Molecular Microbiology (2000) 37(6), 1444±1455
Identification of Streptococcus agalactiae virulence
genes in the neonatal rat sepsis model using
signature-tagged mutagenesis
Amanda L. Jones, Katherine M. Knoll and Craig E.
Rubens*
Division of Infectious Diseases, Department of Pediatrics,
Children's Hospital and Regional Medical Center and
University of Washington, 4800 Sand Point Way NE,
CH-32, Seattle, WA 98105, USA.
Summary
Group B streptococcal (GBS) infections are the most
common cause of bacterial sepsis in the immediate
newborn period. Apart from the capsule, the factors
required for survival of GBS in the host are not well
defined. In this study, signature-tagged transposon
mutagenesis (STM) was used to identify genes
required for growth and survival of GBS in a neonatal
rat sepsis infection model. Approximately 1600 trans-
poson mutants were screened in pools of 80 mutants,
and approximately 120 mutants defective for survival
in the animal host were identified. We successfully
cloned and sequenced DNA flanking the transposon
insertions from 92 of the mutants. Fifty per cent of the
mutants had transposon insertions in genes with
homologues in the public databases, whereas the
remaining 50% had transposon insertions in genes
with unknown function. A significant proportion of
the avirulent mutants had transposon insertions in
genes encoding transport-associated or regulatory
proteins or in genes involved in cell surface metabol-
ism, emphasizing the significance of these functions
for in vivo survival of GBS. Overall, STM analysis
revealed GBS genomic loci that encode a wide variety
of functional gene classes, underscoring the diver-
sity of bacterial processes required for the infection
process. Currently, the function of the genes identi-
fied during the screening can only be inferred by
homology to previously described genes. However, a
number of the genes identified in this study have
been shown to correlate with virulence in other
pathogens. Avirulence of a subset of mutants identi-
fied during the screening was confirmed by perform-
ing competitive index assays and lethal dose assays.
This represents the first report of a genome-wide
Accepted 10 July, 2000. *For correspondence. E-mail cruben@chmc.
org; Tel. (11) 206 526 2073; Fax (11) 206 527 3890.
Q 2000 Blackwell Science Ltd
scan for virulence factors in GBS. The identified
genes will further our understanding of the pathogen-
esis of GBS infections and may represent targets for
intervention or lead to the development of novel
therapies.
Introduction
Streptococcus agalactiae [group B streptococci (GBS)] is
the leading cause of neonatal sepsis and meningitis in the
United States and western Europe, and is an emerging
pathogen in immunocompromised adults (Schuchat,
1998). The early-onset form of GBS disease typically
presents in the first 24 h of life as fulminant pneumonia
and septicaemia, with high mortality (10±15%) despite
intensive supportive therapy (Nizet et al., 2000). The less
common, late-onset (. 7 days of age) form of infection
presents as meningitis, bacteraemia or osteoarthritis.
Prematurity and maternal obstetric complications are
important risk factors for early-onset disease, whereas
late-onset infections most commonly affect term new-
borns with unremarkable perinatal histories (Nizet et al.,
2000). GBS associated with human disease are almost
invariably encapsulated, belonging to one of the nine
recognized capsular serotypes: Ia, Ib or II±VIII. GBS
isolates from neonates with early-onset infection are well
distributed among the capsular serotypes, whereas most
isolates from infants with late-onset disease (and infants
with meningitis regardless of age of onset) are of serotype
III (Nizet et al., 2000). The type-specific capsule of GBS
has been studied extensively and is an important
virulence determinant for both virulence and immunity
(Rubens et al., 1987; Wessels et al., 1989). Studies of
GBS virulence have examined adherence to and invasion
of epithelial and endothelial barriers, interactions with host
immune pathways, using or focusing on a few specific
factors, such as haemolysin, alpha and beta proteins and
the capsular polysaccharide (for a review, see Nizet et al.,
2000). The role of basic biosynthetic, metabolic and
transport pathways in infection has not been well
characterized.
Recognition of the prevalence and severity of human
neonatal and adult disease underscores the importance of
investigating the mechanisms important to the pathogen-
esis of GBS infections. Large-scale screening to identify
 Signature-tagged mutagenesis of group B Streptococcus
attenuated mutants has not been attempted because of
the technical limitations of testing a large number of indi-
vidual mutants in animal models of infection. Signature-
tagged mutagenesis (STM) allows the screening of a
large number of mutants in a single animal (Hensel et al.,
1995). In STM, mutants are created using a uniquely
tagged transposon, and pools of tagged mutants are
screened for attenuation in an animal host. Attenuated
mutants are identified by hybridization analysis of tags
recovered from the animal.
The utility of screening pools of transposon mutants in
animal infection models has been demonstrated in Gram-
negative pathogens, including Salmonella typhimurium,
Yersinia enterocolitica, Vibrio cholerae (Hensel et al.,
1995; Chiang and Mekalanos, 1998; Darwin and Miller,
1999), and Gram-positive pathogens such as Staphylo-
coccus aureus, Streptococcus pneumoniae and Myco-
bacterium tuberculosis (Mei et al., 1997; Coulter et al.,
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
1445
1998; Polissi et al., 1998; Camacho et al., 1999).
Recently, STM has also been applied to Brucella suis in
an in vitro human macrophage infection model (Foulongne
et al., 2000). Here, we describe the application of STM to
the Gram-positive pathogen group B streptococci in a
neonatal rat model of sepsis, in order to identify genes
involved in the growth and survival of GBS in the host.
Results and discussion
Construction and screening of GBS signature-tagged
transposon mutants
Plasmid pAJ010 (Fig. 1), carrying a modified Tn917
(Tn917stm), was constructed for STM of the type Ia
GBS strain A909. A909 was selected for these studies
because of the relative ease of transformation and genetic
manipulation of this strain. The pBR322 ori/rop/tet region
Fig. 1. Construction of pAJ010. Tn917stm is
carried on the temperature-sensitive pTV10K
derivative pAJ010. The pBR322 ori/rop/tet
region was inserted into Tn917 to allow
replication and selection of the plasmid in
E. coli and to facilitate self-cloning of GBS
chromosomal DNA flanking transposon
insertions. Signature tags were incorporated
into the SacII site of Tn917stm. Self-cloning
of GBS chromosomal DNA flanking the
transposon insertions was accomplished by
digesting chromosomal DNA from the mutants
with the unique enzymes NsiI, NcoI or KpnI.
Transposon Tn917stm is shown in black.
aphA3, aminoglycoside phosphotransferase
(kanamycin resistance); repAts, temperature-
sensitive replicon; em, erythromycin
resistance; Tc, tetracycline resistance.
1446 A. L. Jones, K. M. Knoll and C. E. Rubens
allows replication and selection of pAJ010 in Escherichia Mutants with tags that hybridized weakly in both pools
coli, whereas the temperature-sensitive repA (repAts) could not be evaluated (Fig. 2).
gene from pWV01 and the kanamycin resistance gene of
pTV10K permit replication and selection in GBS at 308C. Cloning and analysis of flanking DNA
Oligonucleotide signature tags were incorporated into the
SacII site of Tn917stm. A series of 80 plasmids carrying Approximately 120 mutants were selected for analysis on
unique tags was selected for library construction and the basis of their attenuated virulence in the animal
designated the master set. The master set of 80 tagged model. The DNA flanking the left terminal repeat (LTR)
plasmids was used to generate 20 sets of 80 emR kmS of the Tn917stm insertion point was cloned in 92 of the
transposon mutants as described in Experimental proce- mutants, as described in Experimental procedures. The
dures. The randomness of the transposon mutant library nucleotide sequence of the flanking clones was deter-
was confirmed using Southern hybridization analysis of 20 mined and analysed by searching the public databases for
mutants using Tn917 as a probe. Each mutant had a similar sequences, and the results are presented in
single transposon insertion in a unique location in the Table 1. Flanking DNA from 46 of the avirulent mutants
chromosome (data not shown). had significant homology to known genes or proteins in
the public databases. Several mutants were identified that
had transposon insertions in different locations in the
Screening of mutant pools same gene, although there were no exact duplications of
Previous STM screens had revealed that all the mutants the transposition site. Nineteen mutants had flanking DNA
in a pool used to infect a given animal must be present
in sufficient numbers to establish an infection simulta-
neously (Hensel et al., 1995; Chiang and Mekalanos,
1998). In our initial experiments, we investigated several
infectious doses and harvest time points. The optimal
dose for inoculation of the rat pups was determined to be
approximately 3  105 cfu, approximately 10 times the
LD50 value for A909. The avirulent control strain, COH
31-21, was included in mutant pools during the optimiza-
tion studies as an indicator that sufficient time had
elapsed to allow clearance of avirulent clones. GBS strain
COH31-21 produces a sialic acid-deficient capsule and
has previously been shown to be attenuated in virulence
(Wessels et al., 1989). The optimal recovery time for STM
mutant pools was determined to be 20 h after inoculation.
At this time point, the rat pups had cleared the avirulent
strain, and sufficient numbers of surviving bacteria were
recovered from the spleen for analysis of tags. At later
time points, the pups either succumbed to infection or
began to clear all clones. A minimum number of recovered
cfu was required to ensure that all mutants not cleared by
the pups were represented in the recovered pool.
Recovered pools consisting of less than 5  103 cfu
were not used for subsequent analysis of tags.
Using the optimized conditions, a total of 1600 mutants
were collected and screened in the neonatal rat model of
GBS sepsis. In the sepsis model, the ability of the bacteria
to disseminate systemically and survive maximal expo- Fig. 2. Identification of GBS mutants with attenuated virulence.
sure to host defence mechanisms was evaluated. Surviv- Identical filters arrayed with the 80 tagged plasmids were
ing bacteria were recovered from the spleens of pups hybridized to DIG-labelled tags from the input and output mutant
pools. Mutants whose tags were present in the input pool but
20 h after intraperitoneal inoculation. Each of the 20 pools missing or reduced in the output pool (A12, B10, C12, D7, D9, D11
was inoculated into three or four pups. Mutants with tags and D12) were lost during screening in the animal. Only mutants
present in the input pool but absent or reduced in the whose tags were absent from two out of three or three out of four
output pools were selected as potential avirulent mutants. Plasmid
output pool in at least two out of three or three out of four pAJ010 without a tag is in position G10 and serves as a negative
rat pups were selected as putative virulence mutants. hybridization control.

