Journal of General Microbiology (1 987), 133, 89 1-898. Printed in Great Britain 89 1

Interaction of Lactoferrin and Transferrins with the Outer Membrane of Bordetella pertussis

By K E I T H R E D H E A D , * T E R E S A H I L L A N D H E N R I K C H A R T ? Diuision of’ Bacterial Products, National Znstitute for Biological Standards and Control,

Holly Hill, Hampstead, London N W3 6RB, UK

(Receiced 7 July 1986; reuised 16 December 1986)

Bordetellu pertussis was able to grow in citro under conditions where the only iron present was bound to the iron-binding proteins ovotransferrin, transferrin or lactoferrin. Under these conditions the bacteria produced neither hydroxamate nor phenolate-catecholate siderophores to assist in the procurement of iron. Examination of B. pertussis outer-membrane preparations by SDS-PAGE and immunoblotting showed that the iron-binding protein ovotransferrin was bound directly to the bacterial surface. Assays of the binding of radiolabelled transferrin by the bacteria showed that the association was a specific process and that there was turnover of the bound proteins. Competitive binding assays indicated that lactoferrin could be bound in the same way. It is suggested that B. pertussis obtains iron directly from host iron-binding proteins during infection.

I N T R O D U C T I O N

An important factor in the establishment of a bacterial infection is the ability of the pathogen to multiply in the iron-restricted environment found in host tissues. Although the host contains relatively large amounts of iron, the levels of freely available iron are kept very low by the iron- binding proteins transferrin and lactoferrin (Griffiths, 1985). These proteins have a very high association constant for iron such that the amount of free iron in the host tissues is insufficient for bacterial growth (Bullen et al., 1978). Consequently, to multiply successfully under these conditions pathogenic bacteria must possess mechanisms for utilizing the iron associated with host iron-binding proteins.

When grown under conditions of iron restriction in citro many bacteria secrete high-affinity iron-chelators (siderophores) which can remove iron from the iron-binding proteins (Griffiths, 1983). Under iron restriction, these bacteria also express novel outer-membrane proteins (OMPs), some of which are involved in transporting the siderophore-iron complexes into the bacterial cell (Neilands, 1982). It appears that the induction of high-affinity iron-transport systems is essential for the growth of some micro-organisms during infection (Griffiths, 1983) and both siderophores and iron-regulated OMPs are known to be produced during growth in uiuo (Griffiths & Humphreys, 1980; Griffiths et al., 1983; Lam et a/., 1984). Other bacterial pathogens, however, for example members of the genus Neisseria, while expressing new OMPs in response to iron limitation (Norqvist et al., 1978) do not appear to secrete siderophores and seem to obtain iron directly from the host iron-binding proteins via cell surface mechanisms (Simonson et al., 1982).

f Present address : Division of Enteric Pathogens, Central Public Health Laboratory, Colindale Avenue,

Abbreciutiuns: OMP, outer-membrane protein ; SS, Stainer and Scholte medium ; SSRFe, iron-restricted SS London NW9 5HT, U K .

medium; TSB, trypticase soy broth; TSBRFe, iron-restricted TSB.

0001-3581 0 1987 SGM

892 K . R E D H E A D , T . HILL A N D H . C H A R T

Borderella pertussis, the causative agent of whooping-cough, is a non-invasive pathogen of the human upper respiratory tract, a mucosal surface bathed in secretions containing lactoferrin. The ability of B. pertussis to colonize and multiply in this environment suggests that it also should possess a means of obtaining iron during pathogenesis. In this paper we report our investigation into the mechanism which may enable this organism to obtain iron.

