Contribution of the outer surface proteins of Borrelia burgdorferi s. I. to the pathogenesis of

Lyme disease

O • V

AKADEMISK AVHANDLING

som för avläggande av doktorsexamen i medicinsk vetenskap vid Umeå Universitet, offentligen kommer att försvaras på engelska språket i

föreläsningssalen, byggnad 6L, Institutionen för Mikrobiologi, Umeå Universitet, fredagen den 7 oktober 1994, kl 09.00

av

Maria Jonsson

Umeå 1994

ABSTRACTContribution of the outer surface proteins of Borrelia burgdorferi s. L to the pathogenesis of Lyme disease.Maria Jonsson, Department of Microbiology, Umeå University, Sweden.

Borrelia burgdorferi s. I is a spirochete which causes the multisystemic disorder Lyme disease. As the borreliae lack toxin production, the pathogenicity is thought to involve, at least in part, molecules from the outer surface. Most Lyme disease Borrelia strains express two major outer surface lipoproteins, OspA (31 kD) and OspB (34 kD), on their surface. However, some strains lack the expression of OspA and OspB, but express a smaller 21 to 25 kD OspC protein instead. This thesis focuses on the importance of these proteins in the pathogenesis of Lyme disease.

Biochemical and immunochemical studies of the OspA and OspB proteins from strains of various geographic origins show considerable differences in the apparent molecular weights and in their reactivities to monoclonal antibodies. The cloning and sequencing of the ospAB opérons from strains of different origins has demonstrated that the heterogeneity is found also at the DNA level Comparison of the ospAB sequences allows the classification of the strains into three types, which coincide with the recent species designations, B. burgdorferi sensu stricto, B. afzelii and B. garinii The genes are located on a linear plasmid about 50 kb in size, and are cotranscribed as a single message.

The expression of the osp operon in different strains was studied by Western blot and Northern blot analysis. The ospAB operon of strains expressing varying amounts of the Osp proteins was cloned and sequenced. The DNA sequence was found to be >99% identical. The regulation appears to be primarily at the transcriptional level

In patients who have received incomplete treatment, B. burgdorferi have been isolated several years after the onset of the disease. As mentioned above, the ospAB loci of different strains show considerable heterogeneity, and it has been speculated that the spirochetes evade the host’s immune system by antigenic variation of the Osp proteins. In a mouse model system it was shown that no variation of the osp genes occurs over the course of an infection, and that other escape mechanisms must be used.

The OspB proteins in particular have been shown to be very heterogeneous in different isolates. The MAb 84C recognizes a wide variety of B. burgdorferi strains, and the binding epitope was mapped to a conserved region in the carboxyl terminus of the OspB protein with putative structural and/or functional importance

It is well known that antibodies can kill bacteria in the presence of complement and phagocytes. Some antibodies seem to have a bactericidal effect by themselves. H6831 is a monoclonal antibody recognizing the OspB protein of some B. burgdorferi strains. The bactericidal action of univalent FAb fragments from H6831 was further characterized, and the binding epitope was mapped to a very heterogeneous region of the carboxyl end of the OspB protein

Key words: Borrelia burgdorferi / Lyme disease / Osp proteins / ospAB operon / gene regulation / pathogenicity

ISBN: 91-7174-938-1

UMEÅ UNIVERSITY MEDICAL DISSERTATIONSNew Series No. 413From the Department of MicrobiologyUmeå University, Umeå, SwedenISBN 91-7174-938-1 ISSN 0346-6612

Contribution of the outer surface proteins of Borrelia burgdorferi s. I. to the pathogenesis of

Lyme disease

by

Maria Jonsson

Department of Microbiology Umeå University

Umeå 1994

2

Front cover:

Illustrator:

Borren Borrelia

Anna Winberg

Printed in Sweden by Solfjädern Offset AB

Umeå 1994

3

TABLE OF CONTENTS

Abbreviations 5Abstract 6Papers in this thesis 7Introduction 9

1. General introduction 91.1 History of Lyme disease 91.2. Taxonomy 101.3. General characteristics 111.4. The vector 121.5 Clinical manifestations 131.6. Therapy 14

2. DNA content 142.1 The linear chromosome 162.2 Linear plasmids 18

3. Pathogenicity 193.1. Bacterial virulence factors 193.2. Antigenic variation 203.3 B. burgdorferi s. I. outer membrane proteins 223.4 Regulation of the Osp proteins 243.5 Pathogenicity 253.6 Adhesion of Lyme disease Borrelia to human mammalian cells 273.7 Animal models and vaccination studies 28Aims of the present work 31Summary and discussion of results 33

4. Heterogeneity of the osp operon (Paper I) 334.1 Prologue 334.2 Cloning of the osp operon of AC AI and Ip90 334.3 Genetic organization 344.4 Sequence comparison 34

5. Expression of the osp operon (Paper II) 375.1 Why clone the osp operon from B. afzelii FI ? 375 .2 Characterization of protein expression by Western blot analysis 385.3 Genetic organization 3 85.4 Transcription 41

4

5.5 Model for Osp regulation 6. Molecular characterization of the OspA and OspB proteins during

infection in vivo (Paper III)6.1 Methodology6.2 Results of cultivation6.3 Cloning and sequencing of the osp A and ospB genes6.4 Immune responseMapping of anti-OspB antibody epitopes7.1 Method of selecting mutants7.2 Cloning and sequencing7.3 Mapping of the binding epitope for the highly cross-reactive

OspB MAb 84C (Paper IV)7.4 Characterization of the anti-OspB bactericidal antibody H6831

7.4.1 Borrelicidal activity7.4.2 Epitope mapping

8. Binding studies (unpublished results)8.1 Binding of spirochetes to glycolipids

9. Discussion of the pathogenicity of Lyme disease borreliae and the involvement of the Osp proteinsConclusions Acknowledgements References Papers I-

41

4243434444454545

484949505151

52555659

ABBREVIATIONS

ACA acrodermatitis chronica atrophicansACAI Borrelia afzelii strain ACAIB31 Borrelia burgdorferi sensu stricto strain B31CNS central nervous systemCSF cerebrospinal fluiddnaN gene encoding DNA polymerase IIIDNA deoxyribonucleic acidEM erythema migransFI Borrelia afzelii strain FIFAb fragment antigen bindinggyrA, gyrB genes encoding DNA gyrasesHB19 Borrelia burgdorferi sensu stricto strain HB19HUVE human umbilical vein epitheliumIp90 Borrelia garinii strain Ip90kb kilo basepairk D kilo DaltonLABC lymphadenosis benigna cutisIP linear plasmidMAb monoclonal antibodyMBC minimal bactericidal concentrationMIC minimal inhibitory concentrationN40 Borrelia burgdorferi strain N40ori origin of replicationOsp outer surface proteinPAGE polyacrylamide gel electrophoresisPBS phosphate buffered salinePCR polymerase chain reactionRNA ribonucleic acidRpa Borrelia burgdorferi sensu stricto strain RParRNA ribosomal RNArrf rrl, rrs genes for the 5S, 23S and 16S ribosomal RNA respectivelySCID severe combined immunodeficiencySDS sodium dodecyl sulphates. /. sensu lato.S', s. sensu strictoVmp variable major proteinVSG variable surface glycoproteinZS7 Borrelia burgdorferi strain ZS7

6

ABSTRACTContribution of the outer surface proteins of Borrelia burgdorferi s. L to the pathogenesis of Lyme disease.Maria Jonsson, Department of Microbiology, Umeå University, Sweden.

Borrelia burgdorferi s. /. is a spirochete which causes the multisystemic disorder Lyme disease. As the borreliae lack toxin production, the pathogenicity is thought to involve, at least in part, molecules from the outer surface. Most Lyme disease Borrelia strains express two major outer surface lipoproteins, OspA (31 kD) and OspB (34 kD), on their surface. However, some strains lack the expression of OspA and OspB, but express a smaller 21 to 25 kD OspC protein instead. This thesis focuses on the importance of these proteins in the pathogenesis of Lyme disease.

Biochemical and immunochemical studies of the OspA and OspB proteins from strains of various geographic origins show considerable differences in the apparent molecular weights and in their reactivities to monoclonal antibodies. The cloning and sequencing of the ospAB opérons from strains of different origins has demonstrated that the heterogeneity is found also at the DNA level. Comparison of the ospAB sequences allows the classification of the strains into three types, which coincide with the recent species designations, B. burgdorferi sensu stricto, B. afzelii and B. garinii. The genes are located on a linear plasmid about 50 kb in size, and are cotranscribed as a single message.

The expression of the osp operon in different strains was studied by Western blot and Northern blot analysis. The ospAB operon of strains expressing varying amounts of the Osp proteins was cloned and sequenced. The DNA sequence was found to be >99% identical. The regulation appears to be primarily at the transcriptional level.

In patients who have received incomplete treatment, B. burgdorferi have been isolated several years after the onset of the disease. As mentioned above, the ospAB loci of different strains show considerable heterogeneity, and it has been speculated that the spirochetes evade the host’s immune system by antigenic variation of the Osp proteins. In a mouse model system it was shown that no variation of the osp genes occurs over the course of an infection, and that other escape mechanisms must be used.

The OspB proteins in particular have been shown to be very heterogeneous in different isolates. The MAb 84C recognizes a wide variety of B. burgdorferi strains, and the binding epitope was mapped to a conserved region in the carboxyl terminus of the OspB protein with putative structural and/or functional importance.

It is well known that antibodies can kill bacteria in the presence of complement and phagocytes. Some antibodies seem to have a bactericidal effect by themselves. H6831 is a monoclonal antibody recognizing the OspB protein of some B. burgdorferi strains. The bactericidal action of univalent FAb fragments from H6831 was further characterized, and the binding epitope was mapped to a very heterogeneous region of the carboxyl end of the OspB protein.

7

PAPERS IN THIS THESIS

This thesis is based on the following articles and manuscripts, which will be referred to in the text by Roman numerals.

I. Jonsson, M., L. Noppa, A. G. Barbour, and S. Bergström. 1991.Heterogeneity of outer membrane proteins in Borrelia burgdorferi:Comparison of osp opérons of three isolates of different origins.Infection and Immunity 60:1845-1853.

II. Jonsson, M., and S. Bergström. 1994.Evidence for transcriptional and translational regulation of the expression of the major outer surface proteins OspA, OspB and OspC in Lyme disease Borrelia. Manuscript

III. Jonsson, M., T. Elmros, and S. Bergström. 1994.Genetic stability of the outer surface proteins OspA and OspB of Borrelia burgdorferi in subcutaneous chambers in mice.Manuscript.

IV. Shoberg, R. J., M. Jonsson, A. Sadziene, S. Bergström, and D. D. Thomas.1994.Identification of a highly cross-reactive outer surface protein B epitope among diverse geographic isolates of Borrelia spp. causing Lyme disease.Journal of Clinical Microbiology. 32:489-500.

V. Sadziene, A., M. Jonsson, S. Bergström, R. K. Bright, R. C. Kennedy, andA. G. Barbour. 1994.A bactericidal antibody to Borrelia burgdorferi is directed against a variable region of the OspB protein.Infection and Immunity. 62:2037-2045.

9

INTRODUCTION

1. General background

1.1 History of Lyme diseaseThe first likely case of Lyme disease to be documented was described by Buchwald in Germany , who in 1883 reported on a patient with an idiopathic skin atrophy (Buchwald, 1883). In 1902 similar cases were described, and the disease was named acrodermatitis chronica atrophicans, ACA (Herxheimer and Hartman, 1902). Erythema migrans (EM) was described in Sweden in 1909 (Afzelius,1910) and in Austria in 1913 (Lipschutz, 1914). In 1922, a patient in France was reported with neurological symptoms and an erythema, not recognized as EM (Garin and Bujadoux, 1922). The Swedish physician Hellerström was the first to report the connection between EM and neurological symptoms in 1930 (Hellerström, 1951). The first case of EM in the USA was reported in 1970 (Scrimenti, 1970).

In 1975 there was an outbreak of juvenile arthritis in Lyme, Connecticut, USA.This is normally a very rare condition, affecting only one in 100,000 children. Twomothers of affected children responded to the many cases, mostly children, but also adults, of this severe condition at the same time, and contacted the Health Department officials. Steere and his colleagues found that in Lyme, the occurence of juvenile arthritis was about 100 times the normal incidence (Steere, 1989). Moreover, the cases were associated with heavily wooded areas, and very few victims were from town centers. In some discrete regions the frequency was up to 10,000 times higher than normal. About 25 % of the patients remembered a strange skin rash from one to several weeks before the outbreak of the arthritis. This was in agreement with the findings of Afzelius in Europe in 1909, who described an expanding red skin rash, EM, in patients who had been bitten by the tick Ixodes ricinus. The EM had been successfully treated with penicillin, indicating a bacterial agent, not a virus

In 1977, nine patients affected by the skin rash remembered a tick bite, and one of them had even removed and saved the tick. It was identified to be of the species Ixodes dammini, closely related to the European I. ricinus (Steere, 1989). Finally, in 1981 Burgdorfer found some long, irregularly shaped spirochaete bacteria in the gut of an Ixodes tick, and Barbour managed to grow them in culture (Barbour, 1984a). Sera from patients infected with the mysterious disease were tested by Burgdorfer and colleagues, and they showed a pronounced antibody response to the bacteria (Burgdorfer et al., 1982). In 1983, the spirochete was also isolated from blood, skin

10

lesions and cerebrospinal fluid of patients (Benach et a i , 1983; Steere et al., 1983). Three of Kochs four postulates were then fulfilled. The fourth postulate, that inoculation with the isolated agent should give rise to the same symptoms as was originally found, cannot, of course, be tested for ethical reasons (Habicht et a i , 1987). The disease is now named Lyme disease, or Lyme borreliosis, and the agent of the disease was named Borrelia burgdorferi in honour of Willy Burgdorfer (Johnson et al., 1984b).

1.2 TaxonomyThe Borrelia genus, named after the french bacteriologist A. Borrel (Nationalencyklopedin, 1990), is distinct from the genera of Treponema and Leptospira by differences in guanine plus cytosine content, and the lack of DNA homology. The DNA of Borrelia have mole % guanine and cytosine contents of 27.3 to 30.5, Leptospira of 35.3 to 39.9, and Treponema pallidum of 52 to 53.7. The DNA homology between Borrelia and the other two genera is 0 to 2 % (Hyde and Johnson, 1984)

In the past, differentiation of borrelia was based on identification of the specific vector of transmission of the spirochete. Today, genetic methods are used for the classification. DNA homology studies have shown that three North American tick- borne relapsing fever species, B. hermsii, B. turicatae and B. parkeri, show high homologies (76-100 %) and should therefore belong to the same species (Hyde and Johnson, 1986). There are other cases where reclassification will be needed.

B. burgdorferi was described as a new species based on genotypic, i. e. DNA- DNA hybridization, and phenotypic characteristics such as reactivity of the outer surface protein A, OspA, to the monoclonal antibody H5332 (Johnson et al., 1984b; Barbour et al., 1983). This definition was based on few isolates. Today, several hundred strains from different sources and geographic areas have been isolated and further characterized. Isolates from Europe have been shown to be especially heterogeneous, both antigenically and at the DNA level (Wilske et al., 1986, 1988b ). Based on rRNA gene restriction patterns, protein electrophoresis patterns and reactivity with murine monoclonal antibodies, Lyme disease Borrelia was divided into three genomic species; B. burgdorferi sensu stricto, B. garinii and VS461. VS461 has now been named B. afzelii (Baranton et al., 1992; Marconi and Garon, 1992; Canica et a i , 1993). In this thesis, I will use the terms B. burgdorferi s. i or Lyme disease Borrelia to refer to these three species. A fourth genomic species, B. japonica has

11

been isolated from the Japanese tick I. ovatus (Kawabata et al., 1993; Postic et al., 1993). B. japonica does not appear to be a human pathogen, as no patient isolates of this group have been found. Furthermore, the ticks from which it has been isolated do not use humans as a host. Recent reports indicate that the number of Borrelia species is increasing. At least some of these new groups were isolated from vectors with enzootic cycles different from the ordinary Lyme disease strains, indicating that they may not be human pathogens (Assous et al., 1994; Postic et al., 1994).

1.3 General characteristicsBorrelia burgdorferi belongs to the genus of Borrelia, and the order of Spirochaetales. They are flexuous, thin, gram-negative, chemoheterotrophic, helical-shaped organisms. A fluid outer cell membrane surrounds the protoplasmic cylinder complex which consists of the cytoplasm, the inner cell membrane and the peptidoglycan layer. Isolated outer membranes are made up of 45 to 62 % protein, 23 to 50 % lipid and 3 to 4 % carbohydrate. The outer sheaths of Borrelia species are characterized by a lack of structural detail (Holt, 1978).

In the tick, the large blood meal is held in the midgut. The digestion of blood occurs intracellularly in the epithelium of the gut lining (Akov, 1982). As a consequence, the Borrelia in the tick midgut are not exposed to proteases or acidity, which makes it unnecessary for them to have a defense against proteases, which would be required if they were harboured in a hematophageous insect, like a mosquito. Thus, both the tick borne B. hermsii and B. burgdorferi are susceptible to attack by proteases like trypsin (Barbour et al., 1984b; Barbour, 1985).

The Borrelia differ morphologically from other procaryotes in that they have an axial fibril called the endoflagellum. B. burgdorferi have 7 to 20 endoflagella, around which the cell is coiled (Holt, 1978; Steere, 1989). The flagella are located between the outer cell membrane and the protoplasmic cylinder, attached to the poles. The bacterial diameter is about 0.2 to 1 pm and the length of the cell is 20 to 30 pm, depending on the physiological state of the bacteria. The length increases as cells reach the stationary phase of growth, but the length also depends on the quality of the growth medium or the experimental animal in which the bacteria is harboured. (Holt, 1978; Barbour and Hayes, 1986; Steere, 1989).

