37

Biochimica et Biophysica Acta, 1053 (1990) 37-42 Elsevier BBAMCR 12720

The influence of net surface charge on the interaction of uropathogenic Escherichia coli with human neutrophils Robert Steadman, Janice Knowlden, Monika Lichodziejewska and John Williams Institute of Nephrology, Kidney Research Unit Foundation, University of Wales College of Medicine, Cardiff Royal Infirmary, Cardiff, Wales (U.K.) (Received 7 February 1990)

Key words: Neutrophil; Charge interaction; (E. coli)

Escherichia coil strains, grown to suppress fimbrial expression, synthesised enhanced quantities of polysaccharide capsule, which significantly lessened their binding to heparin sepharose columns. In the presence of poly-L-lysine, these strains were strongly retained on the columns confirming their highly anionic nature. Uropathogenic strains of E. coli expressing type 1 fimbrial adhesins activated the respiratory burst, the degranulation response and the release of leukotrienes from human neutrophils (PMN) to a significantly greater extent than the same strains grown in a medium to suppress this fimbrial expression. The addition of the poly-cation poly-IMysine, however, selectively increased neutrophil activation in response to these non-fimbriate strains. This dose-dependent effect was reversed by the addition of heparin suggesting a mechanism dependent on surface charge. The results of this study suggest that non-specific mechanisms involving the neutralisation of surface charge, in addition to specific receptor and adhesin mediated events could affect neutrophil activation at sites of infection.

Introduction The interaction of unopsonized Escherichia coli with human neutrophils (PMN) is mediated through the binding of specific bacterial adhesins (type 1 fimbriae) to mannoside residues on the PMN surface [1]. This binding results in the activation of the PMN respiratory burst [2] and the release of primary, secondary and tertiary lysosomal granule enzymes [3]. The expression of a polysaccharide capsule by these same bacteria, however, inhibits phagocytosis [4] and abolishes the capacity of the organisms to trigger PMN activation [5]. It has therefore been suggested that physicochemical factors such as net negative surface charge or relative hydrophobicity may influence the response of inflammatory cells to E. coli [4,6,7].

Abbreviations: PMN, human neutrophils; PLL, poly(L-lysine), DAPI, 4'6-diamino-2-phenylindole; PBS, phosphate-buffered saline; CL chemiluminescence. Correspondence: J. Williams, Kidney Research Unit Foundation, University of Wales College of Medicine, Cardiff Royal Infirmary, Cardiff, Wales, U.K.

More recently it has been demonstrated that there is no direct correlation between the relative surface hydrophobicity of E. coli strains and PMN activation [8]. Electrostatic forces, however, have been shown to play an important role in the selective uptake of charged particles by phagocytes and in controlling intercellular contact [9,10]. In addition, macrophages internalise cationic particles (but not neutral or negatively charged particles), at 4 ° C, in coated pits, by a process which is independent of the cytoskeleton [11]. The cationic protein poly(L-lysine) (PLL) has previously been used to neutralise cell surface net negative charges resulting in the activation of the cell in several systems [12-14]. Furthermore, poly-cation neutralisation of membrane charge has been shown to augment the phagocytosis of gram-positive organisms [15]. In contrast the same study also demonstrated that the phagocytic response to strains of E. coli was affected to a lesser extent by poly-cation pre-treatment. There was no detailed surface characterisation, however, of the E. coli used. Gramnegative organisms, such as E. coli, may express a complex variety of surface structures. The present study was undertaken in order to investigate, in more detail, the influence of charge interactions on the PMN re-

0167-4889/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

38 sponse to uropathogenic isolates of E. coli, cultured to control the expression of particular surface characteristics. Materials and Methods

