Genetics-squared: Combining host and pathogen genetics in the analysis of innate immunity and bacterial virulence

Abstract

The interaction of bacterial pathogens with their hosts' innate immune systems can be extremely complex and is often difficult to disentangle experimentally. Using mouse models of bacterial infections, several laboratories have successfully applied genetic approaches to identify novel host genes required for innate immune defense. In addition, a variety of creative bacterial genetic schemes have been developed to identify key bacterial genes involved in triggering or evading host immunity. In cases where both the host and pathogen are amenable to genetic manipulation, a combination of host and pathogen genetic approaches can be used. Focusing on bacterial infections of mice, this review summarizes the benefits and limitations of applying genetic analysis to the study of host-pathogen interactions. In particular, we consider how prokaryotic and eukaryotic genetic strategies can be combined, or "squared," to yield new insights in host-pathogen biology.

Full text

REVIEW
Genetics-squared: combining host and pathogen genetics
in the analysis of innate immunity and bacterial virulence
Jenny Persson & Russell E. Vance
Received: 16 August 2007 / Accepted: 20 August 2007
#
Springer-Verlag 2007
Abstract The interaction of bacterial pathogens with their
hosts’ innate immune systems can be extremely complex
and is often difficult to disentangle experimentally. Using
mouse models of bacterial infections, several laboratories
have successfully applied genetic approaches to identify
novel host genes required for innate immune defense. In
addition, a variety of creative bacterial genetic schemes
have been developed to identify ke y bacterial gen es
involved in triggering or evading host immunity. In cases
where both the host and pathogen are amenable to genetic
manipulation, a combination of host and pathogen genetic
approaches can be used. Focusing on bacterial infections of
mice, this review summarizes the benefits and limitations of
applying genetic analysis to the study of host–pathogen
interactions. In particular, we consider how prokaryotic and
eukaryotic genetic strategies can be combined, or
“squared,” to yield new insights in host–pathogen biology.
Keywords Genetics
.
Mutagenesis
.
Host resistance
.
Mice
.
Bacteria
Introduction
It could be argued that infection is an intrinsic a nd
unavoidable feature of life. Certainly, infectious disease
continues to pose a great burden on humanity, accounting
for approximately a quarter of deaths worldwide each year,
with the bulk of these occurring in developing countries
(World Health Organization). The advantages pathogens
have over their hosts are numerous, including the general
ability of pathogens to replicate much faster than their
hosts. Thus, in the face of natural selective pressures—or
even in the face of unnatural selective pressures applied by
humans, such as antibiotics—pathogens exhibit a virtually
unlimited capacity to adapt, infect, and (too often) cause
disease. There is no single defining structural characteristic
of a pathogen; indeed, the variety of pathogens is
impressive. Yet, the difference between a pathogen and a
commensal organism is often simply a matter of a few
genes on a virulence plasmid.
It is against this backdro p that the remarkable capacity of
vertebrate immune systems becomes evident. Many, if not
most, infections do not cause significant morbidity or
mortality and are limited by the host. Whether, in fact,
reduced host pathology is merely yet another way that
pathogens manipulate their hosts to maximize replication is
still an interesting question (Brown et al. 2006; Portnoy
2005). Regardless, it is clear that host immune systems are
adaptable and sophisticated enough that their interactions
with pathogens are often highly complex and interwoven.
In this paper, we consider what the complexity of host–
pathogen interactions might mean for the geneticist trying
to dissect these interactions. Genetic strategies have been
proven time and again to provide powerful experimental
insight into complex biological processes, and the immune
response is no exception. However, as is widely acknowl-
edged and discussed further below, genetic redundancy and
pleiotropy can often obscure the power of genetics to
determine roles for bacterial or host genes in infections. In
cases where both the host and pathogen are genetically
amenable, one idea that recurs in the literature is that it
might be possible to gain further insight into infectious
disease processes by developing genetic strategies in which
host and pathogen genetic approaches are combined in a
single experimental system. We refer to this id ea as
Immunogenetics
DOI 10.1007/s00251-007-0248-0
J. Persson
:
R. E. Vance (*)
415 Life Science Addition, University of California,
Berkeley, CA 94720, USA
e-mail: rvance@berkeley.edu

“genetics-squared.” Here, focusing on bacterial infections
of mice, we first discuss examples of host- and pathogen-
based genetic strategies that have been applied, and then
consider the potential of “genetics-squared” as an additional
strategy in the study of host–pathogen interactions.
Genetic approaches in the mouse
A wide variety of genetic approac hes have been applied in
the mouse to dissect host–pathogen interaction s. The ability
to generate targeted “gene knockout” mutations in the
mouse genome has been an incredibly valuable “reverse
genetic” tool for obtaining valuable information about the
role of particular (known) host genes in the response to
pathogens (Buer and Balling 2003). Here, we focus on
forward genetic strategies aimed at determining novel host
genes involved in resistance to bacterial pathogens. Tradi-
tionally, the most popular strategy has been to identify
natural genetic variants among inbred mous e stra ins
involved in resistance to pathogens. A contrasting approach
that we also consider is the generation of novel mutations
affecting susceptibility to pathogens using the potent
mutagen N-ethyl-N-nitrosourea (ENU).
Identification of natural polymorphisms
The develo pment of a large number of inbred mouse strains
that exhibit considerable genetic and phenotypic diversity
(Wade and Daly 2005) has permitted the discovery of
several naturally occurring mutations in novel genes that
control susceptibility to bacterial pathogens. The following
examples illustrate both the power and limitations of using
naturally occurring variation as a way to dissect host–
pathogen interactions.
Nramp1 (Slc11a1) The contrasting susceptibility of differ-
ent inbred mouse strains to the intracellula r pathogens
Mycobacterium bovis (BCG), Salmonella typhimurium, and
Leishmania donovani was mapped to the same locus on
chromosome 1, denoted Bcg, Ity, or Lsh in each respective
case (Bradley 1974 ; Gros et al. 1981; Plant and Glynn
1974). Subsequent g enetic analysis indicated that the
phenotype was controlled by a single dominant gene within
the Bcg (Ity, Lsh) locus (Bradley and Kirkley 1977; Gros et
al. 1981; Plant and Glynn 1976; Skamene et al. 1982), and
after narrowing the locus to an ~1-Mb region, the gene
responsible for the phenotype was defined by positional
cloning (Vidal et al. 1993). Analysis of the expression
pattern of the candidate gene revealed a high association
with reticuloendothelial organs (spleen and liver), and in
particular, macrophages from these tissues. Thus, the gene
was named Nramp1 for Natur al resistance-associated
macrophage protein 1. Subsequently,
Nramp1 has been
renamed Solute carrier family 11a member 1 (Slc11a1).
Sequence analysis defined Nramp1 as a previously uniden-
tified protein, structurally similar to membrane-spanning
proteins, and containing a motif conserved in prokaryotic
and eukaryotic transport proteins (Vidal et al. 1993). Analysis
of Nramp1 allelic variants uncovered a strong correlation
between the susceptible phenotype and a single amino acid
substitution in a transmembrane domain (Malo et al. 1994;
Vidal et al. 1993). Notably, the Nramp1 mutation has little or
no effe ct on mouse sus ceptibility to M. tuberculosis
(Kramnik et al. 2000; North et al. 1999). The formal
confirmation of Nramp1 as the responsible gene in differing
susceptibility phenotypes was obtained by the demonstration
that a mouse carrying a null allele at Nramp1 exhibited
increased susceptibility to M. bovis, S. typhimurium, and L.
donovani (Vidal et al. 1995). With the discovery that Nramp
homologs function in transport of divalent cations (Fleming
et al. 1997; Gunshin et al. 1997; Supek et al. 1996), it was
suggested that Nramp1 might also function as a divalent
cation pump. Indeed, subsequent studies showed that Nramp1
is a pH-dependent transporter of divalent cations and that
Nramp1 transports Mn
2+
more efficiently than Fe
2+
(Forbes
and Gros 2003; Jabado et al. 2000). It has been proposed that
Nramp1 functions in antimicrobial resistance by limiting the
phagosomal availability of critical metals and/or by modulat-
ing the phagosomal environment (Fortier et al. 2005). The
increasing number of studies on Nramp1 in diverse contexts
underlines the pleiotropic effects of this gene. Thus far,
Nramp1 has been ascribed distinct roles in different bacterial
infections, such as in infections with Mycobacteria (acidifi-
cation of mycobacterial phagosome; Hackam et al. 1998) or
Salmonella (maturation of Salmonella- containing vacuole;
Cuellar-Mata et al. 2002). Additional roles have also been
suggested, e.g., in Th polarization, MHC class II expression,
and antigen presentation (Caron et al. 2006; Kaye and
Blackwell 1989; Lang et al. 1997; Stober et al. 2007; Zwilling
et al. 1987).
The identification of Nramp1 is an impressive accom-
plishment that illustrates several themes that recur in the
forward genetic identification of pathogen-susceptibility
genes. First, forward genetics permits the identification of
novel genes that are not likely to have been otherwise
discovered. Second, the forward genetic identification of
naturally occurring polymorphisms affecting a phenotype is
often slow and di fficult. Although this proc ess will
undoubtedly be facilitated by the availability of whole
genome sequences for several mouse strains, an underlying
problem is that multiple polymorphisms are often linked to
a phenotype of interest. Even with whole genome sequen-
ces, it remains an extremely difficult problem to identify the
critical, causative polymorphism. As discussed below, this
problem does not arise as significantly for ENU-induced
Immunogenetics

mutations and is in fact one major advantage of ENU
mutagenesis over the identification of naturally occurring
polymorphisms. The third lesson is that the identification of
a gene influencing susceptibility to pathog ens—while time
consuming and difficult—often turns out to be only the
beginning. Even if the phenotype is controlled by a single
gene, that single gene can have pleiotropic effects. Indeed,
understanding the function of pathogen-susceptibility genes
often turns out to be more difficult than mapping them in
the first place.
Naip5 The intracellular bacterium Legionella pneumophila
exhibits little growth in mouse macrophages from the B6
mouse strain but replicate s 3 – 4 logs in macrophages from
the A/J strain (Yamamoto et al. 1988; Yoshida et al. 1991).
Studie s of crosses of restrictive and permissive mi ce
showed that restriction of L. pneumophila is dominant and
segregates in a Mendelian fashion to a single autosomal
locus on mouse chromosome 13, denoted Lgn1 (Beckers et
al. 1995; Dietrich et al. 1995). The Lgn1 locus was found to
contain a tandemly repeated family of genes called neural
apoptosis-inducing protein (Naip) genes (also sometimes
called Birc1 genes). Early genetic mapping of the Naip
(Birc1 ) genes implicated either Naip2 (Birc1b) or Naip5
(Birc1e) in controlling perm issiveness for Legionella
growth in mouse macrophages (Growney and Dietrich
2000; Growney et al. 2000). Ultimately, evidence obtained
by posi tio nal cloning and functional c omp lem ent ati on
(Diez et al. 2003; Wright et al. 2003) suggested that the
Naip5 gene was solely responsible for the Lgn1 phenotype.
The presence of 14 polymorphisms in the A/J allele of
Naip5, and/or the observed lower expression of the Naip5
gene in A/J macrophages, may be responsible for the
permissive phenotype (Wright et al. 2003). However, again
illustrating the confounding difficulties of multiple link ed
polymorphisms, the precise polymorphism(s) responsible
for the Lgn1 phenotype has not been identified. The role of
Naip5 in restriction of Legionella growth is still unclear but
has, together with the hom ologous protein Ipaf, been
suggested to be involved in intracellular sensing of bacterial
flagellin and subsequent activation of a rapid, caspase-1-
dependent macrophage cell death (see below; Molofsky et
al. 2006; Ren et al. 2006; Zamboni et al. 2006). This rapid
cell death may act to limit bacterial spread to new host cells.
Alternatively, Naip5 may promote rapid maturation of the
bacteria-containing vacuole, thereby inhibiting the forma-
tion of the specialized compartment in which Legionella
replicates (Amer and Swanson 2005; Fortier et al. 2007;
Watarai et al. 2001). Again, the identification of the
mechanism by which Naip5 protects cells from Legionella
has been almost as challenging as mapping and cloning the
gene in the first place. As yet, a Naip5 knockout mouse has
not been reported.
Tlr4 One of the most celebrated successes of the applica-
tion of forward genetics to innate immunity has been the
discovery that the unresponsiveness of certain mou se
strains to bacterial lipopolysaccaride (LPS) is controlled
by Toll-like receptor 4 (TLR4; Poltorak et al. 199 8 ). The
story began in the 1970s with the identification of the Lps
locus on chromosome 4. In C3H/HeJ mice, this locus
appeared to be linked to defective LPS responses (Watson
et al. 1977, 1978) and extreme susceptibility to infection
with Salmonella (O’Brien et al. 1980; Robson and Vas
1972; von Jeney et al. 1977). Early backcross analysis of
responsive C57BL/6J and unresponsive C3H/HeJ mice
suggested that a single gene was responsible for the Lps
phenotype (Watson et al. 1977, 1978), and genetic analysis
in C3H/HeJ and C57BL/10ScCr mice, both unresponsive to
LPS (Coutinho and Meo 1978; Watson et al. 1977, 1978),
indicated that this gene was the Tlr4 gene (Poltorak et al.
1998). The generation of a Tlr4 knock out mouse provi ded
confirmatory evidence that LPS unresponsiveness is indeed
due to lack of expression of Tlr4, and it was shown that
peritoneal macrophages from Tlr4
−/−
mouse were as
unresponsive to LPS as macrophages from C3H/HeJ mice
(Hoshino et al. 1999).
Interestin gly, and in c ontrast to the relatively slow
progress seen in identifying the function of other pathogen-
resistance genes identified by forward genetics, we have
learned a tremendous amount about the role of Tlr4 in
immune defense since its discovery. Rapid progress was
likely possible because prior data had already indicated that
Toll-like receptors were crucial in the response to patho-
gens, and there was already wide acceptance of the idea that
innate recognition of conserved microbial substances was
cr itical for immunity (Janeway 1989; Medzhitov and
Janeway 1997). One influential finding was the demonstra-
tion that Drosophila Toll is critical for resistance to fungal
infections (Lemaitre et al. 1996). In addition, the year
before the LPS-responsive phenotype was linked to the Tlr4
gene, Medzithov et al. had clo ned the huma n TLR4
homologue (hToll). By fusing the cytosolic region of
human Toll to the extracellular region of CD4, a TLR that
could signal in the absence of an activating ligand was
created and analyzed. Importantly, hToll exhibited several
characteristics of a receptor involved in innate immune
responses, such as induction of NF-κB, costimulator y
molecules, and proinflammatory cytokines such as IL-1,
IL-6, and IL-8 (Medzhitov et al. 1997). One lesson may be
that the power of genetics is often maximal when combined
with other approaches.
Nalp1b The lethal toxin ( LeTx) of Bacill us anthracis
induces a very rapid necrotic-like death in macr ophages of
certain mouse strains (e.g., C3H, 129; Friedlander et al.
1993; Roberts et al. 1998) but has only delayed or relatively
Immunogenetics

