Regulation of cra, A Regulatory Gene of Glycolytic and Gluconeogenic Pathways in Salmonella enterica serovar Typhimurium

The Cra protein is a global regulator of carbon and energy metabolism for glycolysis and gluconeogenesis. The live oral vaccine candidate, Salmonella typhimurium SR-11 Fa£f, is llllable to utilize gluconeogenic substrates as sole carbon sources and is avirulent and protective in BALB/c mice. Furthennore, the era gene is interrupted in this strain of Salmonella. The era gene was suspected to be regulated by another global regulator, acetyl phosphate. To further investigate the regulation ofthe era gene and to determine if gluconeogenesis is linked to virulence, mutations in the gluconeogenic genesjbp (fructose-1,6-bisphosphatase), maeB (NADP-dependent malic enzyme), and sfeA (NAD-dependent malic enzyme) were constructed in S. typhimurium SR-11. A mutation in the pta (phosphotransacetylase) gene was also constructed to interrupt acetyl phosphate synthesis. Virulence assays in BALB/c mice were performed with these mutant strains. A era promoter-lacZ transcriptional fusion was also inserted into the chromosome of these mutant strains to assay for the era promoter activity during growth on various glycolytic and gluconeogenic substrates. The SR-1 lfbp-, SR-11 maeE, SR-11 sfeAmutant strains were virulent in BALB/c mice, whereas the ptamutant strain was avirulent. The mutation in the maeB gene slightly down-regulated the era promoter activity when grown on either gluconeogenic or glycolytic substrates versus the SR11 cradl craz integrant control strain. A mutation in either the sfeA gene or the pta gene significantly (P = 0.05) up-regulated the era promoter activity for growth·on both gluconeogenic and glycolytic substrates versus the control strain. This up-regulation of era promoter activity, combined with their virulence in BALB/c mice, suggest that the sfrA-encoded NAD-dependent malic enzyme and acetyl phosphate act in concert as a repressor of the era gene. It also appears that the era gene may be subject to regulation by multiple regulatory proteins or multiple forms of regulation. Acknowledgements A sincere thank you to: Dr. Paul S. Cohen for his summer stipends, funding for supplies, guidance, and time. To my committee members, Dr. David C. Laux, Dr. David R Nelson, Dr. Richard C. Rhodes III, and Dr. Alison Roberts for their time, guidance, patience, and effort. Dr. Jay F. Sperry for enduring yet another dissertation defense and the funding for supplies to finish my research project. Dr. Richard E. Koske for his friendship. Mary P. Leatham for her guidance with lab protocols and cloning expertise. Dr. Steven M Denkin for his help, infectious enthusiasm, and countless laughs during our journey through "gradual school." Richard G. Allen for his assistance and expertise using Adobe® Photoshop®. Catherine M Miranda for her love and inescapable endless support. George and Dolores Allen for their love and support during the final phase of this journey. Regina L. Miranda, the "new" graduate student assigned to me by Dr. Cohen for training: I fell in love and married her! I would never have finished without her love, understanding, and gentle reassurance.


Introduction
General introduction. Salmonella species cause a variety of foodbome and waterborne illnesses ranging from localized gastroenteritis to systemic diseases such as typhoid fever in both humans and animals (1-3). The life-threatening diseases such as enteric fever, septicemia, and the focal infections, osteomyelitis and meningitis, caused by Salmonella involve invasion of the bacteria into the blood, the reticuloendothelial system and other organs (3).
How does Salmonella wreak such havoc in humans and animals? This facultative intracellular parasite, containing a 50-90 kbp virulence plasmid, has evolved an arsenal of toxins and mechanisms of pathogenesis to invade and destroy euk:aryotic cells, while evading the immune system of the host cell (4)(5)(6). Fundamentally, the outer membrane of this motile gram-negative bacterium contains lipopolysaccharide (LPS) (7,8). This endotoxin, specifically the lipid A component, is pathogenic in humans and other mammals (8,9).
Most Salmonella infections result from the ingestion of contaminated water or food (2). The ingested microorganisms proceed to the intestinal tract (2). The bacteria then adhere to specialized small intestinal epithelial cells, called microfold cells (M cells), via long polar fimbriae ( 10). These facultative intracellular parasites produce a breach in the intestinal wall via the M cell and ultimately reach the lamina propria of the Peyer's patches (2,10,11). At this site, the bacteria may replicate and establish a local infection, or they may be ingested by macrophages which may disseminate these microorganisms to deeper tissues such as regional lymph nodes, the liver, and the spleen to establish a systemic infection (2). Salmonella possesses the ability to not only survive in and kill macrophages, but also to replicate within them and other constituents of the reticuloendothelial system (2)(3)(4)(5)10).
An estimated 1.4 million cases of salmonellosis, resulting in approximately 500 deaths, occur annually in the United States (12,13). Half of the sahnonellosis cases are caused by two serovars: Salmonella enterica serovar Typhimurium and serovar Enteritidis (12,14). The health care costs associated with human salmonellosis caused by eating contaminated eggs and poultry is estimated at $4 billion annually (15). Typhoid fever, a disease caused by the bacterium Salmonella enterica serovar Typhi, still remains a serious public health problem in developing countries with 16 million cases of typhoid fever causing an estimated 600, 000 deaths annually ( 16,17). Scilmonella typhi appears to be a human-specific pathogen, mtlike S. typhimurium, which colonizes many animals as well as humans (18).
Poultry production also is significantly reduced by Salmonella bacteria (15). The U.S. poultry industry hatches approximately 7.5 billion eggs annually in incubation facilities, and nearly $77 million is lost each year due to Sa1monella outbreaks in poultry farms (15). Vaccination of chickens against Salmonella would reduce losses to the poultry industry and poultry-associated human salmonellosis.
Animal models. Although S. typhimurium can cause disease in numerous non-human animals, many animal carriers of S. typhimurium do not show any signs of disease (19). The rhesus monkey is the animal that most resembles htmlans in its response to a S. typhimurium infection ( 19).
The guinea pig model also appears to closely mimic the disease in humans (19). Yet, these animal models are not used by most investigators ( 19). Most investigators use mice are routinely used as the animal model for reasons of cost and convenience ( 19). More importantly, the disease caused by S.
typhimurium in mice is not the gastroenteritis observed utilizing the other animal models ( 19). The disease caused by S. typhimurium in mice mimics typhoid fever in humans and is well accepted as a model for human typhoid (20). Whatever the reason for the difference between mice and primates, htllllans and other primates usually control the infection within a week, whereas in mice the bacteria spread to the liver and spleen before the animal can mount an effective protective response (21 ).
Acid tolerance response. After ingestion of Salmonella, the acid tolerance response allows the microorganisms to survive passage through the acidic environment of the stomach, a prerequisite for infection (22,23). S. typhimurium can survive well in vitro at pH values down to about pH 4, but below that rapid death of the bacteria is observed (23). However, cells given a chance to adapt (growth for one generation at pH 6), can survive at pH values as low as pH 3 for prolonged periods. The acid tolerance response appears to be regulated by the Fur protein (22,23). Fur is required for the acidinduced activation of atr genes and the expression of iron uptake systems under iron-limiting conditions (22). Fur appears to be sensing and responding to pH as well as iron (23).
More recently, another acid tolerance response has been identified (23). This response is controlled by the PhoP-Ph~ two-component regulatory system, which regulates the expression of many of the virulence genes in S. typhimurium (23). The acid tolerance response seems to be coregulated with the modulation of genes in helping the bacteria survive and invade once they reach the intestine. This is the first time such a clear connection between the acid tolerance response and the general virulence response has been found (23).
Two-component regulatory systems. Two-component regulatory systems (phosphorylationdependent signal transduction systems) act like switches (24 ). One component, referred to as the sensor (such as the periplasmic sensor Ph~, is a histidine kinase protein (24,25). This protein binds ATP and is autophosphorylated at a conserved histidine residue. The second component of the switch formed by two-component regulatory systems is known as the response regulator (such as PhoP) (24,25). The response regulator protein contains a domain that is transiently phosphorylated on a conserved aspartyl residue (24 ). Phosphoryl groups are transferred from the histidine residue of the sensor to the conserved aspartyl residue of the response regulator, which activates the response regulator protein (25). The activated response regulator protein is involved in the regulation of transcription of genes (25).
The expression of many of the virulence genes in S. typhimurium is regulated by PhoP-Ph~ in response to changes in pH and the concentration of divalent cations such as Mg2+ and Ca 2 + (25).
Limiting concentrations of extracellular divalent cations activate the system, resulting in the net phosphorylation of PhoP by Ph~. Phosphorylated PhoP interacts with the promoters of PhoP-Ph~ regulated genes and activates or represses transcription of these genes (24,25).
Examples of PhoP-Ph~ gene regulation include the activation of mgtA, a gene encoding a high affinity Mg2+ transporter and activation of the hi/A and sir A genes, which are also involved the regulation of Salmonella virulence genes (25,26). PhoP-Ph~ is thought to repress Salmonella pathogenicity island 1 (SPI-1; large segments of DNA found in S. typhimurium but not in the nonvirulent E. coli strain K 12) invasion genes while activating expression of pags (PhoP activated genes) (22,26). These pags include the mgtCB operon, which also encodes a high affinity Mg2+ 3 uptake system, and genes on SPI2 (22). PhoP-PhoQ has also been implicated in modulating another putative two-component signal transduction system: PmrA-PrmB. PmrA binds to and activates promoters of certain pags, including several whose products alter LPS in vitro. Members of the PmrA regulon, includingpmrCAB itself: are activated by PhoP-PhoQ (22).