METHODS

Bacterial strains. Borderella pertussis strains Wellcome 28 (W28), Tohama and 18323 were maintained in the medium of Stainer & Scholte (1971) (SS) containing 5 % (v/v) glycerol at -70 "C. Cells were grown in 250 ml Erlenmeyer flasks (100 ml medium per flask) at 35 "Con an orbital shaking incubator for 3 d using the standard SS medium. Escherichia coli strain 01 11 K58 H2 was stored in Brain Heart Infusion broth (Difco) containing 10% (v/v) glycerol at - 70 "C and was grown in trypticase soy broth (TSB, BBL) as previously described (Griffiths & Humphreys, 1978). Bacteria were harvested by centrifugation and washed twice in the appropriate medium without added iron salts before being used to inoculate test media. All glassware was acid rinsed before use to avoid iron contamination.

Assay of the freely available iron content of media. The unsaturated iron-binding capacity of a solution of the iron- binding protein ovotransferrin (4 mg ml-I) was measured according to the method of Bullen et al. (1972). One ml of this ovotransferrin solution was mixed with an equal volume of medium and the iron-binding capacity of the mixture again measured. The freely available iron content of the medium could then be estimated from the reduction in the iron-binding capacity of the ovotransferrin solution. These estimations were confirmed using the 'bleomycin assay' described by Gutteridge et al. (1981).

Cultivation under dijering conditions of iron availability. The test media used were as follows : (1) iron-replete medium (SS + Fe), standard SS medium containing 18 p~ freely available iron plus 0.07 M NaHCO,; (2) iron- limited medium (SS - Fe), SS medium containing <0.5 pM-irOn plus 0.07 M-N~HCO,; (3) iron-restricted medium consisting of either SS medium containing 4 p - i r o n (SSRFe) or TSB containing 6 PM-irOn (TSBRFe), to which was added 0.07 M-NaHCO, and 0.5 mg iron-binding protein ml-I [usually apovotransferrin, or human apotransferrin or apolactoferrin (Sigma)] (Griffiths & Humphreys, 1978). The NaHCO, was included in the media as the iron-chelator proteins used require bicarbonate ions for binding of iron. Assays of the iron-restricted media showed them to contain no detectable free iron and the iron-chelator proteins to be less than 50% saturated.

Washed bacteria were used at an inoculum of lo8 cells ml-' to initiate growth in 500 ml Erlenmeyer flasks (150 ml medium per flask) on an orbital shaking incubator (100 r.p.m.) at 35 "C. The pH of the medium was maintained between 7.6 and 8.0 by passing a humidified gas mixture of 95% air and 5 % C 0 2 over the surface of the agitated liquid at 100 ml min-l. Bacterial growth was monitored by measuring optical density (623 nm). After a maximum of 3 d incubation the cultures were harvested by centrifugation and cells and supernates stored separately at - 30 "C until used.

Detection of siderophores. Culture supernates were assayed for the presence of phenolate and hydroxamate siderophores by the methods of Arnow (1937) and Atkin et al. (1970) respectively.

Preparation of antisera. Washed B. pertussis bacteria were prepared as vaccines by heating cells at 56 "C for 30 min prior to storage at 4 "C (Redhead, 1984). Antisera were prepared in 6- to 8-week-old Balb/c mice (Olac). The mice received two intraperitoneal injections of the vaccine preparations, each of 0.5 ml containing 2 x lo9 bacteria, 14 d apart. They were exsanguinated under ether anaesthesia 7 d after the second injection and the resultant sera were stored at -20 "C until used.

Extraction and electrophoresis of OMPs. Bacterial outer membranes were prepared and analysed by SDS-PAGE as described elsewhere (Redhead, 1983). After electrophoresis, gels were either fixed and stained in a solution of 0.025% (w/v) Coomassie Blue R250 in 50% (v/v) methanol/5% (v/v) acetic acid and destained in 5 % (v/v) methanol/7-5% (vjv) acetic acid, or used for immunoblotting.