mild effects on macrophages from other mouse strains (e.g.,
C57BL/6). Differing susceptibility to LeTx was predicted
to be due to a single gene that mapped to a locus, Ltx1, on
mouse chromosome 11 (Roberts et al. 1998). Initially, the
gene Kif1c was proposed to be the gene responsible within
the Ltx1 locus (Watters et al. 2001). However, because the
“susceptible” allele of Kif1c did not always correlate with a
susceptible cell phenotype, this finding was later reeval-
uated, leading to the discovery that one of three tandem
Nalp (NACHT-, LRR- , and PYD-containing protein;
Martinon et al. 2002) paralogs, Nalp1b, is actually the
gene responsible for the Ltx1 phenotype (Boyden and
Dietrich 2006). Although Nalp1b is closely linked to Kif1c,
Nalp1b was originally not thought to cause the Ltx1
phenotype because expression of Nalp1 genes had not been
detected (Watters et al. 2001). Improved polymerase chain
reaction (PCR) and sequencing analysis of Nalp1b cDNA
from mouse strains with different susceptibility to LeTx
revealed that Nalp1b alleles are highly polymorphic and
that expression of specific alleles correlates with resistant or
susceptible phenotypes (Boyden and Dietrich 2006), again
ill ustrating the p otential pitfalls of forward genetic
approaches. The data suggested that a functional Nalp1b
protein correlates with LeTx-induced macrophage suscep-
tibility. This was confirmed with the gen erati on of a
transgenic mouse , in which expression of the susceptible
Nalp1b
129S1
allele rendered previously LeTx-resistant B6
macrophages fully sensitive (Boyden and Dietrich 2006). In
accordance with previous data showing that human NALP1
activates caspase-1 (Martinon et al. 2002) and that caspase-1
is activated in LeTx-susceptible murine macrophage cell
lines (Cordoba-Rodriguez et al. 2004), it was demonstrated
that LeTx activated caspase-1 in primary macrophages from
susceptible, but not resistant, mice (Boyden and Dietrich
2006). Moreover, Caspase-1
−/−
macrophages carrying a
“susceptible” allele of Nalp1b were virtually unresponsive
to LeTx treatment (Boyden and Dietrich 2006). Together, the
available data suggest that the LeTx-sensitive phenotype
depends on Nalp1b-mediated activation of caspase-1, which
induces macrophage cell death. This finding has received a
lot of attention in light of several studies demonstrating that
Nalp-like genes are critical components of a multiprotein
complex called the “inflammasome” that appears essential
for activation of caspase-1, leading to secretion of important
cytokines such as IL-1β and IL-18 (Meylan et al. 2006).
Thus, forward genetic analysis of even a relatively restricted
phenotype such as susceptibility to anthrax LeTx can result
in the discovery of biological pathways of fundamental
importance. The genetic analysis of the Ltx1 locus was
greatly facilitated by a rapid, robust, and easily assayed cell-
death phenotype. Nevertheless, the studies again illustrate
how difficult and time consuming forward genetic analysis
can be when confronted with the problem of multiple linked
polymorphisms. If this lesson is true of monogenic traits,
such as those discussed above, then it is even more apparent
for polygenic or quantitative traits, such as the Sst1 or Ctrq
loci discussed below.
Ipr1 and Mycobacterium tuberculosis Inbred mouse strains
exhibit major differences in susceptibility to M. tuberculosis
(MTB). Early work using crosses between resistant and
susceptible strains demonstrated that MTB susceptibility was
a complex non-Mendelian, i.e., quantitative, trait (Lurie et al.
1952; Lynch et al. 1965). Among the most resistant mouse
strains is B6, while C3HeB/FeJ inbred mice are extremely
susceptible. Notably, other C3H strains (C3H/HeJ, C3H/
HeSnJ, and C3H/HeOuJ) are relatively resistant to MTB
infection. In spite of the complex nature of the genetic traits
that regulate MTB infection, a major locus regulating MTB
susceptibility was recently mapped to mouse chromosome 1
and was named Sst1 for Supersusceptibility to tuberculosis 1
(Kramnik et al. 2000). This locus was shown to function
independently of infection route (respiratory or i.v.) and
appears to play a role in the lung-specific control of
infection, resulting in the diffe rential development of
necrotic lesions and the ability to control bacterial replica-
tion. However, Sst1 function did not confer full protection
against virulent MTB, underlining the multifactorial nature
of the fully resistant phenotype (Kramnik et al. 2000; Pan
et al. 2005).
Among the 22 gene s comprising the Sst1 locus, a
candidate gene i nvolve d in r egulating MTB infection,
intracellular pathogen resistance 1 (Ipr1), was identified
by positional cloning (Pan et al . 2005). Macroph age
susceptibility correlates with lack of Ipr1 expression, and
the ability to limit bacterial replication in lung tissue and
resist infection is partially restored by expression of a full-
length Ipr1 transgene. Interestingly, upon infection, macro-
phages carryi ng the resistant Sst1 locus succum b to
apoptosis, while Sst1-sensitive macrophages undergo ne-
crotic cell death, which may be reverted to apoptosis by
expres sion of the Ipr1 transge ne (Pan et al. 2005).
Importantly, the h uman homologue of Ipr1, SP110, has
been shown to exhibit polymorph isms associated with
susceptibility to MTB in West Africa (Tosh et al. 2006).
In a recent study by Yan et al., congenic mice ca rrying the
resi stant B6 Sst1 locus on a sensitive C3HeB/FeJ b ack-
ground were intercros sed with B6 mice, and F2 progeny
were analyzed. Infection of these mice g enerated a wide
range of survival times (10% died withi n 70 days, ~25%
survived more than 140 days), strongly indicating that
there are loci, other than Sst1, that det ermine sensitivity to
MTB infect ion. Indeed , whole geno me scans of individual
mice showing radical su rvival phenotypes indicated resis-
tance linkage to chromosomes 7, 12, 15, and 17. In
contrast, in an intercross where all mice carried the
Immunogenetics

sen sitive S st1 locus, 80% of the mice died within 30 da ys,
suggesting that, although there are other loci contributing
to sensitivi ty, they do not exhibit phenotypes powerful
enough to penetrate the Sst1-sensitive phenotype (Yan et
al. 2 006). This study highlight s one of the difficulties of
quantitative trait l oci (QTL) mapping, namely, that dom-
inance of one locus might mask effects of less prominent,
but not necessarily unimportant, loci.
Ipr1 and Listeria Resistance to Listeria monocytogenes is
also a complex quantitative trait (Boyartchuk et al. 2001 ),
and the Sst1 locus and the Ipr1 gene have also been shown
to contro l su sceptibility to Listeria infection in mice
(Boyartchuk et al. 2004; Pan et al. 2005). C3Heb/FeJ mice
carrying the resistant (B6) allele of Sst1 are somewhat more
efficient in controlling replication of Listeria in spleen and
liver than wild-type C3Heb/FeJ mice. To test the role of the
Sst1 locus in innate immunity more sensitively, Listeria
infection was monitored in mice that also carried the Scid
mutation, in which the role of Sst1 is more pronounced
because of defective adaptive immune responses. IFNγ-
producing NK-cells and phagocytic cells such as macro-
phages have previously been shown to contribute to control
of Listeria infection in Scid mice (Bancroft and Kelly 1994;
Dunn and North 1991; Huang et al. 1993; Tripp et al.
1994), yet neither NK-cell activation nor the production of
inflammatory cytokines or the ability to recruit phagocytic
cells differ significantly between C3Heb/FeJ-Scid mice
carrying susceptible or resistant Sst1 alleles (Boyartchuk
et al. 2004). However, at early time points of infection,
macrophages from resistant animals show increased levels
of reactive oxygen intermediates (ROI), indicating a more
efficient activation of these cells. Accordingly, differential
killing of Listeria in bone marrow-derived macrophages de-
pended on IFNγ-dependent production of ROI (Boyartchuk
et al. 2004). As for macrophages infected with MTB,
resistance to Listeria is associated with expression of the
Ipr1 gene, and shift of Listeria-induced necrosis (Barsig and
Kaufmann 1997) to apoptosis (Pan et al. 2005). In view of
available data, it is an attractive suggestion that the Ipr1
gene
may be part of a general pathway protecting against
intracellular pathogens that induce necrotic cell death.
It should be noted that, in crosses of resistant B6/ByJ
and sensitive BALB/cByJ mice (Boyartchuk et al. 2001),
Ipr1, or Sst1, were not identified as major contributors to
the genetic control of Listeria infection. Infected B6/ByJ ×
BALB/cByJ F2 mice did not display normally distributed
survival rates or lengths (required for traditional QTL
mapping; Lander and Botstein 1989), and thus, analysis
was performed using a novel statistical “single-QTL”
model. Using this model, the authors defined two loci on
chromosomes 5 and 13, respectively, while a locus on
chromosome 1, i.e., possibly the Sst1 locus, was observed
as a locus with minor influence on Listeria infection
(Boyartchuk et al. 2001 ). Taken together, the Listeria and
MTB studies illustrate the point that analysis of quantitative
traits is considerably more challenging than monogenic
traits. In addition, it is important to bear in mind the
obvious point that different strain backgrounds or experi-
mental approaches can lead to differing conclusions on the
role of given genes in the response to pathogen. For
instance, B6/ByJ mice are themselves more resistant to
Listeria than are the more commonly employed B6/J mice,
owing to a polymorphism that affects splicing of the
transcription factor IRF-3 (Garifulin et al. 2007).
Ctrq-3 (Igtp, Irgb10) Chlamydia trachomatis is a leading
cause of preventable blindness and infertility in the world.
This obligate intracellular ba cterium is currently not
amenable to genetic manipulation, which limits its acces-
sibility for genetic studies. Dietrich and colleagues devel-
oped a mouse model of systemic Chlamydia infection,
which allowed them to map QTL that segregated with
bacterial load in the spleen during acute infection of B6 ×
C3H/HeJ F2 generation mice. Linkage was found to three
loci on chromosomes 2, 3, and 11, respectively, and these
loci were denoted Ctrq-1 , Ctrq-2, an d Ctrq-3 for C.
t rachomatis resistance QTL (Bernstein-Hanley et al.
2006a). A congenic mouse strain carrying the susceptible
Ctrq-3 allele on a resistant B6 backgroun d exhibited
defective IFNγ-dependent susceptibility to Chlamydia in
fibroblasts (Bernstein-Hanley et al. 2006a). Fine structure
mapping of Ctrq-3 identified Igtp and Irgb10, two members
of the p47 family of IFNγ-inducible GTPases, a family of
genes believed to play important roles in resistance to
several intracellular pathogens (MacMicking 2005). Over-
expression of the susceptible or the resistant allele of Irgb10
suggested that the difference in susceptibility is related to a
difference in expression levels rather than coding poly-
morphisms. The role of Igtp in resistance to Chlamydia
infection is less clear because overexpression of either the
B6 or C3H allele r endered cells more susceptible to
Chylamydia, as does lack of Igtp
expression (Bernstein-
Hanley et al. 2006b). In a new study by Miyairi et al.
(2007), genetic analysis of crosses between B6 and DBA/2J
mice identified Iigp2, yet another p47 GTPase, as a player in
infection with C. psittaci. Iigp2 is located in the same gene
cluster as the previously identified Irgb10 and Igtp loci.
Identification of novel induced mutations
It is interesting to note that many pathogen-susceptibility/
resistance genes are often members of small multigene
families. This is true of the Naip5, Tlr4, Nalp1b, and p47
GTPase genes discussed above, as well as of even more
Immunogenetics