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455

 Signature-tagged mutagenesis of group B Streptococcus 1447
Table 1. Analysis of GBS DNA flanking transposon insertions in STM mutants.

Percentage amino
acid identity
a b
Classification Mutant Homology Hypothesized function (homology)/spanc

Transport/binding AJ3E8 Mal X (S. pneumoniae) Maltose-binding protein 34 (49)/140

AJ2F4 Amino acid-binding protein ABC transporter family 61 (73)/224
(Streptococcus uberis)
AJ7D11 DppA (S. pyogenes) Dipeptide permease subunit A: 93 (96)/241
substrate-binding membrane-associated
lipoprotein
AJ3F3
AJ16D10 ] Hyaluronate-associated protein
(HAP) (Streptococcus equi)
ABC transporter family 69 (85)/136
67(84)/121
AJ5D8 Na1/H1 antiporter homologue Cation antiport 52 (69)/153
(Lactococcus lactis)
AJ5B2 Permease (Bacillus lichenformis) Arginine/ornithine antiporter 65 (79)/49
AJ8G1 PtkB (E. coli) Phosphotransferase enzyme, permease 26 (55)/79

Š
component
Regulatory KK12B8 Member of two-component regulatory system 40 (62)/161
AJ4C4 33 (54)/162
AJ9G5 Putative response regulator 54 (70)/53
AJ9B11 (Lactobacillus sakei) 40 (62)/121
AJ20C8 54 (70)/53
AJ6B1 YvoA (B. subtilis) Transcriptional regulator (GntR family) 37 (58)/74
AJ1E8 LytS (S. aureus) Autolysis sensor, member of two-component 53 (73)/207
regulatory system
Cell envelope AJ4F8 murF (S. pneumoniae) D-alanyl-D-alanine adding enzyme 69 (84)/106
AJ4F10 LrgB (S. aureus) Murein hydrolase 38 (61)/55
AJ3F6 PBP1a (S. pneumoniae) Peptidoglycan transglycosylase-transpeptidase 61 (80)179
Adherence AJ5D1 Lmb (S. agalactiae) Laminin-binding protein 78 (78)/41
Metabolic KK15A2 ScrB (Streptococcus sobrinus) Sucrose hydrolase 70 (81)/153
AJ2F5 LyxK (E. coli) Cryptic l-xylulose kinase 38 (61)/125
AJ10F10 PepX (L. lactis) Dipeptidyl peptidase 44 (61)/210
AJ1A4
AJ2D12 ] PepN (L. lactis)
Amimopeptidase 54 (72)/146
73 (86)/200
AJ18G6
AJ18E7 ] Xylanase (Mycoplasma genitalium)
Putative xylananse, xylose metabolism 28 (51)/176
46 (82)/157
Pyrimidine biosynthesis AJ2F8 CarB (Lactobacillus plantarum) Carbamoyl phosphate synthase 88 (93)/36
Purine biosynthesis AJ20C4 GuaA (L. lactis) GMP synthase 59 (73)/166
Amino acid biosynthesis AJ7B7 Hdh (L. lactis) Homoserine dehydrogenase 54 (70)/165
Cellular processes AJ8D3 RopA (S. pyogenes) Chaperone/regulator of secreted proteins 77 (87)/165
AJ16E7 LepA (B. subtilis) Membrane-associated GTP-binding protein 75 (92)/168
AJ5F7 Clp-like ATP-dependent Chaperone, proteolysis mediator 65 (75)/20
protease-binding subunit

]
KK14G4 Unknown 40 (64)/124
AJ20B5 DNA processing chain A 43 (64)/216
AJ7A3 (Thermatoga maritima) 40 (53)/134
AJ9F12
AJ9D9 ] Nisin resistance protein (L. lactis) Nisin resistance
31
27
(59)/214
(58)/137
Adaptations KK12G2
AJ9D8 ] Carbon starvation protein
homologue (E. coli)
Putative CstA family 40
40
(55)/77
(55)/77
Transcription factors KK13B10 plaC (S. aureus) DNA-directed RNA polymerase sigma factor 47 (68)/87
Translation AJ6E12 Amino group acetyl transferase Protein modification 33 (53)/72
(Chlamydia trachomatis)
Unknown AJ18D3 Membrane protein (B. subtilis) Putative integral membrane protein 72 (90)/48
Plasmid functions AJ10B3 RepJ (S. aureus) Plasmid replication initiation protein 34 (54)/146
Transposon-related
functions
KK15C4
KK13B4
AJ9C3
] ISL2 protein (L. lactis)
61
61
61
(78)/108
(78)/108
(78)/118
19 Unknown/putative (significant homology)
7 Unknown/putative (low homology)
20d No homology in public databases Unknown

a. Homology scores with a sum probability , e210 were considered significant. Mutants with lower homology (e204 to e209) were classified
separately. Homology scores . e203 were classified as having no significant database match. Only individual mutants of particular interest with low
homology are listed separately.
b. Hypothesized function inferred by analogy to other known proteins or gene sequences in the public databases.
c. Span describes the length of the region in amino acids to which percentage similarity and identity refers.
d. Sequence obtained from the flanking DNA of these mutants showed no significant homology to any sequence in the public databases or to any
unfinished genome sequences.

Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455

1448 A. L. Jones, K. M. Knoll and C. E. Rubens
Table 2. Competition assays of selected mutants identified by STM.

LD50a
Strain Mutation (homology) CI (no. of pups)b (cfu/animal)

AJ2D12 PepN , 0.018 (4)

AJ3E8 MalX , 0.0019 (5)
AJ5D1 LmbP 0.21 (3) 1.2  106
AJ3F6 PBP1a , 0.0025 (5)
AJ2F4 Amino acid-binding protein , 0.064 (5)
AJ7D11 DppA 0.14 (4)
AJ8D3 RopA , 0.00022 (3) 6.7  106
AJ9F12 NisR , 0.035 (3)
AJ1E8 LytS , 0.059 (3) 2.8  106
AJ4F10 LrgB 0.012 (4)
AJ4C4 Putative response regulator , 0.057 (5)
AJ4F8 murF , 0.006 (3) 4.0  105
AJ5F7 Clp protease homologue , 0.0025 (5) 1.0  106
KK13B10 Sigma factor , 0.015 (3)
AJ5B2 Permease , 0.0025 (3)
AJ16D10 Hyaluronate-associated protein , 0.0022 (3) 8.4  105

a. LD50 values were calculated using the methods of Reed and Muench (1938) as described in Experimental procedures. LD50 for the parent strain
A909 was 3.2  104 cfu/animal.
b. Competitive indices (CI) are calculated relative to the wild-type strain A909 as described in Experimental procedures. The competitive index is
defined as the output ratio (mutant/wild type) divided by the input ratio (mutant/wild type). CI is the average determined from the number of rat pups
shown in parentheses. In some cases, no mutant bacteria were recovered from one or more rats. In this case, CI was calculated assuming that one
mutant bacterium had been recovered, and the CI is expressed as less than (,) this number.

with significant homology to unknown or putative genes. Classification of mutants