12

1.4 The vectorThe vectors of Lyme borreliosis are hard shelled ticks of the genus Ixodes. In Europe the species is I. ricinus, and in Eastern Europe and Asia, the tick species is primarily I. persulcatus. In the USA, three species of Ixodes ticks are implicated. I. dammini is found in North-eastern and mid-western USA, I. paciflcus in Western USA, and I. scapularis in South-eastern USA (Gustafson, 1993). It should be noted that recent work has shown that I. dammini and I. scapularis are the same species (Oliver et ai, 1993). In this thesis, I will use the name applied in the littérature cited, usually I. dammini.

The Ixodes ticks go through different life stages. Eggs are usually deposited in the spring, and it takes several weeks for the larvae to hatch. At this stage they are very small, barely visible, and have only six legs. They feed once, usually on blood from small mammals like mice, and then the larvae crawl into the ground and moult to become the slightly larger, eight-legged nymphs. These also feed once, on mice or larger mammals like dogs, deers or humans, and then moult into adults. Both sexes have eight legs, but the female is larger than the male. The females feed once more, while the adult males do not. The females die after laying their eggs (Steere, 1989).

The frequency of transovarial transmission of spirochetes from an infected female to her offspring is variable. In one study, the transmission of spirochetes to the progeny of two I. ricinus female ticks was 60 % and 100 % (Burgdorfer et al., 1983). In another study, the occurence was <1 % of infection in I. dammini larvae (Magnarelli et al., 1987). The infection is usually spread by larvae or nymphs becoming infected by feeding on an infected animal, and then transmitting it to their next host (Magnarelli et al., 1987).

B. burgdorferi can be found in the midgut, hindgut and the rectal area of the tick, and eventually in the salivary glands. From the intestinal region the borreliae may be transmitted by the regurgitation of the gut contents during the feeding act (Burgdorfer, 1984), but usually they are transmitted via the infected saliva. To avoid infection, it is best to remove the tick from the skin as soon as possible after detection. The best way to remove an attached tick is to use a pair of forceps, without squeezing the abdomen. Oil covers the tracheas of the tick and suffocates it. If covered with oil or butter, the tick will withdraw from the skin, but while doing so, the contractions of the abdomen will increase the transfer of spirochetes in the midgut into the host. Gasoline or acetone are also often used to remove ticks. These chemicals will kill the tick, with the result that it is often caught fastened in the skin (Magnarelli, personal communication).

In Sweden it was previously assumed that Dalälven was the northern boundary of

13

the tick population. Thus, lacking ticks, there were no Borrelia infections in Northern Sweden. It has now been documented that ticks are more widely spread than previously thought. We have shown that B. burgdorferi infected ticks and Lyme disease exist along the coast in the north of Sweden (Bergström et a i , 1992; Jaenson et al., 1989). Ixodes ricinus has also been found in areas close to the arctic circle (Olsén, personal communication).

1.5 Clinical manifestationsOne of the problems with Lyme borreliosis is that the symptoms of the disease mimic a lot of other clinical conditions (Stechenberg, 1988). B. burgdorferi may spread locally in the skin causing a migrating erythematous lesion, erythema migrans (EM) (Åsbrink et al., 1984a). The period of time between the tick bite and the appearance of EM has been reported to be 3 to 32 days, with a median of 7 days. The most common EM lesion is a circular expanding erythema, which may be homogeneous, but usually shows a central clearing. It is occasionally accompanied by constitutional symptoms such as fever, headache, malaise and fatigue. Secondary erythema, due to spreading in the blood, may occur at other locations than the primary lesion, but these are often smaller. Untreated, the EM usually disappears within weeks to months, however, relapses are known to occur. B. burgdorferi can also cause a tumor-like bluish-red swelling or nodule, borrelia lymphocytoma, BL„ usually on the ear lobes or the nipple- areola mammae area. BL can appear together with EM or acrodermatitis chronica atrophicans, ACA, and be accompanied by neurological and arthritic symptoms (Stechenberg, 1988; Steere, 1989; Karlsson, 1990).

The third characteristic skin lesion is, as mentioned above, ACA. The onset is usually gradual, and characterized by a bluish-red discoloration and swollen skin, predominantly located on the extremities. The inflammatory phase may persist for years or decades, and is gradually replaced by atrophy of the skin and sometimes also underlying tissues.

B. burgdorferi can spread not only in the skin and underlying tissues, but also through the lymph or blood to the skin at other sites or to other organs like the heart, joints and the nervous system (Steere, 1989; Karlsson, 1990). Infection of the heart leads to fatigue and arrythmia. It usually lasts for approximately a week, and is not followed by recurrences, although the arrythmia may persist (Karlsson, 1990). Cardiac manifestations occur in less than 10 % of the patients (Stechenberg, 1988), and so far, only one fatal case has been reported (Marcus et al., 1985).

14

Arthritis affects a few or several small or large joints. It is often accompanied by migrating pains in the tendons and muscles, and can last from about a week to months. Artralgias may recur for several years, interrupted by long, symptom-free periods. Some patients suffer from chronic arthritis in the larger joints, with erosion of the cartilage and bone.

Both the peripheral and the central nervous systems (CNS) may be affected in Lyme disease. The symptoms can be diffuse, including profound fatigue, loss of appetite, slight headache or change of personality. However, they can also be more profound, with facial palsy, double vision and loss of hearing. Neuroborreliosis may resolve itself spontaneously within weeks or months, but a small percent of the cases become chronic and may last for decades. The fatigue symptoms resemble those observed in patients with chronic fatigue symptoms, CFS, which are not triggered by Borrelia infections (Coyle et al., 1994).

1.6 TherapyIn vitro studies show that B. burgdorferi is most susceptible to erythromycin, ceftriaxone and cefotaxime, although erythromycin is less effective in vivo (Preac- Mursic et al., 1987). It was also shown that Penicillin G, which is considered an appropriate antibiotic for the treatment of Lyme disease, has very poor activity both m vitro and in vivo. In EM, LB and AC A, oral treatment with penicillin, amoxycillin or doxycycline is usually recommended. In disseminated stages, per oral treatment with doxycycline and parenteral treatment with Penicillin G, cefotaxime and ceftriaxone has been succesful (Steere, 1989).

There are cases where failure of the antibiotic treatment has been documented by positive culture or serological findings. This can be explained by insufficient dosages of antibiotic, and/or inadequate absorption or penetration to the affected tissues. Another explanation is deficient bacteriocidal effect of the antibiotic (Karlsson, 1990). In late, chronic stages it is possible that the symptoms are caused by autoimmune reactions rather than the presence of spirochetes (Aberer et al., 1989).

2. DNA contentAs mentioned above, the DNA of B. burgdorferi has a low G+C content of 28-30.5 mole % (Hyde and Johnson, 1984; Johnson et al., 1984b; Schmid et al., 1984). The

15

DNA homology to other Borreliae is 30 to 59 % (Hyde and Johnson, 1984; Schmid et al., 1984). The low G+C content alters the codon usage in the spirochetes. A and T rich codons, which are regarded as rare codons in E. coli, are preferred codons in Borrelia (Bergström et al., 1989). Long stretches of poly(dA) is known to cause natural bending in DNA (Koo et al., 1986). In E. coli, stretches of such structures have been found upstream of strong promoters, and are thought to be important for the promoter strength (Bracco et al., 1989). Being very AT rich, stretches of poly(dA) are frequent in the Borrelia genome. The regulation of gene expression in the spirochetes may, therefore, be quite different from what is found in other bacteria with higher G+Ccontents.

vwvywv/

Borrelia burgdorferi s. 1.

: ..-. =;v- - ..... ■ ■■ -------- ■ .................o

Chromosome 1.000 kb

-------------- ------------------o Linear plasmids 15-50 kb

Sb Circular plasmids 5-30 kb

Phage 1-2 kb

Figure 1. The DNA content of B. burgdorferi s. I.

A schematic drawing of the DNA organization in Borrelia burgdorferi s. I. is shown in Figure 1. Like most, if not all, bacteria, B. burgdorferi has supercoiled circular plasmids of 5 to 30 kb (Hyde et al., 1984, 1986; Simpson et al., 1990). However, B. burgdorferi and all other Borrelia investigated, contain linear plasmids which are uncommon in eubacteria (Plasterk et al., 1985; Barbour and Garon, 1987; Pemg and LeFebvre, 1990). Both the size and the number of linear plasmids vary between strains, but the smallest are around 15 kb, and the largest is about 50 kb. Even more unique than the linear plasmids is that the chromosome also is linear. To date, Streptomyces lividans (Lin et al., 1993) and Agrobacterium tumefaciens (Allardet-Servent et al.,

16

1993) are the only other examples of eubacteria having linear chromosomes. Besides the chromosome and plasmids, phage have been visualized inside the bacteria by electron microscopy (Hayes et al., 1983 ).

2.1 The linear chromosomeThe discovery that the B. burgdorferi chromosome is linear was made simultaneously in two separate laboratories (Ferdows and Barbour, 1989; Baril et a i , 1989). High molecular weight Borrelia DNA was subjected to pulse field gel electrophoresis. Under the running conditions used, circular DNA should not enter the gel. Both groups found a distinct band about 950-1 000 kb in size, with little material remaining in the wells. To avoid shearing, DNA was prepared in situ by lysing the spirochetes embedded in agarose. Other members of the Borrelia genus, including B. hermsii, B. coriaceae, B. turicatae, B. parkeri and B. anserina, have linear chromosomes of roughly the same size as B. burgdorferi (Casjens and Huang, 1993; Rosa and Schwann 1992; Kitten and Barbour, 1992 ). Chromosomal DNA from other spirochetes, /. e. Leptospira interrogans, Spirochaeta aurantia and Treponema denticola, does not enter the gel, indicating that linear chromosomes are not common to all spirochetes (Ferdows and Barbour, 1989; Baril et al., 1989).

Figure 2. Map of the linear chromosome. The chromosome is depicted by the horizontal line. The numbers above this line indicate the distance in kb from the left end. The regions to which the genes have been mapped, are indicated by filled triangles. The two putative origins of replication are indicated above the chromosome (Davidson et a l., 1992; Casjens and Huang, 1993).

The linear nature of the chromosome has also been suggested by the physical and genetic maps of the chromosome which have been generated (Davidson et al., 1992; Casjens and Huang, 1993). Figure 2 represents a compilation of the two available

17

maps. If the chromosome were circular, all restriction endonucleases would have to cleave at one location. The probability of this is quite low. The possibility that the chromosome contains one extremely fragile locus that is always broken during preparation cannot be eliminated until the structure of the chromosomal ends has been determined. The ends do not hybridize to each other. Genes homologous to replication genes in other eubacteria as well as genes unique to Borrelia have been found near the telomeric regions (Casjens and Huang, 1993, 1994). Rapid reannealing of the chromosome after dénaturation suggests that the ends of the telomeres are hairpins similar to those of the characterized 16 kb linear plasmid, see below (Ferdows and Barbour, 1989; Barbour 1993).

The rRNA genes, 23S (rrl), 5S (rrf) and 16S (rrs), have been cloned on the basis of their sequence homology to the E. coli genes, and mapped to the Borrelia chromosome, see Figure 2. Normally, rRNA genes are encoded by opérons in the order; promoter, 16S, 23S, 5S. This results in the production of equimolar amounts of the three rRNA species. The number of rRNA opérons varies from one to ten, in different bacterial species. In general, fast-growing bacteria have more copies than slow-growing bacteria (Krawiec and Riley, 1990). B. anserina, B. turicatae and B. hermsii have a single operon with the usual organization (Schwartz, et al., 1992), but in B. burgdorferi, rrl and rrf are tandemly duplicated, while the single rrs gene is located more than two kb upstream of the first rrl gene (Schwartz, et a i , 1992; Davidson et al., 1992; Fukunaga, et al., 1992).

In bacteria with circular chromosomes, like Escherichia coli, Bacillus subtilis and Pseudomonas aeruginosa, the replication of the chromosome is initiated and terminated at specific regions denoted the origin of replication (ori) and the termination of replication (ter). The initial step in replication is the binding of the protein DnaA to the DnaA boxes, which consist of at least four characteristic 9 bp repeats, usually located close to the dnaA gene (Smith, et al., 1991; Holz, et a l , 1992; Ogasawara, et al., 1990; Ogasawara and Yoshikawa, 1992). Subsequently, DnaB and DnaC bind to the DnaA/DNA complex, and replication proceeds bidirectionally around the chromosome until the ter sites are reached. Other genes important for the replication are dnaN encoding the DNA polymerase III, and gyrA and gyrB encoding DNA gyrases. In B. burgdorferi s. I these three genes map together with the dnaA gene next to the rRNA genes in the middle of the chromosome (Old et a l , 1992, 1993a, b; Casjens and Huang, 1993). In other bacteria, gidA, involved in cell division, is also situated by the chromosomal origin. In B. burgdorferi gidA has been mapped near the left end. This suggests that the Borrelia chromosome is either replicated bidirectionally

18

from the centre of the chromosome out to the ends, or from the left end towards the right end. No cluster of DnaA boxes have been identified in B. burgdorferi to date. The two putative ori regions are indicated in Figure 2.

Comparison of the genetic maps of the chromosomes from B. burgdorferi s. s., B. afzelii and B. garinii show a significant degree of conservation of the position of the mapped genes, indicating a very stable chromosome (Saint Girons et al., 1994).

2.2 Linear plasmidsThe first bacterial linear plasmids were described in Streptomyces rochei in 1979 (Hayakawa et al., 1979). Linear plasmids have since been detected in at least 10 other species of Streptomyces, as well as in Nocardia opaca, Ascobulus immersus and Rhodococcus fascians (Kalkus et al., 1990; Francou, 1981; Crespi et al., 1992; Hinnebusch and Tilly, 1993). All members of the genus Borrelia contain linear plasmids (Plasterk et al., 1985; Marconi et al., 1993c). They account for most of the extrachromosomal DNA in B. burgdorferi cells. The B. burgdorferi strain B31 harbours four linear plasmids of 16, 29, 38 and 49 kb, as well as two circular plasmids of 27-30 kb (Barbour and Garon, 1987, Saint Girons et al., 1994). The most studied of the plasmids are the linear 16 kb and 49 kb plasmids. The ends of these plasmids are covalently closed, which can be demonstrated by electron microscopy (Barbour and Garon, 1987; Barbour, 1988; Hinnebusch et al., 1990). Reannealing of the plasmid after alkaline dénaturation is very fast, but this feature is destroyed if the plasmid has been pretreated by SI nuclease. When examining the denatured plasmid, single stranded circles of about 100 kb were observed. Chemical sequencing of both ends of the 16 kb plasmid and the left end of the 49 kb plasmid has revealed 19 bp of almost perfect inverted repeats proceeded by four unpaired nucleotides forming hairpin loops at the ends (Hinnebusch et al., 1990; Hinnebusch and Barbour, 1991; Barbour 1993; Hinnebusch and Tilly, 1993). Some animal viruses have the same kind of covalently closed ends, but, to date similar plasmids have not been observed in any other prokaryotes (Ploeg et al., 1984; Bergström et al., 1990). The plasmids are structurally similar to a vaccinia virus, African swine fever virus, which is carried by the Omithodoros ticks, and also harbours B. duttoni (Hinnebusch and Barbour, 1991; Hinnebusch and Tilly, 1993).

In B. hermsii, the amount of the different DNA species has been studied under various growth conditions (Kitten and Barbour, 1992). The copy numbers of the chromosome and the two linear plasmids containing the silent and expressed forms of

19

one of the vmp genes, vmp7, during in vitro culturing and growth in mice were determined. Spirochetes grown in vitro had approximately 4 chromosomes per bacteria.The same value was found for the plasmid harbouring the silent vmp gene, while the expression plasmid was found in about 6 copies per cell. When grown in mice, with a generation time of 6-8 h compared to 10-12 h in vitro, the copy numbers of the replicons changed both absolutely and relative to one another. The number of chromosomes and the expression plasmids was 15, while the silent storage plasmid was present at about 7 copies per cell. This difference could be a result of the higher gene dosage required by the expressed genes. In B. burgdorferi, only the relative copy numbers of the chromosome and the linear plasmids have been determined. The 16 and 49 kb plasmids have copy numbers of approximately one per chromosome, suggesting that the replication and partitioning of the chromosome and the plasmids are tightly coupled (Hinnebusch and Barbour, 1992).

Méthylation of DNA is considered to be an important regulatory mechanism involved in, for example, protecting the DNA from degradation, DNA replication and recombination, gene expression and development. Some but not all B. burgdorferi strains possess an adenine méthylation system. There is no obvious correlation between strains capable of méthylation and infectivity (Norton Hughes and Johnson,1990).

B. burgdorferi lose their infectivity after many passages in culture media. Some of the plasmids, both linear and circular, are also lost after prolonged cultivation, which has led to the conclusion that at least one of the genes responsible for infectivity should be plasmid encoded (Schwan et al., 1988; Simpson et al., 1990). Recently, an 18 kD protein, EppA, apparently only expressed during infection, was cloned from a 9 kb circular plasmid present only in low-passage, virulent B. burgdorferi strains (Champion et a i , 1994).

3. Pathogenicity

3.1 Bacterial virulence factorsAn infection is established when the body is colonized by a bacterium capable of causing disease. Virulence and pathogenicity describe the ability of a bacterium to cause disease. A virulence factor is a bacterial product or strategy contributing to virulence or pathogenicity (Salyers and Whitt, 1994). As was described in section 1.5,

20

Borrelia burgdorferi can be isolated from various organs in patients suffering from Lyme disease (Steere, 1989; Karlsson, 1990). Natural infections are initiated by a tick bite, where the spirochetes are deposited in the skin (Shih et al., 1992). From this position, the spirochetes enter the circulatory system and are spread to distant parts of the body, where they can invade different organs and cause infection. Since they are able to do this, the spirochetes probably possess several virulence factors.

A very important bacterial virulence factor is adhesion to host cells. This is required both for the colonization of body surfaces, and for the subsequent invasion of the host cells. Different types of adhesion are known. Pili are structures extending from the bacterial surface, which usually bind to carbohydrate moieties of host cell glycoproteins or glycolipids. As the pili structure is quite fragile, the contact formed between the bacterium and the host cell is easily broken. Afimbrial adhesins are bacterial surface proteins that mediate a tighter association between the bacterium and the host cell. Many bacteria bind to extracellular host proteins that, in turn, attach to host cells (Isberg, 1991). Bacteria can also adhere to surfaces by forming biofilms consisting of bacteria held together by a polysaccharide matrix (Salyers and Whitt,1994).