classic immune loci such as Mhc. Immune-related multi-
gene families often exhibit considerable genetic variation,
likely driven by evolutionary interactions with pathogens.
The significant variation is what permits genetic analysis in
inbred strains, but as discussed above, when multiple
polymorphisms in multiple paralogous genes are linked to
a given phenotype, genetic variation can also severely
complicate the identification of the key, causative genetic
variant. Several investigators have therefore realized that
generation of nov el mutations in a uniform genetic
background may afford several advantages. First, one is
not limited to naturally occurring varia tion, and one may
therefore characterize alleles affecting highly conserved
genes as well as variable genes. Bu t perhaps more
importantly, identification of an induced mutation is usually
much more straightforward than identification of a naturally
occurring variant. Under mutagenesis schemes using the
highly efficient point mutagen ENU, the frequency of base
changes is approximately 1 per megabase, or approximately
3,000 per haploid genome. The frequency of loss-of-
function mutations is orders of magnitude lower and has
been estimated to be approximately 20 per haploid genome
(Concepcion et al. 2004; Hitotsumachi et al. 1985). Thus, if
an ENU-induced loss-of-function basepair change is iden-
tified in even a crudely mapped (megabase) inte rval, it is
highly likely to be causing the phenotype of interest. With
t he availability of whole genome sequence and the
decreasing costs of resequencing, most groups have been
able to ident ify causative ENU-induced mutations rapidly
and at reasonable expense (Caspary and Anderson 2006).
Several groups have been applying ENU mutagenesis to
the study of immunity (Papathanasiou and Goodnow 2005).
Bruce Beutler’s group, in particular, has focused on innate
immunity (Beutler et al. 2006). Beutler’s most productive
screen to date has been to identify defects in macrophage
responses to TLR ligands. Several novel genes have been
identified, and many of these affect the response to bacterial
pathogens. As contrasting examples, here, we co nsider two
such mutations, Oblivious (Hoebe et al. 2005) and 3d
(Tabeta et al. 2006). The Oblivious phenotype turned out to
be due to mutations in a previously known gene (Cd36),
whereas the 3d phenotype turned out to be due to mutations
in a previously unstudied gene (Unc93b1). These examples
illustrate the potential of ENU mutagenesis to uncover
“new” genes, as well as to uncover new functions for
previously “known” genes. Although both genes affect
susceptibility to bacterial pathogens, it is interesting that
neither gene was identified in a screen specifically testing
for susceptibility to bacterial pathogens.
Oblivious (Cd36) The Oblivious phenotype was identified
in a mouse strain that exhibited defective responses to
bacterially derived diacylated lipopeptides, such as lip-
oteichoic acid (LTA), a major cell wall constituent of Gram-
positive bacteria. The Oblivious phenotype was found to be
due to a nonsense mutation in the previously identified
scavenger receptor gene, Cd36 (Hoebe et al. 2005). Thus,
CD36 appears to act as an essential coreceptor with TLR2/6
heterodimers in the recognition of diacylated lipopeptides.
Mice homozygous for the Cd36
Obl
allele exhibit increased
sensitivity to infection with Staphylococcus aureus, likely
because of defective signaling though TLR2/6, resulting in
reduced levels of TNFα. The in vivo Obl ivious phenotype
is intermediate between wild-type mice and TLR2 knock
outs, suggesting that CD36 is not absolutely required for
full TLR2/6 respon se to LTA.
3d (Unc93b1) The 3d, or “triple D,” mutation was
identified in a mouse strain defective in signaling through
the intracellular TLRs 3, 7, and 9 (Tabeta et al. 2006). The
phenotype was due to a point mutation in the gene
Unc93b1, encoding a previously uncharacterized trans-
membrane protein located in the endoplasmic reticulum.
Interestingly, 3d mutants also exhibited reduced MHC I and
II presentation of exogenous antigens. Macrophages from
mice homozygous for the 3d mutation demonst rated
reduced production of TNFα and IL-12p40 mRNA in
response to infection with L. monocytogenens. In addition,
3d mice fail to control S. aureus infection (Tabeta et al.
2006). It has been suggested that the Unc-93b protein
directly interacts with the membrane-spanning domain of
TLRs and that this interaction is abolished by the point
mutation causing the 3d phenotype (Brinkmann et al.
2007). The reason that Unc93b1 mutant cells are defective
in antigen presentation remains to be understood.
Bacterial genetic strategies
Although bacteria are general ly though t to be mo re
genetically ap proachable than mice, not all bact erial
pathogens are genetically tracta ble. Several important
bacterial pathogens, e.g., Chlamydia, continue to resist
genetic approac hes. But for many bacteria, a few simple
genetic strategies open whole worlds of possibilities. The
essential toolkit of bacterial genetics includes the ability to
make targeted gene kno ckouts (often termed “allelic
replacement”), the ability to generate random mutations
(usually by use of a transposon), the ability to map or
identify these mutations, and lastly, the ability to comple-
ment these mutations. These essential tools are now
available for many bacterial pathogens. A more recent
addition to the bacterial geneticis t’s toolbox is the avail-
ability of whole genome sequences. More than 400
bacterial genomes have been sequenced, in cluding the
Immunogenetics

genomes of multiple strains for most pathogenic bacterial
species. Whole genome sequences have not only made
traditional genetic approaches more efficient but have also
facilitated new comprehensive approaches to the identifi-
cation of bacterial virulence factors (Burrack and Higgins
2007).
Comprehensive approaches to the identification of bacterial
virulence factors The first whole-genome approaches to the
identification of novel bacterial virulence factors predate
the availability of whole genome sequenc es. Because many
of these techn iques have been re viewed extensively
elsewhere (Chiang et al. 1999), they will be covered only
briefly here. Although each of the following techniques
differs considerably from the others, they share a common
purpose, which is to allow screening of many mutants in a
single host animal. This purpose is borne of necessity, as
screening individual mutants in individual host animals is
time consuming and prohibitively expensive. The draw-
backs of screening multiple mutants in a single host animal
are discussed below.
IVET IVET, or in vivo expression technology, is the name
given to a group of strategies to identify virulence genes
based on the assumption that such genes are likely to be
expressed preferentially in the host, but not in vitro (e.g., on
agar plates). A variety of technical approaches to the
identification of in vivo induced genes have been described,
including brute-force screens (Klarsfeld et al. 1994) and
mo re ele gant genetic or fluorescence-based sele ction s
(Camilli and Mekalanos 1995; Mahan et al. 1993; Mahan
et al. 1995; Osorio et al. 2005; Rietsch et al. 2004; Valdivia
and Falkow 1997). In general, a reporter gene (e.g.,
encoding for antibiotic resistance or GFP) is fused to
random promoters. Bacterial strains carrying individual
promoter–reporter fusions are then tested or selected in
vitro and in vivo for expression of the fusion. Strains
carrying a reporter selectively induced in vivo are selected
for further study. One advantage of IVET is that the fusion
need not disrupt the wild-type copy of the gene, thus
permitting identification of virulence genes that are essen-
tial in vitro. A technical dis advantage is that genes
identified by IVET must generally be confirmed by making
clean deletions, a process that can be fairly laborious and
can be problemat ic if the gene is ess ential. A more
significant technical disadvantage of IVET is that it requires
the establishment of arbitrary cutoffs that determine when a
gene is considered “expressed” or not. Thus, a gene that is
moderately expressed in vitro and highly expressed in vivo
might not be isolated by IVET if a stringent cutoff is
applied in which even genes weakly expressed in vitro are
eliminated from consideration. Modified genetic systems
overcoming some of these problems have been reported,
including systems t o identify in vivo repressed genes
(Hsiao et al. 2006; Osorio et al. 2005).
In general, the experience with IVET has been mixed. One
reason may be that an underlying assumption of IVET—
namely, that virulence genes are preferentially expressed in
vivo—is not likely to be correct for all virulence genes or for
all bacterial species (Rengarajan et al. 2005). For example,
there may be genes important for virulence that are expressed
in vitro. Alternatively, there may be genes that are induced in
vivo that are not required for virulence. The latter category of
genes may include genes that are induced in vivo for purely
spurious reasons, but may also include genes that play
important roles in virulence but which are not strictly
required for growth in vivo because of redundancy or other
reasons. The biological significance of these genes might be
difficult to establish. It might however be considered an
advantage of IVET that it potentially permits identification of
virulence genes that would not be identified by methods
(such as signature-tagged mutagenesis, see below) which
require that the virulence gene confer a selective in vivo
growth defect when mutated. Thus, IVET-based schemes are
a valuable complement to other approaches.
Microarrays It might be thought that, with the increased
availability of microarrays, classical IVET techniques
would be supplanted by microarray analysis. Unlike
IVET-based techniques, microarrays have the advantage of
being able to provide quantitative information about gene
expression levels in vitro and in vivo, allowing identifica-
tion of in vivo induced or repressed genes regardless of
their basal level of expression in vitro (Graham et al. 2006;
Lawson et al. 2006; Revel et al. 2002; Snyder et al. 2004;
Talaat et al. 2004; Xu et al. 2003). While microarrays are
likely to be increasingly used to identify in vivo induced
genes, there are technical limitations of microarrays that
have thus far limited their widespread application. Chief
among these limitations is the need to obtain sufficient
amounts of high-quality bacterial RNA from the infected
host. This problem is particularly difficult to overcome for
infection models in which relatively few bacteria colonize
the host or cause disease, but has been circumvented in
some cases by the use of procedures to specifically amplify
bacterial RNA (Francois et al. 2007; Talaat et al. 2004).
Another potential difficulty for microarray analysis of in
vivo grown bacteria arises when not all bacteria in a host
occupy equivalent niches. For example, in cholera, Vibrio
cholerae bacteria adhering to the intestinal epithelium may be
the relevant disease-causing population, but they may be
relatively outnumbered by bacteria in the intestinal lumen.
Gene expression profiling might not easily detect gene
expression in the relevant disease-causing population. Never-
theless, the comprehensiveness of microarrays is unmatched by
other techniques, and is likely to become more widely applied.
Immunogenetics