Bacterial strains. The E. coli strains used in this study were stored on Dorsett Egg slopes (Oxoid Ltd., Basingstoke, U.K.) and subcultured three times overnight at 37 °C without shaking in either Oxoid Nutrient Broth 2 (NB2) (to enhance fimbrial expression) or modified Vogel Bonner's medium (MVBM) (to enhance capsule synthesis) [16] before harvesting by centrifugation at 2000 × g for 15 min at 4°C. The bacterial cell pellet was suspended and washed two times in phosphate-buffered saline (pH 7.3, PBS) before being resuspended to an absorbance at 560 nm of 2.0 in PBS (1-109 colony forming units (cfu)/ml). The surface characteristics of each strain following subculture in each medium are shown in Table I. Bacterial surface characteristics. The strains were tested for their mannose sensitive agglutination of erythrocytes before each experiment. 100/~1 of bacteria (1 • 108 cfu/ml) were mixed with 100 /~1 of either 3% (v/v) guinea-pig, sheep or human erythrocytes prepared from citrate& peripheral blood and washed × 2 in PBS in the presence or absence of 2.5% (v/v) o-mannose in multi-well plates and gently agitated for 2 min. Agglutination was estimated visually and scored - , +, + + or + + +. The strains were also tested for their P-fimbriation using a specific latex agglutination test kit (PF test, Bradsure Biologicals, Loughborough, U.K.). All the strains capable of mannose resistant haemagglutination of human RBC (MRHA) were P-fimbriate. Fimbriation and encapsulation were confirmed by

TABLE I

E. coli s t r a i n

504 SC DF AB GG

Ks71a 49 NK 56

B

O serotype

K antigen *

Fimbriation b NB2

MVBM d

U 1 U

+ +" -

-

U

-

-

U U

+ +

-

8

U

+

-

4

U

+

-

4

U

+

-

6 1 11 NT 4 6

a

-

-

a Smooth but not 01,02,04,05,06,07,08,09,011,015,017,018,025,075. Fimbriation was assessed using electron microscopy, haemaggl u t i n a t i o n a n d specific i m m u n o f l u o r e s c e n c e . S t r a i n SC e x p r e s s e d o n l y P f i m b r i a e . d S t r a i n s g r o w n in M V B M e x p r e s s e d e x t e n s i v e c a p s u l a r m a t e r i a l . S t r a i n s w e r e tested a g a i n s t p h a g e s f o r K I a n d K 5 . U is n o t K I o r K5.

electron microscopy of the pellets of each organism resuspended in 2% (w/v) phosphotungstic acid, containing 0.2% (w/v) BSA, at pH 7.0 and dried onto formvar/carbon-coated films on copper grids (Emscope Laboratories Ltd., Ashford, U.K.). Indirect immunofluorescence assay. Specific fimbriation was confirmed by indirect immunofluorescence. The rabbit anti-P fimbriae antibody (a kind gift of Professor T. Korhonen, University of Helsinki, Finland) and the rabbit anti-type 1 fimbrial antibody (a kind gift of Dr. S. Parry, Unilever, Sharnbrook, U.K.) have been described previously [17,18]. Bacterial suspensions (10 FI) were dried on plastic tissue culture plates (Nunclon, Gibco Ltd., Paisley, U.K.) at room temperature, fixed with cold 3.5% (w/v) paraformaldehyde (BDH Ltd., Poole, U.K.) in PBS (pH 7.1) for 15 min and washed 3 times with PBS (pH 7.3) for 5 min. The plates were incubated at room temperature for 15 min with the appropriate antiserum at optimal dilution, washed 3 times for 5 min in PBS (pH 7.3) and incubated for 15 min at room temperature with a 1:50 dilution of FITC-conjugated goat-anti-rabbit IgG (ICN Biomedicals Ltd., High Wycombe, U.K.) in PBS (pH 7.3). The plates were again washed 3 times in PBS for 15 min and then stained with 4'6-diamidino-2-phenylindole (DAPI) (Sigma, Poole, U.K.) (diluted 1 : 100 in PBS, pH 7.3) for 8 min in the dark at room temperature. DAPI staining is specific for DNA and thus allows the visualisation of every bacterial cell [18]. After rinsing in PBS (pH 7.3) the plates were mounted in one drop of PBS (pH 7.3) containing buffered glycerol (pH 7.3) (BDH) (9 : 1, v/v), and examined at two different wavelengths. Specific FITC, yellow-green staining was observed at an excitation wavelength of 450-490 nm and for DAPI, blue staining at 340-380 nm. Heparin-Sepharose chromatography. A slurry of heparin sepharose CL-4B (Pharmacia Ltd., Milton Keynes, U.K.) in 50% ethanol was degassed and packed in 10 cm Econo-columns (Bio-Rad Ltd., Watford, U.K.) to the depth of 15 mm. The gels were equilibrated with 15 ml of 10 mM phosphate buffer (pH 6.8) containing 0.25 M ammonium sulphate (ASP buffer) and stored at 4 ° C until required. Bacterial suspensions in PBS (pH 7.3) were adjusted to an absorbance of 1.0 and samples of 100 #1 were diluted with 900 FI ASP buffer. 500/~1 of this suspension was loaded onto each column and eluted with 2 ml of ASP buffer. The remaining 500 ~1 was diluted with 2 ml ASP buffer and served as a control. ATP was extracted from the bacteria in the eluate and control tubes to quantify bacterial numbers. 100 pl of bacterial suspension was mixed for 10 s with 100 FI of a bacterial extraction reagent prepared in this laboratory. (Bac-Extract; British Patent Application No. 851300) and then diluted in 800 #1 of 25 mM Hepes buffer (pH 7.5) containing 2 mM EDTA. 100/~1 of this dilution was added to the firefly bioluminescence re-