The invasion of host tissue is accomplished by bacterial invasins. The invasins are surface proteins that cause actin rearrangements in normally nonphagocytic host cells that lead to pseudopod formation and engulfment of the bacterium in a phagocytic vesicle (Isberg, 1991). The bacterium can remain there or escape the phagocytic vacuole and proliferate in the cell cytoplasm. Through the M cells (macrophage like cells found in different membranes), the bacteria can also pass through the mucous membrane to the underlying tissue (Finlay and Falkow, 1989).

Different mechanisms have evolved which allow certain organisms to evade the host’s immune system. Capsules minimize complement activation and prevent phagocytosis (Cross, 1990). To avoid recognition by the host’s defence system, some bacteria cover the bacterial surface with host proteins, others can undergo antigenic variation (see below) (Finlay and Falkow, 1989).

3.2 Antigenic variationMany pathogenic organisms change their surface antigens in order to escape the host’s immune response. Antigenic variation is used by both bacteria and protozoa to promote the persistence in the host. Neisseria gonorrhoeae has the ability to switch on and off the expression of pili (Bergström et al., 1986; Swanson et al., 1986). Different strains

21

of gonococci also show extreme antigenic variation in both pili and outer membrane proteins. The protein subunit of pili, pilin, consists of constant, variable and hypervariable regions, and genetic rearrangements and recombinations occurring in the repertoire of pilin genes form the basis of the antigenic variation (Sparling et al., 1986; Jonsson, 1991).

The causative agent of sleeping sickness, Trypanosoma brucei, is a unicellular protozoan infecting the bloodstream of mammalian hosts. T. brucei survive by changing their surface coat, which consists of an abundant glycoprotein, called variant surface glycoprotein, VSG. The genome contains as many as 1 000 different vsg genes. Usually only one vsg is expressed at any given time, and about 10 copies of the protein completely cover the outer membrane of the trypanosome. Only genes located at special expression sites adjacent to the chromosomal ends are expressed. The rest of the vsg genes are stored internally on the chromosomes as silent copies. The switch from the transcription of one vsg to another is accomplished by activating a new gene in an expression site. The situation is complicated by the fact that there are more than one potential expression sites, but only one is normally activated at any time (Donelson, 1989).

The bacterial counterparts of the trypanosomes are the etiology agents of relapsing fever, of which B. hermsii is the most studied. The variable proteins in B. hermsii are lipoproteins, called variable major proteins, Vmp’s. The repertoire of genes is probably not as large as in the trypanosomes, but 25 different serotypes have been isolated from a single bacteria (Stoenner et a l , 1982). The organization of the genes is similar, with expressed genes situated at an expression locus near the end of a 28 kb linear plasmid, and silent copies are stored internally on the linear plasmids of around 30 kb in size (Stoenner et a l , 1982; Plasterk et a l , 1985). Formerly silent vmp genes replace the actively expressed gene by non-reciprocal recombination between the linear plasmids (Meier et a l , 1985; Barbour, 1989).The silent and expressed genes of serotypes vmp7 and vmp21 from the B. hermsii strain HS1 have been cloned and sequenced. The upstream region of expressed vmp7 and vmp21 are identical, while the equivalent regions differ when corresponding silent genes are compared (Plasterk et a l , 1985; Barbour, 1989; Burman et a l , 1990). The silent copies lack promoters, and can be activated by a recombination event that places them downstream of a promoter (Barbour et a l , 1991).

There are many indications that antigenic variation occurs in B. burgdorferi as well. Spirochetes have been isolated from an AC A patient who had had the disease for more than 10 years (Asbrink and Hovmark, 1985). Some patients with Lyme

22

borreliosis have developed new antibodies a year or more after the initial infection (Craft et al., 1986). Clonal polymorphism has also been observed in vitro. A European strain of B. burgdorferi isolated from human cerebrospinal fluid (CSF) contained a 32 kD protein that became more dominant after being passaged in culture for an unspecified length of time (Wilske et al., 1986). In another study, a B. burgdorferi strain lost the expression of OspB after only 11-15 passages in BSKII. As the initial population was not clonal, it cannot be excluded that a strain lacking OspB was selected from a mixed Borrelia population. Interestingely, Western blot analysis with two different monoclonal antibodies indicated that changes in certain epitopes of the OspB protein also occurred (Schwan and Burgdorfer, 1987). Clonal B. burgdorferi strains grown in vitro have also shown clonal polymorphism of OspB. Subsequent cloning of different passages from an originally clonal B. burgdorferi, HB19, showed variations both in the size and expression of the OspB protein. These changes at the protein level could not be explained by large DNA rearrangements, as no differences between the distinct clones were detected by Southern or Northern blot analysis (Bundoc and Barbour, 1989). Rearrangements of the ospA and ospB genes during growth in vitro creating chimaeric gene fusions were later shown to occur by homologous recombination, either within or between the two genes. Osp variation was also shown to arise from nonsense mutations and sequence divergence (Rosa et al., 1992).

Antigenic variation during B. burgdorferi infection has been contradicted by recent studies (Barthold, 1993; Jonsson and Bergström, Paper III this thesis). Mice were immunized intradermally, and then the infection was erradicated by antibiotic treatment. The treated mice were immune to a challenge by either the initial isolate, or cultures, as well as biopsies, from animals where the infection was allowed to proceed for as long as 180 days before antibiotic treatment (Barthold, 1993). The persistence of viable spirochetes despite a functional immune response shows that the spirochetes avoid the immune defence in some way. The fact that they were immune to not only the early, but also the late isolates suggests that surviving the immune response is achieved by some other mechanism than antigenic variation.

3.3 B. burgdorferi s. I outer membrane proteinsAs mentioned above, 45 to 62 % of the outer membrane in Borrelia consists of protein. More than 100 polypeptide species have been detected by two-dimensional SDS- polyacrylamide gel electrophoresis (SDS-PAGE) (Luff et al., 1989). Labeling with

23

125] or biotin has identified 13 major surface proteins with apparent molecular weights of 22, 24, 29, 31, 34, 37, 39, 41, 52, 66, 70, 73, and 93 kD. Some of these proteins have been named Osp proteins. So far, six Osp proteins have been identified. They are designated OspA, B, C, D, E and F. The common features of the Osp proteins are that they are all surface exposed, have a hydrophobic leader sequence, and a consensus cleavage sequence(L-X-Y-C) recognized by signal peptidase II. They are also all lipoproteins, and can be labelled by [3H] palmitate (Brandt et al., 1990). The lipid moiety is probably required to anchor the proteins in the fluid outer membrane. Analysis of the fatty acids in B. burgdorferi show that the lipoproteins contain palmitate (Belisle et al., 1994). In E. coli, lipoproteins are synthesized as precursor proteins. The prolipoprotein is covalently modified with diacylglycerol via a thioether linkage to cysteine and with fatty acid by an amide bonde. Thereafter, the signal peptide is cleaved off and the apolipoprotein is further modified at the ot-NH2 group of the glyceride-cysteine residue with palmitate by N-acyl transferase. The same mechanism is probably used in other bacteria, like B. burgdorferi s. I. as well (Hayashi and Wu, 1990).

The most studied proteins in B. burgdorferi are the OspA and OspB proteins, of 31 and 34 kD in size respectively. The osp genes from the American type strain B31 were the first to be cloned and sequenced (Howe et al., 1985, 1986; Bergström et al., 1989). The genes are located in an operon, separated by just a few bases, and transcribed as one transcriptional unit from the 49 kb linear plasmid. The ospAB sequences contain signal sequences similar to those for E. coli lipoproteins.

When the protein profiles from different isolates are compared, the Osp proteins are quite heterogeneous. Both the size and reactivity with various antibodies differ. The OspB proteins are particularly heterogeneous (Barbour et al., 1984; Lane and Pascocello, 1989). The OspA proteins are homogeneous in American isolates, but show a higher degree of polymorphism in European isolates (Barbour et al., 1984, 1985; Barbour and Schrumpf, 1986; Bundoc and Barbour, 1989; Wilske et al., 1985, 1986, 1988b). Most strains express both OspA and OspB proteins, but some isolates express only one of the proteins. Others express neither OspA nor OspB, but express a smaller, 21 to 25 kD protein, OspC, instead (Barbour et al., 1985; Wilske et al., 1988b). In European isolates, more than 90 % of B. burgdorferi isolates express OspA, and about 40 % express OspC (Wilske et al., 1988b). Like the OspA and OspB proteins, OspC is heterogeneous and shows immunological and molecular polymorphism in different isolates (Wilske et al., 1993). The ospC gene was found in all strains tested in one study despite the fact that it is not expressed in all of these

24

strains (Marconi et al., 1993a). In contrast to the ospA and ospB genes, the ospC gene is located on a circular plasmid, 26-27 kb in size in B. burgdorferi (Sadziene et al., 1993a; Marconi et al., 1993a). Amino acid sequencing demonstrated that the N- terminus is blocked. This, together with nucleotide sequence homologies to the cleavage site of E. coli signal peptidase II also found in OspA and OspB, suggest that OspC is a lipoprotein (Fuchs et al., 1992).

OspC is found not only in B. burgdorferi. It was first shown that an OspC specific polyclonal sera recognized a 20 kD Vmp in B. hermsii (Wilske et al., 1993). Further studies revealed ospC homologues also in B. coriaceae, B. anserina, B. turicatae and B. parkeri. Interestingly, all these genes are encoded on linear plasmids, in contrast to the localization in B. burgdorferi (Marconi et al., 1993b). OspC, together with Vmp’s of 19 to 22 kD, form a genus-wide family of 20 kD proteins (Carter et al., 1994).

OspD is a 28 kD lipoprotein expressed only in low-passage B31 strains, but absent in most high-passage, nonvirulent strains tested. The ospD gene is located on a 38 kb linear plasmid, usually lost during prolonged in vitro cultivation (Norris et al., 1992).

OspE and OspF, 19 and 26 kD in size respectively, are the most recent Osp proteins to be reported. Like the ospA and ospB genes, their genes are arranged in tandem as one transcriptional unit, with ospE first, and separated by 27 bp. The whole operon hybridizes to a 45 kb plasmid in B. burgdorferi N40 (Lam et al., 1994).

Using an antiserum from a patient with cutaneous manifestations, another surface exposed, 27 kD lipoprotein, P27, has been isolated from the European B. burgdorferi strain B29 (Reindl et al., 1993). The p27 gene is expressed on the largest, 55 kD, linear plasmid. In B. burgdorferi B31, the gene for P27 is missing.

3.4 Regulation of the Osp proteinsThe expression of OspA and OspB proteins appears to be regulated at the transcriptional level. A B. burgdorferi isolate named CA-11-90 was cloned by plating, and then inoculated as single colonies into BSKII liquid medium. One clone, CA-11 2A, did not express detectable amounts of OspA or OspB on Coomassie stained protein gels. Instead, a major 24-kDa protein could be seen. After 40 in vitro passages, OspA and OspB were detectable on the stained gels. At the same time, the 24-kDa protein decreased in amount. Northern blot analysis using a probe covering the whole ospAB operon showed no detectable ospAB transcript in the clone lacking OspA and

25

OspB, while a strong signal could be seen in the extract from the high passage clone (Margolis and Rosa, 1993). This suggests that the lack of expression is primarily due to lack of transcription of the ospAB operon.

B. burgdorferi B31 usually lacks the expression of OspC, but mutant clones of B31 which had lost a 16 kb linear plasmid, lpl6, started expressing the OspC protein. This together with the finding that only strains lacking lpl6 homologous DNA expresses the protein and vice versa, suggests that OspC expression is negatively regulated by some factor from this plasmid (Hinnebusch and Barbour, 1991; Sadziene et al., 1993a).

It was also found that the plating efficiency was reduced dramatically when Ipl 6 was lost, suggesting that something on this plasmid is essential for growth on solid media (Sadziene et al., 1993a). A conflicting result was reported in another study where OspC expressing clones were obtained from a nonexpressing isolate of B. burgdorferi strain Pka2 by culture on solid agar (Wilske et al., 1993b). These findings suggest that:a) The negative regulation of OspC expression from lpl6 can be turned off without losing the ability to grow on solid media.b) Something in the solid media inhibits the negative effects of the lpl6 factor.c) Some other factors induce the expression of OspC.d) The regulatory mechanisms are different in the two strains.

Sequence comparison between B31 and an OspC expressing strain, 2591, showed that a 54 bp fragment just upstream of the -35 and -10 consensus promoter sequences was absent in strain B31. This fragment may be an enhancing element, promoting the transcription of the ospC gene (Padula et al., 1993). Further sequencing of the 27 kb circular plasmid showed that the guaA and guaB genes were located upstream of the ospC gene, transcribed in the opposite direction. Those genes are involved in the purine synthesis. Mammalian blood has very low purine levels, while one of the major waste products in ticks is guanine. Therefore, the guaAB operon is probably only expressed in mammals. This may also influence the expression of ospC (Rosa and Margolis, 1994), and thereby the expression of the ospAB operon indirectly (see discussion in section 5.5).

3.5 PathogenicityIt now appears that variations in symptoms of Lyme disease are related to the group of Borrelia that caused the infection. Seven patients with ACA all reacted strongly with a

26

skin isolate, now classified as B. afzelii (Wilske et al., 1988a). In another study, an OspA serotyping system was developed (Wilske et a i, 1993a). Seven different serotypes were found. Serotype 1 corresponds to B. burgdorferi s. s., serotype 2 to B. afzelii, and serotypes 3 to 7 to B. garinii. As in the previous study, serotype 2 was most prevalent among European skin isolates. In the heterogeneous B. garinii group, serotypes 4 and 5 were only isolated from patients, while serotype 6 was the most frequent isolate found in ticks. Assous et al performed a study where the serological response of Lyme borreliosis patients was analyzed by Western blot using strains of the three distinct genomic species. They found that about half of the patients suffering from meningoradiculitis showed preferential reactivity with B. garinii. All patients with AC A reacted mainly with B. afzelii, and 50 % of the patients with arthritis reacted preferentially with B. burgdorferi s. s. (Assous et al., 1993). It was only in the arthritis group that a strong reaction against the OspA and OspB proteins was found. In another study, Western blot analysis with patient sera on antigens from the three different genomic groups showed that patients with meningopolyneuritis react with more antigens of B. garinii than the other two groups. Patients with acrodermatitis or arthritis had more antibodies to B. afzelii antigens (Dressier et al., 1994).

Persistent infection of SCID mice with B. turicatae, which normally causes relapsing fever, gave rise to symptoms that resembled human Lyme borreliosis (Cadavid et al., 1994). In that study, two different serotypes of B. turicatae were isolated and further characterized. The only discernable difference between the two variants, was the size of the Vmp proteins expressed. Interestingely, the mice developed different symptoms depending on what serotype they were infected with, suggesting that the outer membrane proteins are indeed important for the clinical manifestations.

The later stages of infection by B. burgdorferi are characterized by the persistence of the organisms despite a strong anti-borreliocidal immune response. This suggests that the spirochetes invade tissues where they are protected. Borrelia may enter and survive inside both human umbilical vein endothelial (HUVE) cells and human skin fibroblasts (Ma et al., 1991; Klempner et al., 1993). The importance of the skin in natural B. burgdorferi infections was established in another study. Excision of the tick attachment site as long as two days after detachment of fully engorged, infected I. dammini was shown to totally abolish the infection. These results suggest that the tick-inoculated spirochetes proliferate where they are deposited, doubling several times during the first two days. The bacteria then remain close to this site for another week, until they start disseminating in the skin and to other parts of the body

27

(Shih et a i, 1992). In this study, it is also implied that no immune response is elicited until the dissemination of spirochetes starts. The saliva of ticks contains anti­inflammatory agents, that may create a protective environment suitable for proliferation of the spirochetes and the establishment of an infection (Ribeiro et al., 1985). That ticks are important for the infection is further emphasized in a study showing that mice infected by the natural route are substantially more infective for ticks than experimentally inoculated animals (Gem et al., 1993).

3.6 Adhesion of Lyme disease Borrelia to human mammalian cellsAs discussed above, spirochetes can adhere to and invade different cell types. Most binding studies of B. burgdorferi have been performed with HUVE cells. Specific adherence was shown to be time and temperature dependent. Pretreatment of the bacteria with heat, immune serum or monoclonal antibodies directed against the outer surface proteins results in reduced attachment, indicating that the binding involvs borrelial surface proteins (Thomas and Comstock, 1989). The spirochetes are not only capable of specific binding to the eukaryotic cells, but also penetration (Comstock and Thomas, 1989). The passage occurs mainly through the cytoplasm of the host cell, and may take place by directional transcytosis or parasite-directed endocytosis. Reorganization of actin and cytoskeletal filaments around the intracellularB. burgdorferi was observed by transmission electron microscopy (Comstock and Thomas, 1991). Further studies have shown that OspA is probably involved in the adhesion, while the OspB as well as the flagellin are important for the invasion. Binding of the FAb fragments of one specific monoclonal antibody directed against OspA reduced the binding by 69 %. FAb fragments of other monoclonal antibodies did not interfere with the binding (Comstock et al., 1993). B. burgdorferi probably has more than one adhesin, as total inhibition of the binding could not be accomplished. This is also suggested by the analysis of an Osp-less mutant totally lacking Osp proteins. This mutant retained some ability to bind HUVE cells but less than the wildtype (Sadziene et al., 1994).

An Sh2-82 strain containing a mutant OspB protein showed a greatly diminished capability of penetrating HUVE cells. It also had reduced infectivity in SCID mice, indicating that this is an important virulence factor (Sadziene et al., 1993b). Analysis of a spontaneous non-motile flagellin mutant showed a 95 % reduction in invasion, displaying that motility is an important factor for spirochetal invasion (Sadziene et a i ,1991).

28

The spirochetes do not only bind to epithelial cells, but also to glial cells and cells of glial origin (Garcia-Monco et al., 1989), as well as human platelets (Cobum et al.,1993). They have been shown to specifically bind to galactocerebroside (Garcia- Monco et al., 1992).

