of sialoglycoconjugates act as ligands to

A. Athamna et al./FEMS

0.0

1.5

1.0

0.6

lnhlbltor

Immunology and Medical Microbiology

2.0

(pglwell)

15 (1996) 13.5-141

139

These results suggest the role of sialic acid residues in the attachment and support the notion that sialic acid residues may act as ligand for the attachment of these bacteria to HA$s, as shown in previous studies on other target cells [ 181. To examine the possibility of involvement of monosaccharides present on the bacteria or on macrophage surfaces in the interaction between the cells, we tested their efficacy as potential inhibitors of M. pneumoniae adherence to the HA+. The attachment of M. pneumoniue strain Ml29 to HA+s in the presence of up to 200 rnh4 of cx-methyl-mannoside, o-methylglycoside and Nacetyl-glucosamine was not affected (data not shown), suggesting that these carbohydrates are most probably not involved as ligands of the adherence process. 4. Discussion

0

20

40 lnhlbltor

60

20

100

(pg/well)

Fig. 4. Inhibition of M. pneumoniae attachment to human alveolar macrophages. Bacterial suspensions in the presence of various concentrations of (A) dextran sulfate, + ; polyanetholsulfonic acid, n ; or dextran, 0 ; or (B) glycoconjugates (a-1 acid glycoprotein, 0 ; fetuin, 0; and .asialo fetuin, n , were added to human alveolar macrophage monolayers for 30 min at 4°C. The non-attached bacteria were removed by washing the wells three times with PBS. The bacterial attachment values were determined by ELISA as described in Section 2, Materials and methods. Percent inhibition of attachment at the indicated concentrations was calculated concerning control attachment detected in monolayers lacking inhibitors. Data are means of triplicates; vertical lines show the standard deviation.

adherence [ 11,181. Fig. 4B shows that sialic acid containing compounds such as fetuin and c-w1acid glycoprotein markedly reduce the attachment of the bacteria to HA+ while asialofetuin had very little effect on the attachment.

Studies on M, pneumoniae interaction with macrophages were previously conducted employing rodent macrophage. This was mainly due to the fact that alveolar macrophages were difficult to get, especially human alveolar macrophages. Despite this difficulty, we here describe the principles of the interaction of the latter with M. pneumoniae. In the present study, we examined the interaction between two phenotypes of M. pneumoniae with HA@. We have shown that M. pneumoniae strain Ml29 glass-adhering phenotype attached to HA@ in a dose responsive manner, while the nonadhering M. pneumoniae phenotype attach weakly to the HA@. It should be noted that this variability in attachment among the two phenotypes may be due to variable surface adhesins in mycoplasma (phase variation). We suggest that M. pneumoniae, as in other mycoplasma species, e.g. M. gallisepticum, [l l] M. hyorhinis [ 191, M. hyopneumoniae [20] and M. fermentuns [21] possess the capability to alter their membrane surface and enable the bacteria to maintain different levels of population diversity needed for survival by evasion of the host defense mechanisms such as the HA+ On the other hand, the differences in the attachment between the two phenotypes to the HA+ may be due to inaccessibility of the adhesins. Pretreatment of the bacteria with human serum enhanced the bacterial attachment to HA@. This

140

A. Athamna et ai./ FEMS Immunology and Medical Microbiology 15 (1996) 135-141

defense mechanism may occur in organ sites with high local opsonin concentration, but is unlikely to occur in certain anatomic sites such as the lung [22,23] or the renal medulla [24,25] where the local opsonin concentration is low. It may be insufficient for effective opsonization of many bacterial species