Adherence and invasion.
After passage and survival through the mouse stomach, S.
typhimurium reaches the ileum, a section of small intestine attached to the cecum (21)(22)(23). Short-chain fatty acids in the distal ileum provide the signal for productive infection by Salmonella (27). The rising concentration of acetate and the low concentration of oxygen in the distal ileum provide the signal for invasion gene expression (27,28).
S. typhimurium has a preference for the Peyer's patches in mice (21). Peyer's patches belong to a group of organized lymphoid tissues known collectively as the mucosal-associated lymphoid tissue which defend vulnerable membrane surfaces (11). The mucosal-associated lymphoid tissue carries an extremely large population of antibody-producing plasma cells. Lymphoid cells are found in three regions of the vulnerable mucus membrane which lines the gastrointestinal tract. The outer mucosal layer contains intraepithelial lymphocytes: T-cells which express unusual T-cell receptors and exhibit limited diversity for antigens (11). Below the outer mucosal layer is the lamina propria, which contains large numbers of B cells, plasma cells, activated helper-T cells, and macrophages in loose clusters. Below the lamina propria, still within the submucosal layer, are nodules consisting of30-40 organized lymphoid follicles, called tlie Peyer's patches (11 ).
The epithelial cells of mucous membranes play an important role in promoting the imrrume response by delivering small samples of foreign antigen from the digestive tract to the tmderlying mucosal-associated lymphoid tissue (11). This antigen transport is carried out by specialized M cells, which contain a deep invagination in the basolateral membrane filled with a cluster ofB cells, T cells, and macrophages (10,11 ). Antigens transported across the mucous membrane by M cells activate B cells. B cells differentiate into plasma cells which secrete the IgA class of antibodies. These antibodies are then transported across the epithelial cells and released as secretory IgA into the hunen where they can interact with antigens present in the lumen (11 ). It is the M cell that is the major site of invasion for S. typhimurium in mice (11,21,29). Other intestinal cells, called enterocytes, are the minor site of invasion for Salmonella in the murine model (28,29).
S. typhimurium produces a number of adhesins for entry via the M cell (30). These include type 1 funbriae (encoded by fim genes), plasmid-encoded fimbriae (pefgenes), long polar fimbriae (/pf genes), and thin aggregative fimbriae (curli; encoded by agf genes). The binding specificity of type 1 fimbriae is tmknown (30). The pef genes are located on the 90 kbp plasmid, called the virulence plasmid (pSL n, which is found in all virulent strains of S. typhimurium. These plasmid-encoded funbriae mediate binding of the bacteria to the microvilli of enterocytes (30). The long polar fimbriae mediate attachment to the Peyer's patches. The thin aggregative fimbriae called curli may aid in attachment to the microvilli of enterocytes, but they also cause the bacteria to become attached to each other (30). The gene rck encodes a surface protein which acts as an adhesin and may be involved in invasion of tissue culture cells. The surface protein also increases the resistance of S. typhimurium to killing by complement (30).
Shortly after coming into close contact with the host cell, Salmonella induces profound changes in the brush border of the intestinal epithelium, characterized by the denaturation ofmicrovilli at the point of contact between the bacteria and the host cell membrane (31 ). Interactions of S.
typhimurium with M cells also leads to membrane ruffi.ing (31 ). Salmonella forces the host cells to engulfit (32). Membrane ru:ffiing and internalization of the bacteria within a membrane-bound vesicle are accompanied by extensive actin-rriediated cytoskeletal and cell surface rearrangements in the vicinity of the invading bacteria (10,32).
Once internalized, S. typhimurium destroys the M-cell and enters macrophages in the mesenteric lymph follicles (10,33,34). Macrophage survival is now important for Salmonella because it enables the evasion of the imrmme system (2,10,22). Salmonella-infected macrophages may disseminate to deeper tissues such as regional lymph nodes, liver, and spleen to establish a systemic infection (2,10,22). Salmonella possesses the ability to not only survive in and kill macrophages, but also to replicate within them and other constituents of the reticuloendothelial system (2-5, 10, 22). For example, S. typhimurium can enter and survive within hlllllan B and T cells, which could play a role in the dissemination of infection (31 ). Furthermore, Salmonella can gain access to non-phagocytic cells of the liver and spleen, which may constitute a "safe site" for replication during the early phases of systemic infection (31 ). Overall, the main sites of replication during systemic infections in mice are the liver and spleen (35). Jn vivo studies have shown that S. typhimurium resides in CDl 8-positive leukocytes in the liver and replicates in red pulp and scavenger receptor-expressing marginal zone macrophages in the spleen (35).
Upon reaching the liver and spleen, the bacteria must continue to survive and replicate in order to cause disease in mice (22). Several genes whose products alter LPS structure and promote polymyxin resistance in vitro may aide S. typhimurium resist killing by PMN-and macrophageproduced cationic peptides at systemic sites. Also, genes on the Salmonella virulence plasmid are required for bacterial replication in these tissues (22). As the bacteria proliferate, they produce lipid A, promoting inflammatory cytokine and inducible nitric oxide responses (iNOS) that may kill the murine host (9,22).
Virulence gene clusters. Many clusters of virulence genes are found in S. typhimurium including three Salmonella pathogenicity islands (SPI-1, 2, 3), two lysogenic phages (Gifsy-1, 2), and a 90 kb virulence plasmid (pSL D (22,36). In addition to the genes which code for a type III secretion system and its secreted proteins, SPI-1 contains genes which code for chaperones that control folding of proteins, and regulatory proteins (PhoP-PhoQ, HilA, InvF., SirA). SPI-2 carries genes for a second type III secretion system along with the genes encoding the proteins secreted by the second type III secretion system (22). Genes encoding chaperones that control protein folding are also found on SPI-2 (36). SPI-2 is involved in inhibition ofphagosome-lysosome fusion and is important during the systemic phase of disease (36). SPI-3 carries genes encoding a high affinity Mg2+ transporter and proteins oftmknown function (26,36). SPI-3 may aid in the survival and growth of Salmonella in the Mg2+-poor vacuole of the invaded cell (36).
The virulence plasmid (pSL n carries a gene for fimbrial adhesion (pef), an adhesion/sennn resistance gene (rck), and genes coding for Spv proteins. The SpvB protein is a toxin that ADPribosylates actin (36,39). Other spv genes encoded within pSLT are required for bacterial replication in the reticuloendothelial system (39). S. typhimurium carries the conjugal transfer gene traTmaking it self-transmissible (39). The virulence plasmid is important in all phases of infection (36). Type m secretion system required for initial invasion. A type III secretion system is an essential basic virulence determinant utilizing a conserved mechanism of protein secretion (34).
However, the secreted proteins themselves are highly divergent. Salmonella is the only genus known to possess two type III secretion systems (34). These two type III systems appear to play different roles during pathogenesis: the first being required for initial penetration of the intestinal mucosa and the second necessary for subsequent systemic stages of infection (22,34).
The inv genes, along with other genes on SPI-1 (spa, prg, and org), encode the first type III secretion system (26,28). The inv genes on SPI-1 are responsible for the membrane rufiling associated with invasion of cells by S. typhimurium (26). SPI-1 also carries the genes that encode the proteins that are injected into the eukaryotic cell by the type in secretion system (26). These genes include sptP, which encodes a tyrosine phosphatase that mimics signal transduction enzymes of eukaryotic cells and may play a role in altering the response of mucosal cells to outside stimuli (26).
SopE and SipA are additional proteins injected into the mucosal cells via the type III secretion system (26,34). These proteins are thought to be responsible for the ruflling response (26). SopE can activate host cell G proteins such as Cdc42 and Rae, which control actin polymerization. Activation of these proteins by SopE initiates the actin rearrangements involved in the ruffling response. Ruffling is a very localized phenomenon; localization of the host cell's response is thought to be the role ofSipA (26). SipA binds actin directly and inhibits depolymerization, thus increasing the amount of 7 polymerized actin in its vicinity. SipA is confined to the immediate area in which the bacteria are found, accounting for the localized nature of the ruflling phenomenon (26).

Survival of Salmonella in macrophages.
Macrophages are phagocytic cells capable of ingesting and digesting exogenous antigens such as whole bacteria ( 40). S. typhimurium enters host macrophages and can induce either an almost immediate cell death or establish an intracellular niche within the phagocytic vacuole ( 41 ). Rapid cell death depends on the SipB effector on Salmonella pathogenicity island SPI-1 and the host protein caspase-1 (35,41). Caspase-1 is a member of the proapoptotic caspase family of proteases, and the process was originally thought to be apoptotic (3 5, 41 ).
Recent studies suggests that it is an tmusual form of necrosis (35 After uptake, Salmonella resides within a unique organelle, the Salmonella-containing vacuole (SCV; Salmonella-containing phagosome) in which it eventually replicates (43). Studies of the S. typhimurium vacuole in macrophages have been hampered by its heterogeneous behavior (35).
Infection of primary macrophages or a macrophage cell line by a culture of s. typhimurium results in variable numbers of bacteria in each host cell (35). For example, twelve hours after bacterial uptake, some macrophages contain clusters of numerous bacteria, while in others only a few dispersed bacteria are visible (35). This probably reflects the simultaneous processes of growth and killing that occur in murine macrophages (35).