Signature-tagged mutagenesis STM, or signature-tagged
mutagenesis, is a strategy to identify genes required for
growth in one condition (e.g., in vivo) as compared to
another (e.g., in vitro; Hensel et al. 1995). Of all bacterial
genetic strategies to screen for virulence factors, STM has
probably been the most widely employed. In STM, pools of
random mutants are generated in which each mutant in the
pool is marked with a specific, rapidly assayable tag. The
pools of mutan ts are usually generated by transposon
mutagenesis. In the original STM scheme (Hensel et al.
1995), the tag was a short oligonucleotide that was present
on the transposon and that could be amplified by PCR and
detected by Southern blotting. If each pool is to contain 96
mutants, then 96 different tran sposons, each with a
uniquely assayable (non-cross hybridizing) tag, need to be
generated before generating the mutant pools. Thus STM
requires considerable upfront effort. Once the pools of
mutants have been generated, how ev er, they can be
relatively easily assayed in multiple different conditions.
For example, the pool of mutants can be passed through in
vivo and in vitro growth conditions. Of particular interest
might be mutants that are able to grow in vitro (e.g., in
broth) but that are unable to replicate in vivo. Several
important virulence factors have been found with this
method (reviewed by Mecsas 2002). The original STM paper
(Hensel et al. 1995), for example, identified the SPI-2 locus
of S. typhimurium, which encodes an important type III
secretion system. An important limitation of STM that has
been recognized is that in vivo competition for niches of
limited size or in vivo “bottlenecks” can sometimes lead to
the spontaneous and random loss of mutants from the pool
(Chiang et al. 1999; Hensel et al. 1995). Thus, the number of
mutants that can be screened by STM per infected animal
may be limited in some infection models.
Transposo n-site hybridization Transposon-site hybridiza-
tion, or TraSH, is an elegant variant of STM that takes
advantage of the availability of microarrays (Badarinarayana
et al. 2001; Sassetti et al. 2001). In TraSH, as in STM, a
library of transposon mutants (the “input” pool) is used to
infect a host. However, instead of tagging each mutant with a
unique “signature tag” or “barcode,” the transposon site itself
serves as a unique tag in TraSH. After a period of growth in
the host, the pool of mutants is recovered (the “output”
pool), and an RNA probe corresponding to the sequences
flanking the transposon insertion sites is transcribed using an
outward-facing T7 promoter built into the transposon. For
comparison, a similar RNA probe is generated from the input
pool. The two probe s are the n labeled with different
fluorescent dyes and competitively hybridized to spotted
microarrays. Genes on the microarray that hybridize to the
input probe but not the output probe are deemed to be genes
in which transposon insertions are detrimental to in vivo
growth. The chief advantages of TraSH over STM are that
many more mutants can potentially be tested in parallel, and
in addition, there is no need to pre-array a library of uniquely
tagged transposon mutants. Conversely, one advantage of
STM is that the mutants of interest are necessarily archived
before the screen and can therefore be retested easily to
confirm their phenotype. In cases where the complexity of
the input pool is limited by the size of the in vivo niche,
STM may be preferred to TraSH. TraSH has been applied by
several labs, working with diverse bacterial genera such as
Bacillus, Francisella, Mycobacterium, and Salmonella (Chan
et al. 2005; Day et al. 2007; Sassetti et al. 2001; Sassetti and
Rubin
2003; Weiss et al. 2007). Ultimately, comprehensive
arrayed libraries of mutants carrying defined loss-of-function
mutations in each nonessential gene may be available for
many bacterial pathogens (Salama and Manoil 2006). Such
libraries require considerable effort in construction and
validation, but will certainly fa cilitate efficient whole-
genome screening.
Transcomplementation: problems and possibilities
STM and TraSH rely on infection of a single host with a
pool of mutants. However, many of the mutants in the pool
will be wild-type, and therefore, the virulence defect of a
particular mutant might be concealed by functions provided
in trans by wild-type bacteria infecting the same host. This
“transcomplementation” might be particularly expected to
prevent isolation of mutants defective in secreted virulence
factors such as toxins, as secreted factors should exhibit
function in trans. Interestingly, howe ver, many secreted
virulence factors have been shown to generate phenotypes
particularly during competitive infections. This seems to be
especially true of type III-secreted effectors (Logsdon and
Mecsas 2006). It has been proposed that one important
benefit of coinfecting wild-type along with mutants is that
the wild-type bacteria will provoke a vigorous immune
re sponse that sharpens th e selection agains t immune-
evasion-defective mutants. Another interesting example of this
phenomenon was reported by Joshi et al. (2006), who found
that M. tuberculosis strains mutated in the Mce1 secretion
system were defective for in vivo growth in competitive
infections, whereas previous work had found that Mce1
mutants were hypervirulent in single infections (Shimono et
al. 2003). It is likely that, in the context of the immune
response triggered by wild-type M. tuberculosis, Mce1 is
required for in vivo growth; but in addition, Mce1 triggers
inflammatory responses that help restrain the infection. Thus,
transcomplementation may be an issue in some screens, but
competitive infections also provide insights that might not
otherwise be observed in infections with single mutants.
Immunogenetics

The “characterization” bottleneck
IVET, STM, TraSH, and other similar approaches have
made possible the comprehensive identification of bacterial
virulence factors. These techniques are still difficult and are
often time consuming, but the fact remains that lists of
putative virulence factors abound, whereas the mechanistic
understanding of these virulence factors lags far behind. In
other words, there is a “characterization” bottleneck. The
bottleneck arises because comprehensive approaches are
designed to identif y genes that affect a function ( for
example, intracellular growth) but shed little light on how
they affect that function. For example, IVET identifies
bacterial genes expressed in vivo but does not tell us which
of the genes expressed are virulence factors. STM and
TraSH identify genes required for growth of a pathogen
under given conditions (e.g., in vivo) but does not often
provide much insight into why a given gene would be
required. The realization that there is a severe characterization
bottleneck first became strikingly evident after the sequencing
of the first bacterial genome, that of Haemophilus influenzae,
revealed that 42% of annotated genes had no known function
(Fleischmann et al. 1995). Twelve years later, the most
current annotation of the Haemophilus genome reveals that
27% of annotated protein-coding genes still have no known
function (http://cmr.tigr.org). In bacteria with larger genomes,
the situation is more dire. For example, ~48% of the protein-
coding genes in the genome of even the highly studied
pathogen, Pseudomonas aeruginosa, are of unknown func-
tion. Although the enzymatic or cell-biological activity of
many virulence factors has been determined, the in vivo
functions of very few virulence factors are actually under-
stood, except in the vaguest terms.
A key question is, therefore, whether there are systematic
ways to ameliorate the characterization bottleneck? Or are the
days of comprehensive strategies waning in favor of the tried-
and-true approach of conducting detailed studies of individual
genes? One reason why many investigators have been attracted
to whole genome approaches is that it has seemed risky to
focus all of one’s efforts on characterization of a single putative
virulence factor, when there is the possibility (or likelihood!) it
will turn out to be of little significance or interest. Compre-
hensive ap proaches also afford the opportunity t o put
individual mutants in context. For example, it can be asked
whether mutations in other genes in the same genetic pathway
are also obtained, and if not, why not? It can also be asked
whether similar screens in other bacterial species identify
similar or unique virulence factors—and of course, either
category might be of interest. Part of the value of comprehen-
sive approaches is therefore likely to be that they will provide
guidance on which of the many uncharacterized virulence
factors are most important to focus individual efforts on.
Camilli and Merrell proposed that one solution to the
characterization bottleneck is simply to perform secondary
screens (Merrell and Camilli 2002; Merrell et al. 2002). The
authors screened 9,600 V. cholerae mutants by STM in an
infant mouse mo del, a nd obtained 164 colonization-
defective mutants. These mutants were then reassembled
into “virulence-attenuated pools” and rescreened in vitro for
defects in acid tolerance (Merrell et al. 2002). Several of the
STM-identified mutants also exhibited defects in acid
tolerance, thus providing clues as to why they might have
scored as attenuated i n the initial STM screen. Acid
tolerance was previously proposed to be an important
virulence trait of V. cholerae. In cases where an educated
guess about the nature of the relev ant in vivo selective
pressure cannot be made (or tested for), additional in vitro
screening may not be particularly useful to alleviate the
characterization bottleneck.
The genetic structure of host–pathogen interactions:
the awesome power of genetics-squared?
A potentially powerful idea is the following: for host –
pathogen interactions in which gen etics of the host and
pathogen are both possible, could a combination of host
and pathogen genetics provide new insights into the host–
pathogen relationship? We call this idea “genetics-squared.”
There are several example s in the literature in which
investigators have combined host and pathogen genetics
to ad dress the in vivo function of a bacterial virulence
factor, but the assumptions and genetic structure underlying
such strategi es have not always been explicitly addressed.
Therefore, we begin our discussion by systematically
describing the logical structure of combined host–pathogen
genetic strategies.
There are several distinct modes in which host and
pathogen genomes can interact (Table 1). In one scenario
(Table 1, Scenario A), a bacterial virulence factor is
specifically required to counteract a particular host defense.
We call this the “Imm une Evasion” mode of host–pathogen
interactions. In this case, mutant 1, defective in the
virulence factor, is unable to counteract a particular host
de fense (host defense A). Thus, m utant 1 would be
attenuated in wild-type hosts, but would be restored to
normal virulence in a host animal deficient in this defense
(Knockout A). The interpretation of such experiments is
complicated by the possibility that Knockout A might be so
severely deficient in host defense that it would be
permissive to virtually any bacterial mutant; therefore, this
experimental strategy ideally requires two importan t con-
trols. First, a second, equally attenuated bacterial mutant
(m utant 2) should not have its virulence restored in
Immunogenetics

Knockout A, and second, mutant 1 should not be restored
to norma l virulence in a host deficient in a secon d,
unrelated aspect of host defense (Knockout B). As
discussed below, such evidence is often difficult to obtain,
or the effects of knockouts are either incomplete or
pleiotropic.
A second scenar io (Table 1, Scenario B) occurs in the
case when a bacterially encoded factor triggers a host
defense that limits pathogen virulence or replication. We
call this the “Immune Triggering” mode of host–pathogen
interactions. In the most simplisti c version of this scenario,
a bacterial mutant (e.g., mutant 1 in Table 1, Scenario B)
lacking an immunologically sensed factor will evade
detection and exhibit increased virulence as compared to
wild-type bacteria. Conversely, host knockout A, unable to
detect wild-type bacteria, should exhibit increased suscep-
tibility to wild-type bacteria, but importantly, should not
exhibit increased susceptibility to the mutant 1. A compli-
cation arises when a particular bacterial mutant not only
evades host detection but is defective in virulence (e.g.,
mutant 2, Table 1, Scenario B). This can occur when the
host senses a bacterial product required for virulence (e.g.,
LPS). Mutant 2 may evade immune detection but may not
exhibit increased virulence in wild-type mice. In addition,
the genetic experiments in Scenario B will not be able to
determine readily whether Knockout A is defective in
detection of bacteria, or in the downstream antimicrobial
response.
The “B” scenario in Table 1 is essentially a restatement
of the classic “Gene-for-Gene” resistance scenario that has
been described to occur frequently in the interaction of
pathogens with plant hosts (Chisholm et al. 2006 ; Jones and
Dangl 2006). In the terminology of the Gene-for-Gene
model in plants, the host factor that senses the pathogen is
called a resistance (R) gene, and the particular bacterial
product sensed by the R protein i s encoded by an
“avirulence” (avr) gene. Although Gene -for-Gene resis-
tance is a common mode of host–pathogen interactions in
plants, it appears to be somewhat less common in animals,
although some specific examples are discussed below.
A third scenario occurs when a pathogen requires a host
function for growth or virulence (Table 1, Scenario C). We
call this mode of host–pathogen interaction “parasitism.”
For example, the host might provide a particular nutrient
required by the pathogen, or, the host might provide a
particular cellular niche that supports pathogen spread or
virulence. A host knockout defective in the production of
this nutrient or niche would exhibit increased resistance to
the pathogen, and a pathogen mutant (e.g., mutant 1,
Table 1, Scenario C) unable to take advantage of the niche
would exhibit decreased virulence. To demonstrate that
mutant 1 is specifically defective in taking advantage of the
niche/nutrient provided by host function A, an important
control is to show that mutant 1 is not more attenuated than
wild-type in knockout A, but is more a ttenuated than wild-
type bacteria in wild-type hosts or in hosts defective in an
unrelated function (e.g., knockout B).
In all three scenarios, interpretation of the genetic
relationship between bacterial and host genes is greatly
facilitated by the presence of multiple bacterial and host
mutants and by the demonstration that interactions between
the host and bacterial genomes are specific.
Table 1 Genetics of host–pathogen interactions
Bacterial
genotype
Growth or virulence of bacteria in
host of genotype
Notes
Wild
type
Knockout
A
Knockout
B
Scenario A: Immune Evasion: bacterial virulence factor required to counteract specific host defenses
Wild-type + ++ ++
Mutant 1 − + − Mutant 1 is unable to counteract host defense A
Mutant 2 − − + Mutant 2 is unable to counteract host defense B
Scenario B: Immune Triggering: bacterial virulence factor triggers a host defense (Gene-for-Gene resistance model)
Wild-type − + +
Mutant 1 + + ++ Mutant 1 is not sensed by host sensor A
Mutant 2 − − + Mutant 2 is not sensed by host sensor A, but mutant 2 is also attenuated for virulence
Mutant 3 + ++ + Mutant 3 is not sensed by host sensor B
Scenario C: Parasitism: bacterial virulence factor requires host factor for function
Wild-type ++ + + Knockouts A and B are deficient in separate host functions required for bacterial
growth/virulence
Mutant 1 + + − Mutant 1 is unable to take advantage of host function A
Mutant 2 + − + Mutant 2 is unable to take advantage of host function B
Immunogenetics