39 agent (ATP bioluminescence reagent, Boehringer Corporation (London) Ltd., Lewes, East Sussex, U.K.) and the luminescence measured in a Lumac Biocounter M2010 (Lumac BV, Landgraaf, The Netherlands) over a 10 s integration period, light emission being recorded as relative light units (rlu) (1 rlu being equivalent to 10 photon counts per second (cps)). Neutrophil preparation. Normal human leukocytes were isolated from citrated peripheral blood by dextran sedimentation in polypropylene tubes and rendered plasma-free and platelet-poor by washing three times with PBS. Neutrophils (PMN) were purified by density gradient centrifugation at 400 × g for 35 min at 23 ° C on cushions of Ficoll-Hypague (Pharmacia, Milton Keynes, U.K.). Following hypotonic lysis of erythrocytes in 0.2% (w/v) NaC1 and washing in PBS, the PMN were counted in a Coulter counter ZM (Coulter Electronics Ltd., Luton, U.K.) and resuspended in PBS to a concentration of 5 • 106 cells/ml. Microscopic examination of Neat stained (Guest Medical Ltd., Sevenoaks, U.K.) Cytospin II (Shandon Southern Products, Runcorn, U.K.) preparations revealed that the cells were more than 98% PMN. Neutrophil chemiluminescence. 100/tl of PBS (pH 7.3) containing 5 • l0 s PMN was incubated with 100 #1 of 10 /xM luminol (5-amino-2,3-dihydro-l,4-phthalazinedione) and 200 /tl of a Krebs-Ringer phosphate buffer containing 11 mM D-glucose, 0.5 mM Ca 2÷, and 1.2 mM Mg 2÷ (KRPG) at 37°C for 6 min, prior to addition of 100 #1 of each bacterial suspension (1.108 cfu), in order to equilibrate temperature and record background chemiluminescence (CL). In some experiments PMN were pre-incubated with PLL (Sigma Chemical Co., Poole, U.K.) at concentrations up to 200 # g / m l before the addition of bacteria. In further experiments, PMN were pre-incubated with PLL in the presence of Heparin (CP Pharmaceuticals, Wrexham, U.K.), at concentrations up to 1000 U / m l before the addition of the E. coli strains. At precise 2 min intervals following addition of E. coli, CL readings were taken in the Lumac Biocounter and peak levels were compared to those of unstimulated incubations.

Neutrophil degranulation and leukotriene B4 release. 100 #1 volumes of KRPG containing 1 • 106 PMN were diluted with 800 #1 KRPG and pre-incubated for 5 min at 37 °C before the addition of stimulus in KRPG (100 ttl). Estimates were carried out in duplicate with 100 #1 of each stimulus at the appropriate concentration for periods up to 60 min, separated by centrifugation for 1 rain at 11 000 × g and the supernatant removed for the enzyme and leukotriene B4 assays. PMN incubated in KRPG for periods up to 60 min without stimulation, bacteria incubated in KRPG without PMN and KRPG alone without PMN or bacteria were included as controis in each experiment. In addition, 5 /~M calcium ionophore A23187 (Cambridge Bioscience, Cambridge,