There are also conflicting results regarding the fusion of the SCV with lysosomal compartments (35). Two groups concluded from studies involving primary macrophages that SCV s fuse with lysosomal compartments (35). Conflicting evidence suggests that Salmonella can inhibit phagosome-lysosome (SCV-lysosome) fusion in both primary macrophages and macrophage-like cell 8 lines (35,44). Despite this, a consensus has emerged over the last few years that is generally consistent with the current llllderstanding ofSCVs (35). According to this, the majority ofSCVs acidify but fail to acquire lysosomal hydrolytic enzymes or reactive oxygen intermediates that would normally accumulate in a phagolysosome (35). There is evidence that both the PhoP-PhoQ twocomponent regulatory system and SPI-2 type III secretion system play important roles in this process, which leads to the establishment of a compartment which is conducive to bacterial replication (35).
Replicating bacteria remain in membrane-bound vacuoles, and this requires a continuous supply of membrane to enclose dividing bacterial cells. The SPI-2 effector protein SifA, and the assembly of an actin mesh work, play major roles in this aspect of intracellular growth (35).
Genes required for the survival and replication of Salmonella in macrophages. The interior of a macrophage can be a very inhospitable environment for a bacterium. However, S.
typhimurium has acquired an arsenal of genes involved in the survival and replication inside macrophages (35). Many of the genes' fimctions are still unknown. Genes encoded within SPI-2 involved with the SPI-2 type ill secretion system and their function are as follows: ssaC (subunit forming the outer membrane secretin porin), ssaV(secretion ofSPI-2 proteins), sseA (required for survival and replication), sseBCD (translocation of SPI-2 effectors), sseFG (interaction of phagosomal membranes after translocation; contributes to Sif formation in epithelial cells), sse.J (regulates the dynamics of the SCV membrane), ssrAB (two-component s~tem regulating SPI-2 gene expression), spiC (inhibition of trafficking; actin polymerization), and NC (the gene has not been characterized but is associated with SPI-2 secretion and involves inhibition of the oxidative burst) (35,(45)(46)(47)(48).
Survival and replication of Salmonella in macrophages requires other genes associated with the SPI-2 type Ill secretion system, which are encoded outside ofSPI-2 (35). The genes and their function are as follows: ompR-erwZ (two-component system regulating the acid-induced ssrAB virulence operon expression), sifA (contributes to Sif formation in epithelial cells and maintains the SCV membrane), srjK (unknown but regulated by ssrAB), and sse.J (required for translocation of effector proteins) (35,(48)(49)(50).
Many genes, required for the survival and replication of Salmonella in macrophages, are controlled by the PhoP-PhoQ regulon (35). These include mgtC (magnesium acquisition), NC (the gene has not been characterized but it is involved in the inhibition of SCV-late endosomal interactions) andpags (PhoP activated genes; LPS modification and resistance to antimicrobial peptides) (35).
Regulation of virulence gen. es by two-component signal transduction systems. Several two-component signal transduction systems regulate virulence genes in vitro (22). PhoP-PhoQ and PmrA-PrmB are two which have already been discussed in detail (see Acid tolerance response, Twocomponent regulatory systems, and Genes required for the survival and replication of Salmonella in macrophages). Evidence indicates that PhoR-PhoB, which regulates genes in response to inorganic phosphate {PJ concentration, can repress SPI-1 invasion genes (22). PhoR-PhoB may also activate SPI-2 genes because their expression is induced by low Pi levels in vitro. OmpR-EnvZ, a twocomponent regulatory system that modulates gene expression in response to osmotic conditions and pH, may also regulate virulence genes because disrupting this regulatory pathway decreases SPI-1 gene expression (22,50). Furthermore, recent evidence indicates that OmpR may play a role in regulation ofSPI-2 genes (10,22). Two other two-component signal transduction systems, SsrA-SsrB and BarA-SirA, respond to l.ll1known environmental cues (22). SsrA-SsrB regulates SPI-2 gene expression of the SPI-2 type ill secretion system as well as its translocated effectors; SsrA-SsrB is in tum regulated by the OmpR-EnvZ two-component system (22,50,56). BarA-SirA is necessary for full expression of invasion genes on SPI-1 and SPI-4 (22~ 27). Another two-component system, CsrA-CsrB, alters RNA stability which may modulate invasion gene expression (22). Finally, the superoxide radical response regulon, SoxR-SoxS, can upregulate the expression of the Tef gene, which confers resistance to reactive oxygen and nitrogen intermediates via NADPH (53).
Regulation of virulence genes by alternative sigma factors. The alternative sigma factor, RpoS, which controls many genes expressed in stationary phase and for stress protection, also regulates virulence genes (57,58). RpoS is necessary for sustaining a log-phase acid tolerance response, induction of a stationary-phase acid survival system, production of thin aggregative fimbriae, and expression of spv genes during stationary phase (22). The rpoE gene, which encodes the extracytoplasmic stress response sigma factor sigmaE (RpoE), is critically important for the virulence of S. typhimurium ( 59). RpoE regulates many genes required for survival in macrophages and proteins involved in protein folding and/or degradation in the periplasm (53,54,59).
Other mechanisms of regulation for virulence genes. The DNA adenine methylase enzyme (dam), which modifies GATC sequences, has pleiotropic effects on the expression of many different virulence genes (22,60,61). To begin with, Dam methylation regulates the expression ofplasmidencoded fimbriae (pef) encoded by the pSL T virulence plasmid (22,36,61 ). The conjugative transfer of this virulence plasmid, which requires thefinP-encoded F-type pili, is regulated by the levels of Dam methylation; the transfer of the pSLT virulence plasmid is elevated by low levels of Dam methylation (61). A dammutation in S. typhimurium revealed reduced secretions of invasion effectors encoded in SPI-1 genes (22,62). In addition, a reduction in the relative amount of peptidoglycanassociated lipoprotein, OmpA (a highly immunogenic, non-specific, outer membrane porin, and murein lipoprotein bound to peptidoglycan was observed in actively growing Dam-mutants, indicating increased cell envelope instability (62). The Danf mutant was wiable to proliferate in target organs in vivo but persisted in low numbers (63). Use of the Heal loop assay revealed that Dam· mutants were less cytotoxic to M-cells and failed to invade enterocytes (63). In the tissue culture model, lack of DNA adenine methylation reduced the ability of the bacteria to invade non-phagocytic cells (63).
Collectively, these results indicate that DNA methylation plays several roles in regulating virulence genes in S. typhimurium (22,62,63).
Other regulators specifically modulate the expression of virulence genes whose products are thought to have related fimctions (22). For example, the regulators RtsA, HilA, InvF, HilC, HilD, and HilE directly or indirectly modulate the expression of genes encoding the SPI-2 type III secretion system and its secreted effectors (22,(64)(65)(66). FimZ and AgfD are two other regulators believed to modulate specific virulence genes of related fimction and are required for in vitro expression of type I :fimbriae and thin aggregative fimbriae (22). .
Many regulators control the expression of multiple virulence genes (22,67). The global regulator CRP ( catabolite repressor protein) regulates a variety of genes in Salmonella in response to the levels of cAMP (67). CRP negatively regulates the spv operon on the Salmonella virulence plasmid, while positively regulating various fimbrial operons (67). ThefliZ gene positively regulates the expression of class II flagellar genes and induces SPI-1 invasion gene expression (22).
These twelve precursor metabolites, except acetyl-CoA, are synthesized by a series of central fueling pathways that are collectively called central metabolism (72,73). Central metabolism includes the Embden-Meyerhof-Pamas (EMP) pathway, which converts glucose-6-phosphate to pyruvate; the tricarboxylic acid (TCA) cycle, which oxidizes acetyl-CoA to C02; and the pentose phosphate cycle, which oxidizes glucose-6-phosphate to C0 2 ( Figure 1) (72)(73)(74). Acetyl-CoA is synthesized by a linker reaction between the EMP pathway and the TCA cycle (72). The Entner-Doudoroffpathway is also utilized by S. typhimurium to metabolize gluconate ( Figure 1) (72,73). Regardless of the fueling reactions that S. typhimurium employs, these twelve precursor metabolites are synthesized and are the metabolic link between fueling and biosynthesis ( 68).
The precursor metabolites, along with inorganic ions, are required for the biosynthesis of cellular components (75,76). For example, glucose-6-phosphate yields the sugar backbone of nucleotides, and fructose-6-phosphate is converted to amino sugars (69). Ribose-5-phosphate is the precursor metabolite for purine and pyrimidine nucleotides as well as the heptose component of LPS (67,74). Erythrose-4-phosphate is utilized for the biosynthesis of aromatic amino acids (69,76).
Phospho( enol)pyruvate is the building block for vitamins and cofactors as well as aromatic amino acids. Pyruvate is utilized for the biosynthesis of the amino acids alanine, valine, leucine, and isoleucine (69,76). Acetyl-CoA is the precursor metabolite ~or fatty acids and the outer membrane component murein. The amino acids glutamate, glutamine, arginine, and proline are all synthesized from a-ketoglutarate (67,74). The precursor metabolite succinyl-CoA is used to form heme (76).
Finally, the oxaloacetate is the precursor metabolite for the biosynthesis of the amino acids aspartate, asparagine, threonine, methionine, and isoleucine (68,69).