Examples of host–pathogen genetic interactions
It is not possible to review here all the instances in which host
and pathogen genetics have been combined in a “genetics-
sq uared” approach. We have th erefore selected a few
illustrative examples of each of the scenarios identified above.
Scenario A: “Immune Evasion” In mammals, the most
commonly described form of host–pathogen genetic inter-
action analysis is Scenario A (Table 1), in which a particular
virulence factor is shown to be required to evade a particular
host defense. One early example (Harvill et al. 1999) studied
the in vivo function of Bordetella bronchiseptica adenylate
cyclase toxin (CyaA). It was found that CyaA-deficient
mutants were attenuated in growth/colonization of the
trachea and lungs of mice. The attenuation of CyaA mutants
was reversed in infections of neutropenic (cyclophosphamide-
treated, or Gcsf
−/−
) mice. CyaA mutants were still attenuated
as compared to wild-type in infections of lymphocyte-
deficient (Scid-beige or Rag1
−/−
) animals. Thus, it was
inferred that a specific function of CyaA is to counteract
immune defense provided by neutrophils. Because neutro-
penic mice are in general extremely susceptible to bacterial
infections, an important control was the demonstration that
neutropenia did not reverse the attenuation of ΔbvgS mutants
of B. bronchiseptica, which are deficient in an unrelated
virulence pathway.
Another example of the potential power of combining
host and pathogen genetics concerns a series of experiments
suggesting links between virulence factors of M. tubercu-
losis (MTB) and host production of antimicrobial reactive
nitrogen and reactive oxygen. These studies took advantage
of mice deficient in the production of reactive nitrogen
(Nos2
−/−
) or reactive oxygen (gp91
phox−/−
). In one study
(Ng et al. 2004), katG MTB mutants were found to be
attenuated in B6 and Nos2
−/−
mice. KatG is a catalase–
peroxidase–peroxynitritase, so it was suspected that perhaps
it was playing a critical role in detoxifying peroxides
produced by oxidative burst. This idea was supported by
the finding that katG mutants exhibited full virulence in
gp91
phox−/−
mice and in gp91
phox−/−
Nos2
−/−
double knock-
outs. The work of Ng et al. is particularly striking because it
was reported that gp91
phox−/−
mice do not exhibit increased
susceptibility to wild-type MTB, thus allaying any worries
that the rescue of the katG mutants by host mutations in
gp91
Phox
was due to a nonspecific weakening of the
immune response. A similar reversal of attenuation was
also demonstrated for infections of superoxide-susceptible
mutants of Salmonella (van Diepen et al. 2002; Vazquez-
Torres et al. 2000) or Aspergillus (Chang et al. 1998) in
Phox
−/−
mice. In addition, a second study on M. tubercu-
losis (Darwin et al. 2003), complementary to the work of
Ng et al., found bacterial mutants that were particularly
susceptible to reactive nitrogen. Three such mutants were
deficient in Rv2115c, an ATPase compo nent of the bacterial
proteasome. However, unlike the katG mutants, whose
phenotype was fully reversed in gp91
Phox−/−
mice, the
Rv2115c mutants were only partially restored to virulence
in Nos2
−/−
mice. It seems likely that proteasome mutations
are pleiotropic, illus trating the key point that genetic
interaction studies of host–pathogen relationships can be
complicated by the lack of a simple one-to-one relationship
between bacterial virulence factors and host resistance
genes. Nevertheless, taken together, the studies on MT B
can be compiled quite nicely into a matrix of the form
shown in Table 1 (scenario A), in which Mutant 1 is a katG
mutant, Mutant 2 is proteasome mutant, Knockout A is a
gp91
Phox−/−
mouse, and Knockout B is a Nos2
−/−
mouse.
The in vivo genetic data strongly supports the conclusion
that KatG specifically counteracts reactive oxygen and that
the proteasome counteracts reactive nitrogen (in addition to
other in vivo stresses).
Another beautiful example of a “scenario A”-type host–
pathogen interaction was reported by Hsiao et al. (2006 ) in
their studies of V. cholerae. In this case, however, evasion of
host defenses was not mediated by expression of a virulence
factor, but was instead mediated by transcriptional repres-
sion of a bacterial type IV pilus, the mannose-se nsitive
hemagglutinin (MSHA) pilus. The authors demonstrated
that the MSHA pilus is normally targete d by host
immunoglo bulins (IgA), leading to agglutination that
prevented V. cholerae from penetrating intestinal mucus.
Enforced expression of the MSHA pilus reduced the ability of
V. cholerae to colonize infant suckling mice by 3 logs, but
had no adverse effects on colonization of suckling IgA
−/−
mice or on colonization of wild-type mice that were not
allowed to suckle. Thus, genetic and behavioral manipulation
of the host permitted the authors to identify a specific function
for MSHA repression in evasion of milk-derived IgA.
Another eleg ant example of the “Immune Evasion”
mode of host–pathogen interaction, conceptually identical
to the MSHA-IgA example, has been presented by
Montminy et al. (2006) in their studies of TLR4 recognition
of Yersinia pestis, the causative agent of plague. It was
previously known that the tetra-acetylated LPS of Y. pestis
was not well recognized by TLR4. To illuminate the
biologic al significance of this observation, the a uthors
engineered a strain of Y. pestis that constitutively produced
a highly stimulatory hexa-acetylated LPS. The modified Y.
pestis strain was at least 1,000-fold less virulent in a mouse
model of plague, despite expression of all other known
virulence factors. Although Y. pestis normally produces a
hexa-acetylated LPS during its life cycle in the flea, it was
important to demonstrate that the virulence defect of the
con stitutive-hexa-acetylated strain was due to immune
recognition rather than to a nonspecific disruption of Y.
Immunogenetics

pestis metabolism or virulence factor function. This is
where the combined application of host and bacterial
genetics proved valuable: Tlr4
−/−
mice were shown to be
highly susceptible to the constitutive-hexa-acetylated strain,
thus proving that the virulence defect of the strain in wild-
type mice was due to TLR4 recognition.
Genetic interaction studies may also be useful for deter-
mining the in vivo functions of bacterial effectors (toxins)
secreted into the cytosol of host cells via specialized type III
secretion systems. Although type III secretion systems are
clearly critical for the virulence of numerous bacterial
pathogens, and although the in vitro activities of numerous
type III-secreted effectors have been described in detail, it has
been significantly more challenging to ascribe in vivo roles to
such effectors. This may be in part because type III secreted
effectors tend to have broad effects on host cells—for
example, disruption of actin polymerization or of MAP kinase
signaling. One group has utilized knockout mice to illuminate
the in vivo function of the type III secreted effectors (Yops) of
Y. pseudotuberculosis (Logsdon and Mecsas 2003; Logsdon
and Mecsas 2006). In vitro work had previously established
the biochemical activities of YopH, a tyrosine phosphatase,
and YopE, a Rac-GTPase-activating protein, and suggested
their function may be to counteract phagocytosis (Viboud
and Bliska 2005). Whether phagocytosis is the relevant in
vivo target of YopH/E was less clear. Interestingly, in vivo
colonization experiments demonstrated that IFN-γ deficien-
cy could largely or entirely reverse the colonization defects
of YopE (but not YopH) mutants previously observed in
wild-type mice (Logsdon and Mecsas 2006). Thus, YopE
appears to function in evading IFN-γ-mediated host
defenses. A limitation of these kinds of studies is that IFN-
γ deficiency results in multiple immunological defects, and
therefore, despite the important contribution of these studies,
the “mechanism” of YopE and YopH function in vivo
remains somewhat unclear. Presumably, experiments with
other knockout mice will help refine the in vivo functions of
these and other secreted effectors. On the other hand, it may
well be that some virulence factors exhibit broad functions
that will not “map” well onto specif ic host immune
functions.
Scenario B: “Immune Triggering” The mode of host–
pathogen interaction in which a bacterial virulence factor
triggers a specific host response is commonly uncovered in
plants, but appears to be considerably more rare in bacterial
interactions with mammalian hosts. One recent example
from mice pertains to the Naip5 locus, discussed above,
that medi ates resistance of C57BL/6 macrophages to L.
pneumophil a. Naip5 was found to e ncode a protein
containing nucleotide-bind ing domain (NBD) and leucine-
rich repeat (LRR) motifs. Similar NBD–LRR proteins were
found in plants and animals to be involved in sensing
molecular structure s associated with a variety of pathogens.
To identify what Naip5 might sense, a genetic selection was
performed to identi fy Legionella mutants that evaded
Naip5-mediated resistance. Twenty-nine indep endent
mutants selected were all defective in expression of
bacterial flagellin (Ren et al. 2006). Targeted deletion of
flagellin, but not of other genes required for assembly or
function of the flagellum, was also found to permit evasion
of Naip5-mediated defenses (Molofsky et al. 2006; Ren et
al. 2006). It was therefore proposed that flagellin itself
might be sensed by Naip5. For reasons outlined above, an
important control experiment demonstrated that flagellin
mutants did not replicate better than wild-type Legionella
in
Naip5-defective (A/J) macrophages. This result suggested
that deficiency in flagellin permits specific evasion of the
Naip5 pathway, rather than of some other, unrelated host
defense. However, flagellin mutants exhibited greater
replication in TNF-deficient macrophages than in Naip5-
de fecti ve (A/J) macrophages. This result implies that
flagellin deficiency synergizes with host mutations in TNF
signaling pathways, and suggests that TNF and flagellin-
sensing are separable components of host defense (Coers et
al. 2007). The above results fit the logical structure outlined
in Table 1 (Scenario B), where bacterial mutant 1 is a
flagellin mutant, knockout A is a Naip5 knockout, and
knockout B is a TNF knockout. The genetic screen that
identified flagellin as a target of the Naip5 pathway would
never had worked if it had turned out that flagellin was
required for viru lence of Legionella in macrophag es
(flagellin and motility, while not essential for virulence in
macrophages, is still likely to play an essential role for
Legionella survival in the environment). In fact, because
vertebrate immune systems tend to sense essential and
conserved targets in bacteria (Janeway 1989), the Naip5-
flagellin story may be fairly unique .
Further evidence strongly suggests that s ensin g of
Legionella flagellin occurs in the cytosol and is indepe n-
dent of the cell-surface flagellin sensor TLR5 (Franchi et al.
2006; Miao et al. 2006; Molofsky et al. 2006; Ren et al.
2006). However, it is im portant to note that genetic
experiments have not established a definitive connection
between Naip5-mediated host defense and flagellin-sensing.
Indeed, there are data that another Naip5-related protein,
Ipaf, is required for flagellin-sensing by host cells (Amer et
al. 2006; Franchi et al. 2006; Lamkanfi et al. 2007; Miao et
al. 2006; Zamboni et al. 2006). Future work will need to
focus on establishing the biochemical basis of flagellin-
sensing by host cells. Thus, the Naip5-flagell in story
illustrates the potential power—and the limitations—of
genetics in dissection of host–pathogen interactions.
A pa rtic ular ly interesting example of the “Immune
Triggering” mode of host–pathogen interactions was pro-
vided by Balachandran et al. (2007). The authors studied
Immunogenetics

the phenotype of P. aeruginosa strains lacking the type
III-secreted toxin ExoT. They found that mice lacking the
ubiquitin ligase Cbl-b were particularly suscept ible to
ExoT
+
but not ExoT-deficient P. aeruginosa. The combi-
nation of host and bacterial genetics to demonstrate the
specificity of the immunodeficiency of Cbl-b
−/−
mice is
one feature that makes this story particularly compelling.
Another compelling feature is that the genetics in the paper
were reinforced with bioche mical evidence that Cbl-b
associates with ExoT, resulting in Cbl-b-dependent degra-
dation of ExoT. Specific host recognition and ubiquitin-
mediated degradation of a bacterial toxin may turn out to be
a fundamental and novel form of immune defense. Unlike
the case of flagellin, where the host senses a bacterial
protein that is not required for bacterial virulence, Cbl-b
detects and specifically eliminates a virulence factor. Thus,
the role of ExoT is not par ticula rly apparent in wild-type
hosts, and can only be revealed in Cbl- b
−/−
mice. In the
scheme por trayed in Table 1 (Scenario B), ExoT can be
represente d by Mutant 2, and Cbl-b
−/−
can be represented
by Knockout A.
Why are there relatively few examples in which genetic
approaches have been able to dissect the “Immune Trigger-
ing” mode of host–bacterial interactions? There are certainly
no shortage of bacterial products that trigger host responses,
e.g., LPS, lipopeptides, flagellin, muramyl dipeptide, CpG
DNA, and various bacterial toxins. In fact, the multiple
redundant pathways for recognition of bacteria might be one
reason that individual mutants (of the host or pathogen) rarely
provide a phenotype. Another factor is that— with the notable
exception of flagellin—many of the bacterial products that
trigger host defenses are essential for bacterial viability. Thus,
it would not be expected to obtain bacterial mutants defective
in production of these products. Thus, the “Immune Trigger-
ing” mode of host–pathogen interactions is not in fact rare; but
useful genetic systems, in which mutants can be generated and
provide phenotypes, seem to be the exception rather than the
rule.
Scenario C: parasiti sm The last mode of genetic interac-
tion that we consider here is the case in which bacteria
require a particular host gene for virulence functions.
Unlike the fi rst two scenarios, in this scenario, ho st
knockouts exhibit increased resistanc e to the pathog en
(Table 1, Scenario C). Several such instances have been
reported (for example, Auerbuch et al. 2004; Carrero et al.
2004; O’Connell et al. 2004; Saleh et al. 2006; Vazquez-
Torres et al. 1999), but it appears that this mode of
interaction is relatively rare. There are undoubtedly many
host functions that are co-opted by bacteria (Portnoy 2005),
but these host functions may be essential for the host as
well, and thus, host knockouts affecting these functions
may not be viable. Neve rtheless, further exploration of
“parasitic”-type host–pathogen interactions will likely be an
area of opportunity for the future.
Squaring genetic screens
The above discussion illustrates the potential power, and the
limitations, of “genetics-squared” in the analysis of host–
pathogen interactions. However, the above examples all
examine genetic systems in which a single known host gene
or bacterial gene are tested for phenotypic interactions. It is
interesting to consider whether combined host–pathogen
genetic approaches could also be applied to high throughput
screens. A few such screens have been performed. For
example, Hisert et al. (2005) performed “differential” STM
to identify bacterial mutants of Salmonella that were
specifical ly a ttenuated i n wild-type mice, but whose
attenuation was reversed in gp91
phox−/−
mice. One such
mutant was identified, harboring a transposon insertion in
cdgR, a regulator of cyclic-di-GMP signaling. The attenu-
ation of the cdgR mutant was not reversed in iNOS
−/−
mice.
Thus, it appears that the cdgR is relatively specifically
required for defense against reactive oxygen as opposed to
reactive nitrogen. Interestingly, cdgR did not appear to
regulate other genes known to provide defense against
reactive oxygen (e.g., catalase or hydroperoxide reductase),
and so the mechanism of action remains to be understood.
Hisert et al. (2004) also applied differential STM to identify
M. tuberculosis “counter-immune” genes specifically in-
volved in resistance to IFN-γ-mediated host defenses.
“Genetics-squared” has also been applied in screens
aimed at identifying genes required for bacterial evasion of
innate immune defenses in the lung. In two separate studies,
Zhang et al. ( 2005, 2007) passaged a signature-tagged
library of P. aeruginosa mutants through wild-type mice as
well as through mice deficient in lung surfactant protein-A
(SP-A). SP-A de ficient mice exhibit increased susceptibility
to P. aeruginosa; thus, the goal was to identify bacterial
mutants that are specifically attenuated in wild-type but not
in SP-A
−/−
mice. In the first study, two such mutants were
reported. One mutant harbored a transposon insertion in
pch, a gene required for salicylate biosynthesis, and the
other mutant was defective in ptsP, encoding phosphoenol-
pyruvate-protein-phosphotransferase (Zhang et al. 2005). In
the second study, flgE (flagellar hook) mutants were also
found to be specifically attenuated in SP-A
−/−
mice (Zhang
et al. 2007). All three mutants exhibited increased sensitiv-
ity to SP-A-med iate d permeabilization, thus providing
valuable insights into the mechanisms by which bacteria
evade innate immune defenses in the lung.
TraSH has also been applied in a differential screening
strategy (Rengarajan et al. 2005). In this study, the authors
identified genes required for growth of MTB in unstimulated
macrophages as compared to macrophages activated by IFN-
Immunogenetics