U.K.) in KRPG was used as a positive control to assess PMN activation. Myeloperoxidase (MPO). 100-/~1 duplicates of supernatant from each stimulation were incubated at 30 °C with 1 ml of substrate consisting of 0.3 mM o-dianisidine, 0.03% v / v HzO z and 0.05% v / v Triton X-100 in 0.1 M citric acid (pH 5.5). The reaction was stopped after 5 min with 1 ml of 3.25 M perchloric acid, and the absorbance at 560 nm was measured. Vitamin B12 binding protein (B12BP). 100 ~1 duplicates of supernatant and 250 ~1 of [57Co]cyanocablamin (4.44 ng/ml in water, 0.067 ttCi/ml) were incubated at room temperature for 30 rnin. 1 ml of activated charcoal (5% w/v) coated with 1% (w/v) BSA was added and left at room temperature for 10 min before centrifugation for 10 rain at 2500 × g to precipitate charcoal-adsorbed, non-protein-bound vitamin B12. 1 ml of supernatant was counted for 1 min in a Kontron gamma counter. The percentage release of enzymes from PMN was calculated after subtraction of the appropriate blank values, as a percentage of that released from cells disrupted by sonication, for two periods of 1 min at 8 ~m peak to peak distance at 4°C in an MSE 150 W ultrasonic disintegrator (MSE Ltd., Crawley, U.K.). Leukotriene B4 (LTB4). The radioimmunoassay (RIA) for LTB4 is a modified version of that described previously [19]. Briefly, 100/~1 of supernatant from each cell incubation was assayed in duplicate by incubating at 37°C overnight with 100 /~1 of antiserum at a final dilution of 1/12000 in RIA buffer (10 mM Tris-HC1 containing 0.9% w / v NaC1, 0.01% (w/v) sodium azide and 0.1% (w/v) gelatin, pH 7.3) and with 100 ~1 (5000 cpm) of tracer (5,6,8,9,11,12,14,15-[(n)-3H]LTB4 (200 Ci/mmol) (Amersham International plc, Aylesbury, U.K.). Standard dilutions of LTB4 from 0.04 ng/ml to 10 ng/ml (4 pg to 1000 pg per tube) were included in each assay. Free LTB4 was separated from antibodybound by adsorption to dextran-coated charcoal (1% w / v dextran T70 (Pharmacia), 1% w / v GSX charcoal) at 4 ° C and centrifugation at 3500 × g for 10 rain. The supernatant was decanted into 4 ml Optiphase MP (LKB) and counted in an LKB rackbeta scintillation counter. Samples of supernatants or cell extracts (500 ~1) containing immunoreactive LTB4 identified in the RIA were injected onto a Nucleosil C18 5/x reversed phase column (25.4 cm x 4.6 mm) (Hichrom Ltd., Reading, U.K.). The samples were eluted for 30 min in a methanol/water/acetic acid (65:35:0.1) mobile phase brought to pH 5.6 with ammonium hydroxide. The flow rate was maintained at 1 ml/min throughout each run, and 1 ml fractions were collected. Each fraction was stored under argon at - 2 0 ° C until analysed. Ultraviolet absorption was monitored throughout each run at 270 nm and peak absorption was integrated on a Gilson

40 620 DataMaster (Anachem Ltd., Luton, U.K.). Duplicate samples of 100/~1 of eluent were dried down under vacuum and resuspended in RIA buffer. The elution time of PMN generated immunoreactive LTB4 and integrated UV absorption were compared to those of authentic LTB4 (a kind gift of Dr. B. Spur, Institute Henri Balfour, Pards) and [3H]LTB4 eluted in the same system.