There is tremendous flexibility in the way the central fueling pathways operate (72). For example, formation of the precursor metabolites during aerobic growth of S. typhimurium on a limiting source of glucose utilizes a major anapleurotic reaction (74,(77)(78)(79). Anapleurotic reactions are the interconnecting, reversing, and bypassing reactions which replenish the pools of precursor metabolites drained by biosynthesis (68). This anapleurotic reaction, catalyzed by the pJX-encoded enzyme 13 phospho( enol)pyruvate carboxylase, forms oxaloacetate by carboxylation of phospho( enol)pyruvate ( Figure 1) (77). Components of the TCA cycle function almost exclusively to provide three precursor metabolites, not as an energy-generating cycle (77).
Aerobic growth of S. typhimurium on a gluconeogenic substrate such as malate is extensively different from growth on a glycolytic substrate such as glucose: TCA components are shunted off into the EMP pathway to form the required precursor metabolites (77). Two routes lead from malate to pyruvate in the EMP pathway; the maeB-and efcA-encoded malic oxidoreductase enzymes catalyze these redundant reactions ( Figure 1) (71,77,80,81). Oxaloacetate is converted to phospho(enol)pyruvate and also shunted off into the EMP pathway ( Figure 1) (77). A reversal of flow in the EMP pathway, called gluconeogenesis, occurs (72,73,77,82). Gluconeogenesis leads to the pentose phosphate cycle and both produce the remaining precursor metabolites (72,77). Tue total cost of fueling reactions operating to produce precursor metabolites is much greater for a gluconeogenic substrate, such as malate, than for a glycolytic substrate such as glucose (83). Therefore, growth on gluconeogenic substrates, such as malate, is slower than growth on glycolytic substrates such as glucose (77).
Levels of nutrients and oxygen at specific locations in the S. typhimurium infection process. The level of nutrients are high and the oxygen level is very low in the lumen and mucus of the ileum (28,33,84). Inside the M cell, or epithelial cell vaeuole, the nutrient and oxygen levels are both low (84). Inside macrophages, the level of nutrients is low and the oxygen level is moderate (35,84 ). The spleen contains high levels of nutrients and oxygen (28,85). S. typhimurium replicates in the CD18-positive leukocytes in the liver which are bathed by hepatic blood vessels (35,85). It is assumed that both the levels of nutrients and oxygen are moderate to high. The levels of oxygen appear to be from microaerophillic to aerobic at all the specific locations of the S. typhimurium infection process (28,33,35,85).
G lycolysis and the utilization of glycolytic substrates. Glycolysis, also called the Embden-Meyerhof-Parnas (EMP) pathway, is the metabolic pathway utilized to break down glucose to pyruvate (73,(87)(88)(89). Four important events occur during glycolysis; { 1} substrate-level phosphorylation which is the synthesis of adenosine triphosphate (ATP) by the donation of a high-energy phosphate to adenosine diphosphate (ADP) from a reaction coupled with the exergonic breakdown of a high-energy substrate molecule, {2} the breaking of the six-carbon glucose molecule into two three-carbon pyruvate molecules, and {3} the transfer of two electrons to the coenzyme nicotinamide adenine dinucleotide (NAD~ to form NADH + H + (87,89).
The usual route for glucose uptake and phosphorylation is the phospho( enol)pyruvate phosphotransferase system (P'fS) (73). The PTS transports and simultaneously phosphorylates glucose in a process called group translocation (90). The PTS in S. typhimurium contains general PTS proteins and proteins specific for glucose transport (91 ). The general PTS proteins, enzyme I (EI) and histidine protein (Hpr) are the soluble cytoplasmic proteins that participate in the phosphorylation of all PTS carbohydrates (91). EI and Hpr in S. typhimurium are encoded by the genes ptsH and ptsl respectively (90)(91)(92). The EIIGlc enzyme is glucose-specific, membrane-bound, encoded by the gene ptsG, and consists of two domains: IIC-IIB (90,91,93). The final component, the err-encoded enzyme IIA Glc, is soluble (91,92,94). All of the PTS enzymes and proteins participate, in sequence, in the transfer of a phosphoryl group to glucose (91 ). Glucose phosphorylation is coupled to glucose translocation across the cytoplasmic membrane, the energy for these processes being provided by the glycolytic intermediate phospho( enol)pyruvate (91 ).
For the purposes of this study, a glycolytic substrate is considered any carbohydrate such as glucose, fructose, and gluconate which enters the glycolytic (EMP) pathway.
For the purposes of this study, a gluconeogenic substrate is considered any substrate which enters the gluconeogenic pathway such as acetate, alanine, glycerol, citrate, fumarate, malate, oleate, phospho( enol)pyruvate, pyruvate, and succinate.
The fate of pyruvate in S. typhimurium metabolism. The oxidation of hexoses, such as glucose, to pyruvate by the EMP pathway generates two molecules of pyruvate and two molecules of NADH (127). To maintain glycolytic flux, the NADH must be oxidized to NAD+ (127). One option for regenerating these reducing equivalents in the absence of oxygen is by a process called fermentation (127,128). Fermentation occurs by depositing the reducing equivalents on partially oxidized metabolic intermediates, which are then excreted from the cell (127). The fermentation products comprise a mixture of ethanol, and acetic, formic, lactic, and succinic acids (127). Since the carbon atoms in the metabolic intermediates are only partially oxidized and the difference in reduction potentials between the primary electron donor and terminal electron acceptor is small, the fermentation processes yield little energy (129).
Another option for regenerating these reducing equivalents which occurs in the presence of oxygen is a process called aerobic respiration; this process involves an electron transport system and oxidative phosphorylation (129)(130)(131)(132)(133)(134). In this process which utilizes 0 2 as the final electron acceptor, all the substrate molecules can be oxidized completely to C0 2 , and a far higher yield of ATP is theoretically possible (129)(130)(131). Two hydrogen atoms, each consisting of one proton and one electron, are transferred from NADH to a series of other carrier compounds (NADH dehydrogenases, flavoproteins, nonheme iron-sulfur proteins, quinones, and cytochromes) embedded within the cell membrane (130)(131)(132)(133)(134). This electron transport chain separates the protons from the electrons during the transport process (131,133,134). The protons are extruded into the periplasm; the net result is the generation of a pH gradient and an electrical potential across the membrane, with the inside of the cytoplasm electrically negative and alkaline, and the outside of the membrane electrically positive and acidic (131,133,134). Membrane-bound catalytic A TPases act as proton channels which drive the formation of ATP from ADP plus inorganic phosphate (131,133,134). This ATP-producing process is known as oxidative phosphorylation (131,133,134).
Aerobic respiration via the TCA cycle and oxidative phosphorylation require that pyruvate first be converted to acetyl-CoA (131,133,134). Acetyl-CoA is synthesized by a linker reaction between the EMP pathway and the TCA cycle (72). The pyruvate dehydrogenase multienzyme complex ( aceEF, lpdA), one of the most complicated enzyme systems known, converts pyruvate to acetyl-CoA (92,135). NAD+ is reduced to NADH and COi is generated in the reaction (135).
The TCA cycle and glyoxylate bypass. The TCA cycle is an inducible pathway with the levels of the enzymes responding primarily to the presence of oxygen and to the carbon source(s) available (7 4 ). The levels of the TCA cycle enzymes are also regulated at the transcriptional level by the interaction of two global regulatory systems: catabolite (glucose) repression and the ArcA-ArcB two-component regulatory system ( 136). The full TCA cycle is fi.mctional only during aerobic growth on acetate or fatty acids ( Figure 1) (74). However, growth on acetate or fatty acids requires the induction and fimction of an anapleurotic pathway, the glyoxylate shunt, to replenish the dicarboxylic acid intermediates consmned in amino acid biosynthesis. (74,137,138).
Citrate synthase (gltA), the rate-limiting step in the TCA cycle, converts oxaloacetate plus acetyl-CoA to citrate ( Figure 1) (74,92,139). The synthesis of the enzyme is subject to catabolite repression and induced by 0 2 when acetate is the carbon source (74,139). The enzyme is activated by acetyl-CoA and K+ ions and inhibited by 2-oxoglutarate, ATP, NADH, oxaloacetate, and NAD+ (7 4, 139).
The enzymes encoded by the three genes, sucA, sucB, and lpdA, are required for the 2oxoglutarate dehydrogenase complex ( Figure 1) (74,92,143). The substrate, a-ketoglutarate, is channeled through the catalytic reactions via a classical swinging arm carrying the substrate molecules to each successive active site (74,143). The final products are succinyl-CoA and C02 (74,143).
Succinate is subsequently converted to fumarate by succinate dehydrogenase (Figure 1) (7 4, 145). Succinate dehydrogenase is encoded by four genes: the sdhA gene encodes a flavoprotein subunit containing a covalently bound flavin adenine dinucleotide (FAD) moiety, the sdhB gene encodes an iron-sulfur protein; the sdh. C and sdhD genes encode two very hydrophobic membrane proteins which serve to anchor the hydrophilic flavoprotein and iron-sulfur protein subunits to the cytoplasmic membrane and also participate in electron transport (74,92,145). Succinate dehydrogenase is made llllder aerobic conditions and FAD is also temporarily reduced to F ADH 2 during the reaction (74,145). Enzyme synthesis is regulated by catabolite repression (74). Activation of the enzyme by covalent attachment of FAD to the SdhA enzyme subunit is promoted by intermediates of the TCA cycle (74).