H41. The results of this study may explain the host defense mechanism in which M. pneumaniae adhere to HA+s in the lung, since the attachment of the bacteria to HA$s also occurs without serum indicating that surface components are exposed on the bacteria or phagocytic cells involved in recognition and attachment. We performed inhibition tests to understand the molecular recognition between the cells. Inhibition experiments, using the sulfated polysaccharide dextran as inhibitor and dextran as control, revealed that the attachment of the bacteria to HA$s was markedly decreased. No inhibition was found with dextran, suggesting that sulfated or negatively charged components may act as receptors on the HA+s surfaces. Moreover, the inhibition of the attachment of the bacteria to HA@ obtained by sialic acid containing compounds revealed that sialic acid residues may act as recognition determinants exposed on the phagocytic cells. The notion that sialic acid residues involved in attachment is supported by previous studies reporting that M. pneumoniae recognize and attach to target cells containing sialic acid residues [ 181. The role of sialic acid residues and sulfate in the interaction between bacteria and phagocytic cells was shown previously [26,27], strengthening our findings. Since host cell surface receptors for many pathogenic bacteria are usually carbohydrate, we eliminated the possibility that other lectins may mediate the interaction between M. pneumoniae and HA@ by testing the effect of monosaccharides in various concentrations on the attachment of the bacteria. It was found that no inhibition effect was observed. This result suggests that various lectins specific to o-methyl-mannoside, o-methyl-glucoside, or to N-acetyl-glucosamine are not involved in the interaction between M. pneumoniae and the alveolar macrophages. The details of the recognition site on both M. pneumoniae and the receptors of the macrophages warrant further studies.

Acknowledgements We thank Dr. E. Rachamim for his skilled assistance in the studying of the samples by SEM. This study was supported by grants from the Golda Meir Foundation and the Deutsche Forschungsgemeinschaft.

References Howard, C.J. and Taylor, G. (1986) Humoral and cell-mediated immunity. In: The Mycoplasmas (Razin, S. and Barile, M.F., Eds.) pp. 259-292. vol. IV. Academic Press, San Diego, London. 121Miihlradt. P.F. and Schade, U. (1991) MDHM, a macrophage-stimulatory product of Mycoplasma fermenrans leads to in vitro interleukin-1 (IL-l), IL-6, tumor necrosis factor, and prostaglandin production and is pyrogenic in rabbits. Infect. Immun. 59, 3969-3974. [31 Gallily, R., Avron, A., Jahns-Streubel, G. and Miihlradt. P.F. (1995) Activation of macrophages and monocytes by mycoplasmas, In: Molecular and Diagnostic Procedures in Mycoplasmology (Razin, S. and Tully, J.G., Eds.) pp. 421-438, vol. I. Academic Press, San Diego, London. 141Bredt. W. (1975) Phagocytosis by macrophages of mycoplasma pneumoniae after opsonization by complement. Infect. Immun. 12, 694-695. [51 Cole, B.C. and Ward, J.R. (1973) Interaction of Mycoplasma arthritidis and other mycoplasmas with murine peritoneal macrophages. Infect. Immun. 7, 681-689. [61 Powell, D.A. and Clyde W.A.. Jr. (1975) Opsonin-reversible resistance of Mycoplasma pneumoniae to in vitro phagocytosis by alveolar macrophages. Infect. Immun. 11, 540550. [71 Davis, J.K., Delozier, K.M., Asa, D.K., Minion, F.C. and Cassell, G.H. (1980) Interaction between murine alveolar macrophages and Mycoplasma pulmonis in vitro. Infect. Immun. 29, 590-599. t81 Collier, A.M. and Clyde. W.A., Jr. (1971) Relationships between Mycoplasma pneumoniae and human respiratory epithelium. Infect. Immun. 3, 694-701. t91 Denny, F.W., Clyde, W.A., Jr. and Glezen, W.P. (1971) Mycoplasma pneumoniae disease: Clinical spectrum, pathophysiology, epidemiology and control. J. Infect. Dis. 123, 74-97. 1101Foy, H., Kenny, G.E., Coony. M.K. and Allan, I.D. (1979) Long-term epidemiology of infections with Mycoplasma pneumoniae. J. Infect. Dis. 139. 248-254. [I11 Lipman, R.D., Clyde, W.A., Jr. and Denny, F.W. (1969) Characteristics of virulent attenuated and avirulent Mycoplasma pneumoniae strains. J. Bacterial. 100, 1037-1043. [121 Bradford, M.M. (1976) A rapid and sensitive method for the quantitiation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. B&hem. 72, 248254.

A. Athamna ef al. / FEMS Immunology and Medical Microbiology 15 (1996) 135-141 [13] Klech, H. and Pohl, W. (1985) Report of the European Society of Pneumology task group on bronchoalveolar lavage: Technical recommendation and guidlines for bronchoalveolar lavage. Eur. Res. J. 2, 561-585. [14] Athamna, A., Ofek, I.., Keisari, Y., Markowitz, S., Dutton, G.G.S. and Sharon, N. (1991) Lectinophagocytosis of encapsulated Klebsiella pneumoniae mediated by surface lectins of guinea pig alveolar rnacrophages and human monocyte-derived macrophages. Infect. Immun. 59, 1673-82. [15] Athamna, A. and Ofek, I. (1988) Enzyme-linked immunosorbent assay for quantitiation of attachment and ingestion stages of bacterial phagocytosis. J. Clin. Microbial. 26, 6266. [16] Blumenstock, E. and Jann, K. (1981) Natural resistance of mice to Salmonella typhimurium: Bacteriocidal activity and chemiluminescence response of murine peritoneal macrophages. J. Gen. Microbial. 125, 173-183. [17] Razin, S., Banai, M., Gamliel, H., Polliack, A., Bredt W. and Kahane, I. (1980) Scanning electron microscopy of mycoplasmas adhering tomred blood cells. Infect. Immun. 30, 538-546. [18] Kahane, I. (1995) Adb-sion of Mycoplasmas. In: Methods in Enzymology: Adhesion of Microbial Pathogens (Doyle, R.J. and Ofek, I., Eds.) pp. 367-373, vol. 253. Academic Press, Inc., Orlando, FL. [19] Yogev, D., Rosengarten, R., Waston-McKown, R. and Wise, K.S. (1991) Molecular basis of mycoplasma surface antigenie variation: a novel set of divergent genes undergo

[20]

1211

[22]

[23]

[24]

[25] 1261

[27]

141

spontaneous mutation of periodic coding regions and 5’ regulatory sequences. EMBO J. 10, 4069-4079. Wise, K.S. and Kim, M.F. (1987) Major membrane surface proteins of Mycoplasma hyopneumoniae selectively modified by covalently bound lipid. J. Bacterial. 169. 5546-5555. Wise, K.S., Kim, M.F., Theiss, P.M. and Lo, S.-C. (1993) A family of variant surface lipoproteins of Mycoplasma fermenmns. Infect. Immun. 61, 3327-3333. Goldstein, E., Lippert, W. and Waschaul, D. (1974) Pulmonary alveolar macrophage defender against bacterial infection of lung. J. Clin. Invest. 54, 519-528. Reynolds, H.Y. and Newball, H.H. (1974) Analysis of protein and respiratory cells obtained from human lungs by bronchial lavage. J. Lab. Clin. Med. 84, 559-573. Silverblatt, F.J., Dreyer, J.S. and Schauer, S. (1979) Effect of pili on susceptibility of Escherichia coli to phagocytosis. Infect. Immun. 24, 218-223. Silverblatt, F.K. and Ofek, I. (1983) Internalization of bacterial pili and leukocytes. Infect. Immun. 11, 235-238. Chmiela, M., Paziak-Domanska, B. and Wadstrijm, T. (1995) Attachment, ingestion and intracellular killing of Hekcobacter pylori by human peripheral blood mononuclear leukocytes and mouse peritoneal inflammatory macrophages. FEMS Immunol. Med. Microbial. 10, 307-316. Chmiela, M., Paziak-Domanska, B., Rudnicka, W. and Wadstrcm, T. (1995) The role of haparan sulphate-binding activity to Helicobacter pylori bacteria in their adhesin to marine macrophages. APMIS 103, 469-474.