40000

b

30000

30OO

E

20000

~E 2ooo

6

~oo~o

6

0

Neutrophil responses to E. coil strains Six of the E. coil strains grown in NB2 expressed type 1 fimbriae (Table I) and generated significant PMN chemiluminescence (CL) in three separate experiments (1.96 • 104 + 1.53 • 104 rlu; mean + S.D.). In the same experiments, all nine strains grown in MVBM and the three strains which did not express type 1 fimbriae when grown in NB2, failed to trigger PMN CL above background (1.13.102 + 3.45.102 rlu compared to 0.67 • 102 + 2.13 • 102 rlu; mean + S.D.). Four of the strains were used to investigate the release of lysosomal enzymes and LTB4 from PMN. Strain 504 (NB2), expressing type 1 fimbriae, released significant MPO and B12 BP(18.2 + 3.6% and 76.2 + 12.1%, respectively; mean + S.D.), following incubation with PMN at an optimum time of 60 min and a bacteria/cell ratio of 100/1. While MPO release in response to 504 (MVBM) was not significantly raised above background and B12 BP release was only 20.9 + 7.2% (mean + S.D.) (Fig. 1). In contrast, strains SC (P-fimbriate only) and D F and AB (non-fimbriate) grown in NB2 or MVBM released significant amounts of the secondary granule marker B12 BP, without releasing primary granule MPO. Optimum release at 60 min and a bacterial/cell ratio of 100/1 was 36.3 + 19.1% for NB2 grown strains and 37.1 +22.9% for strains grown in MVBM; (mean + S.D.). MPO and B12

100

30

~

8o

§

4o

~

2o

20

2

g -~

10

g. •

PMN

•

NB

MVBM

•

"ooo

0

%ENE

Results

4000

PLL PLL~HE~

NCNE

PLL PLL+HEP

Fig. 2. Peak PMN chemiluminescence following incubation at a bacteria/PMN ratio of 100/1 at 37°C in KRPG with individual E. coil strains (a) expressing (n = 6) or (b) not expressing (n = 12) type 1 fimbriae. PMN were preincubated for 6 min in KRPG alone (NONE) or containing 20 p.g/ml PLL with or without 200 units of heparin (HEP) before the addition of each individual strain. Results are the mean_+S.D. for all the strains in three experiments each using PMN from a different donor.

BP release in response to 5 #M A23187 at 60 min was 31.2 + 10.3% and 79.0 + 18.4%, respectively. Strain 504 (NB2) caused significant LTB4 release reaching a plateau of 0.33 + 0.07 ng/1 • 106 PMN at a bacteria/cell ratio of 100/1 by 60 min (mean + S.D., n = 3) in experiments where the ionophore-stimulated release of LTB4 was 10.1 + 2.6 /~g/1-106 PMN. In contrast, non-fimbriate 504 grown in MVBM stimulated only 0.17 + 0.03 ng LTB4/1 • 106 PMN (mean + S.D.) at the same dose and time in the same experiments. P fimbriate strain SC and the nonfimbriate strains DF and AB whether grown in NB2 or MVBM did not stimulate LTB4 release above background levels in these experiments.

E. coil binding to heparin sepharose (HS) E. coli strains grown, to enhance fimbrial expression in nutrient broth 2 (NB2), bound with moderate affinity to HS columns (31.6 + 16.3%; mean + S.D.) loaded and eluted in PBS (pH 6.8) in three separate experiments. In contrast, the same strains grown in MVBM, to inhibit fimbrial expression and enhance capsule expression, bound significantly less well to the columns in each experiment (22.3 _+ 7.6%; mean + S.D.) (t = 2.20, P < 0.05; Student's paired t-test). The binding of NB2 grown organisms was significantly enhanced in the presence of 20 /~g/ml PLL (74.5 + 16.8%; mean + S.D.) (t = 5.84, P < 0.001; Student's paired t-test). Similarly the binding of MVBM grown strains was significantly increased in the presence of 20 /xg/ml PLL to 80.0 +_ 10.4% (mean _+ S.D.) (t = 15.76, P < 0.001; Student's paired t-test).

0

PMN

NB

MVBM

Fig. 1. MPO (a) and B12 BP (b) release from 1.106 PMN incubated at 37°C in KRPG for 60 rnin either alone (PMN) or with 1•108cfu of E. coil 504, subcultured in either nutrient broth (NB) or modified Vogel Bonner's medium (MVBM), to maximally express type 1 fimbriae or capsule respectively. Results are the mean+ S.D. for three experiments each using PMN from a different donor.

Poly(L-lysine) effects on P M N responses There was a dose dependent inhibition of the PMN CL response to all of the type 1 fibriate strains of E. coil by doses of PLL up to 200 # g / m i . At 20 # g / m l PLL this inhibition reached a maximum of 59.7 + 21.8% (mean + S.D.) in three experiments. In contrast, there

41 b

a 2.5

[]

" PLL

2.o~

%.