The last reaction in the TCA cycle is the conversion of malate to oxaloacetate by the dimeric enzyme malate dehydrogenase (mdh) (Figure 1) (74,92,149). Enzyme synthesis is subject to catabolite repression and is co-regulated withfumA; mdh andfamA have a common catabolite repressor protein (CRP) binding site upstream (149). NADH is an allosteric inhibitor of malate dehydrogenase (149). NAD+ is also reduced to NADH in this final reaction in the TCA cycle (149).
The glyoxylate bypass in S. typhimurium is an inducible anapleurotic pathway within the TCA cycle required for growth on carbon sources such as acetate or fatty acids (74,137,138). This pathway allows the net conversion ofacetyl-CoA to metabolic intermediates (74). Strains lacking this pathway fail to grow on these carbon sources since acetate carbon entering the TCA cycle is quantitatively lost as C0 2 with no means to replenish the TCA precursor metabolites consumed for amino acid and heme biosynthesis ( Figure 1) ( 68, 69, 7 4 ).

21
Ins. typhimurium, isocitrate dehydrogenase (icdA) is regulated by phosphorylation; the function of this phosphorylation is to control the flow of isocitrate through the glyoxylate bypass ( Figure 1) (74,92,142). During growth on acetate, approximately 75% of the isocitrate dehydrogenase is converted to the inactive phosphorylated form by isocitrate dehydrogenase kinase/phosphatase (74,151). Inhibition of the isocitrate dehydrogenase slows the TCA cycle and this forces isocitrate through the glyoxylate bypass (74,151). Although the glyoxylate bypass can provide metabolic intermediates, the TCA cycle is more efficient at generating energy (74). The cell must, therefore, precisely balances the flux of isocitrate between these two competing pathways during growth on acetate; the energy requirements of the cell appear to be monitored through AMP levels (74). AMP activates the phosphatase moiety and inhibits the kinase moiety ofisocitrate dehydrogenase kinase/phosphatase (7 4, 151 ). Isocitrate dehydrogenase kinase/phosphatase also controls the glyoxylate bypass during transitions between carbon sources (74,151).
The genes which encode the metabolic and regulatory enzymes of the glyoxylate bypass reside in the same operon, aceBAK, containing a single promoter (74,92,(150)(151)(152). The operon is regulated by a repressor encoded by the ic!R gene (7 4, 138). Derepression of the aceBAK operon occurs upon adaptation to growth on acetate or fatty acids, presumably involving some metabolic intermediate; FadR arepressor of fatty acid degradation, mediates a part of the derepression process when the substrate is a fatty acid (74,138). Expression of th~ operon is also upregulated by the histone-like protein, IHF (7 4 ).
The pentose phosphate pathway. In addition to the formation of the precursor metabolites, ribose-5-phosphate and erythrose-4-phosphate, the pentose phosphate pathway may be required to provide NADPH for biosynthesis and survival within phagocytic cells ( Figure 1) (57,68,69,73).
Another fimction of this pathway is the loog route between glucose-6-phosphate to fructose-6phosphate; this would effectively be a cycle for the complete oxidation ofhexose monophosphate to c~ (73). The pentose phosphate pathway is also needed for growth on pentoses and for the portion of gluconate metabolism not utilizing the Entner-Doudoroffpathway (73).
The Entner-Doudoroff pathway. The inducible Entner-Doudoroff pathway in S. typhimurium is employed for the metabolism of gluconate (73). The Entner-Doudoroff pathway can be properly considered as one of three pathways found in nature, in addition to the Embden-Meyerhof-Parnas and pentose phosphate pathways, that feed into the "bottom half' of glycolysis, which is central to all intermediary metabolism ( Figure 1) (164).

The fatty acid (3-oxidation pathway.
Although enzymes in the aerobic fatty acid J3oxidation pathway degrade both long and short chain fatty acids, it is the long chain compounds that induce the enzymes of this pathway (172). The first step in fatty acid degradation is the activation of the free fatty acid to an acyl-CoA thioester by acyl-CoA synthetase ifadD) ( Figure 1) (93,172,173) This initial activation step requires coenzyme A (CoA-SH) and two high-energy phosphate equivalents from ATP per molecule of fatty acid (173). Each turn of the cycle yields an acetyl-CoA, an NADH, and an F ADH 2 (172,173). The cycle utilizes the enzymes acyl-CoA dehydrogenase (yafe/), cis-~3trans-~2-enoyl-CoA isomerase ifadB), enoyl-CoA hydratase ifadB), 3-hydroxyacyl-CoA epimerase ifadB), 3-hydroxyacyl-CoA dehydrogenase (fadB), and 3-ketoacyl-CoA thiolase (fadA) (92, 173, 174l 76). One of the products of J3-oxidation, acetyl-CoA, is fed into the TCA cycle and glyoxylate bypass, which facilitates its conversion to metabolic intermediates (74,137,173). The reducing equivalents produced, NADH and F ADH 2 , can be regenerated in the electron transport chain producing ATP via oxidative phosphorylation (131,133,134,173).
When even numbered fatty acids are broken down, a two carbon compound remains, acetyl-CoA ( 172). When odd numbered fatty acids are broken down, a three carbon compound remains, propionyl-CoA; this is further catabolized by the reactions of propionate catabolism (172).
Unsaturated fatty acids require additional metabolic reactions by an isomerase, an epimerase, and a reductase reactions to enter the acid J3-oxidation pathway ( 172). Expression of the genes of the J3oxidation pathway are usually subjected to strong catabolite repression (173). ThefadBA operon also is negatively regulated by the FadR repressor; acyl-CoA relieves this repression (173).
Fructose catabolism. The same general PTS system as glucose is utilized by fructose for entry into the cell; fructose is simultaneously transported and phosphorylated by group translocation (73,90). The general PTS proteins, EI (ptsll) and Hpr (ptsl) are the soluble cytoplasmic proteins that participate in the phosphorylation of all PTS carbohydrates ( Figure 1) (90)(91)(92). The Ellfru enzyme complex ifru.A,fruB) is fructose-specific, possesses three domains in the FruA protein (IIB'-IIB-IIC) and three domains in the FruB protein (IIA-IIM-IIH) (91,93,177,178). Domains IIA-IIB-IIB' are localized to the cytoplasmic side of the membrane (177,178). The domain IIB' is required for high affinity binding ofFruB to FruA, but does not participate in phosphoryl transfer (177,178). Domain IIA is the first phosphorylation site, IIM is a central domain of unknown function, and IIH is an HPrlike domain called FPr (fructose-inducible HPr) (177,178). Fructose-I-phosphate is the product of this series of reactions (91,177,178).

Growth of S. typhimurium utilizing a non-limiting glucose source. The branched,
biosynthetic form of the TCA cycle utilizing a non-limiting glucose source was separated from the above pathways because of its importance to this study. The TCA cycle is split into a oxidative branch terminating at a-ketoglutarate and a reductive branch terminating at succinyl-CoA ( Figure 2) (74,78).
The oxidative branch functions as described in the section above titled: The TCA cycle and glyoxylate bypass. However, the reductive branch of the TCA cycle functions backwards utilizing the same reversible enzymes except succinate dehydrogenase (sdhCDAB), which is replaced by fumarate reductase (frdABCD) (14, 78, 93, I80).
Fumarate reductase is composed of four subllllits and two domains (I80). The catalytic domain consists of two subunits: one with a covalently-bound flavin cofactor and the fumarate binding site; the other contains three iron-sulfur clusters (I80). This catalytic domain is attached to the cytoplasmic side of the cytoplasmic membrane by the anchor domain, which consists of two subunits that interact with quinone and contain heme (180). The enzyme also has two quinol-binding sites, Qp and ~ (180). The covalent attachment of FAD to the enzyme (A subunit apoprotein) is stimulated by citrate, isocitrate, succinate, and fumarate , acting as possible allosteric effectors (I 80). Fumarate reductase is inhibited by oxaloacetate and malonate (180).
Oxaloacetate is converted to aspartate by aspartate aminotransferase ( aspC) in this branched, biosynthetic form of the TCA cycle ( Figure 2) (74, 93, I8I). Glutamate is also a rea:ctant and 2oxoglutarate a product of this reaction (I 8 I). The dimeric aspartate aminotransferase enzyme is inhibited by 2-methyl aspartate (181). Aspartate ammonia lyase (aspA) next converts aspartate to fumarate and NH 3 (74,92,182). Aspartate ammonia lyase is activated by aspartate and inhibited by citrate and n-propanol (182).
Tue full TCA cycle is not required for growth of S. typhimurium utilizing a non-limiting glucose source because the bulk of energy is derived from glycolysis (7 4, 78). In this glucose-rich environment, cells produce excess acetyl-CoA which drains through acetyl phosphate, to further produce ATP with the associated secretion of acetate ( Figure 2) (78,79). As cell density increases, the build-up of acetate also serves to increase the size of the acetyl phosphate pool (78).

Regulation by catabolite repression.
Catabolite repression is often called the glucose effect because glucose, which yields the highest return of ATP per unit of expended energy, usually strongly represses operons for the utilization of other carbon sources (183). Catabolite repression in S.
typhimurium involves the cytoplasmic sensor of carbon and energy concentrations, cAMP, and the dimeric CRP protein; cAMP binds to CRP at specific DNA sequences in cAMP-CRP-sensitive promoters, induces bends in the DNA, and interacts with RNA polymerase to promote transcriptional initiation (184,185). However, the mechanism of cAMP-CRP regulation varies; CRP can fimction not only as an activator, but also as a repressor depending where it binds relative to the promoter (184,185).