Electrophoretic immunoblotting. The proteins in SDS-PAGE gels were transferred to nitrocellulose sheets (0.45 pm pore size; Schleicher and Schuell) as described by Johnstone & Thorpe (1982) using an electroblot system EC-420 (EC Apparatus). The nitrocellulose sheets were exposed to murine antisera and antigen-antibody complexes were detected as previously described (Redhead, 1984) using an 2SI-labelled goat anti-mouse F(ab), antibody (10" c.p.m. per lane) kindly supplied by Dr R. Thorpe (National Institute for Biological Standards and Control). After washing and drying, iodine-labelled immunocomplexes were detected by autoradiography using Kodax X-omat S film.

Assay ofbacterial binding ojchelator proteins. Iron-binding proteins were labelled with 2 s I using chloramine T (Greenwood et al., 1963). Labelled and unlabelled proteins were added to iron-restricted medium to give a final chelator concentration of 0.5 mg ml-' and a radioactivity of 2 x lo5 c.p.m. m1-I. The cultures were inoculated and incubated as described above, without gassing. The association of labelled iron-binding proteins with the bacteria was assessed by aseptically removing 1 ml portions of the cultures at various time intervals and filtering out the bacteria on Millipore filters (0.22 pm pore size). The membranes were each washed with 10 ml PBS, and

Iron-binding proteins and B . pertussis 89 3

d 0.8 0

0.6

0.4

0.2

10 20 30 40 50 60 70 Time (h)

Fig. 1. Growth of B. pertussis W28 in iron-limited medium (SS - Fe) (a), iron-restricted medium (SSRFe) (0) and iron-replete medium (SS + Fe) (0). Results are representative of those obtained in three experiments.

inherent radioactivity was quantified using a multigamma counter. The degree of binding of chelator to bacteria was expressed as the counts associated with the bacteria present in 1 ml culture.

Prorein esrimarions. Protein concentrations were determined using the modified Lowry method of Schacterle & Pollack (1973).

R E S U L T S

Availability ojiron and the growth of B . pertussis

The growth of B. pertussis strain W28 when cultivated in iron-limited medium (SS - Fe) was severely reduced during the initial 24 h, compared with that in iron-replete medium (SS + Fe), and then ceased completely (Fig. 1). However, in the iron-restricted medium (SSRFe), containing only iron bound to the chelator protein ovotransferrin, strain W28 showed only a slight reduction in growth rate over the whole 65 h cultivation period (Fig. 1). B. pertussis strain Tohama showed changes in growth rates dependent upon iron availability which were very similar to those seen with strain W28 (data not shown).

Analysis of supernates from strains W28 and Tohama, cultivated in SS - Fe or SSRFe media for 48 to 65 h, failed to detect the presence of phenolate or hydromaxate compounds.

Efiect o j iron-restricted growth on O M P pro$les

When the OMPs of B. pertussis strains W28, 18323 and Tohama grown in SS + Fe medium and SSRFe medium (containing ovotransferrin) were compared by SDS-PAGE (Fig. 2) i t was obvious that a major extra band (arrow 1, molecular mass 77 kDa) was associated with cells of all three strains grown in iron-restricted medium. Some minor differences could also be seen between iron-replete and iron-restricted cultures, e.g. in strain 18323 grown in SSRFe medium there was a reduction of the 82 kDa band (arrow 2) and the appearance of a faint 49 kDA band (arrow 3); in strain Tohama grown in SSRFe medium the 28 kDa band (arrow 4) was severely reduced. However, the appearance of a 77 kDa protein was the only consistent change associated with iron restriction which occurred in all three strains.

In order to examine immunogenic differences between B. pertussis grown under the different conditions, murine antisera were raised to whole-cell vaccines prepared from organisms grown under iron-replete or iron-restricted conditions; antibody interactions with OMP preparations were examined by electrophoretic immunoblotting. The results in Fig. 3 show that antisera raised to both preparations produced identical patterns of antigen-antibody complexes with iron-replete or iron-restricted OMPs, with one exception : antiserum to iron-restricted bacteria