γ pre- or postinfection. Transposon insertions in one gene,
glnB, appeared to be particularly underrepresented in bacteria
obtained from the IFN-γ pre-infection condition, suggesting
that glnB might be important for resisting IFN-γ-mediated
host defense. However, the effect was not large, and the
result was not validated by an independent method. What is
perhaps more striking was that TraSH did not identify more
bacterial genes specifically required for replication in IFN-γ-
treated (as opposed to untreated) macrophages. It may be that
—because of redundancy or essentiality—M. tuberculosis
simply does not encode many genes that can be mutated to
confer increased susceptibility to IFN-γ. In addition,
bacterial genes required for resistance to the effects of IFN-
γ may also be important for resistance to other stresses
encountered in macrophages in the absence of IFN-γ; such
genes would not be uncovered in a differential screen of
IFN-γ -treated vs untreated macropha ges. As discussed
above, the lack of a clear one-to-one correspondence
between host defense genes and bacterial immune evasion
genes is one reason why “genetics-squared” strategies may
sometimes not generate the desired results. The TraSH study
also did not reveal the genes previously suggested by others
to be required for evasion of IFN-γ in vivo (Darwin et al.
2003; Hisert et al. 2004). This may reflect technical differ-
ences between the screens, or possibly, it may be that in vitro
vs in vivo exposure to IFN-γ is distinct in important ways.
Conclusion
The power of forward genetics is derived in part from its
ability to generate novel insight into complex biological
systems in a relatively unbiased manner. Sign ificant
advances in understanding host–pathogen interactions have
been discerned by the use of host and pathogen genetic
systems. However, the interactions between pathogens and
their hosts are highly complex and can be most successfully
understood only by the application of multiple experimental
strategies. “Genetics-squared” is an increasingly applied
approach, in which host and pathogen genetics are
combined in a single experimental system. When success-
ful, genetics-squared can provide novel insights into the
relationships between host immune genes and pathogen
virulence genes that would be difficult to discern by other
methods. As with conventional genetic approaches, genetics-
squared can be limited by genetic redundancy and pleiotropy.
Nevertheless, as technical advances continue to increase the
power of host and pathogen genetics systems, it is likely that
the power of genetics-squared will also increase accordingly.
Acknowledgment The authors would like to acknowledge stimulat-
ing discussions with Greg Barton, Zeke Bernstein-Hanley, Victor
Boyartchuk, Sky Brubaker, Jörn Coers, Bill Dietrich, Joan Mecsas,
Jeff Miller, Jeff Murry, Carl Nathan, Dan Portnoy, and Jun Zhu.
References
Amer AO, Swanson MS (2005) Autophagy is an immediate macrophage
response to Legionella pneumophila. Cell Microbiol 7:765–78
Amer A, Franchi L, Kanneganti TD, Body-Malapel M, Ozoren N, Brady
G, Meshinchi S, Jagirdar R, Gewirtz A, Akira S, Nunez G (2006)
Regulation of Legionella phagosome maturation and infection
through flagellin and host IPAF. J Biol Chem 281:35217–35223
Auerbuch V, Brockstedt DG, Meyer-Morse N, O’Riordan M, Portnoy
DA (2004) Mice lacking the type I interferon receptor are
resistant to Listeria monocytogenes. J Exp Med 200:527–533
Badarinarayana V, Estep PW 3rd, Shendure J, Edwards J, Tavazoie S,
Lam F, Church GM (2001) Selection analyses of insertional mutants
using subgenic-resolution arrays. Nat Biotechnol 19:1060–1065
Balachandran P, Dragone L, Garrity-Ryan L, Lemus A, Weiss A, Engel
J (2007) The ubiquitin ligase Cbl-b limits Pseudomonas aerugi-
nosa exotoxin T-mediated virulence. J Clin Invest 117:419–427
Bancroft GJ, Kelly JP (1994) Macrophage activation and innate
resistance to infection in SCID mice. Immunobiology 191:424–431
Barsig J, Kaufmann SH (1997) The mechanism of cell death in
Listeria monocytogenes-infected murine macrophages is distinct
from apoptosis. Infect Immun 65:4075–4081
Beckers MC, Yoshida S, Morgan K, Skamene E, Gros P (1995)
Natural resistance to infection with Legionella pneumophila:
chromosomal localization of the Lgn1 susceptibility gene.
Mamm Genome 6:540–545
Bernstein-Hanley I, Balsara ZR, Ulmer W, Coers J, Starnbach MN,
Di etr ich WF (2006a) Genetic analysis of susceptibility to
Chlamydia trachomatis in mouse. Genes Immun 7:122–129
Bernstein-Hanley I, Coers J, Balsara ZR, Taylor GA, Starnbach MN,
Dietrich WF (2006b) The p47 GTPases Igtp and Irgb10 map to
the Chlamydia trachomatis sus ceptibility locus Ctrq-3 and
mediate cellular resistance in mice. Proc Natl Acad Sci USA
103:14092–14097
Beutler B, Jiang Z, Georgel P, Crozat K, Croker B, Rutschmann S, Du X,
Hoebe K (2006) Genetic analysis of host resistance: Toll-like receptor
signaling and immunity at large. Annu Rev Immunol 24:353–389
Boyartchuk VL, Broman KW, Mosher RE, D’Orazio SE, Starnbach
MN, Dietrich WF (2001) Multigenic control of Listeria mono-
cytogenes susceptibility in mice. Nat Genet 27:259–260
Boyartchuk V, Rojas M, Yan BS, Jobe O, Hurt N, Dorfman DM,
Higgins DE, Dietrich WF, Kramnik I (2004) The host resistance
locus sst1 controls innate immunity to Listeria monocytogenes
infection in immunodeficient mice. J Immunol 173:5112–5120
Boyden ED, Dietrich WF (2006) Nalp1b controls mouse macrophage
susceptibility to anthrax lethal toxin. Nat Genet 38:240–244
Bradley DJ (1974) Letter: genetic control of natural resistance to
Leishmania donovani. Nature 250:353–354
Bradley DJ, Kirkley J (1977) Regulation of Leishmania populations
within the host. I. the variable course of Leishmania donovani
infections in mice. Clin Exp Immunol 30:119
–129
Brinkmann MM, Spooner E, Hoebe K, Beutler B, Ploegh HL, Kim
YM (2007) The interaction between the ER membrane protein
UNC93B and TLR3, 7, and 9 is crucial for TLR signaling. J Cell
Biol 177:265–275
Brown NF, Wickham ME, Coombes BK, Finlay BB (2006) Crossing the
line: selection and evolution of virulence traits. PLoS Pathog 2:e42
Buer J, Balling R (2003) Mice, microbes and models of infection. Nat
Rev Genet 4:195–205
Burrack LS, Higgins DE (2007) Genomic approaches to understand-
ing bacterial virulence. Curr Opin Microbiol 10:4–9
Immunogenetics