The synthesis of cAMP is controlled through the regulation of activity of adenylate cyclase (186). This enzyme, which catalyzes the formation of cAMP from ATP, is more active when cellular concentrations of catabolites are low and less active when catabolite concentrations are high ( 186).
The synthesis of cAMP is regulated by a protein phosphorylation mechanism that is catalyzed by the PTS system (184). PTS penneases fimction as transmembrane signal transduction devices: when one of the PTS sugars (i.e. glucose, fructose, mannitol) such as glucose is present extracellularly, the crrencoded enzyme IIA Glc protein becomes dephosphorylated as the phosphoryl groups are transferred to incoming sugar molecules via the sugar-specific permease proteins (91,94,184). This process results in both allosteric deactivation of adenylate cyclase, allosteric inhibition of the non-PTS permeases (ie. lactose, maltose, and glycerol permeases ), and catabolic enzymes that generate cytoplasmic inducers ( 91 , 184). Exogenous glucose both inhibits the synthesis of cAMP and stimulates the efllux of cAMP from the cell cytoplasm (184).
Almost all genes that encode enzymes and transport proteins which initiate the metabolism of an exogenous carbon source are tlllder cAMP-CRP control; the expression of the genes is consequently subject to catabolite repression (184). The levels of the TCA cycle enzymes are regulated at the transcriptional level by catabolite repression (136). Phospho(enol)pyruvate carboxylase (ppc) is tlllder cAMP control an well ( 184 ). One of the primary responses to a limitation of a specific nutrient is the activation of the cAMP-CRP regulon; this allows higher-affinity uptake of the nutrient present in low concentration or the utilization of an alternative carbon source ( 187).
Many proteins in addition to carbon catabolic enzymes have been shown to be subject to catabolite repression (183). Flagellar synthesis, as well as chemotaxis are subject to cAMP control (184). Finally, genes involved in pH regulation, extracellular macromolecule degradation, ubiquinone synthesis, intracellular glycogen metabolism, nitrogen utilization, organic phosphate ester utilization, thiolsulfate reduction, iron uptake, drug (antibiotic) resistance, colicin induction and recognition, and toxin production are all regulated directly or indirectly by cAMP-CRP (184). Many of these processes are related to carbon metabolism (184).
Regulation of central carbohydrate metabolism via the csrA gene. The carbon storage regulator gene, csr A, exhibits pleiotropic regulation of central Carbohydrate metabolism in E. coli (82).
The regulatory effects of Csr A peak during the transition from exponential phase into early stationary phase (82). The CsrA regulatory protein dramatically affects the biosynthesis of glycogen through its negative control of two glycogen genes, glgC (ADP-glucose pyrophosphorylase) and glgB (glycogen branching enzyme) (82). Intracellular carbon flux is also directed by CsrA in both the glycolytic (EMP pathway) and gluconeogenic pathways (82). CsrA exerts negative regulation of gluconeogenesis with a decrease in the specific activity of phospho( enol)pyruvate carboxykinase (pckA.), phospho(enol)pyruvate synthetase (pps), and fructoste-1,6-bisphosphatase (jbp) (Figure 1) (82).
Positive regulation of glycolysis by CsrA is reported in the increased specific activity of glucosephosphate isomerase (pgi), 6-phosphofructokinase-I (pjkA), triose phosphate isomerase (tpi), enolase ( eno ), and pyruvate kinase I (pykF) (82). In contrast, the expression of genes in the pentose phosphate pathway is weakly or negligibly affected by CsrA (82). S. typhimurium possesses a nearly identical homologue of this csr A gene (92,188).
Regulation of central carbohydrate metabolism by the ArcA-ArcB two-component signal transduction system. The two component regulatory ArcA-ArcB system functions as a major control system for the regulation of expression of genes encoding enzymes in both aerobic and anaerobic catabolic pathways (189,190). Over forty operons are controlled by the ArcA-ArcB system (189). Almost all of the genes of the TCA cycle and glyoxylate shllllt are subjected to repression by the ArcA-ArcB two component regulatory system ( Figure 1) (191). The exception is the sdhCDAB-sucABCD operon, containing five ArcA binding sites, which is primarily initiated and regulated at the upstream sdh promoter (192). ArcA positively regulates this operon at the two upstream ArcA binding sites and represses the operon at the remaining ArcA binding sites flanking or within the sdh promoter (192). This operon also contains one internal promoter' p sue (192).
The expression of Arc-regulated genes varies in response to environmental 0 2 , although this compound is not thought to be the signal detected by the ArcB sensory kinase (190,193). ArcB probably senses the redox state of the cell through the detection of an electron transport component in reduced form (189,193). The phosphorylated form of the response regulator ArcA (ArcA-P) is predicted to reach peak levels in anoxic cells as the ArcB kinase activity progressively increases dming the transition from aerobic to anaerobkgrowth (193). However, significant levels of ArcA-P are apparent in aerobic cells, and differential patterns of expression of the members of the ArcA modulon can be attributed, at least in part, to the affinity of ArcA-P for DNA binding sites located in the transcriptional regulatory regions of its targeted operons (193).
ArcA plays a vital role in adjusting catabolism to oxygen-restricted growth conditions through regulation of carbon flux via the TCA cycle and electron flux via terminal cytochrome oxidases, rather than to adjust catabolism to fully aerobic or anaerobic conditions (189). Studies of the intracellular redox state (NAD+ /NADH) tlllder varying Di concentrations, of an ArcK mutant, strongly suggest that ArcA is actually a microaerobic redox regulator ( 189).
Acetyl phosphate and the activation of two-component signal transduction systems. It has been estimated that there may be as many as fifty different two-component signal transduction systems in E. coli (24 ). Acetyl phosphate has bee? shown to phosphorylate numerous response regulator proteins in vitro and has been shown to influence several two-component signal transduction pathways in vivo (25,190). Studies have shown that a major secondary source ofphosphoryl groups is from a small molecule phospho-donor, acetyl phosphate (79). When E. coli is incubated with acetyl phosphate, the following response regulator protein components of two-component signal transduction systems, become phosphorylated: CheB, NtrC, PhoB, OmpR, and ArcA (24,79). These phosphorylated response regulator proteins control the expression of genes or operons involved in chemotaxis, nitrogen regulation, phosphate-specific transport systems, outer membrane proteins involved in osmo-regulation, and enzymes in both aerobic and anaerobic catabolic pathways respectively (24,79,189,190).
The demonstration that acetyl phosphate can serve as a substrate for response regulator autophosphorylation revealed some subtle aspects of the two-component signal transduction pathways (24). These pathways can be regulated by controlling the levels ofphospho-donors, phospho-histidine kinases, or by controlling the rates of dephosphorylation of the active response regulators (24 ).
The function of acetyl phosphate in two-component signal transduction systems may be to adjust the sensitivity and/or the magnitude of an adaptive response such as shifts in metabolism and transitions in the expression of virulence factors (79). If acetyl phosphate contributes to the basal levels of phospho-response regulators within a cell, then higher levels of acetyl phosphate would produce a more sensitive system; a smaller stimulus would be required to produce enough phosphorylated response regulator to trigger a response (79).
Salmonella vaccines and vaccine candidates. Two Salmonella typhi vaccines, the live oral S. typhi Ty21 a and a vaccine based on the Vi polysaccharide capsule of S. typhi, currently are licensed in several countries ( 194 ). The S. typhi Ty21 a is a galactose epimerase (galE) mutant: the galactose required for the synthesis of smooth LPS 0 antigen is derived from UDP-galactose, which is synthesized from UDP-glucose by the enzyme galactose epimerase (195). The result is an avirulent strain with rough LPS (195). This property, coupled with the fact that S. typhi Ty2la is also Vi antigen negative, is a factor in the safety of this live vaccine strain in vivo (195). The Vi antigen, a capsular polysaccharide composed of N-acetylglucosamine, is produced by the most virulent S. typhi strains (l96). Each of the above vaccines, while effective, still suffers from drawbacks such as fevers after administration, incomplete protection, and a loss of protection with time (194,195). Therefore, new attenuated S. typhi live oral vaccines are being developed (194,195).
Live oral typhoid fever vaccine candidates, based on deletions in genes necessary for Salmonella pathogenesis, have shown progress (197,198). For example, the live oral S. typhimurium ~c.ya (adenylate cyclase) fl.crp (cAMP receptor protein) vaccine is based on deletions in a global regulatory system and is effective in animals and humans (197,198). The live oral S. typhi fl.eya ~crp vaccine becomes even more effective when a third deletion in the cdt gene (deep tissue colonization) is introduced (199,200).
Another typhoid fever live oral vaccine, avirulent and imnumogenic in mice and cattle, is based on aroC and aroD deletions in S. typhimurium; this strain is auxotrophic for aromatic compounds which are unavailable in mammalian cells (198,201,202). This ~aroC and fl.aroD double mutant is unable to synthesize chorismic acid, a precursor of the aromatic compounds p-aminobenzoic acid and 2,3-dihydroxybenzoate, and no other pathway for their synthesis exists in S. typhimurium (198,201). As a consequence of this auxotrophy, the bacteria Cannot proliferate within mammalian cells, yet they reside and replicate long enough to stimulate protective immune responses (198,201,202). Again, the introduction of a third deletion in the htrA gene (a protease 00 precursor heat shock protein) in a S. typhi strain increase the effectiveness of this vaccine candidate in clinical trials (54,199,200).