894 K . R E D H E A D , T . H I L L A N D H . C H A R T

Fig. 2. SDS-PAGE of OMPs from B. pertussis strains grown in media of different iron availabilities. A, W28, iron-replete; B, W28, iron-restricted; C, 18323, iron-replete; D, 18323, iron-restricted; E, Tohama iron-replete ; F, Tohama, iron-restricted ; G, molecular mass standards. Arrows 14 indicate iron-regulated proteins which are mentioned in the text.

reacted with an extra protein present only in OMP preparations from the iron-restricted organisms. This protein corresponded to the major extra band seen in SDS-PAGE of the outer membranes. On one occasion BDH standard proteins were also subjected to electrophoretic immunoblotting and the antiserum to iron-restricted B. pertussis reacted with the 77 kDa molecular mass marker, which was ovotransferrin. This reaction was confirmed by immunoblotting with purified ovotransferrin (Sigma) which was also detected only by the antiserum to iron-restricted bacteria (Fig. 3, arrow 1). The results strongly suggested that the novel 77 kDa protein was indeed ovotransferrin from the SSRFe medium which had become associated with the bacterial outer membranes. Attempts were made to raise murine antisera to ovotransferrin alone to confirm the results of the immunoblotting experiments; however, ovotransferrin was a poor immunogen and antiserum of sufficiently high titre could not be obtained.

Association of' iron-binding proteins with bacteria

After incubation of B. pertussis W28 for 30 min in SSRFe medium containing '251-labelled ovotransferrin, the radioactivity associated with washed bacteria began to rise above the background level (Fig. 4). The amount of radioactivity associated with the bacteria continued to increase quite rapidly over a period of 90 min, then began to level off. In contrast, when E. coli 0111, a strain which is well adapted to growth in iron-restricted conditions (Griffiths & Humphreys, 1978; Griffiths et al., 1983) was placed in an equivalent iron-restricted medium (TSBRFe) it displayed only background levels of associated 251-labelled ovotransferrin even

Iron-binding proteins and B. pertussis 895

Fig. 3. Demonstration of antibodies to B . pertussis OMPs and ovotransferrin by immunoblotting of SDS-PAGE gels. .Antisera from mice immunized with B. pertussis W28 grown under iron-restricted (lanes A-C) and iron-replete (lanes D-F) conditions were examined as described in the text. Gel samples : A, purified ovotransferrin ; B, W28, iron-restricted ; C, W28, iron-replete; D, purified ovotransferrin; E, W28, iron-restricted ; F, W28, iron-replete.

after incubation for 2 h (Fig. 4). Similar experiments showed that the iron-chelating proteins lactoferrin and transferrin were also bound by B. pertussis, with the same kinetics as for ovotransferrin (data not shown).

Competitive binding studies were performed using a fixed concentration (0.5 mg ml-l) of 251-labelled transferrin mixed with increasing concentrations (0.5-5.0 mg ml-l) of unlabelled

transferrin, lactoferrin or ovotransferrin in portions of SSRFe medium. Washed bacteria from 48 h cultures of B. pertussis W28 were added to the media to a final concentration of lo8 cells ml-I and incubated under standard iron-restricted conditions for 60 min. The amount of labelled transferrin bound was assayed as before and expressed as a percentage of the amount originally bound (Fig. 5) . Labelled and unlabelled transferrin competed against each other in this binding study; the percentage of labelled transferrin bound was reduced as the concentration of unlabelled transferrin was increased. The presence of the other two iron- binding proteins, particularly ovotransferrin, also interfered with the binding of transferrin, indicating that all three proteins could compete against each other during binding to B. pertussis.