Camilli A, Mekalanos JJ (1995) Use of recombinase gene fusions to
identify Vibrio cholerae genes induced during infection. Mol
Microbiol 18:671 –683
Caron J, Lariviere L, Nacache M, Tam M, Stevenson MM, McKerly
C, Gros P, Malo D (2006) Influence of Slc11a1 on the outcome
of Salmonella enterica serovar Enteritidis infection in mice is
associated with Th polarization. Infect Immun 74:2787–2802
Carrero JA, Calderon B, Unanue ER (2004) Type I interferon
sensitizes lymphocytes to apoptosis and reduces resistance to
Listeria infection. J Exp Med 200:535–540
Caspary T, Anderson KV (2006) Uncovering the uncharacterized and
unexpected: unbiased phenotype-driven screens in the mouse.
Dev Dyn 235:2412–2423
Chan K, Kim CC, Falkow S (2005) Microarray-based detection of
Salmonella enterica serovar Typhimurium transposon mutants
that cannot survive in macrophages and mice. Infect Immun
73:5438–5449
Chang YC, Segal BH, Holland SM, Miller GF, Kwon-Chung KJ
(1998) Virulence of catalase-deficient Aspergillus nidulans in
p47(phox)−/− mice. Implications for fungal pathogenicity and
host defense in chronic granulomatous disease. J Clin Invest
101:1843–1850
Chiang SL, Mekalanos JJ, Holden DW (1999) In vivo genetic analysis
of bacterial virulence. Annu Rev Microbiol 53:129–154
Chisholm ST, Coaker G, Day B, Staskawicz BJ (2006) Host-microbe
interactions: shaping the evolution of the plant immune response.
Cell 124:803–814
Coers J, Vance RE, Fontana MF, Dietrich WF (2007) Restriction of
Legionella pneumophila growth in macrophages requires the
concerted action of cytokine and Naip5/Ipaf signalling pathways.
Cell Microbiol 9(10):2315–2541
Concepcion D, Seburn KL, Wen G, Frankel WN, Hamilton BA (2004)
Mutation rate and predicted phenotypic target sizes in ethyl-
nitrosourea-treated mice. Genetics 168:953–959
Cordoba-Rodriguez R, Fang H, Lankford CS, Frucht DM (2004)
Anthrax lethal toxin rapidly activates caspase-1/ICE and induces
extracellular release of interleukin (IL)-1beta and IL-18. J Biol
Chem 279:20563–20566
Coutinho A, Meo T (1978) Genetic basis for unresponsiveness to
lipopolysaccharide in C57BL/10Cr mice. Immunogenetics 7:17–24
Cuellar-Mata P, Jabado N, Liu J, Furuya W, Finlay BB, Gros P,
Grinstein S (2002) Nramp1 modifies the fusion of Salmonella
typhimurium-containing vacuoles with cellular endomembranes
in macrophages. J Biol Chem 277:2258–2265
Darwin KH, Ehrt S, Gutierrez-Ramos JC, Weich N, Nathan CF (2003)
The proteasome of Mycobacterium tuberculosis is required for
resistance to nitric oxide. Science 302:1963–1966
Day WA Jr, Rasmussen SL, Carpenter BM, Peterson SN, Friedlander
AM (2007) Microarray analysis of transposon insertion muta-
tions in Bacillus anthracis: global identification of genes required
for sporulation and germination. J Bacteriol 189:3296–3301
Dietrich WF, Damron DM, Isberg RR, Lander ES, Swanson MS (1995)
Lgn1, a gene that determines susceptibility to Legionella pneumo-
phila, maps to mouse chromosome 13. Genomics 26:443–450
Diez E, Lee SH, Gauthier S, Yaraghi Z, Tremblay M, Vidal S, Gros P
(2003) Birc1e is the gene within the Lgn1 locus associated with
resistance to Legionella pneumophila
. Nat Genet 33:55–60
Dunn PL, North RJ (1991) Early gamma interferon production by
natural killer cells is important in defense against murine
listeriosis. Infect Immun 59:2892–2900
Fleischmann RD, Adams MD, White O, Clayton RA, Kirkness EF,
Kerlavage AR, Bult CJ, Tomb JF, Dougherty BA, Merrick JM et
al (1995) Whole-genome random sequencing and assembly of
Haemophilus influenzae Rd. Science 269:496–512
Fleming MD, Trenor CC 3rd, Su MA, Foernzler D, Beier DR,
Dietrich WF, Andrews NC (1997) Microcytic anaemia mice have
a mutation in Nramp2, a candidate iron transporter gene. Nat
Genet 16:383–386
Forbes JR, Gros P (2003) Iron, manganese, and cobalt transport by
Nramp1 (Slc11a1) and Nramp2 (Slc11a2) expressed at the
plasma membrane. Blood 102:1884–1892
Fortier A, Min-Oo G, Forbes J, Lam-Yuk-Tseung S, Gros P (2005)
Single gene effects in mouse models of host: pathogen
interactions. J Leukoc Biol 77:868–877
Fortier A, de Chastellier C, Balor S, Gros P (2007) Birc1e/Naip5
rapidly antagonizes modulation of phagosome maturation by
Legionella pneumophila. Cell Microbiol 9:910–923
Franchi L, Amer A, Body-Malapel M, Kanneganti TD, Ozoren N,
Jagirdar R, Inohara N, Vandenabeele P, Bertin J, Coyle A, Grant
EP, Nunez G (2006) Cytosolic flagellin requires Ipaf for
activation of caspase-1 and interleukin 1beta in salmonella-
infected macrophages. Nat Immunol 7:576–582
Francois P, Garzoni C, Bento M, Schrenzel J (2007) Comparison of
amplification methods for transcriptomic analyses of low abundance
prokaryotic RNA sources. J Microbiol Methods 68:385–391
Friedlander AM, Bhatnagar R, Leppla SH, Johnson L, Singh Y (1993)
Characterization of macrophage sensitivity and resistance to
anthrax lethal toxin. Infect Immun 61:245–252
Garifulin O, Qi Z, Shen H, Patnala S, Green MR, Boyartchuk V
(2 007) Irf3 polymorphism alters induction of IFNbeta; in
response to L. monocytogenes infection. PLoS Genetics (in press)
Graham MR, Virtaneva K, Porcella SF, Gardner DJ, Long RD, Welty
DM, Barry WT, Johnson CA, Parkins LD, Wright FA, Musser
JM (2006) Analysis of the transcriptome of group A Streptococ-
cus in mouse soft tissue infection. Am J Pathol 169:927–942
Gros P, Skamene E, Forget A (1981) Genetic control of natural
resistance to Mycobacterium bovis (BCG) in mice. J Immunol
127:2417–2421
Growney JD, Dietrich WF (2000) High-resolution genetic and
physical map of the Lgn1 interval in C57BL/6J implicates Naip2
or Naip5 in Legionella pneumophila pathogenesis. Genome Res
10:1158–1171
Growney JD, Scharf JM, Kunkel LM, Dietrich WF (2000) Evolution-
ary divergence of the mouse and human Lgn1/SMA repeat
structures. Genomics 64:62–81
Gunshin H, Mackenzie B, Berger UV, Gunshin Y, Romero MF, Boron
WF, Nussberger S, Gollan JL, Hediger MA (1997) Cloning and
characterization of a mammalian proton-coupled metal-ion
transporter. Nature 388:482–488
Hackam DJ, Rotstein OD, Zhang W, Gruenheid S, Gros P, Grinstein S
(1998) Host resistance to intracellular infection: mutation of
natural resistance-associated macrophage protein 1 (Nramp1)
impairs phagosomal acidification. J Exp Med 188:351–364
Harvill ET, Cotter PA, Yuk MH, Miller JF (1999) Probing the function
of Bordetella bronchiseptica adenylate cyclase toxin by manip-
ulating host immunity. Infect Immun 67:1493–1500
Hensel M, Shea JE, Gleeson C, Jones MD, Dalton E, Holden DW
(1995) Simultaneous identification of bacterial virulence genes
by negative selection. Science 269:400–403
Hisert KB, Kirksey MA, Gomez JE, Sousa AO, Cox JS, Jacobs WR Jr,
Nathan CF, McKinney JD (2004) Identification of
Mycobacterium
tuberculosis counterimmune (cim) mutants in immunodeficient
mice by differential screening. Infect Immun 72:5315–5321
Hisert KB, MacCoss M, Shiloh MU, Darwin KH, Singh S, Jones RA,
Ehrt S, Zhang Z, Gaffney BL, Gandotra S, Holden DW, Murray
D, Nathan C (2005) A glutamate-alanine-leucine (EAL) domain
protein of Salmonella controls bacterial survival in mice,
antioxidant defence and killing of macrophages: role of cyclic
diGMP. Mol Microbiol 56:1234 –1245
Hitotsumachi S, Carpenter DA, Russell WL (1985) Dose-repetition
increases the mutagenic effectiveness of N-ethyl-N-nitrosourea in
mouse spermatogonia. Proc Natl Acad Sci USA 82:6619–6621
Immunogenetics

Hoebe K, Georgel P, Rutschmann S, Du X, Mudd S, Crozat K, Sovath
S, Shamel L, Hartung T, Zahringer U, Beutler B (2005) CD36 is
a sensor of diacylglycerides. Nature 433:523–527
Hoshino K, Takeuchi O, Kawai T, Sanjo H, Ogawa T, Takeda Y,
Takeda K, Akira S (1999) Cutting edge: Toll-like receptor 4
(TLR4)-deficient mice are hyporesponsive to lipopolysaccharide:
evidence for TLR4 as t he Lps gene product. J Immun ol
162:3749–3752
Hsiao A, Liu Z, Joelsson A, Zhu J (2006) Vibrio cholerae virulence
regulator-coordinated evasion of host immunity. Proc Natl Acad
Sci USA 103:14542 –14547
Huang S, Hendriks W, Althage A, Hemmi S, Bluethmann H, Kamijo
R, Vilcek J, Zinkernagel RM, Aguet M (1993) Immune response
in mice that lack the i nterferon-gamma receptor. Science
259:1742–1745
Jabado N, Jankowski A, Dougaparsad S, Picard V, Grinstein S, Gros P
(2000) Natural resistance to intracellular infections: natura l
resistance-associated macrophage protein 1 (Nramp1) functions
as a pH-dependent manganese transporter at the phagosomal
membrane. J Exp Med 192:1237–1248
Janeway CA Jr (1989) Approaching the asymptote? Evolution and
revolution in immunology. Cold Spring Harb Symp Quant Biol
54(Pt 1):1–13
Jones JD, Dangl JL (2006) The plant immune system. Nature
444:323–329
Joshi SM, Pandey AK, Capite N, Fortune SM, Rubin EJ, Sassetti CM
(2006) Characterization of mycobacteria l virulence genes
through genetic interaction mapping. Proc Natl Acad Sci USA
103:11760– 11765
Kaye PM, Blackwell JM (1989) Lsh, antigen presentation and the
development of CMI. Res Immunol 140:810–815 discussion
815–22
Klarsfeld AD, Goossens PL, Cossart P (1994) Five Listeria mono-
cytogenes genes preferentially expressed in infected mammalian
cells: plcA, purH, purD, pyrE and an arginine ABC transporter
gene, arpJ. Mol Microbiol 13:585–597
Kramnik I, Dietrich WF, Demant P, Bloom BR (2000) Genetic control
of resistance to experimental infection with virulent Mycobacte-
rium tuberculosis. Proc Natl Acad Sci USA 97:8560–8565
Lamkanfi M, Amer A, Kanneganti TD, Munoz-Planillo R, Chen G,
Vandenabeele P, Fortier A, Gros P, Nunez G (2007) The Nod-like
receptor family member Naip5/birc1e restricts Legionella pneu-
mophila growth independently of caspase-1 activation. J Immu-
nol 178:8022–8027
Lander ES, Botstein D (1989) Mapping Mendelian factors underlying
quantitative traits using RFLP linkage maps. Genetics 121:185–199
Lang T, Pri na E, Sibthorpe D, Blackwell JM (1997) Nramp1
transfection transfers Ity/Lsh/Bcg-related pleiotropic effects on
macrophage activation: influence on antigen processing and
presentation. Infect Immun 65:380–386
Lawson JN, Lyons CR, Johnston SA (2006) Expression profiling of
Yersinia pestis during mouse pulmonary infection. DNA Cell
Biol 25:608–616
Lemaitre B, Nicolas E, Michaut L, Reichhart JM, Hoffmann JA
(1996) The dorsoventral regulatory gene cassette spatzle/Toll/
cactus controls the potent antifungal response in Drosophila
adults. Cell 86:973–983
Logsdon LK, Mecsas J (2003) Requirement of the Yersinia pseudo-
tuberculo sis effectors YopH and YopE in colon ization and
persistence in intestinal an d lymph tissues. In fect Immun
71:4595–4607
Logsdon LK, Mecsas J (2006) The proinflammatory response induced
by wild-type Yersinia pseudotuberculosis infection inhibits
survival of yop mutants in the gastrointestinal tract and Peyer’s
patches. Infect Immun 74:1516–1527
Lurie MB, Zappasodi P, Dannenberg AM Jr, Weiss GH (1952) On the
mechanism of genetic resistance to tuberculosis and its mode of
inheritance. Am J Hum Genet 4:302–314
Lynch CJ, Pierce-Chase CH, Dubos R (1965) A genetic study of
susceptibility to experimental tuberculosis in mice infected with
mammalian tubercle bacilli. J Exp Med 121:1051–1070
MacMicking JD (2005) Immune control of phagosomal bacteria by
p47 GTPases. Curr Opin Microbiol 8:74–82
Mahan MJ, Slauch JM, Mekalanos JJ (1993) Selection of bacterial
virulence genes that are specifically induced in host tissues.
Science 259:686–688
Mahan MJ, Tobias JW, Slauch JM, Hanna PC, Collier RJ, Mekalanos
JJ (1995) Antibiotic-based selection for bacterial genes that are
specifically induced during infection of a host. Proc Natl Acad
Sci USA 92:669 –673
Malo D, Vogan K, Vidal S, Hu J, Cellier M, Schurr E, Fuks A,
Bumstead N, Morgan K, Gros P (1994) Haplotype mapping and
sequence analysis of the mouse Nramp gene predict susceptibility
to infection with intracellular parasites. Genomics 23:51–61
Martinon F, Burns K, Tschopp J (2002) The inflammasome: a
molecular platform triggering activation of inflammatory cas-
pases and processing of proIL-beta. Mol Cell 10:417–426
Mecsas J (2002) Use of signature-tagged mutagenesis in pathogenesis
studies. Curr Opin Microbiol 5:33–37
Medzhitov R, Janeway CA (1997) Innate immunity: the virtues of a
nonclonal system of recognition. Cell 91:295–298
Medzhitov R, Preston-Hurlburt P, Janeway CA Jr (1997) A human
homologue of the Drosophila Toll protein signals activation of
adaptive immunity. Nature 388:394–397
Merrell DS, Camilli A (2002) Information overload: assigning genetic
functionality in the age of genomics and large-scale screening.
Trends Microbiol 10:571–574
Merrell DS, Hava DL, Camilli A (2002) Identification of novel factors
involved in colonization and acid tolerance of Vibrio cholerae.
Mol Microbiol 43:1471–1491
Meylan E, Tschopp J, Karin M (2006) Intracellular pattern recognition
receptors in the host response. Nature 442:39–44
Miao EA, Alpuche-Aranda CM, Dors M, Clark AE, Bader MW,
Miller SI, Aderem A (2006) Cytoplasmic flagellin activates
caspase-1 and secretion of interleukin 1beta via Ipaf. Nat
Immunol 7:569–575
Miyairi I, Tatireddigari VR, Mahdi OS, Rose LA, Belland RJ, Lu L,
Williams RW, Byrne GI (2007) The p47 GTPases Iigp2 and
Irgb10 regulate innate immunity and inflammation to murine
Chlamydia psittaci infection. J Immunol 179:1814–1824
Molofsky AB, Byrne BG, Whitfield NN, Madigan CA, Fuse ET,
Tateda K, Swanson MS (2006) Cytosolic recognition of flagellin
by mouse macrophages restricts Legionella pneumophila infec-
tion. J Exp Med 203:1093–1104
Montminy SW, Khan N, McGrath S, Walkowicz MJ, Sharp F, Conlon
JE, Fukase K, Kusumoto S, Sweet C, Miyake K, Akira S, Cotter
RJ, Goguen JDm, Lien E (2006) Virulence factors of Yersinia
pestis are overcome by a strong lipopolysaccharide response. Nat
Immunol 7:1066–1073
Ng VH, Cox JS, Sousa AO, MacMicking JD, McKinney JD (2004)
Role of KatG catalase-peroxidase in mycobacterial pathogenesis:
cou ntering the p hagocyte oxidati ve burst. Mol Microbiol
52:1291–1302
North RJ, LaCourse R, Ryan L, Gros P (1999) Consequence of
Nramp1 deletion to Mycobacterium tuberculosis infection in
mice. Infect Immun 67:5811–5814
O’Brien AD, Rosenstreich DL, Scher I, Campbell GH, MacDermott
RP, Forma l SB (1980) Genetic control of susceptibility to
Salmonella typhimurium in mice: role of the LPS gene. J
Immunol 124:20–24
Immunogenetics