Another live oral typhoid fever vaccine candidate, based on deletions in genes of the phoP-phoQ two-component signal transduction system (regulation of virulence genes), is avirulent and immunogenic in mice and humans (200,203,204). A recent Salmonella enterica serovar Typhimurium live oral vaccine contains a deletion in the DNA adenine methylase (dam) gene (205).
DNA adenine methylase plays an important role in the timing and targeting of a number of cellular functions including DNA replicatio~ segregation of chromosomal DNA, and mismatch DNA repair ( 6 l). The Uun-vaccine strain was avirulent in day-of-hatch chicks (broiler chickens) and elicited cross protective immune responses against challenge by homologous serovars ofTyphimurhnn and heterologous serovars ofEnteritidis (205,206). The vaccine strain also proved to be highly attenuated for virulence in mice and conferred protection against murine typhoid fever (206).
The use of attenuated Salmonella strains as live vaccines is a safe and effective means of inducing significant humoral and secretory antibody responses in animal species humans, cattle, sheep, rabbits, fowl, and mice (195). Live Salmonella invade the Peyer's patches, where they present their numerous antigens directly to the T and B lymphocytes of the mucosal-associated lymphoid tissues (195). This elicits the mucosal immune system to produce antigen-specific immune responses (195). Due to the significant humoral and secretory antibody responses, live attenuated Salmonella strains show great promise for carriers of antigenic determinants from other pathogenic microorganisms (195,200).
Other S. typhimurium strains attenuated for virulence. A strain of S. typhimurium, with a mutation inptsH-encoded enzyme I (El) of the phospho(enol)pyruvate: sugar phosphotransferase system (PTS), has been reported as attenuated for virulence in an intraperitoneal mouse model; the log (attenuation) values were 2.3 below the wild-type Salmonella strain (90)(91)(92)207). The general PTS proteins, EI and HPr, transfer phosphoryl groups from phospho(enol)pyruvate the sugar-specific transporters (Ellsu~ in the PTS system (91,207). The PfS system is also an integral part of cAMP-CRP regulation (catabolite repression) and the genes of its regulon (183)(184)(185).
The outer membrane of S. typhimurium contains LPS; lipid A is the major component of this endotoxin that stimulates inflammatory cytokine release and inducible nitric oxide synthase (iNOS) responses in the murine host (7-9, 22, 208). AS. typhimurium strain with a deletion in the waaN gene, which encodes the enzyme that catalyzes one of the two secondary acylation reactions that complete lipid A biosynthesis, was constructed (9). The waaN mutant strain, which synthesize a full-length LPS molecule containing the O antigen and lack only the secondary acyl chain, were intravenously injected into BALB/c mice at a dose of with 102 bacteria per mouse (9). Extremely high bacterial 31 counts of 10 9 in the liver and spleen were observed (the lethal load of the wild-type parental strain is 10 s bacteria per organ), and only approximately 10% of the mice expired (9). Most of the remaining mice carrying these extremely high waaN mutant bacterial loads recovered and eventually cleared the bacteria from the organs (9). lhls study implied that death in a mouse typhoid infection is directly dependent on the toxicity of lipid A and suggested that this is mediated via pro-inflammatory cytokine and/or the iNOS responses (9).
The Vaccine Candidate Salmonella typhimurium SR-11 Fad-. The live oral vaccine candidate, Salmonella typhimurium SR-11 Fad-, was constructed by TnJ Od: :cam transposon mutagenesis of the SR-11 parent strain (209,210). The mutant strain was selected for its inability to catabolize oleate and citrate as sole carbon sources (210). It was phenotypically designated Fad-(Fatty acid) for its inability to catabolize fatty acids as a sole carbon source (210). The SR-11 Fad-strain also was unable to utilize acetate and isocitrate as carbon sources, but could utilize glucose, and glycerol (210).
The disease caused by Salmonella typhimurium in mice mimics typhoid fever in humans and is well accepted as a model for human typhoid (20,21). Salmonella typhimurium SR-11 Fad-, administered peroraly, was completely avirulent in BALB/c mice at a dose as high as 10 9 colony forming units ( cfu) (210). In contrast, the SR-11 parent strain proved lethal at doses of greater than 10 4 cfu (210). The vaccine candidate also was found to be protective in BALB/c mice (33,210) as well as avirulent and protective in chickens (Dr~ Paul S. Cohen: personal communication). A single oral dose of 10 7 SR-11 Fad-cells protected BALB/c mice against a lethal dose of 10 9 cells of the virulent SR-11 S. typhimurium parent strain (33,210).
SR-11 Fad-is a era (fruR) mutant. Southern hybridization with a probe specific for the chloramphenicol resistant cassette in SR-11 Fad-, followed by cloning of a 4.5 kb Pst I fragment containing the mini-TnlOd::cam into pBluescript II SK(+) yielded the plasmid pJHA7 (33).
Sequencing of pJHA 7 revealed the gene interrupted by the mini-transposon was the era gene (33).
Furthermore, complementation with a wild-type era gene in the plasmid pJHA8 (Table 1)  The Cra protein recognizes an imperfect palindromic DNA sequence (TGAA, A or T, C • C or G, any nucleotide, T, A or C, A or C, A or T) which it binds to asymmetrically (217). Ifthe Cra binding site is upstream of the promoter, it activates transcription of the target operon or gene; ifit overlaps or is downstream of the promoter, Cra represses transcription of the target operon or gene (217). Transcription of at least 20 genes is upregulated up to ten-fold and downregulated by a maximum of eleven-fold by the Cra protein (214,217,218). The Cra protein acts independently of cAMP-CRP transcriptional regulation (catabolite repression) (217,219,220). Two effector molecules of the era protein, µM concentrations of fructose-I-phosphate and mM concentrations of fructose-1,6bisphosphate, inactivate the DNA-binding protein ( Figure 4) (212, 217,219,221). Therefore, growth on glucose or fructose may inactivate the Cra DNA-binding protein via these two effectors (222,223).
Derepression of the glycolytic pathway and deactivation of the gluconeogenic pathway by the effectors of era also help to ensure the preferential utilization of carbon sources such as glucose and fructose (223).
The Cra global regulatory protein also represses the pts operon: ptsH (PTS general enzyme El), ptsl (PIS general enzyme Hpr), and err (glucose specific enzyme IIA Glc) constitute the pts operon ( Figure 3) (90,91,94,212,215,222). Cra (FruR) mutants were first isolated as suppressor mutations which allowed ptsH mutants to grow on PTS carbohydrates; the suppression resulting from the constitutive synthesis of the HPr-like domain ofthefruB-encoded diphosphoryl transfer protein because it can substitute for HPr (212). Furthermore, data has shown that llllder certain physiological conditions, Cra modulates the activity of adenylate cyclase, the cAMP biosynthetic enzyme (217).
Since the PTS system is also an integral part of cAMP-CRP regulation ( catabolite repression), the global regulator Cra is intertwined with other global regulators that exert pleiotropic effects on the transcription of operons and genes involved in central carbon ID:etabolism (72,73,(183)(184)(185)217).
The goal of this study. The initial goal of this study was to determine the extent to which gluconeogenesis is linked to virulence and pathogenesis in S. typhimurium SR-11, since S.
typhimurium SR-11 Fad-, is unable to l.llldergo gluconeogenesis and is avirulent in BALB/c mice (33,210). To that end, the construction of several mutations in the gluconeogenic pathway was planned to determine the virulence of these strains. A second goal of this study was to determine the regulation of the era gene by assaying for era promoter activity during growth on various gluconeogenic and glycolytic substrates and how deletions in the gluconoegenic pathway would the alter era promoter activity.

Materials and Methods
Bacterial strains and plasmids. Bacterial strains and plasmids used in this study are listed in and separately supplemented with the various glycolytic and gluconeogenic substrates, were utilized for determination of growth (utilization of substrate by the bacteria) (236). The M9 minimal agar plates were incubated at 37°C and the results were recorded at 24 hours and 48 hours. M9 minimal medium (broth) were used for determination of growth, rate of growth, and p-galactosidase assays (for a detailed protocol see "P-galactosidase assays for era promoter activity" in this section) (232,233,236,237 (240)).
Genomic DNA was isolated by a modified protocol from Current Protocols in Molecular Biology (241 ). Bacterial cells from an overnight culture were pelleted by centrifugation, resuspended in TE  (242).
Genomic DNA also was isolated by using the GNOME® genomic DNA isolation kit (BIO 101, Vista, CA). The protocol from the manufacturer was modified by increasing the initial volume of overnight liquid bacterial cell culture to 10 ml per miniprep. The kit used a "Cell Suspension Solution," "RNase Mboc," "Cell Lysis/Denaturing Solution," "Protease Mixx," and "Salt-Out Mixture" of unspecified composition and replaced the chloroform steps of the standard protocol with incubation at 4°C followed by centrifugation (241). A final precipitation of the genomic DNA in 100% ethanol completed the manufacturer's protocol (242).
Genomic DNA also was isolated by using the Wizard® genomic DNA Prep Kit (Promega Corporation, Madison, WI). The manufacturer's protocol was followed using up to 3 ml of an overnight culture of bacteria. All centrifugation steps were performed at 14,000 x gin an Eppendorf microcentrifuge (#5415C; Brinkmann Instruments, Inc., Westbury, NY). The isolated genomic DNA was incubated in an excess of DNA rehydration solution (10 mMTris-HCl, pH 7.4; 1 mMEDTA, pH s.O) at 65°C for greater than one hour or a combination of incubation at 65°C followed by overnight incubation at room temperature. All genomic DNA was frozen at -20°C for short-term storage and _ 78°C for long-term storage.