Stability ojbinding B. pertussis W28 (lo8 cells ml-l) was incubated for 3 h in SSRFe medium containing

radiolabelled transferrin. The cells were then harvested, washed, resuspended and incubated in

896 K . R E D H E A D , T . H I L L

30 60 90 120 Time (min)

Fig. 4

A N D H . C H A R T

I I I I I J

2 4 . 6 8 10 Ratio of unlabelled to labelled

iron-binding protein

Fig. 5

Fig. 4. Binding of 2SI-labelled ovotransferrin under iron-restricted conditions, by B. pertussis W28 (0) and E. coli 0 1 11 (0). Results are the means of duplicate experiments. Fig. 5 . Binding of 2SI-labelled transferrin by B. pertussis W28, in the presence of other, unlabelled, iron- chelator proteins : transferrin (.), lactoferrin (O), ovotransferrin (0) (see text for details). Results are the means of duplicate experiments.

1 I I I 30 60 90 120

Time (min) Fig. 6. Changes in the amount of '2SI-labelled transferrin bound after transfer of B. pertussis W28 cells from SSRFe medium containing I 2SI-labelled transferrin into fresh SSRFe medium containing 251- labelled transferrin (0) or unlabelled transferrin (0). Results are the mean of duplicate experiments.

identical volumes of fresh SSRFe medium containing the same concentration of either labelled or unlabelled transferrin. The radioactivity associated with the bacteria was measured at intervals over the following 2 h (Fig. 6). Bacteria resuspended in medium containing labelled transferrin showed an initial slight reduction in cell-associated radioactivity, which then stabilized. When the bacteria were resuspended in medium containing unlabelled transferrin, the bound radioactivity fell steadily, 40% of the labelled material having disassociated after 2 h (Fig. 6).

D I S C U S S I O N

To cause the disease whooping-cough, B. pertussis must be able to survive and grow in the iron-restricted environment of the mucosal surface of the human upper respiratory tract. In this report we have shown that B. pertussis can grow in vitro under iron-restricted conditions induced

Iron-binding proteins and B. pertussis 897

by the iron-binding proteins ovotransferrin, transferrin and lactoferrin. Indeed, the organism grew almost as well in the presence of these proteins as it did in iron-replete medium. The bacteria must therefore have obtained iron from the iron-binding proteins as opposed to using intracellular stores, as very little growth occurred in medium chemically deficient in iron.

Unlike several other bacteria, such as E. coli and Salmonella typhimurium, which are also capable of growing under iron-restricted conditions (Griffiths, 1985) B. pertussis does not appear to produce siderophores. Certainly none of the two main bacterial siderophore types were detected in this study.

The major change which occurred in the SDS-PAGE profile of B . pertussis OMPs following growth in the presence of ovotransferrin, and the only one common to all three strains examined, was the apparent appearance of a new 77 kDa protein in iron-restricted cultures. However, it was a surprise to find that the antiserum raised to iron-restricted bacteria recognized the 77 kDa marker, ovotransferrin, as well as the novel protein in immunoblotting, strongly suggesting that the novel protein was indeed ovotransferrin from the medium which had been bound by the bacteria. This is remarkable since the preparation of outer membranes involved multiple centrifugations, several washings, sonication and two treatments with Triton X-100 prior to analysis by SDS-PAGE. The presence of ovotransferrin in the final envelope preparations shows how strongly it was bound by the bacteria and suggests that the association of iron- binding proteins with B. pertussis cells was due to the presence of specific mechanisms rather than the result of physicochemical interactions such as hydrophilicity. Outer membranes prepared from E. coli grown in the presence of iron-binding proteins did not contain detectable levels of these proteins.

Since no novel proteins other than the iron-binding proteins were found associated with the outer membranes of any of the iron-restricted B. pertussis cultures, i t seems likely that an ovotransferrin receptor, if one exists, is constitutively expressed rather than induced by iron restriction. The ability of avirulent phase IV variants of the strains used in this study to bind iron-chelator proteins (data not shown) demonstrates that this property is not subject to phase degradation like several other properties of B. pertussis (Dobrogosz et al., 1979). Furthermore, the reduction in level of bound radiolabglled transferrin when B. pertussis was incubated in the presence of unlabelled chelator suggests that there is turnover of bound iron-chelators.

Some bacteria, such as Neisseria meningitidis, can grow under iron-restricted conditions yet do not appear to produce siderophores (Archibald & DeVoe, 1980). In these cases it has been suggested that iron is removed from the iron-chelating proteins by a mechanism which involves direct interaction between receptors on the bacterial cell surface and the iron-binding proteins (Archibald & DeVoe, 1979; Simonson et al., 1982). The human pathogenic protozoan Trichomonas vaginalis, which is known to bind several host plasma proteins (Peterson & Alderete, 1982), has also been shown to possess specific receptors for human lactoferrin and to acquire iron from this source (Peterson & Alderete, 1984). It is possible that a similar system may be employed by B. pertussis, where the binding of the iron-chelator proteins to the bacteria is a preliminary step to the removal of iron.

One of the most significant features regarding acquisition of iron from iron-binding proteins by Neisseria spp. is the highly specific nature of the process. None of the Neisseria spp. that have been examined was able to obtain iron from all three iron-binding proteins, lactoferrin, transferrin and ovotransferrin (Mickelsen & Sparling, 1981 ; Mickelsen et al., 1982). In contrast it appears that B. pertussis can use more than one iron-binding protein as its iron source. In addition to ovotransferrin it can also bind transferrin and lactoferrin and presumably utilize the bound iron. As all three iron-chelators competed with one another in the binding studies they must attach either to the same bacterial cell surface receptor or to receptors in close enough proximity to cause severe steric hinderance.

We are presently engaged in research to confirm whether B. pertussis does incorporate iron from the bound chelator proteins and are attempting to identify a specific binding receptor in the outer membrane. A surface-exposed outer-membrane component with a virulence potential such as this could prove to be an important protective antigen and might be of value in the new acellular pertussis vaccines which are currently being developed.

898 K . R E D H E A D , T . H I L L A N D H . C H A R T

We thank J. M. C. Gutteridge for performing the bleomycin assays, E. Griffiths for helpful advice and discussion and I. Stephens for typing the manuscript.

R E F E R E N C E S

ARCHIBALD, F. S. & DEVOE, I. W. (1979). Removal of iron from human transferrin by Neisseria meningiti- dis. FEMS Microbiology Letters 6, 159-162.

ARCHIBALD, F. S. & DEVOE, I. W. (1980). Iron acquisition by Neisseria meningitidis in vitro. Infection and Immunity 21, 322-334.

ARNOW, L. E. (1937). Colorimetric determination of the components of 3,4-dihydroxyphenlalanine-tyro- sine mixtures. Journal of Biological Chemistry 228,

ATKIN, C. L., NEILANDS, J. B. & PHAFF, H. J. (1970). Rhodotorulic acid from species of Leucosporidium, Rhodosporidium, Rhodotorula, Sporidiobolus and Sporobolomyces, and a new alanine-containing ferri- chrome from Cryptococcus melibiosum. Journal of Bacteriology 103, 722-733.

BULLEN, J. J., ROGERS, H. J. & LEIGH, L. (1972). Iron- binding proteins in milk and resistance to Escheri- chia coli infections in infants. British Medical Journal

BULLEN, J. J., ROGERS, H. J. & GRIFFITHS, E. (1978). Role of iron in bacterial infection. Current Topics in Microbiology and Immunology 80, 1-35.

DOBROGOSZ, W. J., EZZELL, J. W., KLOOS, W. E. & MANCLARK, C. R. (1979). Physiology of Bordetella pertussis. In International Symposium on Pertussis, pp. 86-93. Edited by C. R. Manclark & J. C. Hill. Washington, DC: US Department of Health, Edu- cation and Welfare.

GREENWOOD, F. C., HUNTER, W. M. &GLOVER, J. S. (1963). The preparation of I 3 l I-labelled human growth hormone of high specific radioactivity. Biochemical Journal %9, 1 14-1 18.

GRIFFITHS, E. (1983). Adaptation and multiplication of bacteria in host tissues. Philosophical Transactions of the Royal Society B303, 85-96.

