Molecular basis of host specificity in human pathogenic bacteria

Pathogenic bacteria display various levels of host specificity or tropism. While many bacteria can infect a wide range of hosts, certain bacteria have strict host selectivity for humans as obligate human pathogens. Understanding the genetic and molecular basis of host specificity in pathogenic bacteria is important for understanding pathogenic mechanisms, developing better animal models and designing new strategies and therapeutics for the control of microbial diseases. The molecular mechanisms of bacterial host specificity are much less understood than those of viral pathogens, in part due to the complexity of the molecular composition and cellular structure of bacterial cells. However, important progress has been made in identifying and characterizing molecular determinants of bacterial host specificity in the last two decades. It is now clear that the host specificity of bacterial pathogens is determined by multiple molecular interactions between the pathogens and their hosts. Furthermore, certain basic principles regarding the host specificity of bacterial pathogens have emerged from the existing literature. This review focuses on selected human pathogenic bacteria and our current understanding of their host specificity.


INTRODUCTION
Host specificity or tropism of a microbial pathogen is defined by its ability to colonize (or infect) a host organism. 1 Pathogenic bacteria display various levels of host specificity. Certain bacteria have extremely diverse host ranges. Bacteria in this category may infect insects, humans, rodents and many other wild/domestic animals, such as Borrelia burgdorferi and Yersinia pestis (the respective etiologic agents of Lyme disease and plague). Many bacteria possess intermediate levels of host specificity. Such pathogens infect human and other mammalian species, but they may exhibit different pathogenicity between human and other hosts. For example, Salmonella Typhimurium causes gastroenteritis in humans after oral ingestion of the bacterium, but in mice, this infection route leads to symptoms resembling typhoid fever of human. 2,3 Listeria monocytogenes can cause listeriosis in human and a number of domestic animals after oral ingestion of the bacterium. Interestingly, oral infection of L. monocytogenes leads to listeriosis-like disease in some model animals (e.g., guinea pig) but not other (e.g., mouse). [4][5][6] These important features have made S. Typhimurium and L. monocytogenes important models for studying many important aspects of bacterial pathogenesis, host immunity, host-pathogen interactions, and host specificity. 2,3,7 In sharp contrast, a number of other bacteria are highly adapted to the human environment and display strict host selectivity for humans, including Haemophilus influenzae, Helicobacter pylori, Neisseria gonorrhoeae, Neisseria meningitidis, Mycobacterium leprae, Salmonella Typhi, Streptococcus pneumoniae, Streptococcus pyogenes, Vibrio cholerae and Treponema pallidum. For the purpose of discussion in this review, we will hereafter refer to these bacteria as human-specific pathogens.
Understanding the genetic and molecular basis of host specificity in pathogenic bacteria is of great importance. First, learning how a pathogen determines its host range at the molecular level can directly enhance our knowledge of pathogenic mechanisms. Second, identifying the molecular determinants of host tropism for a pathogen provides insights that are valuable for the improvement of animal models used to simulate the human diseases caused by natural infection. Finally, defining the molecular mechanisms of bacterial host specificity can provide potential molecular targets and other clues to aid in the design of new strategies and therapeutics for the control of microbial diseases.
Host specificity of microbial pathogens is determined by elaborate molecular interactions between the pathogens and hosts. This principle is best demonstrated in viral pathogens due to the relative simplicity of their genomes and structures. The host specificity of viruses is predominantly defined by the interactions of viral proteins with their cognate cellular receptors. 8 The molecular mechanisms of host specificity are less understood for bacterial pathogens due to the complexity of their molecular compositions and cellular structures. However, important progress has been made toward the complete understanding of the host specificity of pathogenic bacteria over the last two decades. This review focuses on the host specificity of selected human-specific pathogens. Due to space constraints, we will not provide exhaustive coverage of the discoveries in the area of bacterial host specificity. Instead, we will attempt to uncover emerging principles in this area by focusing on several bacterial pathogens for which the host specificity has been extensively characterized.