1.5

~

2

[]

- PLL

1

1.0

?, 0.5

"~

0.0

0 PMN

504

~ STRAINS

AB

PMN

504

~

AB

DF

STRAINS

Fig. 3. LTB4 release from 1 • 106 PMN incubated at 37 ° C for 60 min in KRPG alone or containing 20/~g/ml PLL with each of four E. coil strains subcultured either in NB 2 (a) or MVBM (b), Results are the mean + S.D. for three experiments each using PMN from a different donor.

was a mean 3-fold augmentation of the PMN CL response to non-type 1 fimbriate strains at this PLL dose (Fig. 2). The inhibitory and stimulatory effects of PLL were reversed in a dose-dependent manner by the poly-anion, heparin. A dose of 200 units/mi of heparin neutralised the effect of 20 ~g/ml PLL (Fig. 2). There was no significant effect of heparin alone (up to 1000 units/ml) on PMN CL generation in response to any of the strains tested. PMN LTB4 release was greatly augmented in response to all the strains tested in the presence of 20 /~g/ml PLL (Fig. 3). Similarly 20 # g / m l PLL significantly augmented the release of MPO stimulated by all of the strains tested, except strains 504 (NB2) (Fig. 4). The release of B12 BP in response to the same strains, however, was raised to a lesser degree in the presence of PLL from a mean of 46.3 + 25.3% (S.D.) to 59.3 5: 31.1% (S.D.) for NB grown strains and a mean of 33.1 :t: 20.4% (S.D.) to 49.4 + 11.5% (S.D.) for MVBM grown strains.

sponses to E. coil and the potential importance of the presence of cationic molecules in modifying these responses. The cationic protein PLL enhanced the PMN response to all strains which did not express type 1 fimbrial adhesins including a strain expressing only P-fimbriae. In contrast, responses dependent on type 1 fimbrial expression were inhibited by the same doses of PLL. This effect of PLL was observed for all of the strains used, suggesting that it was independent of the differences in outer membrane or capsule composition associated with different O or K serotypes. The effect of PLL on PMN CL was reversed by the poly-anion heparin, suggesting that the effect was solely related to changes in the net surface charge, rather than to specific irreversible changes in the PMN membrane. Extensive work in animal models has shown that the negatively charged mucin surface of the intact uroepithelium plays a major role in preventing the attachment and subsequent colonisation of the urinary tract by microorganisms [20-22]. The results of the heparin sepharose chromatography of the strains used in the present study confirm the importance of surface charge in controlling the interaction of microorganisms with solid matrices possessing a net negative charge. They further suggest, although reduced fimbrial expres-

Discussion

The present study demonstrates the significance of a negative surface charge layer in limiting the PMN reb [] -PLL

[] -PLL

30

20

20 '°

0

0 PMN

504

~C~ STRAINS

AB

DF

PMN

504

~ STRAINS

AB

DF

Fig. 4. Myeloperoxidase (MPO) release from 1.106 PMN incubated at 3 7 ° C for 60 min in K R P G alone or containing 2 0 / ~ g / m l PLL with each of 4 E. coli strains subcultured either in NB 2 (a) or MVBM (b). Results are the mean + S.D. for three experiments each using PMN from a different donor.

42 sion (favoured by growth in MVBM) may result in reduced binding to the anionic columns that, in the presence of cationic PLL, the resistance to binding shown by MVBM grown strains is overcome and all strains will bind equally well to the columns, irrespective of their expression of different surface structures. In animal models of chronic renal infection, neutrophil infiltration and activation is essential for the initiation of tissue damage leading to kidney scar formation [23,24]. The degree of leukocyte infiltration is similar in response to a variety of strains of E. coli, irrespective of their fimbrial expression [25]. Extensive scarring, however, is dependent on the expression of type 1 fimbriae and is probably due to the release of primary granule contents and reactive oxygen metabolites from PMN in the response to these strains [3,26]. The results of the present study suggest that, while type 1 fimbriate E. coli strains will specifically activate PMN primary granule release, a nonspecific activation, dependent on neutralising PMN surface charges, would also result in similar activation in response to organisms which were expressing P-fimbriae or were non-fimbriate. These observations, while probably not confined to uropathogenic organisms, may have important implications for the pathogenesis of urinary tract infections induced by a variety of E. coli strains.