3.7 Animal models and vaccination studiesMost animal studies of B. burgdorferi have been performed in mice. There are several reasons for this. Mice constitute the main reservoir for B. burgdorferi in nature, and persistent infections with Lyme disease symptoms resembling those found in human patients, can be established. The immune system of mice is very well characterized, which is helpful in studies of the pathogenesis of Lyme disease. Several inbred mouse strains with specific déficiences in their immune systems further facilitates these studies. These immune deficient mice are also very suitable for vaccine studies (Simon et al., 1991b). Other animals used in experimental B. burgdorferi infections are rats (Barthold et al., 1988), Syrian hamsters (Johnson et al., 1984a), rabbits (Komblatt et al., 1984), dogs (Appel et al., 1993), Rhesus monkeys (Philipp et al., 1993) and Guinea pigs (Sonnesyn et al., 1993).

Most of the vaccination efforts have focused on OspA. Both native OspA from B. burgdorferi ZS7 and B31, as well as recombinant ZS7 OspA, can generate hyperimmune sera in immunocompetent mice. Transfer of immune sera or antibodies specific for OspA to severe combined immunodeficient (SCID) mice, totally protects from challenge with strain ZS7 ( Schaible et al., 1990; Simon et al., 1991). In another study, both passive and active immunization with recombinant OspA from B. burgdorferi N40 conferred protection upon subsequent challenge with N40 in immunocompetent C3H mice (Fikrig et al., 1990). The active immunization with purified recombinant N40 OspA also conferred long-lasting protective immunity. All challenged animals were still immune 150 days after the last boost (Fikrig et al., 1992a). The lipid moiety is very important for eliciting a strong immune response. In experiments using both lipidated and de-lipidated OspA, it was shown that the lipid moiety was needed to activate the immune system, while the antibodies produced were directed against the peptide moiety (Erdile et al., 1993). 100 to 1000 fold increases in protective antibody levels were found when the lipidated OspA was fused to the Mycobacterium bovis strain Bacille Calmette-Guerin (BCG) and expressed as a membrane associated lipoprotein (Stover et al., 1993). Not only OspA, but also OspB, OspC and a 39 kD protein, P39, can induce protection by active immunization

29

(Schmitz et al., 1991; Fikrig et al., 1992b; Probert and LeFebvre, 1994; Norton Hughes et al., 1993). Moreover, borreliae within ticks which engorge on mice immunized with OspE and OspF get partially destroyed (Nguyen et al., 1994).

To be protective, it is necessary that the vaccination is performed before or at the time of initiation of infection. Immune sera administered before or at the time of inoculation can protect against the challenge, but not sera administered as little as 24 h after the start of infection (Barthold and Bockenstedt, 1993). Active immunization with OspA after an infection has been established cannot eliminate the infection, but appears to accelerate the resolution phase of the disease (Fikrig et al., 1993). A recent study showed that more severe arthritis was induced in hamsters vaccinated with whole-cell preparations of formalin-inactivated spirochetes than in unvaccinated animals if the challenge was made before sufficient levels of borreliacidal antibodies had developed. Even after high levels of specific antibodies could be detected, challenge with heterogeneous strains induced severe arthritis (Lim et al., 1994).

There are also problems in using only the OspA protein as a vaccine, for example the heterogeneity displayed in different strains (Marconi et al., 1993a; see also section 3.3). Vaccination with the OspA protein from one strain only confers immunity to homologous strains. This is a problem especially in Europe. To circumvent this problem, chimeric proteins containing conserved regions from the OspA of different genomic groups could be used. Another possibility is to construct recombinant proteins consisting of parts from several different immunogenic proteins. Trials of this kind have been initiated (Luft et al., 1994).

So far, the Osp proteins, and in particular the OspA protein, is the best vaccine candidate. The importance of the Osp proteins in the virulence of B. burgdorferi s. /. is clear. Therefore, a molecular and structural characterization of the Osp proteins will be beneficial to our understanding of the pathogenicity of this spirochete. More knowledge will aid in the development of better diagnostic and therapeutic methods. It may also be useful in the development of a reliable human vaccine against Lyme disease.

AIMS OF THE PRESENT WORK

This thesis has examined the outer surface proteins of the Lyme disease Borrelia. The aims have been to understand the heterogeneity of Osp proteins between different strains, the regulation of Osp expression and their role in the pathogenesis of this spirochete.

33

SUMMARY AND DISCUSSION OF RESULTS

4. Heterogeneity of the osp operon (Paper I)

4.1 PrologueWhen I started my work, the osp operon from the North American B. burgdorferi strain B31 had been cloned and sequenced (Howe et al., 1985; Bergström et al.,1989). The first part of my work was to continue the examination of the Osp proteins and their genes in other Borrelia burgdorferi strains. The primary objective was to look for similarities and differences in distinct isolates, in order to gain insight into the pathogenesis of this organism. In addition to B31, the strains AC AI and Ip90 were chosen for this study. ACAI is a Swedish strain isolated from a patient with acrodermatitis chronicum migrans (Åsbrink et al., 1984b). Ip90, like B31, is a tick isolate, but which originates from the Khabarovsk territory in the Asian part of Russia, on the border with China (Kryuchechnikov et al., 1988). These strains were chosen because the bacteria originated in geographically distinct regions, and were isolated from different sources. When total protein extracts from the three strains B31, ACAI and Ip90, were separated by SDS-PAGE, the various strains all had different protein patterns. Different reactivity of the strains to monoclonal antibodies raised against the Osp proteins of strain B31 demonstrated that the Osp proteins differ antigenically between the strains.

4.2 Cloning of the osp operon of ACAI and Ip90A X phage gene library was constructed by EcoRI digestion of genomic ACAI DNA. The ACAI osp operon was cloned from this library by probing with labelled oligonucleotides synthesized from the B31 ospAB DNA sequence. A positive clone containing the entire ospAB operon was isolated, and subcloned into pUC18.

Two different pUC18 plasmid gene libraries were constructed from genomic Ip90 DNA. From the partial Hindlll digested gene library, a clone containing the entire osp operon except for the beginning of osp A and the control region was obtained using an ospA PCR fragment as probe. Clones containing the missing regions were obtained by screening an EcoRI gene library with an oligonucleotide homologous to the beginning of osp A.

34

Most of the sequence data was obtained by sequencing deletion libraries made on the AC AI clone and the shorter of the Ip90 clones. To obtain the complete sequences, internal oligonucleotides were used as primers.

4.3 Genetic organizationThe ospA and ospB genes are organized in one operon, separated by 9 bp. ospA is located at the 5’ end. They are transcribed as one transcriptional unit. The transcriptional start site was shown for B31 to be at position +1 in Figure 3, Paper I. Based on sequence homologies, the same transcriptional start point was assumed for the two other strains. This has subsequently been confirmed for AC AI (Paper II).

Southern blot analysis showed that the osp operon is located on a 49 kb linear plasmid, lp49, in B31. In AC AI and Ip90, the operon hybridizes to homologous plasmids, slightly larger in size (Figure 4, Paper I). The overall plasmid content of the three strains is very different. In B31, only the linear plasmids lpl 6, lp29 and lp49, are observed under the conditions used in the pulsed field gel electrophoresis. In both ACAI and Ip90, several additional plasmids of varying sizes are detected. The only plasmid that is common to all three strains is the one harbouring the osp operon, even if the size of it varies slightly between the strains.

4.4 Sequence comparisonThe only available sequences at the time were the B31 osp operon, and the ospA sequences from B. burgdorferi strains N40, isolated from a tick in the USA (Fikrig et al., 1990), and ZS7, a German tick isolate (Wallich et al., 1989). Comparison of the ospA and ospB DNA sequences and the deduced amino acid sequences gives the percentage of identity shown in Table 1 and Table 2. The previously sequenced ospA genes from B31, ZS7 and N40 are almost identical, there are only single base pair exchanges between the strains. The osp sequences from ACAI and Ip90 show that the Osp proteins are indeed heterogeneous, which is in agreement with earlier immunochemical findings (Wilske et al., 1988b). As can be seen in Figure 3, Paper I, the osp A genes are very conserved in the 5’ region of the gene. The deduced amino acid sequences are compared in Figure 5, Paper I. The N-terminal end of OspA is totally identical in all of the strains studied. This indicates that the beginning of OspA is functionally and/or structurally very important. This conservation of the N-terminal region is not found in OspB, where differences occur even in the leader sequence.

35

Table 1. Percentage of the identity of the asp/l/OspA sequences between the strains

ACAI Ip90 ZS7 N40B31 &5777b 86/19 99/99 >99/99ACAI - 86/ 81 85/11 85/11Ip90 - 86/19 86/19ZS7 - 99/99a) nucleotide sequence identityb) amino acid sequence identity

Table 2. Percentage of the identity of the

ACAI Ip90B31 79766h 79/67ACAI - $7/68

a) nucleotide sequence identityb) amino acid sequence identity

In Figure 3, a cladiogram of the DNA sequences is shown. Phylogenetically, the osp A and ospB genes first diverge into two branches. Within each branch, the strains can be placed into three groups. N40 and ZS7 belong to the B31 group. The ACAI and Ip90 strains are slightly more related to each other than to the B31 group. Attempts to systemize different strains of B. burgdorferi based on other criteria have been made previously (Postic et a i, 1990; Adam et al., 1991; Rosa et al., 1991), and those results are quite consistent with the findings presented here. As mentioned in the introduction, B. burgdorferi has now been split into three genomic species, i. e. B. burgdorferi s. s., B. afzelii and B. garinii (Baranton et al., 1992; Canica et al., 1993). ACAI has been placed into the B. afzelii group, Ip90 into the B. garinii group, and B31, N40 and ZS7 are now classified as B. burgdorferi s. s.When comparing the OspA and OspB proteins to each other, more than 50 % similarity is observed (Bergström et a i , 1989). This indicates that the osp operon has been formed by a duplication of a primitive osp gene. Because of the high similarity,

36

157

41

- ZS7-ospA

N40-ospA

B31-ospA

23

30

32

ACAI-ospA

Ip90-ospA

42

37 ACAI-ospB

36 Ip90-ospB

39 B31-ospB

Figure 3. A phylogenetic tree made from the ospA and ospB sequences of strains B31, AC AI and Ip90, and the ospA sequences from strains N40 and ZS7. Parsimony analysis was performed using the PAUP program for the Macintosh, version 3.0s (Swofford, 1990). The numbers indicate the number of nucleotides diverging between the compared subjects.

37

this is probably quite a recent evolutionary event. The cladiogram in Figure 3 clearly shows that this duplication must have taken place before the strains started to diverge into different species.

As all the ospA DNA sequences published at the time came from strains belonging to B. burgdorferi s. s., it was speculated that ospA was very conserved at the DNA level although Coomassie stained protein gels and Western blot data indicated large variations both in size and immunoreactivity (Fikrig et a i , 1990; Barbour et a i , 1985; Wilske et al., 1986,1988). The results herein show that variation occurs at the DNA level too. This has implications for the use of the osp genes or proteins in diagnostic methods as well as in vaccinations. PCR is becoming a very popular method to use in diagnostics. If the osp genes are to be used as target genes, it is important to find regions that are highly conserved between the strains. The most conserved regions are definitely found in the control region and in the beginning of ospA. Conserved regions suitable for the second primer can also be found further downstream in both the ospA and ospB genes, although these stretches with homology are not as long as those previously mentioned.

Using the Osp proteins as vaccines will probably require a mixed vaccine, using Osp proteins, preferrable OspA, from the different genomic species in order to be universally protective. In the USA, most strains isolated belong to B. burgdorferi s. s, and it may therefore be possible to use a simpler form of vaccine than in Europe where all three genomic species are present.

5. Expression of the osp operon (Paper II)

5.1 Why clone the osp operon from B. afzelii FI?One approach to the examination of the regulation of proteins is to compare systems with different expression patterns. The Osp proteins are differentially expressed in the two B. afzelii strains ACAI and FI. ACAI expresses both OspA and OspB proteins, and the ACAI ospAB operon was characterized in Paper I. FI is a Swedish tick isolate, previously reported to lack the expression of OspA and OspB (Barbour and Schrumpf, 1986). Instead, it expresses the smaller OspC protein in large amounts, as seen on Coomassie stained protein gels. Studying the ospAB operon from strain FI in more detail could give information concerning the regulation of expression of Osp proteins in Lyme disease Borrelia.

38

5.2 Characterization of protein expression by Western blot analysisTo analyse the amount of and structure of the OspA and OspB proteins we have employed Western blot analysis. This required a panel of antisera directed against the Osp proteins. We had access to several monoclonal antibodies raised against B. burgdorferi s. s. antigens. However, these reagents, except for the OspB MAb 84C (see Paper IV), have a low affinity for the Osp proteins of B. afzelii. Therefore, a specific rabbit polyclonal anti-OspA sera was produced by immunizing a rabbit with purified, recombinant OspA protein cloned from the B. afzelii strain ACAI.

Whole cell extracts from strains B31, ACAI and FI were separated on a 12.5 % SDS-PAGE. After staining with Coomassie brilliant blue, major OspA and OspB proteins were detected in B31 and ACAI. In FI, a major protein of approximately 25 kD in size is observed. No major OspA and OspB proteins are detected in FI, but there are some faint bands in the size range of the Osp proteins of ACAI.

In the first Western blot experiments, the anti-OspA MAb 184.1 (Jiang et al.,1990) and the anti-OspB specific MAb 84C (Comstock et al., 1993) were used. Both antibodies reacted with the OspA and OspB proteins of B31 and ACAI. Surprisingly, the 84C antibody reacted with a protein of the size of OspB in strain FI as well. The OspA antibody gave no signal with strain FI. This indicated that the FI strain expresses the OspB protein in low amounts but not the OspA protein. Using the rabbit anti-OspA polyclonal sera, a signal was obtained from the FI strain. Compared to the OspA signal from ACAI, less OspA protein is found in strain FI, but it varies in different preparations. These results show that the FI strain is capable of expressing the osp AB operon in variable amounts.

The differences in the amount in Osp proteins found could be due not only to differences in expression, but also in protein stability. Therefore, a pulse chase expreiment was performed that showed that the OspA proteins in both ACAI and FI are very stable. Therefore, the differences in the protein amount is not due to an alteration of the turnover rate.

5.3 Genetic organizationThe DNA content of strains B31, ACAI and FI is depicted in Figure 4a. Genomic DNA was separated in a 1.2 % agarose gel by pulse field gel electrophoresis using the same parameters as described in Paper I. The DNA is visualized by ethidium bromide staining. The plasmid profiles of the two B. afzelii strains differ although they belong to the same species. Both strains have the 50 kb linear plasmid harbouring the ospAB

39

operon. DNA homologous to ospA in Fl was identified by Southern blot of the pulse field gel (Figure 4b).

co O CD <

co O CD <

- 29 -

- 16 -

Figure 4. Separation of genomic DNA from B. burgdorferi B31 and B. afzelii strains AC AI and FI by pulsed field gel electrophoresis.A. DNA is visualized by ethidium bromide staining.B. Southern blot analysis o f the osp opérons using an oligonucleotide specific for ospA

The ospAB operon from FI was cloned from plasmid libraries as described in Paper II and sequenced using internal primers. Comparison of the DNA sequence with the osp operon from ACAI, demonstrated that they are almost 100 % identical. Within the coding region there are only two base pair substitutions. In the beginning of ospA there is an A to G transition in position 264 giving a glycine instead of a glutamate. At position 1419, an A is substituted for a C in strain FI, resulting in a leucine instead of isoleucine in the end of OspB. As can be seen in Figure 6, Paper II, there is also one difference in the highly conserved region 5’ of the ospA start. Upstream of the -35 region, ACAI has a stretch of 10 T:s. FI has only 9 T:s, and B31 8 T:s in this area. It is possible that this region is involved in the regulation of expression of the Osp proteins in Lyme disease Borrelia. However, this region does not influence the expression of cloned ospAB genes in E. coli, as can be seen in Figure 10, Paper II.

40

Figu

re

5. Sc

hem

atic

m

odel

of

the

regu

latio

n of

the

Osp

prot

eins

41

5.4 TranscriptionTranscription of the ospAB operon was assessed by Northern blot analysis. In order to compare different preparations, the amount of ospAB transcript was normalized to the amount of flagellin transcript. As can be seen in Figure 8, Paper II, the relative amount of ospAB transcript obtained from different cultures of FI varies dramatically. The highest level of ospAB mRNA was found in Flb3. This culture had been stored at 4 °C for some time before the mRNA was prepared. Thus, the transcription of ospAB may be induced by low temperatures, or by the spirochetes entering the stationary phase. It is also possible that the ospAB transcript is particularly stable, and is thereby enriched in the stationary phase. Hardly any ospC mRNA could be detected in this FI culture.

The relative amounts of ospAB transcript together with the amount of OspA, OspB and OspC proteins detected in AC AI and different cultures of FI are shown in Figure 4, Paper II Interestingly, the correlation is not perfect between the amount of transcript and protein observed. Relatively high levels of ospAB transcript are observed in Flb2, but hardly any OspA and OspB protein was detected. More OspA protein is found in the other two FI cultures, where the level of ospAB transcript is lower. The presence of OspA in the culture F iel, which lacks detectable amounts of transcript, could be explained by the stability of the OspA protein. Nothing is known about the rate of protein synthesis in Borrelia. It is possible that there is a delay between transcription and translation, and this results in the low levels of OspA compared to transcript seen in preparations from culture Flb2. Another possibility could be that the expression is not only regulated at the level of transcription, but also at the translational level.

5.5 Model for Osp regulationA schematic representation of the regulation of Osp expression is shown in Figure 5. The Lyme disease Borrelia must be adaptable to differences in their surroundings, as they move between ticks and mammals in their life cycle. Mammalian blood is very low in purines, while guanine is one of the primary waste products in ticks. Upstream of the ospC. gene on the 26 kb circular plasmid, the guaA and guaB genes have been found. These genes, involved in purine biosynthesis, are expressed when the bacteria is present in mammals but not in ticks (Rosa and Margolis, 1994). OspC cannot be detected on the surface of spirochetes in unfed ticks, but is expressed by spirochetes in engorged ticks. Temperature also appears to influence OspC expression (Schwan,1994). It is possible that factors in the ingested warm mammalian blood signal the

42

activation of expression from the 26 kb plasmid. The presence of a 16 kb linear plasmid has been reported to correlate with OspC expression (Hinnebusch and Barbour, 1991). This would suggest that a factor is synthesized from this plasmid that may function as a negative regulator of OspC, and possibly also GuaA and GuaB, expression. High levels of guanine is probably sensed by the guaAB operon, either directly or via some sensor molecule. The ospC and guaAB genes are transcribed in opposite directions from separate promoters, but it is possible that activation of the guaAB promoter also influences ospC transcriptional activity by conformational changes that facilitate binding to the ospC promoter.

In contrast to OspC, OspA is present on the surface of spirochetes in unfed ticks (Schwan, 1994), but is probably not expressed in the mammalian stage as antibodies against OspA are usually absent in sera from Lyme disease patients (Wilske et al., 1986, 1988a). In Paper II, the inverse relationship between expressed OspA and OspB contra OspC exists both at the protein and RNA level, see Figures 5 and 9. Therefore, it is possible that OspAB expression is somehow coupled to the transcription of the 26 kb plasmid. When the 26 kb plasmid is inactive, expression of the ospAB operon is fully induced. Activation of the 26 kb plasmid may produce some kind of repressor for the ospAB operon. This factor could be either a protein or an RNA molecule with anti­sense RNA properties, synthesized from this circular plasmid.

6. Molecular characterization of the OspA and OspB proteins during infection in vivo (Paper III)

It is well documented that B. burgdorferi can cause chronic infections in both humans and other mammals. To be able to survive inside the host, the spirochetes must avoid the immune system in some way. As the closely related B. hermsii use antigenic variation of the vmp genes to achieve this (Stoenner et al., 1982), it has been speculated that B. burgdorferi may use some kind of variation of the Osp proteins to survive. Both the ospA and ospB genes are composed of small, tandem repeats, as can be seen in Figure 3, Paper I. PCR sequencing of the osp genes from B. burgdorferi Sh2-82 has shown that antigenic variation occurs during growth in vitro. The mechanism for this variation has been proposed to be recombination between the repeats, either intra- or intergenetically. This is possible due to the high homology between the osp A and ospB genes. The absolute copy number of the plasmid harbouring the osp operon has not been determined, but it is probable that more than

43

one copy exists per cell. This would allow for recombination between the plasmids (Rosa et al., 1992). To determine whether this is the mechanism used in vivo, we decided to study the ospA and ospB genes in B. burgdorferi at the molecular level during an ongoing infection.

6.1 MethodologyTo perform these studies, we needed an animal model where infection could be established, and sampling once a week could be done in a convenient way. We chose to use mice as they are easy to handle and can be infected by B. burgdorferi. The strain B. burgdorferi Sh2-82 was chosen since it had previously been shown to be infective for mice (Schwan et al., 1988) and in vitro data suggested that antigenic variation could occur in this strain (Rosa et al., 1992) To be able to take sequential comparative samples over several weeks, we decided to use a tissue cage model previously successfully used to study streptococcal infections (Holm et al., 1978; Knöll et al., 1985). Small, cylindrical net cages were placed in subcutaneous pouches on the backs of anaesthesized mice. Infection, and then sampling was performed as described in Paper V. Addition of NaCl or PBS to the cages to replace the volume taken out at sampling, prolonged the time before the cages became dried out. This also made it easier to take out sufficient fluid at each sampling. The presence of spirochetes was demonstrated by the cultivation of chamber fluid.

6.2 Results of cultivationIn mice numbers 1 to 18, an Sh2-82 culture, cloned by limiting dilution and passed more than 10 times in vitro, was used, while a low-passage, non-clonal culture of Sh2- 82 was used in mice numbers 19 to 36. The DNA pattern of the inoculated strains was found to be identical by PCR amplification, indicating that they were monocultures, at least with respect to the osp genes. It is a well known fact that B. burgdorferi lose their infectivity during prolonged cultivation in medium. This is probably because factors required for infectivity, but not for in vitro growth, are lost during the cultivation. As can be seen in Table I, Paper V, the low passage strain was much more infectious than the strain passaged for a longer period of time in vitro. In the first group, only the highest dosage of spirochetes used, 106, could induce an infection, and only in the C3H/Tif mice (No 16 and 18). When the low-passage strain was used, 100 fold less spirochetes were needed to establish an infection, and it was possible to cultivate

44

spirochetes not only from C3H/Tif, but also from the BalbC/cJ mice. Furthermore, in this group, it was easier to establish a long-lasting infection in the C3H/Tif mice.

The growth pattern found in Table 1, Paper V, resembles the variation expected from an organism surviving by antigenic variation: Positive culture indicating iarge” amounts of spirochetes in the circulation, followed by negative cultures resulting from the removal of spirochetes by the immune system. Positive cultures later on indicate that a new serotype, not recognized by the immune system, has arised, and so on.

6.3 Cloning and sequencing of the ospA and ospB genesThe results of the cultivation experiment fit with the antigenic variation found in vitro by Rosa and coworkers (Rosa et al., 1992). To verify that this was responsible for the fluctuations, we PCR amplified the ospA and ospB genes from the obtained cultures. No difference in the size of amplified DNA fragments from early or late cultures compared to the original strains could be detected. To study if recombination within the genes, giving rise to new, antigenically different Osp proteins, had taken place, the fragments were cloned into pET7-vectors and sequenced. Surprisingly, no DNA rearrangements at all could be detected, implying that antigenic variation of the Osp proteins is not a mechanism used to survive inside an immunocompetent host.

6.4 Immune responseTo examine whether the Borrelia infections induced any immune response in the mice, and if there was any difference in mice positive or negative in culture, we performed ELISA with the mouse sera on whole cell lysates of Sh2-82. Recombinant B31 OspA was also used as antigen to test the immune response against the Osp proteins. The osp opérons of B31 and Sh2-82 are almost identical, and polyclonal antibodies to OspA should react with both strains. Figure 1 Paper III show the mean values of the ELISAresults obtained. As can be seen, the most pronounced responses were found in micewhere infection had been initiated as indicated by a positive culture. This shows that persistence of the infection is not because these mice were unable to mount an immune response. Similarly, the lack of detectable infection in mice No 1-15 and No 17 is not because these mice responded with a stronger immune defence. The lack of infection in these mice is probably because not enough of infectious bacteria were injected. That a relevant immune response was raised in the infected mice was further established by the observation that several of the C3H/Tif mice had swollen joints 10 weeks after the

45

onset of infection. Only two mice, numbers 9 and 18, raised an immune response towards OspA that was above the cut off value 0.2. These mice were both injected with 106 spirochetes. In natural infections, immune responses to Osp proteins are rarely found (Wilske et al., 1986, 1988a). But when higher amounts of spirochetes are injected, fast responses to OspA and OspB are seen (Gem et a i , 1993). The data in Paper V indicate that 106 spirochetes is a borderline dose to elicit an OspA response.

7. Mapping of anti-OspB antibody epitopes

7.1 Method of selecting mutantsTo date, there are no genetic methods available for creating mutants in Borrelia. It is not possible to introduce new DNA into the spirochetes since no vector system and no method of transformation has been developed. The situation is further complicated by the fact that the modified BSKII medium used for cultivation is very complex and contains, among other things, rabbit semm and can therefore not be considered defined. This makes it very difficult to screen for mutants in the ordinary way, by plating on selective plates.

An alternative method for the generation of mutants has been developed by Sadziene et al (1992a). Antibody resistant mutants are isolated under selective pressure from anti Borrelia antibodies in complete BSKII medium. After incubation at 34°C, surviving borrelia are cloned by plating with low melting point agarose on solid BSKII plates. By comparing the DNA sequences from such mutants with the wild type sequence, it is possible to get an idea of the binding epitopes of the antibodies used. In combination with functional studies of the mutants, it is possible to characterize the studied proteins.

7.2 Cloning and sequencingThe ospB genes from wild type and antibody resistant mutants were amplified by PCR and cloned into TA-vectors as described in Paper III and IV. Internal, as well as plasmid-encoded, primers were used for sequencing. Normally, sequencing was performed from several clones isolated from separate PCR rounds, to ensure that base pair exchanges were true mutations, and not artefacts due to the PCR procedure. Occasionally, the sequences of the variable regions were confirmed by sequencing of asymmetrical PCR products. The results of the sequencing can be seen in Table 3 and

46

tu o o Ü il in<0 < <MHI H

tO E-

U~) <U H

TS t-H

tO H

«C 11t© <UM CM U

co x ; n<N 0 0

(U u TJ EHto F-

< 15TJ H

0 0N 0 H

<0 < te(U Eh (NJ «Ç 0 CNC U

U HTI Eh tN Ö UTJ U

tO Eh

f0 Eh

Q) P-

m eh

(U E- Xî U

C CJ00 *H 0 CN Ü l!CNJ *J

N Ü HjM Eheu oto fi, M Eh (U u to Eh 0, U en fi < 0

.T J Eh Ü0 ^ 0

É-H0 0 3 H (U Eh ^ U

Figu

re

6. Th

e DN

A se

quen

ce

of the

osp

B ge

ne

of B

. bur

gdor

feri

B31

is sh

own

in ca

pita

l le

tters

belo

w the

am

ino

acid

sequ

ence

in

italic

s.Th

e rib

osom

al b

indi

ng

site

(RBS

) is

unde

rline

d, a

nd

stop

codo

ns

are

indi

cated

by

thre

e sta

rs Th

e m

utat

ions

in

the

diffe

rent

mut

ants

are

indi

cate

d in

bold

.

47

Table 3. Description of the antibodies and genotypesMutant Selecting

antibodybase-pairexchange

amino acid exchange

B31-2 H6831 1738 A-»G 253 Lys—»GluB31-3,4 H6831 1570 G-»A

1802 C-»A197 Valgile 274 Ala—»Asp

B31-7 H6831 1738 A—»G 253 Lys—»GluB31-8 H614 1799 C-»T 273 Thr-»IleB31-9 3A3 1153 T-»C 58 Ser-»ProB31-10 3D3 1331 C-»T 117 Ala-» ValB31-11 84C 1769 A-»G 263 Asp-»GlyB31-12 84C 1730 C-»T 250 Ala-» ValHB191* Polyclonal 1697 G-»A 239 Trp—»Stop* The HB191 sequence was compared to HB19 wild type sequence, not to the

B31 sequence

Figure 6. As can be seen in Figure 6, most of the mutations have occurred in the C- terminal end of the protein. The only exceptions are mutants B31-9 and B31-10, which have mutations in amino acids 58 and 117 respectively. They are selected for by the monoclonal antibodies 3A3 and 3D3. Mutants derived from selection with Mab 84C are further described in Paper IV, and Mab H6831 mutants are described in Paper V.

Analysis of the predicted secondary structures of the mutants revealed that the differences in amino acids 126 and 128, where B31 has an alanine and an aspartate, while UBI9 has a threonine and an asparagine probably confers a slight difference in the secondary structure on the C-terminal side of this area, around amino acid 150. Because of the truncation of the OspB protein in mutant HB191, the secondary structure in the C-terminal end differs from the wild type HB19 OspB sequence.

Among the B31 OspB mutants, it is only in mutants B31-3 and B31-4, together with mutant 84C’, that the amino acid exchanges appears to affect the secondary structure. Therefore, all the other mutants apparently have altered amino acids directly involved in the binding epitopes of the respective antibodies. B31-3 and B31-4 are the only mutants where more than one nucleotide has been changed. Therefore, it is likely that the epitopes of the selecting antibodies are very precise, and very sensitive to even slight changes in the primary structure. Mutant B31-8, generated by selection with Mab H614, is mutated next to one of the amino acids that is changed in the mutants

48

B31-3 and B31-4. The other amino acid which is changed in this double mutant, is situated next to amino acid 198, which is changed in strains Rpa and HB19 compared to the wild type B31 OspB sequence. The mutations in mutants 84C’ and B31-2 and B31-7 are separated by only two amino acids. Thus, the OspB protein contains regions that are apparently more easily mutated than other parts. It is probable that these areas are exposed on the surface of the bacteria, as they are targets for several different antibodies. It is possible that mutations in other parts of the OspB protein also occurs, but that these mutants are not allowed to grow.

7.3 Mapping of the binding epitope of highly cross-reactive OspB MAb 84C (Paper IV)

As was described in Paper I, the OspB protein is particularily heterogeneous among different isolates. In Paper III, an OspB monoclonal antibody, MAb 84C, is described, which, despite the diversity, recognizes a wide variety of B. burgdorferi isolates. Of 38 strains tested by Western blot analysis, 32 reacted with MAb 84C. Strains reactive to this monoclonal antibody belong to all three genomic species, /. e. B. burgdorferi s. s., B. garinii, and B. afzelii. The isolates came from North America, Europe, Russia and Japan, and were isolated from ticks, humans and other mammals. This shows that the epitope of MAb 84C is very conserved, and indicates that it may be functionally important, although binding of MAb 84C does not interfere with Borrelia attachment or the invasion of HU VE cells. MAb 84C is specific for Lyme disease borrelias, and does not react with other Borrelia species , like B. hermsii.

The epitope for MAb 84C is probably surface exposed, as it causes agglutination when mixed with spirochetes. It was also possible to isolate escape variants, which further implies that the recognition site is exposed on the outside of viable bacteria.

Deletion mutagenesis of recombinant B31 OspB mapped the MAb 84C epitope to the carboxy-terminal end of the protein, between amino acids 230 and 273. To further map the binding site, MAb 84C escape variants were selected as described above. Two different mutants, 84C and 84C’, were isolated. Mutant 84C’ still reacts with MAb 84C in Western blot analysis, while mutant 84C had obviously lost the ability to interact with the antibody. When sequenced, both mutants had a single base pair exchange. Aspartate-263 was substituted for a glycine in mutant 84C, and alanin-250 was exchanged for a valine in 84C\ As can be seen in Figure 6, Paper III, the 250 Ala->Val substitution confers a predicted disruption of the secondary structure, which probably hides the MAb 84C epitope from exposure on the surface, allowing the 84C’

49

mutant to survive. Amino acid 263 is situated within the area defined by the deletion studies. When the amino acid sequences between positions 230 and 273 from B31, ACAI, Ip90 and the two mutants were aligned, amino acids 259 to 269 are totally conserved, except in the non-binding 84C mutant. Therefore, it is suggested that the MAb 84C epitope is situated within this area, and that amino acid 263 is particularily important. The epitope is probably linear, as neither dénaturation nor fixation apparently affect the binding. Furthermore, the wild type predicted secondary structure is not changed by the amino acid substitution in the 84C mutant.

7.4 Characterization of the anti-OspB bactericidal antibody H6831 (Paper V)

7.4.1 Borrelicidal activityIt is well known that specific antibodies in the presence of complement or phagocytes can kill bacteria. Antibodies with inherent inhibitory or bactericidal activity are less common. In an previous study (Sadziene et al., 1992b), Fab fragments from the OspB MAb H6831 were shown to inhibit growth of B. burgdorferi in vitro, in the absence of complement. In Paper IV, this activity of H6831 was further characterized.

To establish whether H6831 simply inhibits growth,or if it has true bactericidal activity as well, a time-kill study was performed. The MAb H4825, specific for serotype C of B. hermsii was used as a negative control in all experiments. This antibody has previously been shown to inhibit the growth of B. hermsii HS1 at a 10 fold lower concentration than H6831 is effective on strain B31. MIC, the minimal inhibitory concentration, was 0.2 and 2.0 pg/ml, respectively (Sadziene et al., 1992b). In this time-kill study, B31 was mixed with 20 jug/ml of either H6831 or H4825, and samples were plated at different time intervals to measure the concentration of live spirochetes. As can be seen in Figure 1A, Paper IV, there is a rapid decline in viable B31 when exposed to H6831, but no effect at all could be seen when B31 was incubated with the B. hermsii antibody H4825. In Figure 2, Paper IV, electron photomicrographs depict the fate of the spirochetes. The bactericidal antibodies apparently disrupt the target cells, and numerous blebs are formed.

The MBC is the minimal bactericidal concentration needed for killing 99.9 % of the inoculum. To determine the MBC of H6831, bacteria were incubated with different concentrations of antibody for 2 h, and then plated as described above. Figure IB,

50

Paper IV, shows that the MBC of H6831 for B. burgdorferi B31 is 10 pg/ml, /'. e. 5 times higher than the MIC.

7.4.2 Epitope mappingThe extraordinary effect of H6831 on the viability of the spirochetes suggests that its epitope on the OspB protein may be important. In an previous study (Barbour and Schrumpf, 1986), it was shown that H6831 bound only to some B. burgdorferi s. s. strains, and not at all to strains from the other genomic species. In that study, H6831 was shown to bind to B31 and RPa, a human skin isolate from New York, but not to the North American blood isolate HB19. In a first attempt to delimit the binding epitope, ospB sequences from binding and nonbinding strains were compared. The ospB gene of HB19 was cloned from a phage library and sequenced. The RPa ospB was PCR amplified, cloned and sequenced as described above. These sequences were also compared to the non-binding sequences of strains N40 and Sh2-82. All sequences were >99 % identical, but strains binding to H6831 had a lysine in position 253, while non-binding strains had a threonine in this position.

To further map the epitope, four H6831 B31 OspB mutants, which had all lost the ability to bind the antibody, were isolated, and the ospB genes were cloned and sequenced. The results can be seen in Table 4 and Figure 6. Mutants B31-3 and B31-4 had two missense mutations, resulting in amino acid exchanges in positions 197 and 274. These mutations may interfere with the secondary structure of the epitope. The other two mutants, B31-2 and B31-7, had substituted the lysine at position 253 with a glutamate.

To define the role of the linear epitope around amino acid 253, a synthetic 21-mer peptide covering this area was synthesized. Sera raised from this peptide could block the binding of H6831, showing the importance of this area for the binding of H6831.

The mechanism of the bactericidal activity of the antibodies is not known. By electron microscopy, the effect of H6831 on B31 cells resembles the action of penicillin. Penicillin causes lysis of actively growing cells binding to special penicillin binding proteins in the cell wall (Boyd and Hoerl, 1986). Binding of H6831 probably does not interfere with the cell wall in the same way as the antibiotic, but it is possible that binding of H6831 may interfere with the anchoring of OspB in the membrane. Detachment of OspB may introduce holes that cause cell lysis.

51

8. Binding studies (unpublished results)

8.1 Binding of spirochetes to glycolipidsB. burgdorferi s. /., like many other microorganisms, binds specifically to certain cell types. Bacterial surface lectins recognizing glycolipid and glycoprotein receptors present on the surface of host cells are known determinants of this cell-specificity (Strömberg et al., 1992). To investigate the involvement of carbohydrate receptors also in B. burgdorferi adherence, reference glycolipids separated on thin-layer plates were tested for their ability to mediate binding of 35S-metabolically-labelled B. afzelii strain ACAI (Strömberg et al., 1990).

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

■ ■ 1

• • À ,ilir illf: jjp

im ÆÈL [

. ' J f c l !A B

Figure 7. Binding of radioactively labelled B. afzelii to glycolipids separated on thin-layer chromatograms. Nine lanes are shown after spray detection with anisaldehyde (A) and after autoradiography following overlay with labeled B. afzelii. Lanes: 1 through 6, non-acidic glycolipids (20-40 ug) isolated from human erythrocytes (lane 1), monkey intestine (lane 2), dog intestine (lane 3), guinea pig intestine (lane 4), mouse faeces (lane 5) and rat faeces (lane 6); lanes 7 to 8, acidic glycolipids (gangliosides) isolated from calf brain (lane 7) and human erythrocytes (lane 8); lane 9, non-acidic glycolipids from a japanese sea urchin.

The results are shown in Figure 7 and Table 4. Selective binding to some but not all glycolipids was observed (see Table 4). Lactosylceramide (number 3 in Table 4) was the smallest receptor-active glycolipid. Certain terminal substitutions, such as Gala 1-3 (number 4) and GalNAcßl-4 (number 8) were tolerated, while others seemed to block receptor activity.

52

Table 4. Binding of Borrelia afzelii AC AI to purified and structurallycharacterized glycolipids.

Number Glycolipid Binding1 GalßCer -2 GlcßCer -3 GalßMGlcßCer +4 Galal -3Galß 1 -4GlcßCer +5 Galal -4Galß 1 -4GlcßCer -6 GalNacß 1 -3Gala 1 -4Galß 1 -4GlcßCer -7 NeuAca2-3Galß 1 -4GlcßCer -8 Galß 1 -3GalNAcß 1 -4Galß 1 -4Glcß 1 -Cer +9 Galß 1 -3GalNac(NeuAca2-3) ß 1 -4Galß 1 -4GlcßCer -

Lectin-carbohydrate interactions mediate the specificity of host-parasite interactions as well as cell-to-cell interactions. The specific interaction between lactosylceramide and related glycolipids by B. afzelii could be implicated in the association of Borrelia with brain tissues or in other adherence-related events associated with this organism. The delineation of a recognition site related to lactose is in line with the previous report of lactosylceramide being a receptor molecule (Garcia-Monco et al., 1992). A large number of additional bacteria have previously been reported to interact with lactosylceramide, which is closely associated with the membrane. In this respect, it should be noted that the HIV virus have been reported to interact specifically with another membrane associated glycolipid, galactosylceramide (Strömberg et al., 1992). It may be speculated that lactosylceramide and galactosylceramide may be involved in direct interactions with the eucaryotic cellular membrane.

9. Discussion of the pathogenicity of Lyme disease borreliae and the involvement of the Osp proteins

The data in Paper III and mouse immunization experiments indicate that the Lyme disease Borrelia do not survive inside the host by antigenic variation of the OspA and OspB proteins as was earlier believed (Barthold and Bockenstedt, 1993; Barthold, 1993). Instead, it appears that the spirochetes invade parts of the body that are not accessible to the immune response. Bacteria can for example invade and survive

53

antibiotic treatments inside human skin fibroblasts (Klempner et al., 1993). From these immunological cryptic pools, the bacteria can reenter the blood, and from there access more distant places in the body.

Immunization with OspA, OspB or OspC proteins confers protection against infection in mice (Probert and LeFebvre, 1994). Another study has shown that only patient sera containing anti-Osp antibodies could confer resistance to infection in mice (Fikrig et a i , 1994). Those data all indicate that the Osp proteins are important for the initiation of the infection. The blocking of OspA with antibodies reduces the binding while blocking of OspB seems to interfere with the invasion of HUVE cells (Comstock and Thomas, 1989). Perhaps OspA mediates the binding to lactosylceramide in the eukaryotic cell membrane (this thesis), and OspB the entry of spirochetes into the cells where they are protected from the immune defence system. The presence of anti-Osp A and/or anti-OspB antibodies would inhibit these processes, and the spirochetes could therefore be erradicated.

Not all anti-OspB antibodies protect from invasion. Binding to the highly conserved 84C epitope in the C-terminal end of the protein does not affect the entry into HUVE cells (Comstock et a i , 1993; Shoberg et al., 1994). Thus, this region of the protein is probably not directly interacting with other molecules involved in the invasion. The high degree of conservation of this region between different strains indicates that it must be essential for a functional OspB protein, and maybe involved in the tertiary structure.

In patient sera, antibodies towards OspA and OspB are rarely found, particularily in the beginning of the infection (Wilske et a i , 1986, 1988a). Recent studies show that B. burgdorferi transmitted to mice by ticks induce an antibody response to a 39 kD protein, but not to OspA (Golde et al., 1994). This, together with data presented in Paper II, indicates that the expression of OspA and OspB can be modulated, and may be turned off upon entry into the mammalian host. The stability of the OspA, and probably also the OspB, protein mean that these highly immunogenic proteins not directly vanish from the surface of the bacteria. They are probably present long enough to mediate the invasion process. The anti-inflammatory components of the tick saliva may inhibit the immune response towards the OspA and OspB proteins at this stage of infection. After the initial entry, the OspA and OspB proteins may no longer be needed for the subsequent infection, and this is perhaps the reason why antibodies against these proteins are not found in the sera of Lyme disease patients.

Another explanation is that the primary function of the OspA and OspB proteins are not within the mammalian host but in the tick, possibly for the transport from the

54

midgut and entry into the salivary glands. The reason that anti-Osp antibodies appear to be protective is then that not all OspA and OspB proteins have disappeared from the surface of the bacteria upon entering the mammalian host, although the production of the OspA and OspB proteins may terminate when the ticks are filled with blood.

55

CONCLUSIONS

• The ospA and ospB genes are located on linear plasmids about 50 kb in size, and are transcribed as one transcriptional unit.

• The ospA genes of the B. burgdorferi strains B31, AC AI and Ip90, are 85 % identical to each other, and the ospB genes show approximately 80 % identity. Nucleotide sequence comparisons of the ospAB genes allow the classification of the strains into three groups that coincide with the recent species designations of B. burgdorferi sensu stricto, B. afzelii and B. garinii.

• The expression of the Osp proteins is regulated primarily at the transcriptional level, but may also be adjusted at the translational level.

• The expression of the OspA and OspB proteins is probably shut off in favour of the OspC protein upon, or soon after, entry into the mammalian host.

• No antigenic variation of the OspA and OspB proteins occurs during infection. The spirochetes probably evade the immune system by migrating to places within the body not reached by the immune system.

• The monoclonal antibody 84C recognizes a highly conserved region of the carboxyl end of the OspB protein. The high degree of conservation indicates that this part of the protein is important for the structure and/or function of the protein.

• The monoclonal antibody H6831 binds to a variable region of the OspB protein. FAb fragments of the MAb H6831 have borrelicidal activity that mimics the action of ß-lactam antibiotics.

56

ACKNOWLEDGEMENTS

This is the part of the thesis that I have worried most about, as this is the only section that people read. But please, don’t have any expectations!

This work had not been possible to do, if I had not been surrounded by wonderfully helpful and talented people, both at the Department of Microbiology, where most of the work has been performed, and ‘outside’ (there actually exists a world on the other side of the door!).

First of all, I am grateful to my supervisor, Sven “Bärka” Bergstrom, for persuading me to enter the interesting field of Borrelia. There have been times when I’ve kind of regretted this, but I don’t ‘seem to’ do that any more. In the end, I think I’ve realized how much You have taught me.

During this 6 year period (is it really that long?!) I’ve had the pleasure to work with several present and former members of the Borrelia team; Nisse, Laila, Björn, Jonas, Anna-Karin, Sara, Jonas II, Annika and Ingela, in the order you showed up. I am especially grateful to Laila, who, besides being a very good friend, has helped me with a lot of different things during the years. Thanks also to Ingela for taking over the lab-work in the end. A special thanks also to Jonas, for struggling through one of the premature versions of my thesis and trying to convert them into some sense. Björn has also helped in trying to improve my medical knowledge, and Nisse has saved the life of some computers.

The binding studies presented in these thesis were performed in colaboration with Niklas Strömberg at the Department of Cariology. I especially want to thank You for your helpful cooperation to finish that section in time for deadline.

Many people have helped to make Micro a pleasant place to work at. Unfortunately, 1 don’t have the space to acknowledge you all by name. But Berit must be mentioned, as being a wonderful person and a close friend that you can always rely on.

A big thanks to the technical personnel; P-A, who can fix anything, Marita for being able to hold track of all administration, Gerd and Doris at the media, and the ‘wash-ladies’ Beth, Ulla, Maj-Britt and Karin. I will probably not realize how much You do before I’ve left this place.

And thanks Amanda for converting my thesis (except for this part) and the two manuscripts into something that hopefully will make sense, and are written in a readable international language. I’m really impressed with your linguistic talents and your speed!

57

The time at Micro has not only been spent by the lab bench, but also performing different sport activities. The first thing happening after I had entered the front door, was Tord coming asking me what 4km-time’ I had, which I, at that time, had no idea of. Thanks to P-A and Micke for being good coaches, even if You always have a tendency to highly overestimate my abilities!

I also want to mention Åsa, Micke and Anna, who all have been close friends during all this time, and never hesitated to listen to my complaints. Anna is not only a good friend, but she also took care of my problems with the front cover despite a very short notice.

Not many students have had the privilege of having such a supporting and understanding family as I have. My parents Bjarne and Inga-Britt, my sister Lena, and my grand parents Rinaldo and Ingrid, have always shown a great interest in what I’ve been doing, supported me and understood that studies take time and are not always performed during normal working hours.

Last, but definitely not least, I want to acknowledge Göran, for all Your love, understanding, encouragement and a million other things like drawing the computer illustrations in these thesis. During this last period, it has been wonderful living together with someone who knows what it is all about. Thanks a lot!

This work was financially supported from the Swedish Medical Research Council, the Swedish Society for Medical Research, the Swedish Research Council for Engineering Science, the National Institute of Health, USA, the Medical Faculty, the Kempe Foundation and the Lenander Foundation.

59

REFERENCES

Aberer, E. C. Brunner, G. Suchanek, H. Klade, A. Barbour, G. Stanek, and H. Lassmann. 1989. Molecular mimicry and Lyme borreliosis: a shared antigenic determinant between Borrelia burgdorferi and human tissue. Ann. Neurol. 26: 732- 737.

Adam, T., G. S. Gassmann, C. Rasiah, and U. B. Göbel. 1991. Phenotypic and genotypic analysis of Borrelia burgdorferi isolates from various sources. Infect. Immun. 59: 2579-2585.

Afzelius, A.. 1910. Verhandlungen der dermatologischen Gesellschaft zu Stockholm. Arch. Deramatol. Syph. 101:404.

Akov, S.. 1982. Blood digestion in ticks. In F. D. Obenchain and R. Galun ( eds ). Physiology of ticks, pp 197-212.

Allardet-Servent, A., S. Michaux-Charachon, E. Jumas-Bilak, L. Karayan, and M. Ramuz. 1993. Presence of one linear and one circular chromosome in the Agrobacterium tumefaciens C58 genome. J. Bacteriol. 175:7869-7874.

Appel, M. J., S. Allan, R. H. Jacobson, T. L. Lauderdale, Y. F. Chang, S. J. Shin, T. W. Thomford, R. J. Todhunter, and B. A. Summers. 1993. Experimental Lyme disease in dogs produces arthritis and persistent infection. Infect. Dis. 167:651-654.

Åsbrink, E., A. Hovmark, and B. Hederstedt. 1984a. Serologic studies of erythema chronicum migrans Afzelius and acrodermatitis chronica atrophicans with indirect immunofluorescence and enzyme-linked immunosorbent assay. Acta Derm. Venereol. 64: 506-512.

Åsbrink, E., B. Hederstedt, and A. Hovmark. 1984b. The spirochetal etiology of erythema chronicum migrans Afzelius. Acta. Derm. Venereol. (Stockh.). 64:291-295.

Åsbrink, E., and A. Hovmark. 1985. Successful cultivation of spirochetes from skin lesions of patients with erythema chronicum migrans Afzelius and acrodermatitis chronica atrophicans. Acta. Path. Microbiol. Immunol. Scand. Sect. B. 93: 161-163.

60

Assous, M. V., D. Postic, G. Paul, P. Névot, and G. Baranton. 1993. Western blot analysis of sera from Lyme borreliosis patients according to the genomic species of the Borrelia strains used as antigens. Eur. J. Clin. Microbiol. Infect. Dis. 12:261-268.

Assous, M. V., D. Postic, G. Paul, P. Névot, and G. Baranton. 1994. Individualisation of two genomic groups among American strains using PCR-based rapid non-radioactive typing method. Program and abstracts, VI International Conference on Lyme borreliosis, Bologna, Italy.

Baranton, G., D. Postic, I. Saint Girons, P. Boerlin, J-C. Piffaretti, M. Assous, and P. A. D. Grimont. 1992. Delineation of Borrelia burgdorferi sensu stricto,B. garinii sp. nov., and Group VS461 associated with Lyme borreliosis. Int. Joum. of Syst. Bacteriol. 42:378-383.

Barbour, A. G.. 1984a. Isolation and cultivation of Lyme disease spirochetes. Yale J. Biol. Med. 57:521-525.

Barbour, A. G.. 1984b. Immunochemical analysis of Lyme disease spirochetes. Yale J. Biol. Med. 57: 581-586.

Barbour, A. G.. 1985. Clonal polymorphism of surface antigens in a relapsing Borrelia species. In G. G. Jackson and H. Thomas (eds). Bayer symposium VIII: the pathogenesis of bacterial infections. Springer-Verlag, Heidelberg, Federal Republic of Germany, pp 235-245.

Barbour, A. G.. 1988. Plasmid analysis of Borrelia burgdorferi, the Lyme disease agent. J. Clin. Microbiol. 26: 475-478.

Barbour A.. 1989. Antigenic variation in relapsing fever Borrelia species: Genetic aspects. In Berg D. E., and Howe M. M. (eds). Mobile DNA, ASM press, Chapter 36: 783-789.

Barbour, A. G.. 1993. Linear DNA of Borrelia species and antigenic variation. Trends in Microbiol. 1:236-239.

61

Barbour, A. G., N. Burman, C. J. Carter, T. Kitten, and S. Bergström. 1991. Variable antigen genes of the relapsing fever agent Borrelia hermsii are activated by promoter addition. Mol. Microbiol. 5: 489-493.

Barbour, A. G., and C. F. Garon. 1987. Linear plasmids of the bacterium Borrelia burgdorferi have covalently closed ends. Science 237:409-411.

Barbour, A. G., and S. F. Hayes. 1986. Biology of Borrelia species. Microbiol. Rev. 50: 381-400.

Barbour, A. G., R. A. Heiland, and T. R. Howe. 1985. Heterogeneity of major proteins in Lyme Disease Borreliae: A molecular analysis of North American and European isolates. J. Infect. Dis. 152:478-484.

Barbour, A. G., and M. E. Schrumpf. 1986. Polymorphisms of major surface proteins of Borrelia burgdorferi. Zentralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt 1 Orig. Reihe A. 263: 83-91.

Barbour, A. G., S. L. Tessier, and S. F: Hayes. 1984. Variation in a major surface protein of Lyme disease spirochetes. Infect. Immun. 45: 94-100.

Barbour, A. G., S. L. Tessier, and W. J. Todd. 1983. Lyme disease spirochetes and Ixoded tick spirochetes share a common surface antigenic determinant defined by a monoclonal antibody. Infect. Immun. 41:795-804.

Baril, C., C. Richaud, G. Baranton, and I. Saint Giron. 1989. Linear chromosome of Borrelia burgdorferi. Res. Microbiol. 140:507-516.

Barthold, S. W.. 1993. Antigenic stability of Borrelia burgdorferi during chronic infections of immunocompetent mice. Infect, and Immun. 61:4955-4961.

Barthold, S. W., and L. K. Bockenstedt. 1993. Passive immunizing activity of sera from mice infected with Borrelia burgdorferi. Infect, and Immun. 61:4696-4702.

62

Barthold, S. W., K. D. Moody, G. A. Terwilliger, P. Duray, R. O. Jacoby, and A.C. Steere. 1988. Experimental Lyme arthritis in rats infected with Borrelia burgdorferi. J. Infect. Dis. 157:842-846.

Belisle, J. T., M. E. Brandt, J. D. Radolf, and M. V. Norgard. 1994. Fatty acids of Treponema pallidum and Borrelia burgdorferi lipoproteins. J. Bacteriol. 176: 2151- 2157.

Benach, J.L., E. M. Bosler, J. P. Hanrahan, J. L. Coleman, G. S. Habicht, T. F. Bast, D. J. Cameron, J. L. Ziegler, A. G. Barbour, W. Burgdorfer, R. Edelman, and R. A. Kaslow. 1983. Spirochetes isolated from the blood of two patients with Lyme disease. N. Engl. J. Med. 308: 740-742.

Bergström, S., A. G. Barbour, N. Burman, M. Ferdows, J. Hinnebusch, M. Jonsson, and T. Kitten. 1990. Molecular biology of Borrelia species. In O. Olsvik and G. Bukholm ( eds. ). Application of Molecular biology in diagnosis of infectious diseases, Norwegian College of Veterinary Medicine, pp 130-134.