The flanking DNA from seven mutants had low homology
to known or putative genes. Flanking DNA cloned from
the remaining 20 mutants did not display any homology to
sequences in the public databases, including the unfinished
microbial genome databases.
Competition assays and LD50 determination of mutants
identified by STM
The virulence attenuation of a subset of mutants was
confirmed and quantified in competition assays and LD50
assays, as described in Experimental procedures. Com-
petitive index assay results and LD50 values are shown in
Table 2. Overall, the mutants displayed a range of attenu-
ation in the competitive index assays, ranging from subtle
attenuation (AJ5D1 CI ˆ 0.21) to severe attenuation
(AJ8D3 CI ˆ 0.00022). Additionally, 50% lethal doses
were determined for six mutants. All the mutants tested
had LD50 values greater than the parent strain, ranging
from approximately 10-fold higher to 200-fold higher than
the parent strain A909. The LD50 values obtained con-
firmed the virulence attenuation of the mutants observed
in the competitive index assays.
General growth defects causing attenuated virulence
were ruled out by generating growth curves of individual
transposon mutants grown in parallel with the parent
strain A909. The growth rate of all strains listed in Table 2
was found to be essentially identical to that of the parent
strain A909, except for KK13B10, which exhibited a mild
in vitro growth deficit (data not shown).
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
Fifty per cent of the mutants identified had homology to
known genes or proteins in the public databases. On
the basis of homology searches, where possible, these
mutants were assigned to a gene class using the micro-
bial genomic classification system outlined by Bult et al.
(1996). Only key examples are described below.
Binding/transporter mutants. The largest class of genes
identified in the screen that impacted virulence included
homologues of binding or transporter genes.
Mutant AJ7D11 had a transposon insertion immediately
downstream of a dipeptide permease (dppA) gene homo-
logue. The Dpp systems are complexes of four or five
membrane-associated proteins (DppA±E) (Podbielski
and Leonard, 1998). Based on BLAST analysis, the GBS
dppA homologue has 93% identity to the Streptococcus
pyogenes (group A streptococci; GAS) dppA gene. The
virulence attenuation seen in this mutant is probably
caused by the polar effects of transposon insertion, as
the dpp genes are usually organized in an operon. Aside
from providing a bacterial cell with dipeptides as a source
of essential amino acids or nitrogen and energy, the
dipeptide permease has been shown to be involved in
haem production, chemotaxis and sporulation (Podbielski
et al., 1998). GBS is naturally auxotrophic for several
essential amino acids; thus, it is possible that this mutant
is defective in the import of peptides from which growth-
essential amino acids are salvaged. Podbielski et al.
(1998) correlated levels of essential amino acids and the
 Signature-tagged mutagenesis of group B Streptococcus
regulation of the virulence factor SpeB (major cysteine
protease) expression. SpeB expression was reduced
eightfold in dpp mutants (Podbielski and Leonard,
1998). The DppA mutant was examined in the competitive
index assay. As shown in Table 2, AJ7D11 displayed only
subtle attenuation in the competitive index assay
(CI ˆ 0.14). It is possible that GBS contain multiple
peptide transporters that can partially complement the
function of a mutated Dpp operon.
Mutants AJ3F3 and AJ16D10 had transposon inser-
tions in a hyaluronate-associated precursor (HAP) gene
homologue. This function of this gene was initially and
erroneously assigned as a hyaluronate synthase in group
C streptococci (Lansing et al., 1993). Upon re-examination,
the HAP protein was found to be a member of the ABC
transporter family and contains an ATP-binding domain
and several hyaluronan binding sites. HAP was found to
be tightly bound to cell-associated hyaluronate and to co-
purify with hyaluronate (Lansing et al., 1993). The precise
role of HAP is uncertain; however, Nickel et al. (1998)
demonstrated that HAP functions as an ecto threonine
kinase in group C streptococci and postulated that HAP
plays a role in the regulation of hyaluronate capsule syn-
thesis and shedding of hyaluronate from the cell surface
through autophosphorylation. Chanter et al. (1999) recently
reported that recombinant forms of the HAP protein
served as a protective immunogen against challenge
with group C streptococcal strains in mice. As seen in
Table 2, mutant AJ16D10 displayed a severely attenu-
ated phenotype in the competitive index assay and had an
LD50 approximately 25-fold greater than the parent strain.
GBS do not produce a hyaluronate capsule; however, it is
possible that the HAP homologue regulates either capsule
types other than hyaluronate or some other important cell-
associated carbohydrate in GBS. Further studies are
under way to characterize these mutants.
In addition to the ABC transporters, other transport func-
tions impacting in vivo survival were also identified, includ-
ing a maltose transporter homologue, a sodium/proton
antiporter homologue, an arginine/ornithine antiporter
homologue and a component of the phosphoenolpyru-
vate-dependent sugar phosphotransferase system, a
major carbohydrate active transport system. The assort-
ment of independent mutations recovered in transporter
homologues emphasizes the biological significance of
these loci for GBS in vivo survival.
Regulatory mutants. A number of genes encoding
putative regulatory genes were identified in the screen,
suggesting the importance of regulation of bacterial gene
expression for in vivo survival of GBS.
Mutant AJ1E8 had a transposon insertion in a gene
encoding a sensor kinase from a two-component regula-
tory system homologous to S. aureus lytS. Two-component
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
1449
regulatory systems are the most widespread means of
signal transduction used by bacteria to sense and
respond to their environment (Throup et al., 2000).
These systems have been shown to regulate a wide
variety of cellular responses, including osmoregulation,
photosynthesis, chemotaxis, sporulation, antibiotic pro-
duction and pathogenicity, in a number of different
bacteria (Appleby et al., 1996). Additional sequencing of
the flanking DNA from this mutant resulted in the iden-
tification of a lytR response regulator homologue located
immediately downstream of lytS. The sequence has been
deposited in GenBank under the accession number
AF250046. lytSR two-component regulatory system
homologues have been described in a number of Gram-
positive organisms, although their precise function is not
known. Throup et al. (2000) systematically deleted
putative two-component regulatory systems in S. pneu-
moniae. Inactivation of a lytSR homologue pair resulted in
a dramatic attenuation of bacterial growth in a mouse
respiratory tract infection model, suggesting that this
signal transduction system is important for the in vivo
adaptation and pathogenesis of S. pneumoniae. The role
of this two-component system in the virulence of
S. pneumoniae is not known (Throup et al., 2000). In
S. aureus, LytS and LytR have been implicated in con-
trolling the rate of autolysis by affecting the intrinsic
murein hydrolase activity associated with the cell (Brunskill
and Bayles, 1996). Autolysis has not been reported in
GBS, and the contribution of this two-component regula-
tory system to virulence requires further investigation.
As shown in Table 1, five mutants (KK12B8, AJ4C4,
AJ9G5, AJ9B1 and AJ20C8) had transposon insertions in
a putative response regulator gene. Sequence analysis of
flanking DNA cloned from these mutants revealed that the
five transposon insertions are clustered together in the
same region of the GBS chromosome. Despite occurring
in the same putative gene, the transposon insertion site
was unique for each mutant. One putative response
regulator mutant (AJ4C4) was examined in a competitive
index assay and found to have an attenuated phenotype
(Table 2).
Cell envelope mutants. Three mutants were identified with
transposon insertions in genes involved in cell wall
metabolism. Mutant AJ3F6 had a transposon insertion in
a putative gene homologue of penicillin-binding protein 1a
(PBP1a), encoded by ponA, from S. pneumoniae. Mutant
AJ4F8 had a transposon insertion in a putative gene with
significant homology to murF from S. pneumoniae and
Bacillus subtilis. murF encodes the UDP-MurNAc-penta-
peptide synthetase (D-alanyl-D-alanine adding enzyme),
which catalyses the final step in the synthesis of UDP-N-
acetylmuramoyl-pentapeptide, the precursor of murein
(Massida et al., 1998). Mutant AJ4F10 had a transposon
1450 A. L. Jones, K. M. Knoll and C. E. Rubens
insertion in a gene with homology to a putative murein
hydrolase (lrgB) from S. aureus.
Enzymes involved in cell wall metabolism, particularly
amidases and peptidoglycan-modifying enzymes in Gram-
positive bacteria, have been identified as differentially
expressed and only necessary for in vivo growth in various
animal infection models (Mei et al., 1997; Lowe et al., 1998;
Polissi et al., 1998; Graham and Clark-Curtiss, 1999).
Murray et al. (1998) reported that B. subtilis PBP1 had
redundant functions in rich media. However, when divalent
cations are limiting, such as in the environment experi-
enced by bacteria inside host cells, PBP1 is required for
cell growth. Graham and Clark-Curtiss (1999) used a
positive cDNA selection method to identify RNAs synthe-
sized by M. tuberculosis in response to phagocytosis by
cultured human macrophages and reported that ponA
was expressed by bacteria grown in macrophages but not
in bacteria grown in broth. As seen in Table 2, each of
these mutants had a severely attenuated phenotype in the
competitive index assay, suggesting that cell wall meta-
bolic functions are necessary for the in vivo survival of
GBS.
Adherence mutants. Mutant AJ5D1 had a transposon
insertion in the previously described GBS adhesin gene
lmb (laminin-binding protein) (Spellerberg et al., 1999).
This mutant displayed subtle attenuation in the competi-
tive index assay (CI ˆ 0.21) and had an LD50 value
significantly higher than the parent strain (Table 2). The
Lmb protein displays significant homology to the strep-
tococcal LraI (lipoprotein receptor antigen I) protein
family. The LraI proteins are thought to serve a dual role
in adhesion and transport; they are located in ABC
transporter-type operons and code for lipoproteins. The
LraI family of adhesins/transporters has been shown to
mediate adherence in several species of streptococci and
to contribute to virulence in these organisms (Spellerberg
et al., 1999). Spellerberg et al. (1999) demonstrated that
Lmb mediated the attachment of GBS to human laminin.
Adhesion of GBS to basement membrane components
such as laminin may be essential for the bacterial
colonization of damaged epithelium and translocation of
bacteria into the bloodstream.
Protein secretion mutants. Mutant AJ8D3 had a transpo-
son insertion in a homologue of the ropA gene (regulator
of proteinase), a regulator of protein secretion in GAS.
The GBS RopA protein homologue is predicted to be very
similar to GAS RopA displaying 77% identity over 165
amino acids. In GAS, RopA contributes post-transcrip-
tionally to the secretion and processing of the secreted
cysteine proteinase (SCP or SpeB) and encodes a
homologue of trigger factor, a peptidyl-prolyl isomerase
and putative chaperone that is highly conserved in most
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
bacterial species but of unknown function (Lyon et al.,
1998). RopA is proposed to have at least two functional
domains: one that participates in targeting SpeB to the
secretory pathway; and a second chaperone-like domain
that participates in protein folding (Lyon et al., 1998).
SpeB is thought to decrease both the resistance of GAS
to phagocytosis and the dissemination of bacteria in the
host (Lukomski et al., 1998). A SpeB-like protein has not
been reported in GBS; however, it is possible that ropA
regulates secretion of GBS proteins necessary for viru-
lence. Interestingly, AJ8D3 was the most severely attenu-
ated mutant examined in both competitive index assays
and lethal dose assays (Table 2). Additional studies are
planned to characterize this mutant further and to deter-
mine whether defects in protein secretion are present in
AJ8D3.
Auxotrophic/metabolic mutants. The nutritional environ-
ment of the host's cells imposes a requirement for de
novo biosynthesis of various amino acids, cofactors and
nucleotides in many pathogens.
The identification of mutants with transposon insertions
that induce auxotrophy demonstrated that, as in other
systems, auxotrophic mutants are cleared from the host.
Mammalian blood contains very low extracellular levels of
purines and pyrimidines, requiring infectious bacteria to
synthesize their own nucleotides (Mahan et al., 1993). De
novo purine biosynthesis is essential for infectivity, growth
and virulence of many bacteria in mammals (Mahan et al.,
1993). A mutant (AJ20C4) with a transposon insertion in
guaA homologue, encoding GMP synthetase, was iden-
tified. Mutant AJ2F8 had a transposon insertion in a
pyrimidine biosynthesis gene homologue (carB). A puta-
tive amino acid auxotroph (AJ7B7) with a transposon
insertion in a homoserine dehydrogenase gene homo-
logue was also identified.
Other mutants. The DNA flanking the transposon insertion
in mutant KK13B10 had homology to two loci listed in the
public databases that are associated with sigma factors.
The highest homology was to plaC in S. aureus. The plaC
gene encodes a protein with high similarity to the vege-
tative sigma factor of B. subtilis (sigA) and is predicted to
act as an RNA polymerase sigma factor. The disrupted
locus also has significant homology to a Listeria mono-
cytogenes open reading frame (ORF) of unknown function
(orfA1, U17284). This ORF is situated 3 0 to rpoD, a major
sigma factor of L. monocytogenes. Klarsfeld and Cossart
(1995) reported that disruption of the operon downstream
from the major sigma factor gene affects the growth and
virulence of L. monocytogenes. Lowe et al. (1998) also
identified a S. aureus gene homologous to the L. mono-
cytogenes orfA1 using in vitro expression technology
(IVET). The S. aureus orfA1 homologue was induced
 Signature-tagged mutagenesis of group B Streptococcus
during infection in a murine renal abscess model. A
mutant was constructed in the orfA1 homologue and
found to be significantly attenuated in virulence compared
with the parent strain (Lowe et al., 1998). Mutant KK13B10
has an attenuated phenotype in the competitive index
assay; however, it also exhibits a growth defect when
grown in vitro in complex media (Table 2).
Mutant AJ5F7 had a transposon insertion in a region of
the GBS chromosome with homology to a region of a Clp
protease regulatory subunit (65% identity over a 20-
amino-acid span). Clp proteases are a family of highly
conserved proteins involved in ATP-dependent proteo-
lysis (Gottesman et al., 1990). The Clp protease is
involved in degradation of denatured proteins and con-
trolling the half-life of unstable proteins to modulate gene
expression (Porankiewicz et al., 1999). The Clp stress
response system for intracellular protein degradation is
widely conserved in bacteria, and the Clp proteases have
been implicated in virulence in an increasing number of
organisms. STM of S. aureus identified an avirulent
mutant with a transposon insertion in a ClpX homologue
(Mei et al., 1997). Components of the Clp system are
also important for the virulence of L. monocytogenes
(Rouquette et al., 1996) and Salmonella typhimurium
(Hensel et al., 1995). This mutant was severely attenu-
ated in the competitive index assay (Table 2).
Conclusions
To understand the mechanisms by which pathogens
cause disease, it is necessary to identify the microbial
genes that are specifically required for the establishment
and maintenance of an infection. STM has been applied
successfully to other Gram-positive organisms such as
S. pneumoniae (Polissi et al., 1998) and S. aureus (Mei
et al., 1997). We sought to apply STM to GBS in a
neonatal rat sepsis model. Neonatal rats have been
shown to possess the same age-related susceptibility to
GBS infections as human neonates and have been used
widely as a model system for GBS infection (Zeligs et al.,
1982; Rubens et al., 1987; Wessels et al., 1989). In the
sepsis model, the ability of the bacteria to disseminate
systemically and survive significant exposure to host
defence mechanisms is evaluated. However, this model
bypasses the early stages of the infection process
including adherence to and colonization of epithelial
surfaces, followed by subsequent invasion of host cells;
hence, some bacterial factors important to these stages in
pathogenesis will not be identified using this model.
Given the estimated 2.2 Mb genome size of GBS, an
assumption of a 1 kb average gene size suggests that
complete coverage by transposon mutagenesis would
require in excess of 2200 transposon mutants. Despite
this fact, duplication of mutants indicates that the system
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
1451
is approaching saturation or, alternatively, that these are
insertion hot-spots. Although Tn917 is reported to integrate
more randomly than other Gram-positive transposons, it
is thought that all transposons exhibit some sequence
bias for transposition (Coulter et al., 1998). Although we
did obtain multiple transposon insertions in a number of
genes, we did not observe exact duplication of transposon
insertion sites, suggesting that hot-spots for Tn917
insertion are less likely.
We were surprised that mutations in the capsule
polysaccharide locus were not recovered, particularly as
the animals invariably cleared the asialo capsule mutant
included in our study as a control strain. Similarly, Polissi
et al. (1998) did not identify any capsule mutants of S.
pneumoniae during STM screening. Using STM screening
of S. aureus, Coulter et al. (1998) found only a small
proportion of transposon insertions in classical virulence
genes, such as proteases, toxins or adhesins, that are
involved in bacterial colonization. This was attributed to
the required use of high-titre inocula to ensure non-biased
STM sampling of the surviving mutants. Additionally,
virulence factors are not constitutively expressed and are
often transcriptionally activated in response to the growth
phase of a culture. A variety of hypotheses has been
proposed to explain the bias in the genes identified during
STM screening. Coulter et al. (1998) hypothesized that
the absence of many of these products may reflect the in
vivo growth state achieved at the time of mutant recovery.
STM screening may be biased against the identification of
genes such as secreted toxins, adhesins and binding
proteins involved in colonization because of its inherent
limitations. Conversely, it may favour the identification of
genes impacting longer term in vivo growth and persis-
tence. (Coulter et al., 1998). Additionally, the model
system used for screening mutants may significantly
influence or bias the genes that are identified.
DNA sequence analysis of regions flanking transposon
insertions in the GBS mutants revealed that approxi-
mately 50% of them represent genes with unknown
function. Our results are in agreement with those obtained
by Mei et al. (1997), who reported similar results in S.
aureus. Insertions in genes lacking homology to known
genes in the databases suggest that several genes
encode novel factors that are potentially important during
the infection process. The identification of the entire ORF
and searches for specific motifs or domains may give
an indication of the function of these genes. Additional
studies are under way to characterize the role of these
unknown genes in GBS infection.
The successful application of STM to GBS establishes
a basis for the application of this technique to other animal
models of infection, such as the pulmonary colonization
model (Martin et al., 1988). Presumably, the lung colon-
ization model will identify mutants that fail to survive
1452 A. L. Jones, K. M. Knoll and C. E. Rubens
during the early stages of infection, including the inability
to adhere to mucosal surfaces, invasion of epithelial
surfaces and evasion of lung clearance mechanisms.
Comparing the profiles of attenuated mutants obtained in
different animal models of infection will identify genes
required for specific stages of infection, as well as those
required for all infection states.
STM analyses revealed GBS genomic loci that encode
a wide variety of functional gene classes, underscoring
the diversity of bacterial processes required for the
infection process. Currently, the function of the genes
identified in the screen can only be inferred by homology;
however, a number of genes that were identified correlate
with virulence in other pathogens. Multiple mutants were
obtained in gene homologues of the transporter, regulator
and cell envelope families, emphasizing the importance of
these functions for GBS infection. These results represent
the first report of a genome-wide scan for virulence factors
in GBS, and a number of important putative virulence
factor genes worthy of further study have been identified.
Experimental procedures
Bacterial strains and plasmids
S. agalactiae strain A909 is a type Ia capsular polysaccharide
clinical isolate (Madoff et al., 1991). GBS strain A909 was
grown in Todd±Hewitt broth (THB) at 378C with aeration or on
Todd±Hewitt agar plates (THA) at 378C. GBS strain COH31-
21 is a tetracycline-resistant transposon mutant that pro-
duces a sialic acid-deficient (asialo) capsule (Wessels et al.,
1989). COH31-21 was grown in THB containing 10 mg ml21
tetracycline (tet). A909 containing pAJ010 was grown at 308C
on THA containing 800 mg ml21 kanamycin (km) to allow
plasmid replication. Transposon mutants of A909 were
maintained on TH medium with 5 mg ml21 erythromycin
(em) at 378C. E. coli strain DH5a was grown in Luria±Bertani
(LB) medium with or without tetracycline (10 mg ml21) at
378C.
DNA manipulations, polymerase chain reaction (PCR),
digoxigenin labelling and hybridizations
Chromosomal DNA from GBS was isolated as described by
Framson et al. (1997). Digestion and modification of plasmid
and chromosomal DNA was performed as described by
Sambrook et al. (1989). Plasmid DNA from GBS A909 was
isolated using a Wizard Miniprep kit (Promega) according to
the manufacturer's directions, except that the bacterial cells
were converted to spheroplasts by mutanolysin (Sigma)
digest (10 U ml21) at 378C for 40 min before plasmid
isolation.
Signature tags were amplified and labelled in two succes-
sive rounds of PCR using a modification of the methods of
Hensel et al. (1995). First-round PCRs were performed using
primers P5 (5 0 -GCGGTACAACCTCAAGCT-3 0 ) and P6 (5 0 -
GCGGCATTCTAACCAAGC-3 0 ). The PCR product was gel
purified and used as a template for round two PCR. In round
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
two PCR, digoxigenin (DIG)-11-dUTP was incorporated into
the tags as described by the manufacturer (Roche) using
primers P7 (5 0 -CGGTACAACCTCAAGCTT-3 0 ) and P8 (5 0 -
GGCATTCTAACCAAGCTT-3 0 ). For dot-blot hybridizations,
1 mg of plasmid DNA was transferred to MagnaGraph
membrane (Micron Separations) using a Minifold I micro-
sample filtration manifold (Schleicher and Schuell). Plasmid
DNA on the membranes was denatured according to the
manufacturer's directions (Micron Separations) and UV
cross-linked using a Stratalinker (Stratagene).
Southern hybridization analysis was performed on 20
transposon mutants. A 4 kb KpnI fragment of Tn917stm
was random-prime labelled with DIG-11-dUTP using a DIG
DNA labelling and detection kit (Roche) and used to probe
HindIII-digested chromosomal DNA.
Construction of plasmid pAJ010
Plasmid pAJ010, carrying a modified Tn917 (Tn917stm), was
constructed for STM of GBS strain A909. The temperature-
sensitive vector pAJ010 was constructed from pTV10K, a
temperature-sensitive derivative of the broad-host-range
plasmid pWV01 (Gutierriez et al., 1996). Plasmid pAJ010
contains the temperature-sensitive repA (repAts) gene from
pWV01, the kanamycin resistance gene of pTV10K and
Tn917stm.
Plasmid pTV10K was digested with PstI and HpaI to
remove 2 kb of the plasmid backbone. The LTR of Tn917
was regenerated, and suitable restriction sites for insertion of
the tags were added. PCR primers AJP1 (5 0 -CTGCAGGG-
GGTCCCGAGCGCCTA-3 0 ) and AJP2 (5 0 -GTTAACCCGC-
GGGAATTCCATACGCAAGACCAATCACTC-3 0 ) were used
to PCR amplify a 300 bp fragment from pTV10K consisting of
the left flank of Tn917, including the LTR. PCR primers were
designed to add a PstI site immediately 5 0 to the Tn917 LTR
and EcoRI, SacII and HpaI sites at the 3 0 end. The 300 bp
PCR product was gel purified, digested with PstI and HpaI
and ligated to pTV10K digested with PstI and HpaI. The
resulting 9.1 kb plasmid was designated pAJ005. To create
the modified Tn917, Tn917stm, a 3.2 kb SspI±DraI fragment
containing the ori/rop/tet region of pBR322 was blunt ended
and ligated to HpaI-digested, phosphatase-treated pAJ005.
The resulting plasmid, pAJ010, allows replication and
selection of the plasmids in E. coli at 378C and self-cloning
of GBS chromosomal DNA flanking transposon insertions.
Construction of pAJ010 is shown in Fig. 1.
Cloning and selection of tags
Double-stranded signature tags (89 bp) were generated by
PCR using the variable oligonucleotide pool AJ01 (5 0 -
TCCCCGCGGTACAACCTCAAGCTT-(NK)20-AAGCTTGGT-
TAGAATGCCGCGGGGA-3 0 ) as template DNA and primers
P3 (5 0 -TCCCCGCGGTACAACCTC-3 0 ) and P4 (5 0 -TTCCC-
GCGGCATTCTAAC-3 0 ) according to the methods of Hensel
et al. (1995). The double-stranded tags were digested with
SacII, gel purified and ligated with SacII-digested, dephos-
phorylated pAJ010. Transformations were performed directly
into GBS strain A909 when possible or through an E. coli
DH5a as an intermediate host. A909 transformants were
 Signature-tagged mutagenesis of group B Streptococcus
21
selected on THB with 800 mg ml km at 308C, and E. coli
DH5a transformants were selected at 378C on LB agar
containing 40 mg ml21 km. Transformants were screened
using PCR and primer AJP1 (5 0 -CTGCAGGGGGTCCC-
GAGCGCCTA-3 0 ), which is located in the plasmid backbone
extending into the LTR, and primer AJP6 (5 0 -GCGGCAT-
TCTAACCAAGC-3 0 ), which is internal to the tag, for the
presence of a single tag that amplified well. Tags were then
tested by dot-blot hybridization for cross-hybridization, and a
subset was sequenced to ensure that the tags were unique.
Tagged plasmids were isolated from each clone, and restric-
tion analysis was performed to confirm that the structure of
pAJ010 had been maintained during manipulations. Eighty
tagged plasmids were selected for library construction and
designated the master set. The A909 clones containing the
master set of tagged plasmids were arrayed in 96-well micro-
titre plates generating a master plate. A series of identical
membranes for dot-blot hybridization was prepared by
transferring 1 mg of each of the selected tagged plasmids
onto MagnaGraph (Micron Separations) membrane using the
Minifold I microsample filtration manifold (Schleicher and
Schuell).
Generation of mutant pools
Using the methods of Mei et al. (1997), the master plate was
used for each round of mutagenesis, such that the same 80
tags were used to generate multiple pools of mutants. To
generate 80 different transposon mutants, each well from the
master plate was stamped into a new 96-well plate using a
48-well stamper and grown overnight at 378C in 200 ml of
THB media to allow plasmid loss. This plate was then serially
subcultured 1:20 and grown overnight at 378C in THB
1 mg ml21 em and THB 5 mg ml21 em in 96-well plates to
select for transposon mutants. Individual colonies were
recovered from each well and replica plated on THA
5 mg ml21 em (378C) plates and THA 800 mg ml21 km
1 mg ml21 em (308C) plates to identify transposon mutants
that had lost the plasmid. One emR kmS colony from each
well was selected and picked to a 96-well plate containing
THB 5 mg ml21 em. Each pool of 80 mutants was stored at
2708C in THB 20% glycerol. A total of 20 pools of 80 mutants
(1600 individual mutants) was prepared.
Animal infection studies
Time-mated, barrier-sustained, Caesarean-delivered female
Sprague±Dawley rats were obtained from Charles River
Laboratories. Each pool of mutants was grown overnight at
378C in THB 5 mg ml21 em, subcultured 1:20 and grown until
the OD600 of individual wells was between 0.20 and 0.25.
Each pool of 80 mutants was spiked with an asialo mutant of
GBS (COH31-21), previously shown to have reduced viru-
lence in the neonatal sepsis model, as an indicator that
sufficient time had passed for the pups to clear avirulent
clones. COH 31-21 was included in the input pool as < 1/80th
of the total cfu, similar to all other mutants in the pool. The 80
transposon mutants from a single pool and the control strain
were combined and adjusted to a concentration of < 5  106
cfu ml21. This represented the input pool. One hundred
microlitres of the pool was injected intraperitoneally into three
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
1453
to five neonatal (20±24 h old) rat pups. At 20 h after inocu-
lation, the pups were sacrificed and the spleens removed.
Spleens were homogenized in 500 ml of sterile phosphate-
buffered saline (PBS), and an aliquot was removed for
quantitative culture. To ensure that sufficient time had passed
for the pups to clear the internal control strain COH 31-21, an
aliquot of the spleen homogenate was plated on THA
containing 10 mg ml21 tet. The remainder of the homogenate
was plated on a single THA plate and grown overnight. The
bacteria were harvested from the THA plate into PBS and
labelled as the output pool. Chromosomal DNA was isolated
from both input and output pools. The tags present in both
pools were amplified, labelled in two successive PCR reac-
tions and used to probe membranes arrayed with the 80
tagged transposons as described above.
Cloning of flanking DNA and sequence analysis
To clone the chromosomal DNA flanking the LTR of the
transposon insertions in mutants identified in the screen,
2 mg of chromosomal DNA was digested to completion with
NcoI, NsiI or KpnI, precipitated and self-ligated overnight at
168C. The ligated products were transformed into E. coli
DH5a, and the transformation mixture was plated on LB agar
containing tetracycline and incubated overnight at 378C.
Plasmid DNA from transformants was obtained using a
QIAprep spin miniprep kit (Qiagen). Flanking DNA sequence
was obtained by either directly sequencing the plasmid or
sequencing a PCR product of the flanking DNA region using
primer Tn917 LX (5 0 -AATGTACAAAATAACAGCGAA-3 0 ),
corresponding to the LTR from the em-proximal end of
Tn917. PCR products of flanking DNA were generated using
Tn917-specific primers Tn917L (5 0 -AGAGAGATGTCACCG-
TCAAGT-3 0 ) and Tn917N (5 0 -GGTAAGGCAAATCCAGAA-
AGC-3 0 ) for clones obtained using NsiI or NcoI, or Tn917L
and Tn917K (5 0 -CGGGGAATTTGTATCGATAAG-3 0 ) for
clones obtained by self-cloning with KpnI. DNA sequencing
was performed using a ThermoSequenase kit (Amersham) or
an Applied Biosystems BigDye sequencing kit according to
the manufacturer's directions. Sequences were analysed
using BLAST analysis (Altschul et al., 1997) and the University
of Wisconsin GCG package (Devereux et al., 1984). Homol-
ogy scores with a sum probability , e210 were considered
significant. Mutants with lower homology (e204 to e209) to
previously described genes were classified separately.
Homology scores , e203 were classified as having no
significant database match (Coulter et al., 1998).
Infection studies and competitive index assays
For rat pup competition assays, mutant strains and the wild-
type strain A909 were grown separately in THB at 378C
overnight. Overnight cultures were subcultured 1:100 in THB
and grown to mid-log phase (OD600 ˆ 0.3). Bacteria were
washed in PBS, and mutant strains and A909 were adjusted
to a concentration of 2  106 cfu ml21 and mixed together in
a 1:1 ratio (4  106 cfu ml21 total bacteria). Rat pups (20±
24 h old) were inoculated intraperitoneally with 0.1 ml of the
mixture. Dilutions of each mutant and A909 were plated on
THA to determine the exact input ratio of mutant to wild
1454 A. L. Jones, K. M. Knoll and C. E. Rubens
type. After 20 h, spleens were recovered and homogenates
plated onto selective media to determine the output ratio of
mutant to wild-type colonies. Recovery of mutant em-
resistant colonies required induction of the em gene in
Tn917. Spleen homogenates were split into two aliquots, and
0.01 mg ml21 em was added to one aliquot. Both aliquots
were incubated for 30 min at 378C. The cfu of the mutant
strains were determined by plating the induced aliquot on
THA 1 mg ml21 em, and total cfu were determined by plating
the non-induced aliquot on THA plates. The competitive
index was calculated by dividing the output ratio (mutant/wild-
type) by the input ratio (mutant/wild-type).
For lethal dose (LD50) studies, neonatal rat pups (20±24 h
old) were inoculated intraperitoneally with 0.1 ml of each
strain in PBS. Randomized groups of five pups were
inoculated with serial log dilutions of mid-logarithmic phase
bacteria, and the LD50 was calculated after 72 h using the
methods of Reed and Muench (1938).
The growth rate of each mutant relative to A909 was
assessed by inoculating 5 ml of THB with 50 ml of an
overnight culture of each mutant. Cultures were grown at
378C with shaking, and the OD600 was recorded hourly.
Acknowledgements
The authors wish to thank Anne Clancy and Christiane
Beckmann for critical review of the manuscript and helpful
discussions. This work was funded by the National Institutes
of Health ± The Streptococcal Initiative, grant no. BWH
811501/N01-AI-75326.
References
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J.,
Zhang, Z., Miller, W., et al. (1997) Gapped BLAST and PSI-
BLAST: a new generation of protein database search
programs. Nucleic Acids Res 25: 3389±3402.
Appleby, J.L., Parkinson, J.S., and Bourret, R.B. (1996)