Isolation of plasmid DNA. Quantities of plasmid DNA greater than 10 µg were isolated by modified alkaline lysis (243). A minimtnn of 100 ml of bacterial cells from an overnight liquid culture were pelleted by centrifugation and resuspended in GTE Buffer ( 50 mM  Precipitated DNA was pelleted by centrifugation (12,000 x g), lyophilized or air-dried at 50°C, and resuspended in TE buffer or Molecular Biology Grade water (5 Prime~ 3 Prime, Inc., Boulder, CO).
All pmified and concentrated DNA was frozen at -20°C for short-term storage and plasmid DNA was frozen at -78°C for long-term storage.
Purification and concentration of DNA from agarose gels. DNA, including amplicons, was also pmified and recovered from agarose gels. Bands of electrophoresed genomic or plasmid DNA were cut from agarose gels and added to GenElute™ Minus EtBr Spin Columns (Supelco, Bellefonte, PA) pre-washed in 1E Buffer. The columns were inserted onto microcentrifuge tubes and centrifuged (12,000 x g) for 10 minutes. The columns retained the agarose and ethidium bromide while the DNA was eluted by the TE buffer. The DNA was then concentrated, when necessary, by ethanol precipitation.
A second method was also employed for the recovery of DNA from low melt agarose gels.
After electrophoresis, the band containing the DNA was cut from of the gel and placed in a 0. was added. The aqueous solution containing the DNA now remained liquid at room temperatlll'e and was utilized directly in restriction endonuclease digests and ligase reactions (244,245). The plll'ified and/or concentrated DNA isolated from agarose gels was frozen at -20°C for storage.  (250).
The temperatures and times of the DNA denaturation, primer annealing, and primer extension steps (30-35 cycles) varied for each specific PCRreaction (251). ODsoo readings (23 7). 44 p-galactosidase assays for era promoter activity. SR-11 strains, containing the era promoter-lacZ transcriptional fusion, were adapted to M9 minimal media separately supplemented with various carbon sources with the appropriate antibiotic(s). The cultures were back-diluted to an ODsoo of~ 0.1 into fresh homologous media containing 150 µglml of ampicillin. The cultures were grown at 37°C with aeration to an ODsoo of~ 1.0, then 1 ml samples were removed in triplicate for each culture.

Quantification of DNA
p-galactosidase assays were performed by modified protocols by Miller (232) and Ausubel (233). Tue OD 600 of the 1 ml samples was measured then centrifuged@l4,000 x gin microcentrifuge for 5-10 minutes. The supernatant was removed and the pellet was resuspended in 1 ml ofZ buffer Louis, MO) was added to each sample to stop the reaction and the elapsed time was recorded. The samples were centrifuge for 5 minutes@ 14,000 x g and the OD 420 and OD 550 ofthe supernatant was measured. The p-galactosidase activity was calculated using the formula:
Tue mice were fed Prolab RMH2000 pellets (PMI Nutrition International, Inc., Brentwood, MO) and sterile deionized, distilled water. The cages were placed in an HEPA-filtered isolation tmit (Britz-Heidbrink, Inc., Wheatlands, WY) to prevent contamination. Cages, food, water, and bedding were changed every other day throughout the entire experiment. All experiments were performed in accordance with institutional guidelines for animal care (IACUC Approval Number: A98-03-030).
All S. typhimurium SR-11 strains utilized in the virulence assays were grown in Luria broth, with the appropriate antibiotic( s ), at 3 7°C, with aeration, for 18-19 hours. One ml of each strain was centrifuged (10,500 x g) for 5 minutes (certain strains required a second centrifugation to completely pellet the cells) in a microcentrifuge and the supernatant was carefully removed. 46 Fisher Scientific, Fair Lawn, NJ) in 1 ml of Luria broth. The homogenates were serial diluted into ml saline dilution tubes and spread plated onto MacConkey agar containing the appropriate antibiotic(s). The cfU/organ were recorded. Surviving SR-11 infected BALB/c mice were sacrificed by carbon dioxide asphyxiation 9-21 days post peroral inoculation. The cfU/liver and cfU/spleen were also counted and recorded as above.
typhimurium SR-11 cradl craz integrant, containing a era promoter-lacZ transcriptional fusion, was obtained from Mary P. Leatham (256). The strain was constructed by conjugative-dependent allele replacement utilizing the suicide vector pML55cradl-craz (Table 1, Figure 5) (256,257) The control strain carries a partial upstream ilvl gene, an upstream llvH gene, the upstream intergenic sequence, and the era promoter ligated to the promoterless lacZ gene from the plasmid pCB267 ( Figure 5) (256,258). Approximately 94 7 bp of DNA upstream of the era promoter were engineered into this construct to retain any possible regulatory sequences (256,259). The 3' end of the lacZ reporter gene is ligated to approximately 200 bp of the remaining truncated era gene, followed by approximately 534 bp of downstream DNA (a partialyabB gene) ( Figure 5) (256). Since SR-11 cradl craz integrant is a partial diploid, containing both a wild-type era gene and a era promoter ligated to a reporter gene ( Figure 5), it is able to utilize fatty acids and gluconeogenic substrates as sole carbon sources (231,256,257).
The SR-11 cradl craz integrant strain was used as a control to ~easure the era promoter activity (J3galactosidase assays) during growth on glycolytic and gluconeogenic substrates.
Construction of S. typhimurium SR-11 mutants by allelic exchange. An S. typhimurium SR-11 fbpmutant was constructed by conjugative-dependent allele replacement (239,257). Since the genome was not fully sequenced and the homology between S. typhimurium and E coli was approximately 85% identical at the nucleotide level, the E coli fructose-1,6-bisphosphatase (jbp) DNA sequence was obtained using PubMed and Entrez Protein (66,(260)(261)(262). The S. typhimurium DNA sequence of the jbp gene was then pieced together from contig518 and contig 1460 of the S.
typhimurium LT2 llllfinished fragments of the complete genome using BLASTn (263,264). ORF Finder was utilized to determine the translational start and stop codons of the jbp gene (265). The 47 DNA sequence of the fop, as well as the gene directly upstream (yi.fG) and directly downstream (yijF), were processed in Webcutter to find the restriction endonuclease sites (266). Primers for PCR were designed to make a permanent deletion in the fop gene in which a chloramphenicol resistance gene could be inserted for selection purposes. The primers were also designed for the amplicon to be ligated into both the phagemid cloning vector pBluescript II SK ( +) and the suicide vector pLD55 for conjugation and allelic exchange (Table 2) (257,267,268) Using S. typhimurium SR-11 wild-type genomic DNA as a template and the yi,jF/forward and yijG/reverse primers ( and restriction mapping ( Figure 6) (269).
A deletion of878 bp from thefop gene was made by amplifying around pJHA:fbp+ ( Figure 7) by PCR using the fop-DEL/forward and fop-DEL/reverse deletion primers ( Table 2). The amplicon was digested with Bam HI, ligated to itself, and electroporated into E coli XL I-Blue MRF' competent cells. The result of the transformation yielded the plasmid p!1fop (Figure 8). The chloramphenicol resistance cassette from pJHA 7 (Table 1) was digested with Barn HI, ligated into Bam HI-digested p4fbp, and electroporated into E coli XLl-Blue MRF' to produce the plasmid p/1fopCam ( Figure 9).
The plasmid p4fbpCarn (Figure 9) was digested with Bam HI and Not I and ligated into the Bam HI/Not I-double digested suicide vector pLD55 (Table 1). This ligase reaction was electroporated into the host strain E. coli S 17-1 A. pir and the suicide vector p55!1fopCarn was subsequently isolated and verified by PCR and restriction mapping ( Figure 10) (257,273).

48
The temperature-sensitive donor strain£. coli  was grown overnight at 30°C in LB containing ampicillin and chloramphenicol to retain the plasmid (273). An overnight culture of the recipient strain, S. typhimurium SR-11 wild-type, was grown in LB containing nalidixic acid. A conjugation was performed by mixing 200 µ1 each of the donor and recipient strains in 5 ml of 10 mM MgS0 4 (273,274). select for partial diploid integrants (275). PCR, using colonies as template DNA, was utilized to screen for the correct exconjugant which successfully achieved homologous recombination with integration of the suicide vector p55~.fbpCam into its genome (257). The exconjugant was designated S. typhimurium SR-11~jb[£am1-3 integrant ( Figure 10).
Deletion primers for PCR ( Table 2) were designed with 45 nucleotides of homology within the wild-type maeB, sfrA and pta genes. The primers were also engineered with 20 nucleotides of homology to the kanamycin resistance gene in pKD4 (Table 1 ). Using pKD4 as template DNA and the deletion primers, five 100 µl PCR r~ctions for each gene to be deleted were performed. The amplicons now contained a kanamycin resistant gene flanked by sequences homologous to the wildtype maeB, sfrA or pta genes ( Figure 13). The PCRreactions were pooled (500 µI/gene to be deleted), purified, and concentrated to approximately 10 µI.
S. typhimurium  was grown in SOB media, containing 150 µglml of ampicillin and 20 mM L-arabinose (the gene encoding ' A Red recombinase, which greatly enhances the rate of homologous recombination, is located the plasmid pKD46 and is controlled by an arabinoseinducible promoter), at 30°C to an OD 600 'of~ 0.6 and then made electrocompetent (254,255). