This manuscript has been reproduced 6om the microfihn master. UMI
fihns the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, v ^ e others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely afreet reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., m ^s, drawings, charts) are reproduced by sectioning the original, hymning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
UMI
A Bell & Bbwell Infomiation Conqany
300 North Z e* Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600
intracellular pathogen Francisella
novicida
b y
Gerald Stephen Baron, B .Sc., University of V ictoria, 1993
A D issertation Submitted in Partial Fulfillm ent o f the Requirements fo r the Degree of
DOCTOR OF PHILOSOPHY
in the Department of B iochem istry and M icrobiology
W e accept this dissertation as conform ing to the required standard:
Dr. F.E. Nano, Supervisor (Department of Biochemistry and Microbiology)
Dr. J.T. Buckley, Departmental Member (Department of Biochemistry and
Dr.'^.T^T OlafsoTir ^ ^ t m e n t a l Member (Department of Biochemistry and M icrobiology)
ton.
Dr. C. Upton, Departmental Member (Department of Biochemistry and M icrobiology)
Dr. N.M^ Sherwood, Outside Member (Department of Biology)
Dr. G.A. McClarty, External Examiner (University of Manitoba)
© Gerald Stephen Baron, 1998, University of V ictoria
All rights reserved. This dissertation may not be reproduced in whole or in part, by photocopying or other means, w ithout the perm ission o f the author.
Supervisor: Dr. Francis E. Nano
ABSTRACT
Francisella novicida is a facultative intracellular pathogen capable
of growing in m acrophages. A spontaneous m utant of F. novicida defective for grow th in macrophages was isolated on LB media containing the chrom ogenic phosphatase substrate 5-brom o-4-
chloro-3-indolyl phosphate (X-p) and designated GB2. Using an in cis com plem entation strategy, four strains were isolated which are
restored for growth in macrophages. A locus isolated from one of these strains com plem ents GB2 for the intracellular growth defect, colony morphology on LB (X-p) media, and virulence in mice. The locus consists of an apparent operon of two genes, designated m g lA B , for m acrophage grow th locus. Both m g lA and mglB tra n sp o so n
insertion mutants are defective for intracellular growth and have a phenotype similar to GB2 on LB (X-p) media. Sequencing of m g IA cloned from GB2 identiHed a missense mutation, providing evidence that both m g IA and m g IB are required for the intram acrophage growth of F. novicida. Preliminary studies have also identified a convergently transcribed gene, tentatively designated m g lC ,
immediately dow nstream o f mg IB. m g lC null m utants are defective for intracellular grow th and show the same phenotype on LB (X-p) agar as GB2. m g lB expression in GB2 was confirmed using antiserum against recom binant M glB. Western blot analysis revealed the
absence of MglA in an m g lB null mutant^ indicating MglB may
influence MglA levels. Analysis o f the regulation o f m glA expression during growth in broth culture shows a decrease in expression upon entering late log-early stationary phase. m glA is also expressed
during culture in macrophages. C ell fractionation studies revealed several differences in the protein profiles of m g l m utants com pared with wild-type F. novicida, most notably the absence of a 70 kDa secreted protein. A candidate clone for the gene encoding this 70 kDa protein has been isolated. The deduced amino acid sequences of
m g lA and m g lB show similarity to the SspA and SspB proteins of Escherichia coli and H aem ophilus spp. In E. coli, SspA and/or SspB
influence the levels o f multiple proteins under conditions of
nutritional stress, and SspA can associate with the RNA polymerase holoenzyme. Taken together, these observations suggest that in
F rancisella MglA and MglB may control the expression of genes
whose products contribute to survival and growth within
macrophages. Roles for the putative MglC and possibly the 70 kDa secreted protein in this activity are also indicated.
Acid phosphatases capable o f inhibiting the respiratory burst of neutrophils have been identified in certain intracellular pathogens. The gene encoding AcpA, a respiratory burst-inhibiting acid
phosphatase of F ra n cisella , was cloned and sequenced. The deduced a m ino acid sequence of AcpA showed limited sim ilarity to
phospholipase C proteins present in Pseudom onas aeruginosa and
found to exhibit w ild-type growth kinetics in both cell-line and inflam m atory m ouse m acrophages as w ell as rem aining v iru len t for mice. These data suggest that AcpA is not essential for the
intracellular growth or virulence o f F. novicida, and that any role it may play in virulence is subtle.
E xam iners:
Dr. F.E. Nano, Supervisor (Department of Biochemistry and Microbiology)
D r J.T. Buckley, Departmental Member (Department of Biochemistry and M icrobiology)
Dr: kTW. OlafsoK Departmental Member (Department of Biochemistry and M icrobiology)
oh.
Dr. C. Uptdh, Departmental Member (Department of Biochemistry and M icrobiology)
Dr. N.M. Sherwood, Outside Member (Department of Biology)
ABSTRACT... i i
TABLE OF CONTENTS... v
LIST OF TABLES...v iii U S T OF FIGURES... i x LIST OF ABBREVIATIONS... -xii
ACKNOWLEDGMENTS... jcvi INTRACELLULAR PATHOGENS...1
S a lm o n e lla ...4
INTRODUCTION... .4
INVASION... 5
The secretion apparatus... 7
S ecreted proteins I 2 R egulatory proteins 1 5 INTRACELLULAR REPLICATION... 1 7 phoP Q ... 2 2 Resistance to antim icrobial p e p t id e s ... 2 5 Adaptation to M g^^-lim iting e n v iro n m e n ts ... 2 7 Survival w ithin m acro p h ag es 2 8 m s g A ...2 9 S P I - 2 3 0 s l y A ... 3 1 f k p A ... 3 3 F rancisella... 3 4 INTRODUCTION... 3 4 TULAREMIA - CLINICAL AND TRANSMISSION 3 6 MORPHOLOGY... 3 8 INTRACELLULAR GROWTH...3 8 Host c e lls ...3 9 In tra c e llu la r c o m p a rtm e n t... 4 1 V irulence factors affecting intracellular g r o w th ...42 m in D ... .43 v a lA B ... .43 LPS phase variation...4 4 Stress response... .45 23 kDa protein... .46
HOST RESPONSE AND IMMUNITY TO F. TU LAREN SIS... A 6 M a cro p h ag e s...4 7 In n ate factors... .4 9 H um oral factors... 3 2 C ellular factors... ^ 4 RATIONALE OF RESEARCH...J 8 COLLABORATIONS...6 0
MATERIALS AND METHODS...6 1
B acterial stra in s...6 1
Isolation o f a spontaneous mutant o f F. novicida...6 1
Intracellular growth assay ...6 5
Mice and experim ental infections...6 6
D eoxycholate and com plem ent sensitivity a ssa y s...6 6
Complementation of F. novicida GB2 for growth in
m a c ro p h a g e s ...6 7
Recombinant DNA techniques and DNA sequencing...6 8
Plasm id construction... 7 0 Isolation o f plasmids which complement GB2 for
in tracellu lar g ro w th ... 7 7 SDS-PAGE and W estern im m unoblotting...7 7 Overexpression and puriHcation o f recom binant MglB and
M glA...7 8
Preparation of an tisera ... 7 9 Construction of an m g lA ’-cat fusion strain of F. novicida...8 0
Regulation o f mg IA expression...8 1
i) preparation of cell extracts...8 1
ii) assay for CAT activity...8 3
Cell fractionation...8 5
Ammonium sulfate fractionation o f F. novicida culture
s u p e r n a ta n ts ...8 6
Southern blotting...8 6
Protein m icrosequencing...8 7
CHAPTER 1. Identification of a locus of Francisella novicida encoding two putative global regulators o f intram acrophage
g r o w th ...8 9
INTRODUCTION...8 9
RESULTS... 9 3 Isolation of a spontaneous mutant o f F. novicida... 9 3 GB2 is defective for growth in m acrophages... 9 3
Complementation of GB2 fo r growth in m acrophages 9 9 Isolation of plasmids w hich com plem ent GB2 for
growth in m acrophages...1 0 2
Cloning of the intracellular growth locus defective
in G B 2...10 5 DNA sequencing and analysis of m g lA B ... 10 7
mglAB insertion mutants are defective for growth
in m acrophages... 1 1 6
mglA is required for intracellular grow th... 117 mglA is required for virulence in m ic e ... 121 W estern blot analysis o f m g lB expression... 1 2 4 W estern blot analysis o f m g lA ex p ressio n ... 12 7 Regulation of m g l A expression...1 3 0 Cell fractionation of m g l m utants... 13 5 Evidence for a third mgl gene - m glC ... 1 4 0 N-terminal sequencing o f a 70 kDa secreted protein 1 4 4 Isolation of a candidate clone for the gene encoding
the 70 kDa protein... 1 4 4 Analysis of mutants defective for secretion of the
70 kDa protein...1 4 9 DISCUSSION...15 1 CHAPTER 2. Analysis of AcpA, a respiratory burst-inhibiting
acid phosphatase of F ra n cisella ... 16 1 INTRODUCnON... 161 RESULTS... 1 6 4 Cloning and sequencing o f acpA ... 1 6 4 Growth of an F. novicida acpA m utant in
m a c ro p h a g e s...17 2
Virulence of F. novicida acpA m utant in mice... 17 7 DISCUSSION...1 7 9 CONCLUSIONS... 1 8 4 REFERENCES... 18 6
LIST OF TABLES
Table 1. B acterial strains, plasm ids, and bacteriophages. 6 2
LIST OF FIGURES
Figure 1. Growth o f F. novicida mutants on LB (X-p) agar. 9 4
Figure 2. Growth of m g l mutants in macrophages and in TSB-C. 9 7
Figure 3. Com plem entation of GB2 for growth in m acrophages. 1 0 0
Figure 4. Growth o f complemented GB2 strain in J774A.1
m a c ro p h a g e s . 1 03
Figure 5. Com plem entation of GB2 for growth in J774A.1
m acrophages with pC3-20. 1 0 4
Figure 6. Southern blot analysis o f GB2 m utant and
com plem ented strain from library o f GB2
t r a n s f o r m a n ts . 1 0 6
Figure 7. Southern blot analysis of F. novicida m gl m utants
and com plem ented GB2 strain. 1 0 8
Figure 8. Restriction m ap and ORF organization of the m g lA B
re g io n . 1 1 0
Figure 9. N ucleotide and deduced am ino acid sequence o f the
1309 bp m g lA B region o f pGB40. 1 1 2
Figure 10. Amino acid alignment between MglA and SspA (A)
and between MglB and SspB (B). 1 1 5
Figure 11. Growth o f F. novicida m gl mutants in J774A.1
m a c ro p h a g e s . 1 1 8
Figure 12. Growth o f F. novicida mutants on LB (X-p) agar. 1 1 9
Figure 13. Com plem entation of GB2 for growth in J774A.1
m acrophages with pGB48Em. 1 2 2
Figure 14. Growth o f complemented GB2 strain in spleens o f
Figure 16. Analysis of MglA expression in F. novicida 128
Figure 17. Construction o f F. novicida m g lA ’-cat transcriptional
fusion strain (GB7). 131
Figure 18. Southern blot analysis o f F. novicida m g lA ’-cat fusion
s tra in . 1 3 4
Figure 19. Analysis of the regulation o f m g lA expression. 1 36
Figure 20. Cell fractionation of m g l m utants. 13 8
Figure 21. Growth o f F. novicida m g lC m utant in J774A.1
m a c ro p h a g e s . 141
Figure 22. Southern blot analysis of F. novicida mglC mutant. 1 4 2
Figure 23. Ammonium sulfate fractionation of F. novicida
culture supernatant p ro tein s. 145
Figure 24. Southern blot analysis o f F. novicida with
oligonucleotides corresponding to the N-terminus of
the 70 kDa protein. 147
Figure 25. A nalysis o f culture supernatants o f putative
F. novicida 70 kDa protein m utants. 1 5 0
Figure 26. Restriction map and ORF organization of the a cp A
reg io n . 165
Figure 27. Nucleotide and deduced am ino acid sequence of the
1798 bp acpA region o f GB3. 166
Figure 28. Amino acid alignm ent between AcpA and PLC
proteins from Af. tuberculosis (MpcA, MpcB) and P.
Figure 29. Southern blot analysis of F. novicida acpA m utant. 1 7 4
Figure 30. Growth of F. novicida acpA null mutant in
m a c ro p h a g e s . 1 7 5
Figure 31. Growth o f F. novicida acpA null mutant in spleens of
ABC A TP-binding cassette
AGP acid phosphatase
A p am p icillin
ATP adenosine trip h o sp h a te
A T P ase adenosine trip h o sp h a ta se
b p base pairs
CAT ch lo ra m p h en ic o l a c e ty ltra n sfe ra se
Cb c a rb e n ic illin
O E cell-free ex tra ct
c fa colony form ing units
CHA-B cystine heart agar w ith horse blood Cm c h lo ra m p h e n ic o l
CoA coenzym e A
dA TP d eo x y ad en o sin e trip h o sp h a te dCTP d eo x y cy to sin e trip h o sp h a te dGTP d eo x y g u an o sin e trip h o sp h a te
DMEM D ulbecco’s ModiAed E agle M edium
DNA deoxyribonucleic acid
DTK d elay e d -ty p e h y p e rs e n s itiv ity dTTP d eo x y th y m id in e trip h o sp h a te
ELISA enzym e-linked im m u n o so rb an t assay
Em e ry th r o m y c in
FKBP FK 506-binding protein
ΠD gam m a interferon knockout GTP g u an o sin e triphosphate h h o u r ( s ) id in tr a d e r m a l IFN in te r f e r o n Ig A im m unoglobulin A IgG im m unoglobulin G IgM im m unoglobulin M in in tr a n a s a l ip in tr a p e r ito n e a l
IPTG isopropyl P -D -th io g alacto p y ra n o sid e I P3 in o sito l 1,4,5-trisphosphate
iv in tr a v e n o u s
k b kilobases or kilobase pairs
kD a k ilo d a lto n
Km k a n a m y c in
LAM P lyso so m e-asso ciated m em brane glycoprotein
LAP lysosom al acid phosphatase
LB L u ria -B e rta n i
LD50 50% lethal dose
LVS live vaccine strain
LPS lip o p o ly sa c c h a rid e
mCi m illic u rie
MES 2- [N -m orpholino] eth an esu lfo n ic acid
pM m ic ro m o la r
m M m illim o la r
m m o l m illim o le
M 6PR m annose 6-phosphate receptor
MUP 4 -m e th y lu m b e llife ry lp h o s p h a te
NK natural k ille r
NMMA iV G -m onom ethyl-L -arginine
ORF open reading fram e
PBMC peripheral blood mononuclear cell PBS p h o sp h a te -b u ffe re d saline
PGR polym erase chain reaction
P IP : p h o sp h a tid y lin o sito l 4 ,5 -b isp h o sp h ate
PLC phospholipase C
PM polym yxin B
PM A phorbol 12-m yristate 13-acetate
PMN p o ly m o rp h o n u clear leukocyte PMSF p h en y lm eth y lsu lfo n y l fluoride pNPPC p -n itro p h e n y lp h o s p h o ry lc h o lin e P P Ia s e peptidyl-prolyl c is-tra n s isom erase P T P ase protein tyrosine phosphatase
rDNA ribosom al deoxyribonucleic acid
s se co n d s
sc s u b c u ta n e o u s
scid severe com bined im m unodeficient
SD stan d ard d e v ia tio n
SDS-PAGE sodium d o d ecy l sulfate-polyacrylam ide gel e le c tro p h o re s is
SP spacious phagosom e
S P I S a lm o n e lla pathogenicity island
TBS-T T ris-buffered saline with T w een-20
TCR T cell receptor
TDL thoracic d u ct lym phocyte
TNF tum or necrosis factor
T ris T ris [h y d ro x y m e th y l] a m in o m e th a n e TSB-C tryptic soy broth with cysteine
ACKNOWLEDGMENTS
I would like to thank the members of the Departm ent o f Biochem istry and M icrobiology for years of helpful advice and support. Special thanks to A lbert Labossiere and Scott Scholz for their superb technical support on countless occasions. T he m em bers of the U niversity of V ictoria Animal Care Facility provided quality care for the animals used in my experiments.
Thanks to all members o f the Nano Lab for being such great teammates on a daily basis.
Finally, many thanks to Fran for his patience and for being an unrelenting source of optim ism and encouragem ent.
Host organisms possess many defense mechanisms designed to resist infection. Primary barriers to infection are both physical and chemical in nature. Physical barriers such as skin and mucous membranes, which are generally impermeable to m ost infectious agents. Both are constantly sloughing cells from the surface, and in the mucous membrane, produce a mucous layer to trap foreign particles and block adherence to epithelial cells. O ther inhibitory factors of the skin include low pH and dryness. Availability o f essential nutrients, such as iron, is limited at certain locations by high affinity iron binding proteins. Also, a normal microbial flora exists at many locations w hich can block potential colonization sites and compete for nutrients. Secondary barriers are encountered on entry into the blood or underlying tissues and include com plem ent and professional phagocytic cells (e.g. neutrophils, m acrophages). At later time points, products are generated by effectors o f a specific acquired immune response (activated m acrophages, lym phocytes).
Successful pathogenic microbes have the capacity to infect a host organism, often by overcom ing epithelial barriers to infection,
becoming localized in an environm ent suitable to support replication, and proliferating in this environm ent prior to transm ission of the amplified microorganisms to new hosts. One class of bacterial pathogens, intracellular pathogens, accomplish these tasks by entering into, and surviving within, eukaryotic cells. The
invading microbe. Invasion o f epithelial cells can allow pathogens to cross epithelial barriers. The host cell provides a potential source of nutrients to support replication in the absence of com petition with normal flora. W ithin the host cell, the bacterium gains some
protection from the innate and acquired im m u n e responses in the extracellular environment. F inally, infection of some cells m ay facilitate the spread o f the organism to new host tissues.
An intracellular lifestyle does not come without complications. Both non-professional and professional phagocytic cells
(macrophages) can serve as hosts to intracellular pathogens.
However, growth within non-professional phagocytic cells requires the bacteria to have m echanism s to induce their own uptake by these host cells. In both host cell types, the invading microbe must possess strategies to deal w ith the potential consequences o f
phagosome-lysosome fusion, w hich could result in exposure to a number of toxic substances including hydrolytic enzymes (lysozym e, proteases, glycosidases, lipases), small cationic peptides (e.g.
defensins), and lactoferrin (M oulder, 1985). Phagocytosis by
macrophages is associated w ith the production of other antim icrobial molecules called reactive oxygen interm ediates, and survival and replication within these cells requires additional specialized
m ech an ism s.
The strategies that intracellular pathogens have evolved to permit survival and growth within eukaryotic cells can be broadly classihed
prior to phagosom e-lysosom e fusion; 2) adaptation to survival within the phagolysosome; and 3) inhibition o f phagosom e-lysosom e fusion. Bacteria in the first category replicate in the nutrient-rich
environm ent of the cytoplasm and spread intercellularly to new host cells, thereby avoiding an extracellular phase. Pathogens in the
latter two categories replicate within specialized vacuoles resulting from altered interactions w ith normal phagocytic pathways. This review will sum m arize the virulence factors contributing to the intracellular parasitism o f one of the m ost w ell-studied model intracellular pathogens. Salm onella enterica.
INTRODUCTION
Salm onella en terica is recognized as a pathogen of a wide variety
m am m als from hum ans to mice. Infection can be manifested in the form of a self-lim iting gastroenteritis, enteric fever, or a severe systemic infection known as typhoid fever. M any S a lm o n e lla
serovars exhibit a high degree of host specificity. Hence, the type of disease resulting from infection is dependent on the species o f the infected host and/or the serovar o f the infecting bacteria. For exam ple, 5. enterica serovar ty p h i (S. typhi) specifically infects
humans to cause typhoid fever. S. enterica serovar ty p h im u riu m (S. ty p h im u r iu m ) causes a typhoid-like disease in mice but self-lim iting
gastroenteritis in hum ans.
S. typhim urium infection o f mice has been used extensively as a
model to study S. typhi pathogenesis. S a lm o n e lla infection is
initiated by ingestion o f contaminated food or water. Bacteria breach the intestinal epithelial barrier by adhering to and invading the M cells of Peyer’s patches (Jones e t al., 1994). M cell invasion leads to cell death within 60 minutes o f addition o f bacteria to ligated ileal loops. Two hours post-infection, bacteria are observed within adjacent enterocytes and penetrating the underlying follicle dome where they are phagocytosed by m acrophages. Replication within these cells is follow ed by a bacterem ia w hich results in the spread of
numbers in these organs, apparently w ithin m acrophages (Richter- Dahlfors et al., 1997), prior to being released into the bloodstream, resulting in septicem ia and death in susceptible m ice. Invasion of epithelial cells and intracellular replication are considered im portant factors in S a lm o n e lla pathogenesis, and the genes shown to
contribute to these activities are the focus of the follow ing discussion.
INVASION
Invasion of M cells in the ileal Peyer's patches is the earliest detectable interaction between S a lm o n e lla and host intestinal tissue (Jones et al., 1994). PhenotypicaUy, entry into M cells resembles that observed into cultured epithelial cells, and many of the host and bacterial factors involved have been identiried using the tissue culture cell m odel. Bacterial contact with the host cell induces m acropinocytosis, a process characterized by the form ation of large membrane ruffles on the host cell mem brane which enclose the bacteria in a m em brane-bound com partm ent (Francis et al., 1992; 1993). M em brane ruffling is associated w ith dram atic
rearrangem ents o f the actin cytoskeleton, which returns to normal shortly after bacterial entry.
The host cell signalling events which accompany S a lm o n e lla invasion include intracellular calcium and inositol phosphate fluxes, but the exact signalling pathway rem ains to be determ ined
rearrangements m ay be mediated by affecting the activity o f CDC42, one of three sm all guanosine triphosphate (G TP)-binding proteins involved in controlling the formation o f actin-based structures, as
Salm onella requires CDC42 for invasion (Chen et al., 1996a). S a lm o n e lla entry into macrophages by ruffling results in the
induction of apoptosis in a high proportion of the host cells (Chen e t
al., 1996b; M onack et al., 1996). O ther intracellular pathogens, including S h ig e lla (Zychlinsky et a i , 1992) and L e g io n e lla
p n e u m o p h ila (M üller et a i , 1996), also induce apoptosis in
macrophages. M utants defective for invasion o f epithelial cells invade a m urine monocyte-macrophage cell line (RAW 264.7) 7- to 10-fold less efficiently than wild-type S. typhim urium (Monack et a i , 1996). Invasive strains defective for intram acrophage growth retain the ability to induce apoptosis, indicating intracellular replication is not necessary for cytotoxicity. Reports are conflicting as to w hether bacterial uptake is necessary to induce apoptosis. Although S.
ty p h im u riu m has recently been shown to induce apoptosis in
phagocytes of infected mice (Richter-Dahlfors et a i , 1997), the role of this activity in the interaction of S a lm o n e lla w ith macrophages is unclear as a non-invasive mutant (BJ66) is still capable of replicating in macrophages (M onack et a i , 1996) and resides in an intracellular com partm ent identical to that occupied by a w ild-type strain
(Rathman et a i , 1997). In addition to invasion-associated genes, genes regulated by the OmpR/EnvZ tw o-com ponent regulatory
m utants fail to induce apoptosis in m acrophages (Lindgren et a i, 1 9 9 6 ).
S a lm o n ella -in d u ced m odifications to h o st cell signalling pathways
are predicted to be carried out by a set o f effector proteins
translocated to the surface o f the bacterium or into the host cell by a type i n or contact-dependent protein secretion system (Galan,
1996). The genetic loci encoding the secretory apparatus, secreted proteins, and regulatory proteins com prising this invasion-associated system are largely clustered in a 40 kb segm ent at 58 to 60 min on the S a lm o n ella chrom osom e known as S a lm o n e lla pathogenicity island 1 (SPI-1) (M ills et al., 1995). This region constitutes one of three pathogenicity islands in S a lm o n e lla , which are thought to be acquired by horizontal gene transfer fro m other microorganisms (Groisman and Ochman, 1996). The m ajor genetic elem ents involved in S a lm o n ella entry into host cells are discussed below.
T h e s e c r e tio n a p p a r a t u s
O f at least 28 genes present in SPI-1, 17 have been shown or are predicted to encode com ponents of the secretion apparatus (Galan, 1996). These com ponents are encoded in the sp a , inv, and p rg loci and have been identified through a variety o f approaches. Putative functions have been assigned based m ainly on amino acid sequence similarities to proteins in other systems such as S h ig e lla and Y e rsin ia spp. Homologs o f the sp a , in v , and p r g genes in other bacteria are
sim ilarly grouped (G roism an and Ochman, 1993; Pegues et aL, 1995) and, in some cases, functionally interchangeable (Groisman and Ochman, 1993; R osqvist e t aL, 1995). F or example, a non-invasive S.
ty p h im u riu m s p a P m utant is complemented for invasion by sp a 2 4 ,
the corresponding locus in Sh ig ella (Groism an and Ochman, 1993). The first SPI-1 genes identified were isolated by
com plem entation o f a spontaneous non-invasive m utant of S.
typ h im u riu m with a cosm id clone (Galan and Curtiss, 1989).
Analysis of subclones o f the cosmid led to the identification of a
group of four genes, designated inv A , -B, -C , -D, w hich complemented the mutant strain, in v A , -B , and -C are organized w ithin an operon
while invD is located dow nstream in an independent transcriptional unit. Both invA and in v C have been shown to be required for entry into epithelial cells (Galan et aL, 1992; Eichelberg et aL, 1994). in v A mutants exhibit reduced virulence in m ice following oral infection, but retain w ild-type virulence when inoculated by the
intraperitoneal route (G alan and Curtiss, 1989). Using a ligated ileal loop model, Jones and Falkow (1994) reported S. typhim urium in v A mutants are defective fo r invasion and destruction o f M cells in the follicle-associated epithelium o f murine Peyer’s patches. These data suggest that invasion of host cells by S a lm o n e lla plays an important role in establishing infection by the oral route. This has been
supported by m ore rec e n t studies using non-invasive strains with mutations in other loci (Penheiter et aL, 1997). Analysis of the predicted secondary structure of InvA suggests it is a membrane
terminal half o f the protein and a hydrophilic carboxyl terminus oriented into the cytoplasm (G alan et aL, 1992). Cell fractionation studies and analysis of translational fusions o f in v A to alkaline
phosphatase ip h o A ) are consistent with InvA being a poly topic inner membrane protein (Galan, 1996). It has been suggested that InvA may form a channel to allow transport of secreted proteins across the inner membrane. Based on sequence sim ilarity to the P subunit of the FqFi proton-translocating ATPase, InvC is suggested to energize the secretion o f proteins through this type HI secretion pathway (Eichelberg et aL, 1994). Purified InvC has been shown to exhibit ATPase activity, and this activity is essential for invasion (Eichelberg
et aL, 1994). Both InvA and InvC are required for the proper assem bly and shedding/retraction o f surface appendages termed invasomes, which are formed shortly after bacterial contact with epithelial cells but disappear immediately preceding entry (Ginocchio
et aL, 1994).
Transposon mutagenesis o f loci adjacent to in vA B C resulted in the isolation of two new genes required for invasion, inv F (described below) and invG (Kaniga et aL, 1994). The deduced sequence of InvG contains a putative signal sequence and is sim ilar to the PulD family of protein translocases. This family includes components of type III secretion systems in other Gram-negative bacteria, such as MxiD of
Shigella fle x n e ri, required fo r the surface presentation o f the Ipa
of the Y ep proteins (Kaniga et al., 1994). By analogy with the role of PulD in pullulanase secretion in K le b sie lla , InvG is postulated to be an outer mem brane p ro tein and assist in the transport o f targ e t proteins across this mem brane. Evidence of a role for InvG in protein
secretion com es from the observations that inv G m utants are defective for the secretion of at least six proteins (Penheiter et aL,
1997) and in the assem bly o f surface appendages which are induced immediately after contact with epithelial cells (Ginocchio e t aL,
1994).
The S. typhim urium PhoP/Q two-com ponent regulatory system affects the expression o f several phenotypes including virulence in mice and survival in m acrophages (discussed in m ore detail below). PhoP/Q activate and repress the expression of various genes, termed
pa g (^ A o f-activ ated ^enes) and p r g (g.A of-repressed ^ e n e s ),
respectively (M iller and M ekalanos, 1990). A strain expressing a mutant PhoQ (p h o P Q ^) with increased net kinase activity (PhoPc phenotype) has constitutive expression of p a g and repression o f p r g , and is attenuated fo r virulence in mice, survival w ithin m acrophages, and invasion of epithelial cells, suggesting p rg are virulence genes (M iller and M ekalanos, 1990; G unn et aL, 1996). Using the
transposon T n p h o A , Behlau and M iller (1993) identified a p r g locus
(p rg H ) which contributes to m ouse virulence by the oral or
intraperitoneal routes and to entry into epithelial cells. T he data suggest this locus plays a role in both invasion o f the m ucosa and survival during interactions w ith phagocytic cells. H ow ever,
p rg H ::T n p h o A m utants exhibit w ild-type survival in macrophages.
Characterization of the region adjacent to p r g H revealed that it is in an operon of four genes designated p r g H U K (Pegues et al., 1995). The predicted sequences of PrgH, PrgI, PrgJ, and PrgK are sim ilar to proteins required for secretion o f Ipa (invasion jgrotein ^ t i g e n s ) and Yop (Yersinia fluter jgroteins) virulence factors. PrgH and PrgK are both predicted to be lipoproteins on the basis o f potential lipoprotein processing sites in the N-termini. The Y e rsin ia (YscJ) and S h ig e lla (MxiJ) homologs of PrgK are m em brane-associated lipoproteins (Pegues et a l , 1995). Homologs for PrgH and PrgJ have only been found in S h ig e lla (Lee, 1997). The localization and functions o f PrgI and PrgJ are unknown. A prgH w H nphoA m utant was shown to be defective for secretion of several proteins, suggesting at least one of the proteins in the p r g H U K operon has a role in protein secretion.
The ability of S a lm o n e lla to enter m am m alian cells is regulated by a number of environm ental conditions including oxygen tension (Lee and Falkow, 1990), growth state (Lee and Falkow , 1990), and
osmolarity (Galan and Curtiss, 1990). Jones and Falkow (1994) identified a group o f invasion genes by screening for oxygen- regulated la c Z Y transcriptional fusions. One gene, o rg A , was sequenced and characterized in more detail. Sim ilar to the in v A
mutant phenotype, an orgA m utant (B J66) is attenuated for virulence in mice (>60-fold) by the oral route o f infection but remains fully virulent by the intraperitoneal route. T his correlates with the inability o f BJ66 to invade and destroy M cells in a murine ligated
ileal loop model, and with the observation that it is defective in the secretion of several proteins (Penheiter et al. y 1997). The o rg A gene is predicted to encode a 48 kD a hydrophilic protein which lacks a signal sequence. OrgA is sim ila r to MxiK of S h ig ella spp., and is one of five putative secretion m achinery components for w hich homologs have been found only in S h ig e lla (Lee, 1997).
Secreted
proteins
Several proteins secreted by the SPI-1 type HI secretion system have been identified. These proteins may participate in a variety of activities including form ation o f invasomes, regulation o f the
secretion process, translocation of secreted proteins into host cells, and execution of putative effector functions within host cells. In addition, two secreted proteins, InvJ and SpaO, are necessary for the export of all proteins known to be secreted through this system to date, suggesting they may be components o f the export apparatus (Collazo et al., 1995; Collazo and Galan, 1996).
Among the first secreted proteins identified were the products of the sip (or ssp) operon (Kaniga et al., 1995a; Kaniga et al., 1995b). This operon encodes four proteins, SipB, SipC, SipD, and SipA, which show sequence sim ilarity to the S h ig e lla IpaB, IpaC, IpaD, and IpaA proteins, respectively. The s ip operon is located ju st downstream o f the spa operon. O f the four proteins, only SipA is not essential for invasion of epithelial cells and (with the exception of sip B which has not been tested) full virulence in mice by the oral route (Kaniga e t
aL, 1995a; Kaniga e t al., 1995b; Penheiter et al., 1997). SipB and SipC
have recently been shown to be translocated into cultured intestinal epithelial cells (Collazo and Galan, 1997). SipB, SipC, and SipD are required for the translocation o f SipB and SipC into host cells. Also, it has been proposed SipD plays a role in m odulating the export of
some secreted proteins as culture supernatants from a sip D m utant contain increased levels of SipA, SipB, and SipC (Kaniga et al., 1995a) as well as SopE (see below; Hardt et al., 1998). Although SipA is detected on the surface of the bacteria after infection, it is not required for the translocation process (C ollazo and Galan, 1997).
A second set o f secreted proteins called Sops { ^ I m o n e l la ^ u t e r proteins) were identified in a S. dublin double fliM /p o la r s ip B
mutant (Wood et at., 1996). These proteins, designated Sop A to SopE, form large filam entous aggregates in the culture medium. One
protein, SopE, has been independently characterized by two groups (Wood et al., 1996; Hardt et al., 1998). SopE is necessary for efficient invasion of HeLa cells and Henle-407 cells during short infection times (15 min). A s o p E mutant induces less extensive cytoskeletal rearrangem ents and diffuse m em brane ruffles as com pared to those induced by a w ild-type strain (Hardt et al., 1998). SopE may
function as an effector o f host cell responses as it is translocated into host cells by a m echanism dependent on a product(s) of the s ip
operon (Wood e t al., 1996) but is not essential for secretion o f other targets of the in v !s p a type III secretion system or translocation of SipC into host cells (Hardt et al., 1998). Interestingly, the so p E gene
is only present in a small subset o f S a lm o n e lla serovars and, in S.
typ h im u riu m ^ maps w ithin the genom e of a cryptic bacteriophage
located outside SPI-1 (H ardt et aL, 1998).
ECaniga et al. (1996) identified a secreted effector p ro tein , denoted SptP, requiring the SPI-1 type HI secretion apparatus fo r secretion. The encoding gene, s p tP , is located downstream o f the s ip operon in SPI-1. The deduced am ino acid sequence of SptP suggests the
protein has a m odular structural organization. The am ino-term inal region shows similarity to the toxins ExoS of P se u d o m o n a s
a eru g in o sa and YopE of Y ersin ia spp., both o f which are targets of
type i n secretion systems and are involved in host cell dam age. The carboxy-term inal region is sim ilar to the catalytic dom ain found in eukaryotic tyrosine phosphatases (PTPases) and another ta rg e t o f the
Y ersin ia type IH secretion system, the tyrosine phosphatase YopH.
Purified SptP exhibits tyrosine phosphatase activity using phosphorylated peptide substrates. SptP is required fo r fu ll virulence in mice, but is not required for invasion of ep ith elial or macrophage cell lines. Very recently, SptP was shown to be
translocated into host cells in a fip-dependent fashion (Y u and Galan, 1998). M icroinjection o f purified SptP into cultured ep ith elial cells results in disruption o f the actin cytoskeleton and disappearance o f stress fibers. Either the am ino-term inal or carboxy-term inal dom ain is sufficient to effect this activity, suggesting the two putative
effector dom ains m ay function independently b u t affect sim ilar cellular functions.
Regulatory
proteins
The regulation o f S. typhim urium invasion is com plex and m odulated in vitro by several environmental factors. Conditions such as oxygen tension, osmolarity, bacterial grow th state, and pH have all been shown to influence S a lm o n e lla entry into host cells (Galan and Curtiss, 1990; Lee and Falkow, 1990; Behlau and Miller,
1993). The m echanism by which these conditions are sensed in order to modify the expression of invasion genes is poorly
understood. Changes in the level of DNA supercoiling can affect in v gene expression (Galan and Curtiss, 1990). C urrent data suggest a model whereby the expression of invasion genes is coordinately regulated by activating the synthesis of a cascade o f transcription factors in response to appropriate environm ental cues. Sub-optim al conditions for any particular environm ental or regulatory factor dram atically represses invasion gene transcription (B ajaj et a i,
1996). Invasion gene expression is under the control o f at least four factors, InvF, HilA, SirA, and PhoP/Q, whose encoding genes are located both inside and outside SPI-1.
The in v F gene is located immediately upstream o f in v G in SPI-1 and is required fo r entry into cultured epithelial cells (K aniga et aL,
1994). The predicted sequence identities InvF as a m em ber of the AraC family of transcriptional activators, which includes the VirF invasion regulators o f S h ig e lla and Y ersin ia . InvF induces the
transcription o f the secreted proteins encoded in the sip operon (Johnston et aL, 1996).
The h ilA locus fhyperinvasion locus) was identified in a search for hyperinvasive S. typhim urium mutants able to enter cultured
epithelial cells even after growth under invasion-repressing
conditions (Lee et aL, 1992). As for InvF, HilA is necessary both for invasion o f epithelial and M cells, and virulence in mice by the oral route o f infection (Bajaj et aL, 1995; Penheiter et aL, 1997). The amino-terminal region o f HUA is similar to the DNA binding and transcriptional activation dom ain o f the OmpR/ToxR family of
transcriptional activators. H ilA coordinately regulates the expression of at least six genes including sipA , sipC , p rg H , p rg K , o rg A , invF, and
orgA (Bajaj et aL, 1995; Bajaj et aL, 1996). The activation of inv F and p rg H expression in Escherichia coli by HilA suggests HilA may
activate transcription by directly binding to target prom oters (Bajaj
et aL, 1996). Since the expression of hi I A is inhibited under
conditions which repress invasion gene expression (acidic pH,
aerobiosis, low osm olarity, presence of phoPQ<^ mutation), it has been suggested that regulation of h ilA expression plays a key role in
controlling the invasion phenotype.
Johnston et aL (1996) have shown that h it A expression is regulated by another protein called SirA {S a lm o n ella in v a sio n
regulator). SirA is also required for the induction of p r g H expression and secretion of SipA, SipB, SipC, and SipD. N ot surprisingly, sirA mutants are defective for invasion of epithelial cells. Similar to the
p h o P Q locus, the sirA gene is located outside SPI-1. The deduced
amino acid sequence of SirA is sim ilar to UvrY o f E. coli and GacA of
P. flu o r esc ens, both of w hich are members of the FixJ subfamily of
response regulators o f tw o-com ponent regulatory system s. The sir A m utant phenotype is suppressed by two unlinked loci called sirB and
sirC . sirC is located in a previously uncharacterized region of SPI-1.
These observations have led to the proposal o f a m odel for the
regulation o f S a lm o n ella invasion (Johnston e t a/., 1996). Detection of "invasion signals" by an unidentified sensor kinase results in
activation of SirA by phosphorylation. Phosphorylated SirA then activates hilA expression, either directly or in d irectly through
activation of sirB and/or s ir C transcription, to continue the regulatory cascade leading to the synthesis of invasion genes. Since activation o f PhoP/Q occurs under conditions thought to resem ble the
intracellular environm ent, PhoP/Q -m ediated rep ressio n o f invasion gene expression at an unknow n point in the regulatory cascade inhibits the expression o f these genes, which are presum ably unnecessary for intracellular survival.
INTRACELLULAR REPLICATION
The ability of S. typhim urium to survive and replicate within epithelial cells and m acrophages is essential for virulence (Fields e t
a l., 1986; Leung and Finlay, 1991). In both professional and non professional phagocytes. S a lm o n e lla resides and replicates within a
m em brane-bound com partm ent. C haracterization o f this
com partm ent has been a subject o f intense study, occasionally with conflicting results. Based on different criteria, some groups have reported th at in m urine m acrophages, 5. typhim urium is localized w ithin phagosom es that fuse with lysosom es (Carrol et a l , 1979; Oh
et a l , 1996), w hile others indicate that S. typhim urium inhibits
phagosom e-lysosom e fusion (Ishibashi and A rai, 1990; Buchmeier and Hefffon, 1991). However, it is now accepted that S a lm o n ella - containing vacuoles (SCVs) are specialized vacuoles that differ from true phagolysosom es (Finlay and Falkow , 1997).
It is now understood that the form ation o f a phagolysosome is not the result o f a single fusion event between a phagosome and a
lysosome, but occurs through a continuum o f fusion reactions with vesicles o f the endosom al-lysosom al pathw ay (D esjardins et a l , 1994; Beron et a l , 1995). U nique com partm ents have been identified along the endocytic pathw ay, and m arkers speciHc for these compartments are used to follow the transform ation o f phagosom es into
phagolysosom es. The three com partm ents o f the endocytic pathway include early endosom es, late endosom es/prelysosom es, and
lysosom es. Transform ation o f prelysosom es into lysosomes involves the accum ulation and/or processing of high levels o f lysosomal acid hydrolases, creating a degradative com partm ent.
The fate o f SCVs after entry into host cells has been characterized using endocytic markers. Trafficking o f SCVs in epithelial cells is sim ilar to that observed in cultured and bone m arrow-derived
m acrophages (Garcia-del Portillo and Finlay, 1995; Rathman et aL, 1997). W ithin 15-30 min after bacterial internalization, SCVs fuse w ith vesicles containing lysosom al membrane glycoproteins (LAMP-1 and LAM P-2), which are m arkers for late endosomes and lysosomes (in higher levels). Lysosomal acid phosphatase (LAP) is also
incorporated into SCVs with the same kinetics as LAMPs. In contrast, phagosom es containing latex beads or heat-killed bacteria progress along a degradative pathway, acquiring cathepsins and m annose 6- phosphate receptors (M6PRs), in addition to LAMPs and LAP. M6PRs function in the delivery of soluble lysosomal enzymes to
prelysosom al com partm ents (G arcia-del Portillo and Finlay, 1995). Also, as opposed to latex bead-containing phagosomes, SCVs show lim ited interaction with fluid phase endocytic tracers, indicating SCVs do not readily fuse with incom ing endocytic traffic. Collectively, the data show that S. typhim urium resides in an intracellular
com partm ent which diverges from the normal degradative endocytic p a th w a y .
It is important to note that the above observations represent the pathway followed by the m ajority of SCVs. A certain percentage of SCVs (approxim ately 25-30% ) are trafficked along a degradative pathway (Buchm eier and H effron 1991; Garcia-del Portillo and Finlay, 1995; Rathman et aL, 1997). Others have reported the presence o f at least two bacterial populations within m acrophages, one static and the other rapidly dividing (Abshire and N eidhardt, 1993b). This heterogeneity in the intracellular population o f bacteria
may in part explain som e of the earlier conflicting data regarding the fate of SCVs.
As described above, S. typhim urium can enter both epithelial cells and m acrophages by m acropinocytosis, although the membrane
ruffling induced in epithelial cells is localized to the point of bacterial contact with the host cell (Alpuche-Aranda et al., 1994; Garcia-del Portillo and Finlay, 1994). Roughly half o f the bacteria initially reside in unusually large vacuoles called spacious phagosomes (SPs), which are m orphologically similar to m acropinosom es. However, while macropinosomes shrink com pletely w ithin 15 min, SPs persist in the cytoplasm for as long as 45 min, som etim es enlarging by fusion with macropinosomes or other SPs.
At least two lines o f evidence suggest form ation and maintenance of SPs contributes to the intracellular survival o f S a lm o n e lla . First, host-adapted S a lm o n e lla serotypes rarely isolated from mice and humans fail to form or maintain SPs in mouse macrophages, and do not survive in macrophages from mice or hum ans (Alpuche-Aranda
et al., 1995). Second, a phoPQ*^ mutant presum ably containing a
constitutively active PhoP transcriptional reg u lato r induces
significantly fewer SPs in mouse macrophages and enters in close- fitting phagosomes (Alpuche-A randa et al., 1994). This mutant is avirulent and defective for survival in m acrophages early after phagocytosis, suggesting a p h o P -lep iessed ^ene(s) (prg) is involved in SP formation (Miller and Mekalanos, 1990). SPs are postulated to facilitate intracellular survival by diluting toxic compounds delivered
to SCVs or by reducing the rate o f SCV acidification, allowing the bacteria tim e to adapt by expressing other genes necessary for intracellular survival (A lpuche-A randa et al. y 1994).
SCVs in m urine macrophages become acidified, although the rate and level o f acidification is in dispute, perhaps due to technical
differences in the protocols em ployed by d ifferen t groups. Alpuche- Aranda et al. (1992) reported SCVs acidify slow ly over a 4-5 h
period com pared to vacuoles containing heat-killed organisms.
Others have found that the m ajority of SCVs acidify to a pH between 4.0 and 5.0 w ithin 20 to 30 m in after form ation (Rathman et aL,
1996). T reatm ent w ith inhibitors o f phagosom e acidification reduced the num ber o f bacteria recovered from infected m acrophages,
suggesting S a lm o n e lla may require an acidic environm ent for
intracellular replication and survival (Rathm an et al., 1996). Other intracellular pathogens, including Coxiella b urnetii, L eish m a n ia spp., and F rancisella tularensis, also replicate w ithin acidified host cell com partm ents (A ntoine et a l , 1990; Maurin e t al., 1992; Fortier et al.,
1995). However, unlike F ra n cisella , acidification of S a lm o n ella phagosomes is not linked to iron acquisition.
Several genes have been identified which affect the intracellular survival and replication of S a lm o n e lla . The follow ing section
summarizes the genes which have been characterized to date. Through a variety o f techniques, a number o f genes have been shown to be upregulated after entry into h o st cells or during
S a lm o n ella fitness in vivo (reviewed in H eithoff et aL, 1997b and
Valdivia and Falkow, 1997). However, in many cases the speciHc roles of these loci in virulence rem ains to be defined. The discussion here is lim ited to genes shown to make a m easurable contribution to phenotypes associated w ith in tracellu lar survival.
phoPQ
PhoP/PhoQ com prise a two-com ponent regulatory system (Groisman et aL, 1989; Miller et aL, 1989) w hich regulates many virulence-associated phenotypes including spacious phagosom e formation (A lpuche-A randa et aL, 1994), invasion of epithelial cells (Behlau and M iller, 1993), inhibition o f processing and presentation of antigens by macrophages (W ick et aL, 1995), resistance to
antimicrobial peptides (Fields et aL, 1989), intram acrophage survival (Fields et aL, 1986), survival in infected mice (Fields et aL, 1986), adaptation to M g^+-limiting environm ents (Soncini et aL, 1996; Blanc-Potard and Groism an, 1997), and m odification of lipid A (Guo
et aL, 1997). PhoP/Q -m ediated m odulation o f these phenotypes occurs through the activation and repression o f the production of over 40 proteins (M iller and M ekalanos, 1990). Although many PhoP/Q-regulated loci have been isolated, in m ost cases, the genes contributing to each phenotype are unknown. In fact, several PhoP/Q-regulated genes are not required for virulence in m ice,
suggesting the regulatory role of PhoP/Q is not lim ited to S a lm o n e lla pathogenesis (Fields et aL, 1989; Guim et aL, 1995). Genes shown to
affect some o f the above PhoP/Q-regulated phenotypes are described below .
PhoQ is a membrane-bound sensor-kinase predicted to contain two transm embrane regions, a long cytoplasmic tail, and a large periplasmic dom ain (M iller et aL, 1989; Gunn et aL, 1996). PhoP is a 224 amino acid cytoplasm ic response regulator belonging to the OmpR subgroup o f two-com ponent response regulators. By analogy with other tw o-com ponent system s, detection o f appropriate
environmental signals by the periplasmic dom ain o f PhoQ triggers the autophosphorylation o f PhoQ on a conserved histidine residue. PhoQ then phosphorylates PhoP on a conserved aspartate residue in the am ino-term inus. Phosphorylated PhoP then activates the
transcription of p a g (g A o f-activ ated genes). Gunn et aL (1996) have shown evidence o f phosphotransfer between PhoQ and PhoP.
Recent studies have shown that PhoQ is a sensor o f extracellular divalent cations, specifically Mg2+ and Ca^+ (Garcia Véscovi et aL,
1996; Garcia Véscovi et aL, 1997). Micromolar concentrations of M g 2+ or Ca2+ activate the transcription of at least 25 PhoP-activated loci, while growth in millimolar levels of these cations represses transcription o f PhoP-activated genes (Garcia V éscovi et aL, 1996; Garcia Véscovi et aL, 1997). The regulatory effect o f the cations appears to be m ediated by altering the conform ation o f the PhoQ periplasmic dom ain. Support for this hypothesis com es from several experiments. First, a protein chimera (ZhoQ) in which the PhoQ
related sensor protein which responds to osm olarity, allow s the expression of a PhoP-activated gene in both high and low Mg2+ media in the absence of PhoQ (Garcia Véscovi et a l, 1996). Second, M g 2+ alters the trypsin sensitivity of PhoQ but not ZhoQ in vitro at the same concentration required to repress expression o f PhoP- activated genes (G arcia Véscovi et a i , 1996). Since the modified
trypsin sensitivity is detected in spheroplasts prepared from bacteria expressing PhoQ, soluble com ponents from the periplasm ic space are not necessary for this effect. M ore recently, Garcia V éscovi et al. (1997) have show n that the tryptophan intrinsic fluorescence pattern of a purified polypeptide corresponding to the PhoQ
periplasmic dom ain is significantly altered in the presence of Mg2+ or Ca2+ but not Ba2+, which is unable to repress PhoP-activated genes. They also provided evidence fo r distinct, independent binding sites for Mg2+ and Ca2+. Repression of PhoP-activated gene expression is achieved at low er concentrations of Mg2+ than Ca2+, although both cations are necessary for maxim al repression. Finally, a m utant PhoQ protein (PhoQc) which m ediates the constitutive overexpression of several PhoP-activated genes displays a reduced affinity for Ca2+ but an unchanged affinity for Mg2+ (Garcia V éscovi et al., 1997). The mutation in PhoQc is a Thr to lie substitution at amino acid 48 which is located in the periplasmic dom ain im m ediately follow ing a
predicted m em brane spanning segm ent (G unn et al., 1996). Amino acids involved in cation binding in the PhoQ periplasm ic domain may include a region w ith several acidic residues (amino acids 135-154).
There are experim ental data supporting this suggestion using the E ,
co li PhoQ protein (Garcia Véscovi et a i , 1997).
R e s is ta n c e to a n tim ic r o b ia l p e p tid e s . Fields e t ai. (1989)
attem pted to determ ine the m icrobicidal m echanism o f m acrophages responsible for decreased intracellular survival o f p h o P m utants. Crude granule extracts from human neutrophils and rab b it peritoneal macrophages were found to have a strong m icrobicidal effect on the m utants. Fractions o f the extracts with the highest activity against the m utants w ere enriched in low m olecular w eight proteins that could correspond to defensins. Defensins are small m olecular w eight antim icrobial cationic peptides capable o f adopting am phipathic
alpha-helical structures and form ing pores in m em branes (Parra- Lopez et at., 1993). Defensins are present in large am ounts in the granules of neutrophils and macrophages o f several m am m als. Both
p h o P and p h o Q mutants are hypersusceptible to p u rified defensins as
well as to related antim icrobial peptides from frogs (m againins), insects (cecropin, m astoparan, m ellitin), and pigs (cecropin) (Fields e t
aL, 1989; Groisman et a i, 1992).
The work of Fields et al. (1989) prom pted a search for loci involved in resistance to antim icrobial peptides. T his search identified several unlinked genes affecting virulence in mice and resistance to subsets of 6 different cationic peptides, b u t none of these genes have been shown to be regulated by PhoP (Groisman e t
survival in m acrophages in vitro (Parra-Lopez et aL, 1993). The deduced amino acid sequences o f sa p A B C D F are m ost sim ilar to the components o f two oligopeptide uptake systems in S. typhim urium and Bacillus subtilis. SapA shows sequence sim ilarity to periplasmic solute binding proteins. The deduced sequences of SapD and SapF are similar to the sequences of several members o f the ATP binding cassette (ABC) family o f transporters. It was suggested the
SapABCDF system may facilitate resistance to som e antim icrobial peptides by transporting them into the cytoplasm where they could be degraded by pro teases.
The only PhoP-regulated locus shown to control resistance to antimicrobial peptides is encoded w ithin p m rC A B (Gunn and Miller, 1996; Soncini and Groisman, 1996). PnurC (also known as PagB) encodes a putative membrane protein o f unknown function. PmrA- PmrB function as a tw o-com ponent regulatory system controlling resistance to the cationic peptide antibiotic polymyxin B (PM). PM resistant mutants of S. typhim urium have m utations which map to the response regulator, p m r A , and are also resistant to other
antim icrobial peptides, including protam ine and the neutrophil peptides CAP37 and CAP57 but not defensins (Gunn and M iller,
1996). However, p m rA null m utants do exhibit w ild-type virulence in mice. Pm rA/B-dependent resistance to antim icrobial peptides is suggested to result from the induction o f genes w hose products covalently modify lipid A by substituting the 4’ phosphate with 4- am inoarabinose (Helander et a l , 1994; Guo et al., 1997). This
substitution decreases the negative charge o f the LPS, which may reduce the accessibility o f cationic peptides to lipid A. Since PhoP/Q also regulate a Pm rA /B-independent lipid A modiHcation, it has been proposed that other lipid A alterations, or a combination o f lipid A alterations and expression o f Pag proteins, contribute to a different cationic peptide resistance pathw ay effective against other peptides including defensins (Gunn and Miller, 1996).
A d a p ta tio n to Mg^+ lim itin g e n v iro n m e n ts . Both p h o P and
p h o Q are necessary for growth in low Mg2+ media (Garcia Véscovi e t aL , 1996). Given that Mg^+ acts as a signal controlling the PhoP/Q
regulatory system, this suggested the PhoP/Q regulon m ay allow adaptation to Mg^+ lim iting environm ents. Consistent w ith this hypothesis, several PhoP-activated genes have been shown to be essential for growth in low Mg^+ liquid or solid media (Garcia Véscovi
et a l , 1996; Soncini et al., 1996; Blanc-Potard and Groisman, 1997).
Two o f these genes, m g tC B , form an operon located in a 17 kb pathogenicity island designated SPI-3 and encode a high affinity M g 2+ uptake system (Suavely et at., 1991; Blanc-Potard and
Groisman, 1997). The function of MgtC is not understood, but MgtB and MgtC may act independently in Mg2+ transport. m g tC is required for growth in macrophages (Blanc-Potard and Groisman, 1997).
Addition of exogenous Mg2+ partially rescues m gtC B and p h o P mutants for growth in macrophages. Since expression o f PhoP- activated genes is induced in host cells, the data suggest the
S a lm o n ella phagosome contains a lim iting concentration o f Mg^+
(A lpuche-A randa et aL, 1992; G arcia-del Portillo e t aL, 1992; H eithoff
et aL, 1997a).
S u rv iv a l w ith in m a c ro p h a g e s . As described above, p r g loci appear to play a role in intracellular survival early after
phagocytosis. A switch to expression of p a g loci occurs at a later stage and is thought to perm it continued survival and replication. Analysis of the induction o f PhoP-activated genes after entry into epithelial cells or m acrophages indicates transcription is m inim al within 1 h after infection and reaches a maximal level at 4-6 h post infection (Garcia-del P ortillo et aL, 1992; Alpuche-Aranda et aL,
1992). This roughly coincides with the time at which bacterial
numbers begin to increase inside both types o f host cells (Leung and Finlay, 1991; Abshire and N eidhardt, 1993b), and the tim e when
p h o P mutants begin to show a defect in intram acrophage survival
(Fields et aL, 1986; M iller and M ekalanos, 1990).
Of the p a g loci identified to date, only p a g C has been clearly shown to be required fo r virulence in mice and survival within
macrophages (Pulkkinen and M iller, 1991). It rem ains possible that some of the PhoP-regulated phenotypes may be the result o f a
cumulative effect of m ultiple p ag -en co d ed proteins, but strains with mutations in m ultiple p a g loci have yet to be created to test this hypothesis. p a g C encodes an 18 kD a outer membrane protein of unknown function (Pulkkinen and M iller, 1991). p a g C m utants have
w ild-type sen sitiv ity to defensins, lysozym e, com plem ent, and cationic peptides derived from m ouse intestine. PagC shows
sim ilarity to oth er outer membrane proteins including Ail o f Y e r s in ia
en tero co litica ^ Lom, a bacteriophage lambda-encoded protein of
unknown function, Rck o f S. typhim urium , and OmpX o f E n te r o b a c te r
c lo a ca e (Pulkkinen and Miller, 1991). Both Ail and Rck have been
shown to affect serum resistance (Bliska and Falkow, 1992;
H effem an et aL, 1992). Ail also mediates attachment to and invasion of epithelial cells by Y ersinia (M iller and Falkow, 1988). There is no evidence supporting a role for PagC in invasion as p a g C m utants display w ild-type invasion of epithelial cells (Galan and Curtiss,
1989), and p a g C is insufficient to confer an invasive phenotype on E .
coli (M iller, 1991).
m s g A
Analysis o f the region adjacent to p a g C led to the discovery o f
ms g A (m acrophage survival gene) (Gunn et aL, 1995). m s g A m utants
have m ore than 300-fold reduced virulence in mice and are defective fo r survival in macrophages. The intracellular survival defect is quantitatively equivalent to that of p h o P mutants. MsgA is predicted to be a 79 amino acid, hydrophilic protein containing a large num ber o f acidic amino acids at the C-terminus and lacking sim ilarity to any known proteins. The deduced amino acid sequence lacks a putative signal sequence and any large stretches of