Signal transduction via the multi-step phospho-relay: not
necessarily a road less traveled. Cell 86: 845±848.
Brunskill, E., and Bayles, K. (1996) Identification and
molecular characterization of a putative regulatory locus
that affects autolysis in Staphylococcus aureus. J Bacteriol
178: 611±618.
Bult, C.J., White, O., Olsen, G.J., Zhou, L., Fleishmann, R.D.,
Sutton, G.G., et al. (1996) Complete genome sequence of
the methanogenic archaeon, Methanococcus janaschii.
Science 273: 251±272.
Camacho, L.R., Ensergueix, D., Perez, E., Gicquel, B., and
Guilhot, C. (1999) Identification of a virulence gene cluster
of Mycobacterium tuberculosis by signature-tagged muta-
genesis. Mol Microbiol 34: 257±267.
Chanter, N., Ward, C., Talbot, N., Flanagan, J., Binns, M.,
Houghton, S., et al. (1999) Recombinant hyaluronate
associated protein as a protective immunogen against
Streptococcus equi and Streptococcus zooepidemicus
challenge in mice. Microb Pathog 27: 133±143.
Chiang, S.L., and Mekalanos, J.J. (1998) Use of signature-
tagged transposon mutagenesis to identify Vibrio cholerae
genes critical for colonization. Mol Microbiol 27: 797±805.
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
Coulter, S.N., Schwan, W.R., Ng, E., Langhorne, M.H.,
Ritchie, H.D., Westbrock-Wadman, S., et al. (1998)
Staphylococcus aureus genetic loci impacting growth and
survival in multiple infection environments. Mol Microbiol
30: 393±404.
Darwin, A.J., and Miller, V.L. (1999) Identification of Yersinia
enterocolitica genes affecting survival in an animal host
using signature-tagged transposon mutagenesis. Mol
Microbiol 32: 51±62.
Devereux, J., Haeberli, P., and Smithies, O. (1984) A
comprehensive set of sequence analysis programs for
the VAX. Nucleic Acids Res 12: 387±395.
Foulongne, V., Bourg, G., Cazevielle, C., Michaux-Charachon,
S., and O'Callaghan, D. (2000) Identification of Brucella
suis genes affecting intracellular survival in an in vitro
human macrophage infection model by signature-tagged
transposon mutagenesis. Infect Immun 68: 1297±1303.
Framson, P.E., Nittayajarn, A., Merry, J., Youngman, P., and
Rubens, C.E. (1997) New genetic techniques for group b
Streptococci high-efficiency transfomation, maintenance of
temperature-sensitive pwv01 plasmids, and mutagenesis
with Tn917. Appl Environ Microbiol 63: 3539±3547.
Gottesman, S., Squires, C., Pichersky, E., Carrington, M.,
Hobbs, M., Mattick, J.S., et al. (1990) Conservation of the
regulatory subunit for the Clp ATP-dependent protease in
prokaryotes and eukaryotes. Proc Natl Acad Sci USA 87:
3513±3517.
Graham. J.E., and Clark-Curtiss, J.E. (1999) Identification
of Mycobacterium tuberculosis RNAs synthesized in
response to phagocytosis by human macrophages by
selective capture of transcribed sequences (SCOTS).
Proc Natl Acad Sci USA 96: 11554±11559.
Gutierriez, J., Crowley, P., Brown, D., Hillman, J., Youngman,
P., and Bleiweis, A. (1996) Insertional mutagenesis and
recovery of interrupted genes of Streptococcus mutans by
using transposon Tn917: preliminary characterization of
mutants displaying acid sensitivity and nutritional require-
ments. J Bacteriol 178: 4166±4175.
Hensel, M., Shea, J.E., Gleeson, C., Jones, M.D., Dalton, E.,
and Holden, D.W. (1995) Simultaneous identification of
bacterial virulence genes by negative selection. Science
269: 400±403.
Klarsfeld, A.D., and Cossart, P. (1995) Disruption of the
operon downstream from the major sigma factor gene
affects growth and virulence of Listeria monocytogenes.
Unpublished Genbank submission LMU17284.
Lansing, M., Lellig, S., Mausolf, A., Martini, I., Crescenzi, F.,
O'Regan, M., et al. (1993) Hyaluronate synthase: cloning
and sequencing of the gene from Streptococcus sp.
Biochem J 289: 179±184.
Lowe, A.M., Beattie, D.T., and Deresiewicz, R.L. (1998)
Identification of novel staphylococcal virulence genes by in
vivo expression technology. Mol Microbiol 27: 967±976.
Lukomski, S., Burns, E.H., Wyde, P.R., Podbielski, A.,
Rurangirwa, J., Moore-Poveda, D.K., et al. (1998) Genetic
inactivation of an extracellular cysteine protease (SpeB)
expressed by Streptococcus pyogenes decreases resis-
tance to phagocytosis and dissemination to organs. Infect
Immun 66: 771±776.
Lyon, W.R., Gibson, C.M., and Caparon, M.G. (1998) A role
for trigger factor and an RGG-like regulator in the
 Signature-tagged mutagenesis of group B Streptococcus
transcription, secretion and processing of the cysteine
proteinase of Streptococcus pyogenes. EMBO J 17: 6263±
6275.
Madoff, L.C., Michel, J.L., and Kasper, D.L. (1991) A
monoclonal antibody identifies a protective C-protein
alpha-antigen epitope in group B streptococci. Infect
Immun 30: 204±210.
Mahan, M.J., Slauch, J.M., and Mekalanos, J.J. (1993)
Selection of bacterial virulence genes that are specifically
induced in host tissues. Science 259: 686±688.
Martin, T., Rubens, C., and Wilson, C. (1988) Lung
antibacterial defense mechanisms in infant and adult rats:
implications for the pathogenesis of group B streptococcal
infections in the neonatal lung. J Infect Dis 157: 91±100.
Massida, O., Anderluzzi, D., Friedl, L., and Feger, G. (1998)
Unconventional organization of the division and cell wall
gene cluster of Streptococcus pneumoniae. Microbiology
144: 3069±3078.
Mei, J., Nourbakhsh, F., Ford, C.W., and Holden, D.W.
(1997) Identification of Staphylococcus aureus virulence
genes in a murine model of bacteraemia using signature-
tagged mutagenesis. Mol Microbiol 26: 399±407.
Murray, T., Popham, D.L., and Setlow, P. (1998) Bacillus
subtilis cells lacking penicillin-binding protein 1 require
increased levels of divalent cations for growth. J Bacteriol
180: 4555±4563.
Nickel, V., Prehm, S., Lansing, M., Mausolf, A., Podbielski,
A., Deutscher, J., et al. (1998) An ectoprotein kinase of
group C streptococci binds hyaluronan and regulates
capsule formation. J Biol Chem 273: 23668±23673.
Nizet, V., Ferrieri, P., and Rubens, C. (2000) Molecular
pathogenesis of group B streptococcal disease in new-
borns. In Streptococcal Infections, Clinical Aspects, Micro-
biology and Molecular Pathogenesis. Stevens, D.L., and
Kaplan, E.L. (eds). New York: Oxford University Press, pp.
180±221.
Podbielski, A., and Leonard, B. (1998) The group A
streptococcal dipeptide permease (Dpp) is involved in the
uptake of essential amino acids and affects the expression
of cysteine protease. Mol Microbiol 28: 1323±1334.
Polissi, A., Pontiggia, A., Feger, G., Altieiri, M., Mottl, H.,
Q 2000 Blackwell Science Ltd, Molecular Microbiology, 37, 1444±1455
1455
Ferrari, L., et al. (1998) Large-scale identification of
virulence genes from Streptococcus pneumoniae. Infect
Immun 66: 562±5629.
Porankiewicz, J., Wang, J., and Clarke, A.K. (1999) New
insights into the ATP-dependent Clp protease: Escherichia
coli and beyond. Mol Microbiol 32: 449±458.
Reed, L.J., and Muench, H. (1938) A simple method of
estimating fifty percent endpoints. Am J Hyg 27: 493±497.
Rouquette, C., Ripio, M.T., Pellegrini, E., Bolla, J.M., Tascon,
R.I., Vazquez-Boland, J.A., et al. (1996) Identification of a
ClpC ATPase required for stress tolerance and in vivo
survival of Listeria monocytogenes. Mol Microbiol 21: 977±
987.
Rubens, C.E., Wessels, M.R., Heggen, L.M., and Kasper, D.
(1987) Transposon mutagenesis of type III group B
streptococcus: correlation of capsule expression with
virulence. Proc Natl Acad Sci USA 84: 7208±7212.
Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989).
Molecular Cloning: A Laboratory Manual. Cold Spring
Harbor, NY: Cold Spring Harbor Laboratory Press.
Schuchat, A. (1998) Epidemiology of group B streptococcal
disease in the United States: shifting paradigms. Clin
Microbiol Rev 11: 497±513.
Spellerberg, B., Rozdzinski, E., Martin, S., Weber-Heyne-
mann, J., Schnitzler, N., Lutticken, R., et al. (1999) Lmb, a
protein with similarities to the LraI adhesin family, mediates
attachment of Streptococcus agalactiae to human laminin.
Infect Immun 67: 871±878.
Throup, J.P., Koretke, K.K., Bryant, A.P., Ingraham, K.A.,
Chalker, A.F., Ge, Y., et al. (2000) A genomic analysis
of two-component signal transduction in Streptococcus
pneumoniae. Mol Microbiol 35: 566±576.
Wessels, M.R., Rubens, C.E., Benedi, V., and Kasper, D.
(1989) Definition of a bacterial virulence factor: sialylation
of the group B streptococcal capsule. Proc Natl Acad Sci
USA 86: 8983±8987.
Zeligs, B.J., Armstrong, C.D., Walser, J.B., and Bellanti, J.A.
(1982) Age-dependent susceptibility of neonatal rats to
group B streptococcal type III infection: correlation of
severity and response of myeloid pools. Infect Immun 37:
255±263.