Bergström, S., B. Olsén, N. Burman, L. Gothefors, T. G. T. Jaenson, M. Jonsson, and H. A. Mejlon. 1992. Molecular characterization of Borrelia burgdorferi isolated from Ixodes ricinus in Northern Sweden. Scand. J. Infect. Dis. 24: 181-188.

Bergström, S., V.G. Bundoc, and A. G. Barbour. 1989. Molecular analysis of linear plasmid-encoded major surface proteins, OspA and OspB, of the Lyme disease spirochaete , Borrelia burgdorferi. Mol. Microbiol. 3:479-486.

Bergström, S., K. Robbins, J. M. Koomey, and J. Swanson. 1986. Piliation control mechanisms in Neisseria gonorrhoeae. Proc. Natl. Acad. Sci. USA 83:3890-3894.

Boyd, R. F., and B. G. Hoerl. 1986. Chapter 11: Chemotherapy. In Basic medical microbiology, third edition. Little, Brown and Company, Boston/Toronto, USA.

Bracco, L., D. Kotlarz, A. Kolb, S. Diekmann, and H. Bue. 1989. Synthetic curved DNA sequences can act as transcriptional activators in Escherichia coli. EMBO J. 8:4289-4296.

63

Brandt, M. E., B. S. Riley, J. D. Radolf, and M. V. Norgard. 1990. Immunogenic integral membrane proteins of Borrelia burgdorferi are lipoproteins. Infect. Immun. 58: 983-991.

Buchwald, A. 1883. Ein fall von diffuser idiopathischer Haut-Atrophie. Arch. Derm. Syph. ( Berlin ). 15:553-556.

Bundoc, V. G., and A. G. Barbour. 1989. Clonal polymorphism of outer membrane protein OspB of Borrrelia burgdorferi. Infect, and Imm. 57:2733-2741.

Burgdorfer, W.. 1984. Discovery of the Lyme disease spirochete and its relation to tick vectors. Yale J. Biol. Med. 57:515-520.

Burgdorfer, W., A. G. Barbour, S. F. Hayes, J. L. Benach, E. Grunwaldt, J. P. Davis. 1982. Lyme disease - atickbome spirochetosis? Science 216:1317-1319.

Burgdorfer, W, A. G. Barbour, S. F. Hayes, O. Peter, and A. Aeschlimann. 1983. Erythema chronicum migrans- a tick-bome spirochetosis. Acta. Trop. ( Basel ) 40:79- 83.

Burman, N., S. Bergström, B. I. Restrepo, and A. G. Barbour. 1990. The variable antigens Vmp7 and Vmp21 of the relapsing fever bacterium Borrelia hermsii are structurally analogous to the VSG proteins of the African trypanosome. Mol. microbiol. 4:1715-1726.

Cadavid, D., D. D. Thomas, R. Crawley, and A. G. Barbour. 1994. Variability of a bacterial surface protein and disease expression in a possible mouse model of systemic Lyme borreliosis. J. Exp. Med. 179:631-642.

Canica, M. M., F. Nato, L. duMerle, J. C. Mazie, G. Baranton, and D. Postic.1993. Monoclonal antibodies for identification of Borrelia afzelii sp. nov. associated with late cutaneous manifestations of Lyme borreliosis. Scand. J. Infect. Dis. 25: 441- 448.

64

Carter, C. J., S. Bergström, S. J. Norris, and A. G. Barbour. 1994. A family of surface-exposed proteins of 20 kilodaltons in the genus Borrelia. Infect. Immun. 62:2792-2799.

Casjens, S., and W. M Huang. 1993. Linear chromosomal physical and genetic map of Borrelia burgdorferi, the Lyme disease agent. Mol. Microbiol. 8:967-980.

Casjens, S., and W. H. Huang. 1994. The linear chromosome of the Lyme Borrelia. Program and abstracts, VI International Conference on Lyme borreliosis, Bologna, Italy.

Champion, C. I., D. R. Blanco, J. T. Ska re, D. A. Haake, M. Giladi, D. Foley, J. N. Miller, and M. A. Lovett. 1994. A 9,0-kilobase-pair circular plasmid of Borrelia burgdorferi encodes an exported protein: Evidence for expression only during infection. Infect. Immun. 62:2653-2661.

Coburn, J., J. M. Leong, and J. K. Erban. 1993. Integrin <Xiibß3 mediates binding of the Lyme disease agent Borrelia burgdorferi to human platelets. Proc. Natl. Acad. Sci. 90:7059-7063.

Comstock, L. E., E. Fikrig, R. J. Shoberg, R. A. Flavell, and D. D. Thomas. 1993. A monoclonal antibody to OspA inhibits association of Borrelia burgdorferi with human endothelial cells. Infect, and Immun. 61:423-431.

Comstock, L. E., and D. D. Thomas. 1989. Penetration of endothelial cell monolayers by Borrelia burgdorferi. Infect, and Immun. 57:1626-1628.

Comstock, L. E., and D. D. Thomas. 1991. Characterization of Borrelia burgdorferi invasion of cultured endothelial cells. Microb. Path. 10:137-148.

Coyle, P. K., L. B. Krupp, C. Doscher, and K. Amin. 1994. Borrelia burgdorferi reactivity in patients with severe persistent fatigue who are from a region in which Lyme disease is endemic. Clin. Infect. Dis. Suppl. I 18: S24-S27.

65

Craft, J. E., D. K. Fischer, G. T. Schimamoto, and A.C. Steere. 1986. Antigens of Borrelia burgdorferi recognized during Lyme disease. Appearance of a new Immunoglobulin M response and expansion of the Immunoglobulin G response late in the illness. J. Clin. Invest. 78:934-939.

Crespi, M., E. Messens, A. B. Capian, M. van Montagu, and J. Desomer. 1992. Fasciation induction by the phytopathogen Rhodococcus fascians depends upon a linear plasmid encoding a cytokinin synthase gene. EMBO J. 11:795-804.

Cross, A.. 1990. The biological significance of bacterial encapsulation. Curr. Top. Microbiol. Immunol. 150:87-95.

Davidson, B., J. MacDougall, and I. Saint Girons. 1992. Physical map of the linear chromosome of the bacterium Borrelia burgdorferi 212, a causative agent of Lyme disease and localization of rRNA genes. J. Bacteriol. 174:3766-3774.

Donelson, J. E.. 1989. DNA rearrangements and antigenic variation in African Trypanosomes. In Berg D. E., and Howe M. M. (eds). Mobile DNA, ASM press, Chapter 35:763-781.

Dressier, F., R. Ackermann, and A. C. Steere. 1994. Antibody responses to the three genomic groups of Borrelia burgdorferi in European Lyme borreliosis. Joum. of Infect. Dis. 169:313-318.

Erdile, L. F., M-A. Brandt, D. J. Warakomski, G. J. Wes track, A. Sadziene, A. G. Barbour, and J. P. Mays. 1993. Role of attached lipid in immunogenicity of Borrelia burgdorferi OspA. Infect. Immun. 61: 81-90.

Ferdows, M. S., and A. G. Barbour. 1989. Megabase- sized linear DNA in the bacterium Borrelia burgdorferi, the Lyme disease agent. Proc. Natl. Acad. Sci. 86:5969-5973.

Fikrig, E., S. W. Barthold, and R. A. Flavell. 1993. OspA vaccination of mice with established Borrelia burgdorferi infection alters disease but not infection. Infect. Immun. 61:2553-2557.. 1

66

Fikrig, E., S. W. Barthold, F. S. Kantor, and R. A. Flavell. 1990. Protection of mice against the Lyme disease agent by immunizing with recombinant OspA. Science 250:553-556.

Fikrig, E., S. W. Barthold, F. S. Kantor, and R. A. Flavell. 1992a. Long-term protection of mice from Lyme disease by vaccination with OspA. Infect. Immun. 60:773-777.

Fikrig, E., S. W. Barthold, N. Marcantonio, K. Deponte, F. S. Kantor and R. A. Flavell. 1992b. Roles of OspA, OspB and flagellin in protective immunity to Lyme borreliosis in laboratory mice. Infect. Immun. 60:657-661.

Fikrig, E., L. K. Bockenstedt, S. W. Barthold, M. Chen, H. Tao, P. Ali-Salaam, S. R. Telford, and R. A. Flavell. 1994. Sera from patients with chronic Lyme disease protect mice from Lyme borreliosis. Joum. of Infect. Dis. 169:568-574.

Finlay, B. B., and S. Falkow. 1989. Common themes in microbial pathogenicity. Microbiol. Rev. 53:210-230.

Francou, F.. 1981. Isolation and characterization of a linear DNA molecule in the fungus Ascobolus immersus. Mol. Gen. Genet. 184:440-444.

Fuchs, R., S. Jauris, F. Lottspeich, V. Preac-Mursic, B. Wilske, and E. Soutschek.1992. Molecular analysis and expression of a Borrelia burgdorferi gene encoding a 22 kDa protein ( pC ) in Escherichia coli. Mol. Microbiol. 6:503-509.

Fukunaga, M., Y. Yanagihara, and M. Sohnaka. 1992. The 23S/5S ribosomal RNA genes (rrl/rrf) are separate from the 16S ribosomal RNA gene (rrs) in Borrelia burgdorferi, the aetiological agent of Lyme disease. Joum. of Gen. Microbiol. 138: 871-877.

Garcia-Monco, J. C., B. Fernandez-Villar, and J. L. Benach. 1989. Adherence of the Lyme disease spirochete to glial cells and cells of glial origin. J. Infect. Dis. 160:497-506.

67

Garcia-Monco, J. C., B. Fernandez Villar, R. C. Rogers, A. Szczepanski,C. M. Wheeler, and J. L. Benach. 1992. Borrelia burgdorferi and other related spirochetes bind to galactocerebroside. Neurology 42:1341-1348.

Garin, C., and R. Bujadoux. 1922. Paralysie parles Tiques. J. Med. Lyon. 71:765- 767.

Gern, L., U. E. Schaible, and M. M. Simon. 1993. Mode of inoculation of the Lyme disease agent Borrelia burgdorferi influences infection and immune responses in inbred strains of mice. J. Infect. Dis. 167:971-975.

Golde, T. W., K. J. Kappel, G. Dequesne, C. Feron, D. Plainchamp, C. Capiau, and Y. Lobet. 1994. Tick transmission of Borrelia burgdorferi to inbred strains of mice induces an antibody response to P39 but not to outer surface protein A. Infect. Immun. 62: 2625-2627.

Gustafson, R.. 1993. Epidemiological studies of Lyme borreliosis and Tick-borne encephalitis. PhD thesis ISBN 91-628-0958-X, Stockholm, Sweden.

Habicht, G. S., G. Beck, and J. L. Benach. 1987. Lyme disease. Sci Am. July: 60-65.

Hayakawa, T., T. Tanaka, K. Sakaguchi, N. Otake, and H. Yonehara. 1979. A linear plasmid-like DNA in Streptomyces sp. producing lankacidin group antibiotics. J. Gen. Appi. Microbiol. 25:255-260.

Hayashi, S., and H. C. Wu. 1990. Lipoproteins in bacteria. J. Bioenerg. Biomembr. 22:451-471.

Hayes, S. F., W. Burgdorfer, and A. G. Barbour. 1983. Bacteriophage in the Ixodesdammini spirochete, etiologic agent of Lyme disease. J. Bacteriol. 154:1436-1439.

Hellerström, S.. 1951. Erythema chronicum migrans Afzelius with meningitis. Acta Derm. Venereol. ( Stockholm ). 31:227-234.

68

Herxheimer, K., and K. Hartman. 1902. Uber Acrodermatitis chronica atrophicans. Acta Dermatol. ( Berlin ). 61:57-76.

Hinnebusch, J., and A. G. Barbour. 1991. Linear plasmids of Borrelia burgdorferi have telomeric structure and sequence similar to those of eukaryotic virus. J. Bacteriol. 173:7233-7239.

Hinnebusch, J., and A. G. Barbour. 1992. Linear and circular-plasmid copy numbers in B. burgdorferi. J. Bacteriol. 174:5251-5257.

Hinnebusch, J., S. Bergström, and A. G. Barbour. 1990. Cloning and sequence analysis of linear plasmid telomeres of the bacterium Borrelia burgdorferi. Mol. Microbiol. 4:811-820.

Hinnebusch, J., and K. Tilly. 1993. Linear plasmids and chromosomes in bacteria. Mol. Microbiol. 10:917-922.

Hirochika, H., and K. Sakaguchi. 1982. Analysis of linear plasmids isolated from Streptomyces: Association of protein with the ends of the plasmid DNA. Plasmid 7:59-65.

Holm, S. E., C. Ekedahl, and A-M. Bergholm. 1978. Comparison of antibiotic assays using different experimental models and their clinical significance. Scand. J. Infect. Suppl. 14: 214-220.

Holt, S. C.. 1978. Anatomy and chemistry of Spirochetes. Microbiol. Rev. 42: 114- 160.

Holz, A., C. Schaeter, H. Gille, W-R. Jueterbock, and W. Messer. 1992. Mutations in the DnaA binding sites of the replication origin of Escherichia coli. Mol. Gen. Genet. 233:81-88.

Howe, T. R., F. W. LaQuier, and A. G. Barbour. 1986. Organization of genes encoding two outer membrane proteins of the Lyme disease agent Borrelia burgdorferi within a single transcriptional unit. Infect. Immun. 54: 207-212.

69

Howe, T. R., L. W. Mayer, and A. G. Barbour. 1985. A single recombinant plasmid expressing two major outer surface proteins of the Lyme disease spirochete. Science 227: 645-646.

Hyde, F. W., and R. C. Johnson. 1984. Genetic relationship of Lyme disease spirochetes to Borrelia, Treponema, and Leptospira spp. J. Clin. Microbiol. 20:151- 154.

Hyde, F. W., and R. C. Johnson. 1986. Genetic analysis of Borrelia. Zentralbl. Bakteriol. Mikrobiol. Hyg. Ser. A. 263:119-122.

Isberg, R.. 1991. Discrimination between intracellular uptake and surface adhesion of bacterial pathogens. Science 252:934-938.

Jaenson, T. G. T., S. Bergström, N. Burman, J. Chirico, and M. Jonsson. 1989. Spiroketinfekterade fastingar {Ixodes ricinus)- risk för angrepp även i Norrland. Läkartidningen 86:2584.

Jiang, W., B. J. Luft, P. Munoz, R. J. Dattwyler, and P. D. Gorevic. 1990. Cross- antigenicity between the major surface proteins (OspA and OspB) and other proteins of Borrelia burgdorferi. J. Immunol. 144:284-289.

Johnson, R. C., N. Marek, and C. Kodner. 1984a. Infection of Syrian hamsters with Lyme disease spirochetes. J. Clin. Microbiol. 20:1099-1101.

Johnson, R. C., G. P. Schmid, F. W. Hyde, A. G. Steigerwalt, and D. J. Brenner.1984a. Borrelia burgdorferi sp. nov.: Etiologic agent of Lyme Disease. Intern. J. of Syst. Bacteriol. 34: 496-497.

Jonsson, A-B.. 1991. Phase switch control of pili in Neisseria gonorrhoeae. PhD thesis ISSN 0346-6612, University of Umeå, Sweden.

Kalkus, J., M. Reh, and H. G. Schlegel. Autotrophy of Nocardia opaca strains is encoded by linear megaplasmids. 1990. J. Gen. Microbiol. 136:1145-1151.

70

Karlsson, M.. 1990. Aspects of the diagnosis of Lyme borreliosis. PhD thesis ISBN 91-628-0033-7, Stockholm, Sweden.

Kawabata, H., T. Masuzawa, and Y. Yanagihara. 1993. Genomic analysis of Borrelia japonica sp. nov. isolated from Ixodes ovatus in Japan. Microbiol. Immunol. 37:843-848.

Kitten, T., and A. Barbour. 1992. The relapsing fever agent Borrelia hermsii has multiple copies of its chromosome and linear plasmids. Genetics 132:311-324.

Klempner, M. S., R. Noring, and R. A. Rogers. 1993. Invasion of human skin fibroblasts by the Lyme disease spirochete, Borrelia burgdorferi. J. Infect. Dis. 167:1074-1081.

Knöll, H., S. E. Holm, D. Gerlach, O. Kuhnemund, and W. Köhler. 1985. Tissue cages for study of experimental streptococcal infection in rabbits. II. Humoral ans cell- mediated immune response to erythrogenic toxins. Immunobiol. 169:116-127.

Koo, H-S., H. M. Wu, and D. M. Crothers. 1986. DNA bending at adenine/thymidine tracts. Nature 320: 501-506.

Kornblatt, A. N., A. C. Steere, and D. G. Brownstein. 1984. Infection in rabbits with the Lyme disease spirochete. Yale J. Biol. Med. 57:613-614.

Krawiec, S., and M. Riley. 1990. Organization of the bacterial chromosome. Microbiol. Rev. 54: 502-539.

Kryuchechnikov, V. N., E. I. Korenberg, S. V. Scherbakov, Yu. V. Kovalevsky, and M. L. Levin. 1988. Identification of Borrelia isolated in the USSR from Ixodes persulcatus schulze ticks. J. Microbiol. Epidemiol. Immunobiol. 12: 41-44.

Lam, T. T., T-P. K. Nguyen, R. R. Montgomery, F. S. Kantor, E. Fikrig, and R. A. Flavell. 1994. Outer surface proteins E and F of Borrelia burgdorferi, the agent of Lyme disease. Infect, and Immun. 62:290-298.

71

Lim, L. C. L., D. M. England, B. K. DuChateau, N. J. Glowacki, J. R. Creson, S. D. Lovrich, S. M. Callister, D. A. Jobe, and R. F. Schell. 1994. Development of destructive arthritis in vaccinated hamsters challenged with Borrelia burgdorferi. Infect. Immun. 62: 2825-2833.

Lane, R. S., and J. A. Pascocello. 1989. Antigenic characteristics of Borrelia burgdorferi isolates from Ixodid ticks in California. J. Clin. Microbiol. 27:2344-2349.