GRIFFITHS, E. (1985). Candidate virulence markers. In The Virulence of Escherichia coli, pp. 193-226. Edited by M. Sussman. London: Academic Press.

GRIFFITHS, E. & HUMPHREYS, J. (1978). Alterations in tRNAs containing 2-methylthio-N6-(A2-isopen- teny1)-adenosine during growth of enteropathogenic Escherichia coli in the presence of iron binding proteins. European Journal of Biochemistry 82,

GRIFFITHS, E. & HUMPHREYS, J. (1980). Isolation of enterochelin from the peritoneal washings of guinea- pigs lethally infected with Escherichia coli. Injection and Immunity 28, 286-289.

GRIFFITHS, E., STEVENSON, P. & JOYCE, P. (1983). Pathogenic Escherichia coli express new outer mem- brane proteins when growing in vim. FEMS Micro- biology Letters 16, 95-99.

GWTERIDGE, J. M. C., ROWLEY, D. A. & HALLIWELL, B. (1981). Superoxide dependent formation of hydroxyl radicals in the presence of iron salts.

531-537.

1, 69-75.

503-513.

Detection of 'free' iron in biological systems by using bleomycin-dependent degradation of DNA. Bio- chemical Journal 199, 263-265.

JOHNSTONE, A. P. & THORPE, R. (1982). Immunoche- mistry in Practice, pp. 184-1 89. Oxford : Blackwell Scientific Publications.

LAM, C., YURNOWSKY, F., SCHWARZINGER, E. & NERUDA, W. (1984). Bacteria recovered without sub- culture from infected human urines express iron- regulated outer membrane proteins. FEMS Microbi- ology Letters 24, 255-259.

MICKELSEN, P. A. & SPARLING, P. F. (1981). Ability of Neisseria gonorrhoeae. Neisseria meningitidis and commensal Neisseria species to obtain iron from transferrin and iron compounds. Infection and Immunity 33, 555-564.

MICKELSEN, P. A., BLACKMAN, E. & SPARLING, P. F. ( 1982). Ability of Neisseria gonorrhoeae, Neisseria meningitidis, and commensal Neisseria species to obtain iron from lactoferrin. Infection and Immunity

NEILANDS, J. B. (1982). Microbial envelope proteins related to iron. Annual Reciew of Microbiology 36,

NORQVIST, A., DAVIES J., NORLANDER, L. & NOR- MARK, s. (1978). The effect of iron starvation on the outer membrane protein composition of Neisseria gonorrhoeae. FEMS Microbiology Letters 4,

PETERSON, K. M. & ALDERETE, J. F. (1982). Host plasma proteins on the surface of pathogenic Trichomonas oaginalis. Infection and Immunity 31,

PETERSON, K. M. & ALDERETE, J. F. (1984). Iron uptake and increased intracellular enzyme activity follow host lactoferrin binding by Trichomonas caginalis receptors. Journal of' Experimental Medicine 160, 398410.

REDHEAD, K. (1983). Variability in the surface expo- sure of the outer membrane proteins of Borderella pertussis. FEMS Microbiology L-etters 17, 35-39.

REDHEAD, K. (1984). Serum antibody responses to the outer membrane proteins of Bordetella pertussis. Infection and Immunity 44, 724-729.

SCHACTERLE, G . E. & POLLACK, R. L. (1973). A simplified method for the quantitative assay of small amounts of protein in biological material. Analytical Biochemistry 51, 654-655.

SIMONSON, C., BRENER, D. & DEVOE, I. W. (1982). Expression of a high affinity mechanism for acquisi- tion of transferrin iron by Neisseria meningitidis. Infection and Immunity 36, 107-1 13.

STAINER, D. W. & SCHOLTE, M. J. (1971). A simple chemically defined medium for the production of phase I Bordetella pertussis. Journal of General Microbiology 63, 2 1 1-220.

35, 915-920.

285-309.

71-75.

755-762.