O’Connell RM, Saha SK, Vaidya SA, Bruhn KW, Miranda GA,
Zarnegar B, Perry AK, Nguyen BO, Lane TF, Taniguchi T, Miller
JF, Cheng G (2004) Type I interferon production enhances
susceptibility to Listeria monocytogenes infection. J Exp Med
200:437–445
Osorio CG, Crawford JA, Michalski J, Martinez-Wilson H, Kaper JB,
Camilli A (2005) Second-generation recombination-based in vivo
express ion technology for large-scale screening for Vibrio
cholerae genes induced during infection of the mouse small
intestine. Infect Immun 73:972–980
Pan H, Yan BS, Rojas M, Shebzukhov YV, Zhou H, Kobzik L,
Higgins DE, Daly MJ, Bloom BR, Kramnik I (2005) Ipr1 gene
mediates innate immunity to tuberculosis. Nature 434:767–772
Papathanasiou P, Goodnow CC (2005) Connecting mammalian genome
with phenome by ENU mouse mutagenesis: gene combinations
specifying the immune system. Annu Rev Genet 39:241–262
Plant J, Glynn AA (1974) Natural resistance to Salmonella infection,
delayed hypersensitivity and Ir genes in different strains of mice.
Nature 248:345–347
Plant J, Glynn AA (1976) Genetics of resistance to infection with
Salmonella typhimurium in mice. J Infect Dis 133:72–78
Poltorak A, He X, Smirnova I, Liu MY, Van Huffel C, Du X, Birdwell
D, Alejos E, Silva M, Galanos C, Freudenberg M, Ricciardi-
Castagnoli P, Layton B, Beutler B (1998) Defective LPS
signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in
Tlr4 gene. Science 282:2085–2088
Portnoy DA (2005) Manipulation of innate immunity by bacterial
pathogens. Curr Opin Immunol 17:25–28
Ren T, Zamboni DS, Roy CR, Dietrich WF, Vance RE (2006)
Flagellin-deficient L egionella mutants evade caspase-1- an d
Naip5-mediated macrophage immunity. PLoS Pathog 2:e18
Rengarajan J, Bloom BR, Rubin EJ (2005) Genome-wide require-
ments for Mycobacterium tuberculosis adaptation and survival in
macrophages. Proc Natl Acad Sci USA 102:8327–8332
Revel AT, Talaat AM, Norgard MV (2002) DNA microarray analysis
of differential gene expression in Borrelia burgdorferi, the Lyme
disease spirochete. Proc Natl Acad Sci USA 99:1562–1567
Rietsch A, Wolfgang MC, Mekalanos JJ (2004) Effect of metabolic
imbalance on expression of type III secretion genes in Pseudo-
monas aeruginosa. Infect Immun 72:1383–1390
Roberts JE, Watters JW, Ballard JD, Dietrich WF (1998) Ltx1, a
mouse locus that influences the susceptibility of macrophages to
cytolysis caused by intoxication with Bacillus anthracis lethal
factor, maps to chromosome 11. Mol Microbiol 29:581–591
Robson HG, Vas SI (1972) Resistance of inbred mice to Salmonella
typhimurium. J Infect Dis 126:378–386
Salama NR, Manoil C (2006) Seeking completeness in bacterial
mutant hunts. Curr Opin Microbiol 9:307–311
Saleh M, Mathison JC, Wolinski MK, Bensinger SJ, Fitzgerald P,
Droin N, Ulevitch RJ, Green DR, Nicholson DW (2006)
Enhanced bacterial clearance and sepsis resistance in caspase-
12-deficient mice. Nature 440:1064–1068
Sassetti CM, Rubin EJ (2003) Genetic requirements for mycobacterial
survival during infection. Proc Natl Acad Sci USA 100:12989–
12994
Sassetti CM, Boyd DH, Rubin EJ (2001) Comprehensive identifica-
tion of conditionally essential genes in mycobacteria. Proc Natl
Acad Sci USA 98:12712–12717
Shimono N, Morici L, Casali N, Cantrell S, Sidders B, Ehrt S, Riley
LW (2003) Hypervirulent mutant of Mycobacterium tuberculosis
resulting from disruption of the mce1 operon. Proc Natl Acad Sci
USA 100:15918– 15923
Skamene E, Gros P, Forget A, Kongshavn PA, St Charles C, Taylor
BA (1982) Genetic re gulation of resistance to intracellular
pathogens. Nature 297:506–509
Snyder JA, Haugen BJ, Buckles EL, Lockatell CV, Johnson DE,
Donnenberg MS, Welch RA, Mobley HL (2004) Transcriptome
of uropathogenic Escherichia coli during urinary tract infection.
Infect Immun 72:6373–6381
Stober CB, Brode S, White JK, Popoff JF, Blackwell JM (2007)
Slc11a1 (formerly Nramp1) is expressed in dendritic cells and
influences MHC class II expression and antigen presenting cell
function. Infect Immun (in press)
Supek F, Supekova L, Nelson H, Nelson N (1996) A yeast manganese
transporte r related to t he macroph age protein involved i n
conferring resistance to mycobacteria. Proc Natl Acad Sci USA
93:5105–5110
Tabeta K, Hoebe K, Janssen EM, Du X, Georgel P, Crozat K, Mudd S,
Mann N, Sovath S, Goode J, Shamel L, Herskovits AA, Portnoy
DA, Cooke M, Tarantino LM, Wiltshire T, Steinberg BE,
Grinstein S, Beutler B (2006) The Unc93b1 mutation 3d disrupts
exogenous antigen pres entation and signaling via Toll-like
receptors 3, 7 and 9. Nat Immunol 7:156–164
Talaat AM, Lyons R, Howard ST, Johnston SA (2004) The temporal
expression profile of Mycobacterium tuberculosis infection in
mice. Proc Natl Acad Sci USA 101:4602–4607
Tosh K, Campbell SJ, Fielding K, Sillah J, Bah B, Gustafson P,
Manneh K, Lisse I, Sirugo G, Bennett S, Aaby P, McAdam KP,
Bah-Sow O, Lienhardt C, Kramnik I, Hill AV (2006) Variants in
the SP110 gene are associated with genetic susceptibility to
tuber culosis in West Africa. Proc Natl Acad Sci USA
103:10364–10368
Tripp C S, Gately MK, Hakimi J, Ling P, Unanue ER (1994)
Neutralization of IL-12 decreases resistance to Listeria in SCID
and C. B-17 mice. Reversal by IFN-gamma. J Immunol
152:1883–1887
Valdivia RH, Falkow S (1997) Fluorescence-based isolation of bacterial
genes expressed within host cells. Science 277:2007–2011
van Diepen A, van der Straaten T, Holland SM, Janssen R, van Dissel
JT (2002) A superoxide-hypersusceptible Salmonella enterica
serovar Typhimurium mutant is attenu ated but regains virulence
in p47(phox−/−) mice. Infect Immun 70:2614–2621
Vazquez-Torres A, Jones-Carson J, Baumler AJ, Falkow S, Valdivia
R, Brown W, Le M, Berggren R, Parks WT, Fang FC (1999)
Extraintestinal dissemination of Salmonella by CD18-expressing
phagocytes. Nature 401:804–808
Vazquez-Torres A, Xu Y, Jones-Carson J, Holden DW, Lucia SM,
Dinauer MC, Mastroeni P, Fang FC (2000) Salmonella pathoge-
nicity island 2-dependent evasion of the phagocyte NADPH
oxidase. Science 287:1655–1658
Viboud GI, Bliska JB (2005) Yersinia outer proteins: role in
modulation of host cell signaling responses and pathogenesis.
Annu Rev Microbiol 59:69–89
Vidal SM, Malo D, Vogan K, Skamene E, Gros P (1993) Natural
resistance to infection with intracellular parasites: isolation of a
candidate for Bcg. Cell 73:469–485
Vidal S, Tremblay ML, Govoni G, Gauthier S, Sebastiani G, Malo D,
Skamene E, Olivier M, Jothy S, Gros P (1995) The Ity/Lsh/Bcg
locus: natural resistance to infection with intracellular parasites is
abrogated by disruption of the Nr amp1 gene. J Exp Med
182:655–666
von Jeney N, Gunther E, Jann K (1977) Mitogenic stimulation of
murine spleen cells: relation to susceptibility to Salmonella
infection. Infect Immun 15:26–33
Wade CM, Daly MJ (2005) Genetic variation in laboratory mice. Nat
Genet 37:1175–1180
Watarai M, Derre I, Kirby J, Growney JD, Dietrich WF, Isberg RR
(2001) Legionella pneumophila is internalized by a macro-
pinocytotic uptake pathway controlled by the Dot/Icm system
and the mouse Lgn1 locus. J Exp Med 194:1081–1096
Immunogenetics

Watson J, Riblet R, Taylor BA (1977) The response of recombinant
inbred strains of mice to bacterial lipopolysaccharides. J
Immunol 118:2088–2093
Watson J, Kelly K, Largen M, Taylor BA (1978) The genetic mapping of
a defective LPS response gene in C3H/HeJ mice. J Immunol
120:422–424
Watters JW, Dewar K, Lehoczky J, Boyartchuk V, Dietrich WF (2001)
Kif1C, a kinesin-like motor protein, mediates mouse macrophage
resistance to anthrax lethal factor. Curr Biol 11:1503–1511
Weiss DS, Brotcke A, Henry T, Margolis JJ, Chan K, Monack DM (2007)
In vivo negative selection screen identifies genes required for
Francisella virulence. Proc Natl Acad Sci USA 104:6037–6042
Wright EK, Goodart SA, Growney JD, Hadinoto V, Endrizzi MG,
Long EM, Sadigh K, Abney AL, Bernstein-Hanley I, Dietrich
WF (2003) Naip5 affects host susceptibility to the intracellular
pathogen Legionella pneumophila. Curr Biol 13:27–36
Xu Q, Dziejman M, Mekalanos JJ (2003) Determination of the
transcriptome of Vibrio cholerae during intraintestinal growth
and midexponential phase in vitro. Proc Natl Acad Sci USA
100:1286–91
Yamamoto Y, Klein TW, Newton CA, Widen R, Friedman H (1988)
Growth of Legionella pneumophila in thioglycolate-elicited perito-
neal macrophages from A/J mice. Infect Immun 56:370–375
Yan BS, Kirby A, Shebzukhov YV, Daly MJ, Kramnik I (2006)
Genetic architecture of tuberculosis resistance in a mouse model
of infection. Genes Immun 7:201–210
Yoshida S, Goto Y, Mizuguchi Y, Nomoto K, Skamene E (1991)
Genetic control of natural resistance in mouse macrophages
regulating intracellular Legionella pneumophila multiplication in
vitro. Infect Immun 59:428–432
Zamboni DS, Kobayashi KS, Kohlsdorf T, Ogura Y, Long EM, Vance
RE, Kuida K, Mariathasan S, Dixit VM, Flavell RA, Dietrich
WF, Roy CR (2006) The Birc1e cytosolic pattern-recognition
receptor contributes to the detection and control of Legionella
pneumophila infection. Nat Immunol 7:318–325
Zhang S, Chen Y, Potvin E, Sanschagrin F, Levesque RC, McCor-
mac k FX, Lau GW (2005) Comparativ e signature-tagged
mutagenesis identifies Pseudomonas factors conferring resistance
to the pulmonary collectin SP-A. PLoS Pathog 1:259–268
Zhang S, McCormac k FX, Leve sque RC, O’Too le GA, Lau GW
(2007) The flagellum of Pseudomonas aeruginosa is required
for res istance to clearance by surfactant protein a. PLoS ONE 2:
e564
Zwilling BS, Vespa L, Massie M (1987) Regulation of I-A expression
by murine peritoneal macrophages: differences linked to the Bcg
gene. J Immunol 138:1372–1376
Immunogenetics