Lin, Y-S., H. M. Kieser, D. A. Hopwood, and C. W. Chen. 1993. The chromosomal DNA of Streptomyces lividans 66 is linear. Mol. Microbiol. 10:923-933.

Lipschutz, B.. 1914. Uber eine seltene Erythemform (Erytme chronicum migrans). Arch. Dermatol. Syph. 118:349-356.

Luft, B. J., W. Jiang, P. Munoz, R. J. Dattwyler, and P. G. Gorevic. 1989. Biochemical and immunological characterization of the surface proteins of Borrelia burgdorferi. Infect. Immun. 57:3637-3645.

Luft, B. J., B. Johnson, B. McGrath, R. J. Dattwyler, and J. J. Dunn. 1994. Chimeric protein vaccine for Lyme borreliosis. In Advances in Lyme borreliosis, VI International Conference on Lyme borreliosis, Bologna, Italy.

Ma, Y., A. Sturrock, and J. J. Weis. 1991. Intracellular localization of Borrelia burgdorferi within human endothelial cells. Infect. Immun. 59:671-678.

Magnarelli, L. A., J. F. Anderson, and D. Fish. 1987. Transovarial transmission of Borrelia burgdorferi in Ixodes dammini (Acari: Ixodidae). J. Infect. Dis. 156:234-236.

Marconi, R. T., M. E. Konkel, and C. F. Garon. 1993a. Variability of osp genes and gene products among species of Lyme disease spirochetes. Infect. Immun. 61:2611- 2617.

Marconi, R. T., D. S. Samuels, and C. F. Garon. 1993b. Transcriptional analysis and mapping of the ospC gene in Lyme disease spirochetes. J. Bacteriol. 175:926-932.

72

Marconi R. T., D. S. Samuels, T. G. Schwan, and C. F. Garon. 1993c. Identification of a protein in several Borrelia species which is related to OspC of the Lyme disease spirochetes. J. Clin. Microbiol. 31:2577-2583.

Marcus, L. C., A. C. Steere, P. H. Duray, A. E. Anderson, and B. Mahoney. 1985. Fatal pancarditis in a patient with coexisting Lyme disease and babesiosis. Ann. Intern. Med. 103:374-376.

Margolis, N., and P. A. Rosa. 1993. Regulation of expression of major outer surface proteins in Borrelia burgdorferi. Infect. Immun. 61:2207-2210.

Meier, J. T., M. I. Simon, and A. G. Barbour. 1985. Antigenic variation is associated with DNA rearrangements in a relapsing fever borrelia. Cell. 41:403-409.

Nationalencyclopedin. 1990. Bra böcker AB, Höganäs, Sweden.

Nguyen, T-P, K., T. T. Lam, S. W. Barthold, S. R. Telford III, R. A. Flave», and E. Fikrig. 1994. Partial destruction of Borrelia burgdorferi within ticks that engorged on OspE- or OspF-immunized mice. Infect, and Immun. 62:2079-2084.

Nielsen, S. L., K. K. Y. Young, and A. G. Barbour. 1990. Detection of Borrelia burgdorferi DNA by the polymerase chain reaction. Mol. Cell. Probes. 4: 73-79.

Norris, S. J., C. J. Carter, J. K. Howell, and A. G. Barbour. 1992. Low-passage- associated proteins of Borrelia burgdorferi B31: Characterization and molecular cloning of OspD, a surface exposed, plasmid-encoded lipoprotein. Infect. Immun. 60: 4662-4672.

Norton Hughes, C. A., and R. C. Johnson. 1990. Methylated DNA in Borrelia species. J. Bacteriol. 172:6602-6604.

Norton Hughes, C. A., S. M. Engstrom, L. A. Coleman, C. B. Kodner, and R. C. Johnson. 1993. Protective immunity is induced by a Borrelia burgdorferi mutant that lacks OspA and OspB. Infect. Immun. 61:5115-5122.

73

Ogasawara, N., and H. Yoshikawa. 1992. Genes and their organization in the replication origin region of the bacterial chromosome. Mol. Microbiol. 6: 629-634.

Ogasawara, N., M. Fujita, S. Moriya, T. Fukuoka, M. Hirano, and H. Yoshikawa.1990. Comparative anatomy of oriC of the eubacteria. In The Bacterial Chromosome. Drlica, K., and M. Riley (eds), Washington, D. C.: American Society for Microbiology, pp. 287-295.

Old, I. G., J. MacDougall, I. Saint Girons, and B. E. Davidson. 1992. Mapping of genes on the linear chromosome of the bacterium Borrelia burgdorferi: possible locations for its origin of replication. FEMS Microbiol. Lett. 99:245-250.

Old, I. G., D. Margarita, and I. Saint Girons. 1993a. Unique arrangement in the dnaA region of the Borrelia burgdorferi linear chromosome: nucleotide sequence of the dnaA gene. FEMS Microbiol. Lett. 111:109-114.

Old, 1. G., D. Margarita, and I. Saint Girons. 1993b. Nucleotide sequence of the Borrelia burgdorferi dnaN gene encoding the ß subunit of DNA polymerase III. Nucl. Acids Res. 21:3323.

Oliver, J. H. Jr., M. R. Owsley, H. J. Hutcheson, A. M. James, C. Chen, W. S. Irby, E. M. Dotson, and D. K. McLain. 1993. Conspecificity of the ticks Ixodes scapularis and I. dammini (Acari: Ixodidae). J. Med. Entemol. 30:54-63.

Padula, S. J., A. Sampieri, F. Dias, A. Szczepanski, and R. W. Ryan. 1993. Molecular characterization and expression of p23 ( OspC ) from a North American strain of B. burgdorferi. Infect, and Immun. 61:5097-5105.

Perng, G-C., and R. B. LeFebvre. 1990. Expression of antigens from chromosomal and linear plasmid DNA of Borrelia coriaceae. Infect. Immun. 58:1744-1748.

Philipp, M. T., M. K. Ayding, R. P. Bohm Jr., F. B. Cogswell, V. A. Dennis,H. N. Lanners, R. C. Lowrie Jr., E. D. Roberts, M. D. Conway, M. Karacorlu, G. A. Peyman, D. J. Gubler, B. J. B. Johnson, J. Piesman, and Y. Gu. 1993. Early and late disseminated phases of Lyme disease in the Rhesus monkey: a model for infection in humans. Infect. Immun. 61:3047-3059.

74

Plasterk, R. H. A., M. I. Simon, and A. G. Barbour. 1985. Transposition of structural genes to an expression sequence on a linear plasmid causes antigenic variation in the bacterium Borrelia hermsii. Nature (London) 318:257-263.

Ploeg, L. H. T. van der, A. Y. C. Liu, and P. Borst. 1984. Structure of the growing telomeres of trypanosomes. Cell. 36:459-468.

Postic, D., M. Assous, and G. Baranton. 1994. Restriction polymorphism of PCR products as an epidemiological and diagnostic tool for characterization of Borrelia associated with Lyme borreliosis. Program and abstracts, VI International Conference on Lyme borreliosis, Bologna, Italy.

Postic, D., J. Belfaiza, E. Isagai, I. Saint Girons, P. A. D. G ri mont, and G. Baranton. 1993. A new genomic species in Borrelia burgdorferi sensu lato isolated from Japanese ticks. Res. Microbiol. 144:467-473.

Postic, D., C. Edlinger, C. Richaud, F. Grimont, Y. Dufresne, P. Perolat,G. Baranton, and P. A. D. Grimont. 1990. Two genomic species in Borrelia burgdorferi. Res. Microbiol. 141: 465-475.

Preac-Mursic, V. B. Wilske, G. Schierz, M. Holmburger, and E. Süss. 1987. In vitro and in vivo susceptibility of Borrelia burgdorferi. Eur. J. Clin. Microbiol. 6:424- 426.

Probert, W. S., and R. B. LeFebvre. 1994. Protection of C3H/HeN mice from challenge with Borrelia burgdorferi through active immunization with OspA, OspB or OspC, but not with OspD or the 83-kilodalton antigen. Infect, and Immun. 62:1920- 1926.

Reindl, M., B. Redl, and G. Stöffler. 1993. Isolation and analysis of a linear plasmid- located gene of Borrelia burgdorferi B29 encoding a 27 kDa surface lipoprotein (P27) and its overexpression in Escherichia coli. Mol. Microbiol. 8:1115-1124.

Ribeiro, J. M. C., G. Makoul, J. F. Levine, D. Robinson, and A. Spielman. 1985. Antihemostatic, antiinflammatory and immunosuppressive properties of the saliva of a tick, Ixodes dammini. J. Exp. Med. 161:332-344.

75

Rosa, P. A., D. Hogan, and T. G. Schwann. 1991. Polymerase chain reaction analyses identify two distinct classes of Borrelia burgdorferi. J. Clin. Microbiol. 29: 524-532.

Rosa, P., and N. Margolis. 1994. Unique plasmid location and linkage of genes encoding purine biosynthetic enzymes and outer surface protein C in Borrelia burgdorferi. In Advances in Lyme borreliosis research. VI International Conference on Lyme borreliosis, Bologna, Italy 1994.

Rosa, P., and T. Schwan. 1992. Molecular biology of Borrelia burgdorferi. In Lyme Disease. Coyle, P. (ed). Philadelphia: B. C. Decker, Mosby Year Book, Inc. pp. 8-77.

Rosa, P. A., T. Schwann, and D. Hogan. 1992. Recombination between genes encoding major outer surface proteins A and B of Borrelia burgdorferi. Mol. Microbiol. 6:3031-3040.

Sadziene, A., P. A. Rosa, P. A. Thompson, D. M. Hogan, and A. G. Barbour.1992a. Antibody-resistant mutants of Borrelia burgdorferi: in vitro selection and characterization. J. Exp. Med. 176:799-809.

Sadziene, A., B. Wilske, M. S. Ferdows, and A. G. Barbour. 1993a. The cryptic OspC gene of Borrelia burgdorferi B31 is located on a circular plasmid. Infect. Immun. 61:2192-2195.

Sadziene, A., A. G. Barbour, P. A. Rosa, and D. D. Thomas. 1993b. An OspB mutant of Borrelia burgdorferi has reduced invasiveness in vitro and reduced infectivity in vivo. Infect. Immun. 61:3590-3596.

Sadziene, A., D. D. Thomas, and A. G. Barbour. 1994. An Osp-less mutant of Borrelia burgdorferi, biological and immunological characterization. In press.

Sadziene, A., D. D. Thomas, V. G. Bundoc, S. C. Holt, and A. G. Barbour. 1991. A flagella-less mutant of Borrelia burgdorferi structural, molecular and functional characterization. J. Clin. Invest. 88:82-92.

76

Sadziene, A., P. A. Thompson, and A. G. Barbour. 1992b. In vitro inhibition of Borrelia burgdorferi growth by antibodies. J. Infect. Dis. 167:165-172.

Saint Girons, I., I. G. Old, and B. E. Davidson. 1994. Molecular biology of the Borrelia bacteria with linear replicons. Microbiol. 40:

Salyers, A. A., and D. D. Whitt. 1994. Bacterial pathogenesis: A molecular approach. American Society for Microbiology Press, Washington D. C..

Schaible, U. E., M. D. Kramer, K. Eichmann, M. Modolell, C. Museteanu, and M. M. Simon. 1990. Monoclonal antibodies specific for the outer surface protein A (OspA) of Borrelia burgdorferi prevent Lyme borreliosis in severe combined immunodeficiency (seid) mice. Proc. Natl. Acad. Sci. (USA). 87: 3768-3772.

Schmid, G. P., A. G. Steigerwalt, S. E. Johnson, A. G. Barbour, A. C. Steere,I. M. Robinson, and D. J. Brenner. 1984. DNA characterization of the spirochete that causes Lyme disease. J. Clin. Microbiol. 20: 155-158.

Schmitz, J. L., R. F. Schell, S. D. Lovrich, S. M. Callister, and J. E. Coe. 1991. Characterization of the protective antibody response to Borrelia burgdorferi in experimentally infected LSH hamsters. Infect, and Immun. 59:1916-1921.

Schwan, T. G.. 1994. Oral presentation at the VI International Conference on Lyme borreliosis, Bologna, Italy.

Schwan, T. G., and W. Burgdorfer. 1987. Antigenic changes of Borrelia burgdorferi as a result of in vitro cultivation. J. Infect. Dis. 156:852-853.

Schwan, T. G., W. Burgdorfer and C. F. Garon. 1988. Changes in infectivity and plasmid profile of the Lyme disease spirochete, Borrelia burgdorferi, as a result of in vitro cultivation. Infect. Immun. 56:1831-1836.

Schwartz, J. A. Gazumyan, and I. Schwartz. 1992. rRNA gene organization in the Lyme disease spirochete, Borrelia burgdorferi. J. Bacteriol. 174: 3757-3765.

Scrimenti, R. J.. 1970. Erythema chronicum migrans. Arc. Dermatol. 102:104-105.

77

Shih, C-M., R. J. Pollack, S. R. Telford III, and A. Spielman. 1992. Delayed dissemination of Lyme disease spirochetes from the site of deposition in the skin of mice. J. Infect. Dis. 166:827-831.

Simon, M. M., U. E. Schaible, M. D. Kramer, C. Eckerskorn, C. Museteanu,H. K. Muller-Hermelink, and R. Wallich. 1991. Recombinant outer surface protein A from Borrelia burgdorferi induces antibodies protective against spirochetal infection in mice. J. Infect. Dis. 164:123-132.

Simon, M. M., U. E. Schaible, R. Wallich, and M. D. Kramer. 1991. A mouse model for Borrelia burgdorferi infection: approach to a vaccine against Lyme disease. Immunol. Tod. 12: 11-16.

Simpson, W. J., C. F. Garon, and T. G. Schwan. 1990. Analysis of supercoiled circular plasmids in infectious and non-infectious Borrelia burgdorferi. Microbiol. Path. 8: 109-118.

Smith, D. W., T. W. Yee, C. Baird, and V. Krishnapillai. 1991. Pseudomonad replication origins: a paradigm for bacterial origins? Mol Microbiol. 5: 2581-2587.

Sonnesyn, S. W., J. C. Manivel, R. C. Johnson, and J. L. Goodman. 1993. A Guinea pig model for Lyme disease. Infect. Immun. 61: 4777-4784.

Sparling, P. F., J. G. Cannon, and M. So. 1986. Phase and antigenic variation of pili and outermembrane protein II of Neisseria gonorrhoeae. Joum. Inf. Dis. 153:196-201.

Stechenberg, B. W.. 1988 Lyme disease: the latest great imitator. Pediatr. Infect. Dis. J. 7: 402-409.

Steere, A. C.. 1989. Medical progress: Lyme disease. N. Engl. J. Med. 321: 586-596.

Steere, A. C., R. I. Grodzicki, A. N. Kornblatt, J. E. Craft, A. G. Barbour, W. Burgdorfer, G. P. Schmid, E. Johnson, and S. E. Malawista. 1983. The spirochetal etiology of Lyme disease. N. Engl. J. Med. 308: 733-740.

78

Stoenner, H. G., T. Dodd, and C. Larsen. 1982. Antigenic variation of Borrelia hermsii. J. Exp. Med. 156:1297-1311.

Stover, C. K., G. P. Bansal, M. S. Hanson, J. E. Burlein, S. R. Palaszynski, J. F. Young, S. Koenig, D. B. Young, A. Sadziene, and A. G. Barbour. 1993. Protective immunity elicited by recombinant Bacille Calmette-Guerin (BCG) expressing outer surface protein A (OspA) lipoprotein: A candidate Lyme disease vaccine. J. Exp. Med. 178: 197-209.

Swanson, J., S. Bergström, K. Robbins, O. Barrera, D. Corwin, and J. M. Koomey. 1986. Gene conversion involving the pilin structural gene correlates with pilus+ pilus' changes in Neisseria gonorrhoeae. Cell 47:267-276.

Swofford, D.. 1990. PAUP: phylogenetic analysis using parsimony. Computer program distributed by the Illinois Natural History Survey, Campaign, version 3.0.

Thomas, D. D., and L. E. Comstock. 1989. Interaction of Lyme disease spirochetes with cultured eucaryotic cells. Infect, and Immun. 57:1324-1326.

Wallich, R., U. E. Schaible, M. M. Simon, A. Heiberger, and M. D. Kramer.1989. Cloning and sequencing of the gene encoding the outer surface protein A (OspA) of a European Borrelia burgdorferi isolate. Nucl. Acids Res. 17: 8864.

Wilske, B., V. Preac-Mursic, U. B. Göbel, B. Graf, S. Jauris, E. Soutschek, E. Schwab, and G. Zumstein. 1993. An OspA serotyping system for Borrelia burgdorferi based on reactivity with monoclonal antibodies and OspA sequence analysis. J. Clin. Microbiol. 31:340-350.

Wilske, B., V. Preac-Mursic, S. Jauris, A. Hofmann, I. Pradel, E. Soutschek, E. Schwab, G. Will, and G. Wanner. 1993. Immunological and molecular polymorphisms of OspC, an immunodominant major outer surface protein of Borrelia burgdorferi. Infect. Immun. 61:2182-2191.

Wilske, B., V. Preac-Mursic, and G. Schierz. 1985. Antigenic heterogeneity of European Borrelia burgdorferi strains isolated from patients and ticks. Lancet I: 1099.

79

Wilske, B., V. Preac-Mursic, G. Schiera, and K, V. Busch. 1986. Immunochemical and immunological analysis of European Borrelia burgdorferi strains. Zbl. Bakt. Hyg. A. 263:92-102.

Wilske, B., V. Preac-Mursic, G. Schiera, W. Gueye, P. Heraer, and K. Weber.1988a. Immunochemical analysis of the immune response in late manifestations of Lyme borreliosis. Zbl. Bakt. Mikrobiol. Hyg. A. 267: 549-558.

Wilske B., V. Preac-Mursic, G. Schiera, R. Kuhbeck, A. G. Barbour, and M. Kramer. 1988b. Antigenic variability of Borrelia burgdorferi. Ann. N. Y. Acad. Sci. 539: 126-143.