HUMAN-SPECIFIC PATHOGENS
Pathogenic Neisseria N. gonorrhoeae and N. meningitidis are two human pathogens within the genus Neisseria. N. gonorrhoeae is the causative agent of gonorrhea (pelvic inflammation), a sexually transmitted disease. N. meningitidis causes invasive infections, such as septicemia and meningitis. Both pathogens have strict host tropism for humans. Like many other human-specific pathogens, the narrow host specificity of these pathogenic Neisseria has made it difficult to study the pathogenic mechanisms and develop therapeutics due to the lack of appropriate animal models that simulate the clinical presentations of the human diseases. [9][10][11] The molecular basis for the host specificity of pathogenic Neisseria is not fully understood, but the existing evidence indicates that multiple factors are involved. On the pathogen side, a number of surfaceexposed proteins appear to be associated with the human specificity of pathogenic Neisseria, including immunoglobulin A1 (IgA1) protease, 12,13 type IV pili, [14][15][16] complement factor H binding proteins (FHBP), 17,18 gonococcal porin, 19 transferrin-binding proteins 20,21 and lactoferrin-binding proteins. 21,22 On the host side, several human proteins have been implicated as host specificity determinants, including IgA1, 12,13 cell surface complement regulator CD46, 10 complement regulator factor H (FH), 17,18,23 complement regulator C4b-binding protein (C4BP), 19 transferrin 9,24,25 and lactoferrin. 24,25 As illustrated in Figure 1, N. gonorrhoeae and N. meningitidis produce an extracellular serine-type protease that specifically cleaves human IgA1, a subclass of the dominant mucosal antibody, IgA. 12,13 IgA1 is the most abundant antibody in upper respiratory secretions. 26 IgA serves as a 'non-inflammatory mucosal protector' at mucosal surfaces, performing multiple important functions, such as inhibition of microbial adherence, neutralization of microorganisms and inhibition of antigen penetration. 27 S. pneumoniae and H. influenzae, two additional obligate human pathogens and colonizers of the human nasopharynx, possess similar capabilities to cleave human IgA1 ( Figure 1). 13,28,29 These IgA1 proteases specifically cleave the hinge region of IgA1, but not that of IgA2 due to the sequence differences of the two IgA subclasses in this region, with 26 amino acids in IgA1 and 13 in IgA2. 26 Mouse and many other mammals possess only one IgA antibody with significant sequence differences from human IgA1, which explains why the IgA1 proteases of pathogenic Neisseria cannot digest the IgA produced by many nonhuman species. 30 Cleavage of IgA1 in the hinge region leads to the separation of the Fc region from the antigen-binding Fab fragment, abolishing the crosslinking activity and all secondary effector functions mediated by the Fc region. 27 The loss of IgA-mediated crosslinking and microbial agglutination may result in higher levels of bacterial adherence to mucosal surfaces and colonization. 13 Type IV pili on the cell surface of pathogenic Neisseria mediate the initial adhesion to the mucosal epithelium through binding to human membrane cofactor protein (MCP or CD46). 14-16 CD46 is a widely distributed cell membrane protein, which protects the host cells from accidental damage by complement through inactivating the alternative pathway of the complement system. 31,32 The complement system is a key component of host immunity against N. meningitidis, and complement-deficient individuals have a significantly elevated risk of developing meningococcal disease. 33 Previous studies have also identified CD46 as a host receptor for multiple viral and bacterial pathogens, such as measles virus, 34-36 herpes virus 6, 37 adenovirus [38][39][40][41] and S. pyogenes. 42 The extracellular segment of CD46 consists of four tandem complement control protein (CCP) modules and one or two heavily O-glycosylated serine/threonine/proline-rich domains. 43 While measles virus and human herpesvirus 6 recognize the CCP1-2 and CCP2-3 modules, respectively, [44][45][46] pathogenic Neisseria and S. pyogenes bind to the serine/threonine/proline-rich and CCP3-4 regions of CD46, respectively. 15,47 Interaction of N. meningitidis and N. gonorrhoeae with CD46 triggers signaling events in the host cells, leading to intimate attachment and bacterial entry into host cells in cell culture models. 14,[48][49][50] The significance of this molecular interaction in the pathogenesis of N. meningitidis has been further implicated in transgenic mice expressing human CD46. 10 When infected intraperitoneally with piliated N. meningitidis, CD46 transgenic mice displayed higher levels of bacteria in the bloodstream and cerebrospinal fluid; similar infection via intranasal inoculation with N. meningitidis also led to bacteremia and mortality only in CD46 transgenic mice, but not in the control mice. These in vitro and in vivo studies strongly suggest that CD46 is a host factor that contributes to the host specificity of N. meningitidis. Because the type IV pili of N. gonorrhoeae and N. meningitidis interact with CD46 in a similar manner, 14,48-50 it is likely that this pathogenhost interaction also contributes to the host tropism of N. gonorrhoeae for humans. Transgenic mice expressing human CD46 are more susceptible to S. pyogenes infection (see below). 51,52 Both N. gonorrhoeae and N. meningitidis are able to recruit FH to the cell surface in a human-specific fashion. [53][54][55] FH is one of the most abundant proteins in the blood (commonly present in human plasma at concentrations of 300 to 500 mg/mL). FH is composed of 20 short consensus repeats (SCRs), each containing approximately 60 amino acids. 56 FH inhibits the alternative pathway of the complement system by preventing the binding of factor B to C3b, enhancing the decay of the C3-convertase (C3bBb) and acting as a cofactor for the cleavage of Host specificity of bacteria X Pan et al 2 C3b by complement factor I. 57 FH deposition on host tissue and cell surfaces prevents non-specific damage and avoids the wasteful consumption of complement components. 57 Many microbial pathogens have been shown to bind to FH as a common mechanism for the evasion of complement-mediated host immunity ( Figure 2). 58,59 This is exemplified by the binding of FH to pathogenic Neisseria, 17,18,23,71 Candida albicans, 70,72 Borrelia burgdorferi, 68,69,73 H. influenzae, 62 S. pneumoniae, 63,74 S. pyogenes, 75 S. agalactiae [76][77][78] and Streptococcus suis. 79 Notably, S. pneumoniae, another resident of the human upper airway, also interacts with FH in a human-specific manner (see below). 80 While N. gonorrhoeae binds to FH via its porin protein, a dominant component of the outer membrane, 17 the FH binding activity of N. meningitidis has been localized to a ,29-kD FHBP. FHBP is expressed as a lipoprotein in the meningococcal outer membrane of all strains of N. meningitidis. 18,81,82 The recruitment of FH helps both pathogens evade complement-mediated lysis in human plasma. 17,18,23 Moreover, FH-binding activity has been shown to mediate gonococcal attachment to human complement receptor 3 on the surface of epithelial cells, suggesting that this interaction may facilitate gonococcal entry into nonprofessional phagocytes and thereby aid the evasion of host immunity. 83 The molecular preference of N. gonorrhoeae and N. meningitidis for human FH leads to a species-specific resistance to the bactericidal activity of human serum. 18,23,53,54 Finally, the expression of human FH leads to significantly enhanced N. meningitidis bacteremia in transgenic rats. 84 These findings are consistent with a clinical observation that polymorphisms in the promoter of the human FH gene are associated with increased risk for meningococcal disease in UK Caucasian patients. 85 Prior to the discovery of its FH binding activity, the FHBP of N. meningitidis was identified as a vaccine candidate. 81,82 Two FHBPbased vaccines have since advanced to clinical trials. 86 The immunization of transgenic mice expressing human FH with the natural form of FHBP led to significantly decreased protective immune responses, as assessed by serum bactericidal activity; a mutant FHBP lacking the FH-binding activity showed the opposite vaccination result. 87,88 This finding suggests that engineering FH binding-negative FHBP vaccine antigens may enhance the immunoprotective efficacy against meningococcal disease in humans.
N. gonorrhoeae and N. meningitidis bind to C4BP via the allelic variants of the porin protein. 19,89 C4BP is a plasma glycoprotein that downregulates the classical complement pathway. 32 It has been shown that N. gonorrhoeae binds only to human C4BP but not the counterparts of rodent, lagomorph and primate species. 53 The species-specific binding pattern with human C4BP is consistent with gonococcal resistance to human serum complement. 53 A number of other pathogens also recruit C4BP to evade complement-mediated immune responses, such as Bordetella pertussis, 90 Escherichia coli K1, 91 H. influenzae, 92 Moraxella catarrhalis, 93 S. pneumoniae 94,95 and S. pyogenes. 96 Interestingly, S. pneumoniae, also interacts with C4BP in a human-specific manner (see below). 94 Iron acquisition appears to contribute to the host specificity of N. gonorrhoeae and N. meningitidis for humans 25 . Iron is an essential nutrient for survival for virtually all bacteria. Pathogenic Neisseria species are able to extract iron from host iron-containing proteins through specific receptors, such as transferrin, lactoferrin and hemoglobin. 97 N. gonorrhoeae and N. meningitidis preferentially bind to human transferrin and lactoferrin and utilize them as iron sources when compared with the transferrins and lactoferrins of other hosts. 25,98 While the majority of transferrin circulates in human serum to deliver iron into cells and sequester free iron, lactoferrin is mostly present in phagocytic cells and in secretions such as milk, mucus and tears. 97 Exogenous administration of human transferrin and lactoferrin through intravenous or intraperitoneal inoculation significantly enhances meningococcal bacteremia and mortality in mice. 24 In addition, transgenic mice expressing human transferrin are significantly more susceptible to systemic infection by N. meningitidis. 9 N. meningitidis utilizes human and bovine hemoglobins at a similar level in vitro and in mice, 24 suggesting that the utilization of hemoglobin does not contribute significantly to the host specificity of pathogenic Neisseria for humans. In sharp contrast, the preferential binding of Staphylococcus aureus to human hemoglobin does contribute to its host specificity. 99 Another interesting observation is that the human-specific binding of lactoferrin by pneumococci is not a means of iron acquisition 100 . These lines of evidence indicate that different pathogens interact with host-containing proteins for various purposes.
Streptococcus pneumoniae S. pneumoniae is an obligate human pathogen and causes numerous infections, such as pneumonia, otitis media, meningitis and sinusitis in humans. 101 The molecular mechanisms behind this strict host specificity are unclear, but several human-specific pathogen-host interactions have been recently revealed using biochemical approaches.
The metallo-type IgA1 protease of S. pneumoniae cleaves the hinge region of human IgA1 but not those of the IgA molecules from many other mammalian species (Figure 1). 28  Host specificity of bacteria X Pan et al 3 the Fc region. 27 S. pneumoniae IgA1 protease has been shown to enhance bacterial adherence to epithelial cells by digesting IgA1, 102 suggesting that this human-specific activity is a host specificity determinant of S. pneumoniae. S. pneumoniae interacts with a list of host factors in a human-specific manner through choline-binding protein A (CbpA). CbpA, also called PspC, 103 is a major surface-exposed protein of S. pneumoniae. 104 CbpA binds to the polymeric immunoglobulin receptor (pIgR), 105 secretory component (SC), 106 secretory IgA (SIgA), 106 FH, 107,108 C4BP, 95 sialic acid, 104 complement C3 protein 109 and vitronectin. 110 The middle repeat region of CbpA is responsible for binding to domains 3 and 4 of human pIgR and the same region in human SC and SIgA. 105,[111][112][113] While the FH binding activity has been mapped to the amino terminal region of CbpA and its allelic variants, 77,107,114 multiple regions of human FH have been reported to bind to CbpA, including SCR 6-10, 65 8-14 63 and 13-15. 64 The CbpA-pIgR interaction mediates the epithelial adhesion and transmigration of S. pneumoniae; 105,115-117 the recruitment of FH to the pneumococcal surface enhances the evasion of complement-mediated phagocytosis 74,80,107 and epithelial invasion. 118 Studies from our laboratory and others have demonstrated that CbpA exclusively binds to the pIgR, SC, SIgA and FH of humans but not their counterparts in common model animals (e.g., mouse, rat and rabbit). 80,105,112,119 These species-specific interactions strongly suggest that CbpA is a major molecular determinant of pneumococcal host specificity.
The enolase of S. pneumoniae binds to C4BP, a negative regulator of the classical complement pathway. 94 This interaction leads to decreased C3b deposition on pneumococci. Interestingly, this binding occurs only with human C4BP, not mouse C4BP. 94 This observation is reminiscent of the human-specific recruitment of C4BP by N. gonorrhoeae, a colonizer of mucosal surfaces in humans. 53 Thus, evasion of complement-mediated immunity appears to be a common mechanism underlying the host specificity of many pathogenic bacteria that colonize the mucosal surfaces of humans.
Pneumococcal surface protein A (PspA) is another protein that may contribute to the host tropism of S. pneumoniae. PspA is a major surface-exposed protein and a protective antigen of S. pneumoniae. 120 Previous biochemical studies have shown that the C-terminal region of PspA binds to human lactoferrin, an iron-sequestering glycoprotein (predominately in mucosal secretions). 121 Unlike pathogenic Neisseria, S. pneumoniae cannot obtain iron from lactoferrin. 122 Instead, PspA binding to lactoferrin protects pneumococci from killing by apolactoferrin, a form of lactoferrin that does not carry iron. 123,124 Interestingly, this interaction is human specific because PspA does not bind to bovine lactoferrin, 100 suggesting that the PspA-lactoferrin interaction contributes to the host specificity of S. pneumoniae.
These species-specific pathogen-host interactions correlate very well with the host specificity of S. pneumoniae. It is thus tempting to conclude that host specificity of S. pneumoniae is defined by multiple molecular determinants. Among the bacterial factors, CbpA mediates multiple molecular species-specific interactions with human factors (e.g., pIgR, SC, SIgA and FH). The lack of appropriate animal models has been a major challenge for determining the biological impacts of these pathogen-host interactions on the host specificity and pathogenesis of S. pneumoniae. However, as exemplified in L. monocytogenes, 125,126 these biochemical findings have provided critical molecular targets to establish genetically modified (humanized) animals and/or (murinized) pneumococci. These tools will be important for further elucidating the mechanisms of pneumococcal host specificity and evaluating new therapeutics and vaccines for the control of pneumococcal disease.

Streptococcus pyogenes
S. pyogenes (group A streptococcus) is an obligate human pathogen that causes pharyngitis (strep throat), localized skin infections, scarlet fever, streptococcal toxic shock syndrome and autoimmune-mediated complications. 127 Although the molecular basis for the human specificity of S. pyogenes remains unclear, a study reported by Sun et al. 128 indicates that the streptokinase secreted by this bacterium is an important determinant of its host specificity. Streptokinase is able to bind to and activate human plasminogen, a blood clot-dissolving protein, but not the plasminogens from other mammalian species. 129 The plasminogen-activating activity of streptokinase has been implicated in the accelerated clearance of host fibrin, which may promote the dissemination of S. pyogenes in host tissues. 128,130 Mice typically possess high resistance to skin infection by S. pyogenes, but Sun et al. 128 showed that the expression of human plasminogen in mice led to markedly increased mortality in the transgenic mice when infected with streptococci.
The molecular interaction of M protein with CD46 also contributes to the host specificity of S. pyogenes. 42 The M protein of S. pyogenes is a major surface-exposed protein and one of its most important virulence factors. 131 The specific interaction between M protein and human CD46 mediates streptococcal binding to keratinocytes 42,132 and the bacterial invasion of epithelial cells. 133 Transgenic mice expressing human CD46 are more susceptible to streptococcal disease. 51,52 In a systemic infection model, transgenic mice expressing human CD46 displayed higher levels of bacteremia, arthritis and mortality compared with the non-transgenic mice. 51 Matsui et al. 52 corroborated this finding in a subcutaneous infection model, in which the CD46 transgenic mice exhibited more severe forms of bacterial growth in deep tissues, necrotizing fasciitis at the infection sites (footpads) and mortality. These studies indicate that M protein is a molecular determinant of the host specificity of S. pyogenes through the targeting of human CD46.
S. pyogenes M protein also interacts with a wide range of other host factors, such as the complement regulator proteins C4BP 134 and FH, 75 as well as fibrinogen 135,136 and IgG. 137,138 S. pyogenes interactions with C4BP and FH are associated with bacterial evasion of complementmediated immunity. 130 It remains to be determined whether these pathogen-host interactions also contribute to the host specificity of this pathogen.

Gram-positive bacteria
Host specificity has been studied to various extents in other pathogenic bacteria. Among them, L. monocytogenes is the best-characterized bacterial pathogen in terms of our knowledge of the molecular mechanisms of its host specificity. L. monocytogenes is a facultative intracellular bacterium that causes foodborne diseases in humans and domestic animals. Infection occurs when ingested bacteria cross the intestinal barrier and disseminate to various organs and the bloodstream, leading to listeriosis. However, many laboratory animals do not develop typical listeriosis through the oral ingestion of L. monocytogenes, such as mouse, rat, rabbit and guinea pig. [4][5][6] Previous studies have revealed multiple molecular interactions that can now explain the host restriction of L. monocytogenes, as discussed in several excellent reviews. 125,139,140 The host selectivity of L. monocytogenes depends on two surfaceexposed bacterial proteins, internalin A (InlA) and internalin B (InlB). InlA and InlB promote the entry of L. monocytogenes into nonphagocytic cells by interacting with their cognate cellular receptors. InlA Host specificity of bacteria X Pan et al 4 binds to E-cadherin, a major cell-cell adhesion glycoprotein; 141,142 this interaction occurs with human E-cadherin but not the mouse counterpart due to a single amino-acid difference between the human and mouse E-cadherin proteins. 143 The importance of the InlA-Ecadherin interaction to the host specificity of L. monocytogenes has been demonstrated in mice expressing human E-cadherin, 144 'humanized' mouse E-cadherin 6 and the 'murinized' InlA of L. monocytogenes. 126 The local expression of human E-cadherin in the small intestine of mice was shown to lead to InlA-dependent bacterial invasion of enterocytes, dissemination across the intestinal barrier and significantly higher mortality following oral infection of the transgenic mice with L. monocytogenes. 144 Based on the structural knowledge of the InlA-human E-cadherin complex, 141 Wollert et al. 126 recently constructed a recombinant InlA that could interact with mouse Ecadherin by introducing two sequence substitutions in InlA, leading to dramatic enhancement in its binding affinity for mouse E-cadherin. The 'murinized' L. monocytogenes expressing the engineered InlA was able to induce typical listeriosis in wild-type mice by oral infection. 126 InlB is also a host specificity determinant of L. monocytogenes. 5 It is essential for the entry of L. monocytogenes into hepatocytes and some epithelial and fibroblast cell lines. [145][146][147] Three different host receptors have been reported to interact with InlB: C1q-binding protein gC1q-R/p32, 148 Met (the receptor for hepatocyte growth factor) 149 and glycosaminoglycans. 150 InlB interacts with Met in a species-specific manner because it binds to the Met proteins of human and mouse but not those of rabbit and guinea pig. 5 Consistently, only the Met proteins from human and mouse support the InlB-dependent entry of L. monocytogenes into mammalian cells; the InlB-deficient mutant does not show attenuated virulence in guinea pigs and rabbits. Among the common model animals, the gerbil is the only species in which both InlA/E-cadherin and InlB-Met interactions are functional. 6 In agreement, the oral infection of gerbils with L. monocytogenes leads to typical listeriosis. 6,151 These advancements have not only explained the pathogenic mechanisms of L. monocytogenes, but have also laid a new foundation for the rational design of better small animal models in studying human infections caused by this pathogen. 151 S. aureus captures hemoglobin for heme-iron acquisition through the hemoglobin-binding protein IsdB. 152 Pishchany et al. 99 showed preferential binding of S. aureus to human hemoglobin when compared with murine hemoglobin. Consistently, human hemoglobinexpressing transgenic mice were significantly more susceptible to systemic infection by S. aureus, suggesting that iron availability in human contributes to the host specificity of S. aureus. 99 Preferential binding to human hemoglobin is also observed with other primarily human-associated bacteria, including Staphylococcus lugdunensis, Staphylococcus simulans and Corynebacterium diphtheriae. 99 Gram-negative bacteria Salmonella enterica can cause a wide spectrum of diseases in many hosts. Different serovars of S. enterica display either a restricted host range (e.g., S. Typhi for human, S. Dublin for cattle, swine and human) or a broad host range (e.g., S. Enteritidis and S. Typhimurium for human and many other hosts). The mechanisms of Salmonella host adaptation and specificity have been exclusively discussed in two recent excellent reviews. 2,3 The major factors that affect the host specificity of Salmonella serovars include differences in the bacteria themselves (e.g., the ability to survive host immunity, to grow in a given host environment, and to transmit to other individuals of the same host species) and host environments (e.g., pH, temperature, immune recognition and response, and the microbiota of a given host). 153,154 Pathogenic E. coli strains can cause numerous intestinal or extraintestinal infections. E. coli strains from human and domestic animals/ birds often display host specificity. 155 The host specificity of pathogenic E. coli can be influenced by the molecular interactions of bacterial surface-exposed proteins with host cellular receptors. [155][156][157][158] Septicemic E. coli strains from chickens and lambs express host-specific adhesins, such as avian-specific AC/I pili and lamb-specific K99 fimbriae. [155][156][157] have demonstrated that the distinct recognition of Gala 1-4Gal-containing glycolipid receptors on host cells contributes to the host specificity of uropathogenic E. coli.

EMERGING PRINCIPLES
In contrast to viruses, bacteria clearly depend on much more complex molecular interactions with their natural hosts to determine host specificity. For this reason, the existing information is not sufficient to paint a full picture for the molecular mechanisms of bacterial host specificity, even for the most extensively studied bacterial pathogens (e.g., L. monocytogenes and pathogenic Neisseria). However, several principles have emerged from previous studies. As illustrated in Figure 3, the strong selectivity for humans of bacterial pathogens can be explained by the following: (i) specific recognition of a human receptor by pathogens for colonization and/or dissemination; (ii) specialized ability to evade or overcome immune mechanism(s) for bacterial survival; and (iii) availability of an essential nutrient for bacterial growth in humans.

Colonization and dissemination
Certain surface-associated or secreted molecules contribute to host specificity by promoting bacterial colonization and/or dissemination. As well-characterized examples for bacterial adhesion and colonization, the pili of pathogenic Neisseria 14-16 and E. coli [155][156][157][158] mediate bacterial adherence to mucosal epithelia in a host-specific manner. The surface proteins of S. pneumoniae (CbpA) and S. pyogenes (M protein) bind to their human receptors for mucosal adhesion and dissemination. 42,105 The surface-exposed proteins InlA and InlB of L. monocytogenes promote bacterial invasion across cellular/tissue barriers for dissemination into remote sites (Table 1). 5 The streptokinase secreted by S. pyogenes promotes bacterial dissemination by fibrinolysis through interacting with human plasminogen. 129 Immune evasion Evasion of host immune mechanisms, particularly complement-and IgA1-mediated immunity, is a common mechanism underlying the host specificity of many mucosal colonizers in humans. This is exemplified by the host-specific interactions with human C4BP, a negative regulator of the classical complement pathway, of many bacteria such as N. gonorrhoeae, 19 N. meningitidis, 89 S. pyogenes 134 and S. pneumoniae. 94,95 Similarly, many pathogens interact with negative regulators of the alternative complement pathway CD46 14,42 and FH 53,54,80 in a human-specific manner. Overcoming human IgA1-mediated immunity by bacterial proteases is a common mechanism underlying the host specificity of bacteria that naturally colonize the mucosal surfaces of the reproductive tract (e.g., N. gonorrhoeae) or the upper airway (e.g., S. pneumoniae, N. meningitidis and H. influenzae) (Table 1). 13 These lines of evidence imply that complement-and IgA-mediated immunity is crucial for host defense against these pathogens in non-permissive hosts other than humans.

Nutrient acquisition
Bacterial preference for host-specific nutrients is another common mechanism underlying host specificity. Although the precise nutrient  Host specificity of bacteria X Pan et al 6 requirements of individual bacteria in their hosts are complex, iron is one essential nutrient for bacterial growth in vivo. 97 Many bacteria acquire iron from host iron-containing proteins by expressing specific receptors for these host proteins. 97 Thus, preferential interactions of pathogenic bacteria with human iron-containing proteins have been demonstrated to contribute to the host specificity of N. meningitidis (transferrin and lactoferrin), 159 N. gonorrhoeae (transferrin) 159 and S. aureus (hemoglobin) (Table 1). 99,152 Similarly, PspA-mediated binding to lactoferrin contributes to the human tropism of S. pneumoniae, although this notion remains to be validated by in vivo studies.

PERSPECTIVES
Recent technologic advancements in the transgenic expression of human genes in model animals have led to the generation of many 'humanized' mouse strains for the study of human-specific infectious diseases. 160 As discussed above, some of the transgenic animals have been used to evaluate the contributions of species-specific biochemical interactions to the host specificity of individual bacterial pathogens. These include mouse strains expressing human E-cadherin for L. monocytogenes, 144 human plasminogen for S. pyogenes, 128 human CD46 for N. meningitidis 10 and S. pyogenes, 51,52 human transferrin for N. meningitidis, 9 and rats expressing human FH for N. meningitidis. 84 Furthermore, as exemplified in L. monocytogenes, 126 the increasing availability of structural information on many pathogen-host molecular interactions has made it possible to engineer murinized bacterial strains that can simulate human disease in mice. Therefore, the revelations regarding species-specific pathogen-host interactions in previous studies have already established a foundation for the establishment of genetically engineered animals and bacterial strains for studying the host specificity and pathogenesis of obligate human pathogens in the future.
Finally, insightful information on the molecular mechanisms underlying bacterial host specificity may be applied to the control of human infections caused by bacterial pathogens. For example, the host specificity determinants of bacterial pathogens may be targeted for therapeutic and vaccine development in the future because these proteins tend to be highly conserved within individual bacteria. This is exemplified by the recent development of meningococcal vaccines based on the FH-binding protein of N. meningitidis. In this case, selectively inactivating the FH-binding activity of the FH-binding protein led to enhanced immunoprotective efficacy of the vaccine. [86][87][88] Because the CbpA, PspA and IgA1 proteases are among the protein antigens that are immunoprotective against S. pneumoniae infections, 161,162 selective mutations in the functional sites of these proteins may be used to improve the protective efficacies of these proteins as vaccine candidates.