Acknowledgements We are grateful to Professor M. Sussman, Department of Microbiology, University of Newcastle upon Tyne, Professor T. Korhonen, Department of Microbiology, University of Helsinki, Finland and Dr. A. Roberts, Department of Medical Microbiology, West London Hospital for supplying the strains used in this study. Dr. Roberts also serotyped the strains for us. We also thank Miss C. Patterson and Miss R. Carter for their help in preparing this manuscript.

References 1 Blumenstock, E. and Jann, K. (1982) Infect. lmmun. 35, 264-269.

2 Svanborg Eden, C., Bjorksten, L.M., Hull, R., Hull, S., Magnusson, K.E., Leffler, H., Moldavano, Z. (1984) Infect. immun. 4.4, 672-680. 3 Steadman, R., Topley, N., Jenner, D.E., Davies, M. and Williams, J.D. (1988) Infect. Immun. 56, 815-822. 4 Ohman, L., Hed, J. and Stendahl. O. (1982) J. Infect. Dis. 146, 751-757. 5 Steadman, R., Knowlden, J.M. and Williams, J.D. (1989) Biochem. Soc. Trans. 17, 756-757. 6 Stendahl, O., Dahlgren, C., Edebo. M. and Ohman, L. (1981) Monogr. Allergy 17, 12-27. 7 Topley, N., Steadman, R., Mackenzie, R.K., Williams, J.D., Davies, M. and Asscher, A.W. (1989) Rev. Infect. Dis., in press. 8 Steadman, R., Topley, N., Knowlden, J.M., Mackenzie, R.K. and Williams, J.D. (1989) Biochim. Biophys. Acta 1013, 21-27. 9 Papadimitriou, J.M. (1982) J. Pathol. 138, 17-24. 10 Mutsaers, S.E. and Papdimitriou. J.M. (1988) J. Leukocyte Biol. 44, 17-26. 12 Shier, V.T., Dubordieu, D. and Durbin, J.P. (1984) Biochim. Biophys. Acta 793, 238-250. 13 Pugliesi, F., Singh, A.K., Kasinath, B.S. and Lewis, E.J. (1987) Kidney Int. 32, 57-61. 14 Pugliesi, F., Menr, P., Anania, M.C. and Linotti, G.A. (1989) Kidney Int. 35, 817-823. 15 Peterson, P.K., Gekker, G., Shapiro, R., Freiberg, M. and Keane, W.F. (1984) Infect. Immun. 43, 561-566. 16 Lam, J., Chan, R., Lam, K. and Costerton, J.W. (1980) Infect. lmmun. 28, 546-556. 17 Pere, A., Nowicki, B., Saxen, H., Siitonen, A. and Korhonen, T.K. (1987) J. Infect. Dis. 156, 567-573. 18 Lichodziejewska, M., Topley, N., Steadman, R., Mackenzie, R., Verrier Jones, K. and Williams, J.D. (1989) Lancet ii, 1414-1418. 19 Rokach, J., Hayes, E.C., Girard, Y., Lombardo, D.L., Maycock, A.R., Rosenthal, A.S., Young, R.N., Zambani, R. and Zweerink, H.J. (1984) Prostaglandins, Leukotrienes and Medicine 13, 21-25. 20 Parsons, C.L., Mulholland, S.G. and Anwar, H. (1979) Infect. Immun. 24, 552-557. 21 Elliott, T.S.J., Slack, R.C.B. and Bishop, M.C. (1984) Brit. J. Urol. 56, 38-43. 22 Parsons, C.L., Stauffer, C.W. and Schmidt, J.D. (1988) Infect. lmmun. 56, 1341-1343. 23 Glauser, M.P., Lysons, J.M. and Braude, A.I. (1982) J. Clin. Invest. 61,403-407. 24 Slotki, I.N. and Asscher, A.W. (1982) Nephron 30, 262-268. 25 Topley, N., Steadman, R., Mackenzie, R.K., Knowlden, J. and Williams, J.D. (1989) Kid. Int. 36. 609-616. 26 Meylan, P.R., Marbert, M., Bille, J. and Glauser, M.P. (1989) Infect. Immun. 57, 2196-2202. 27 Movat, H.Z., Cybulsky, M.I., Colditz, I.G., Chan, M.K.W. and Dinarello, C.A. (1987) Federation Proc. 46, 97-104.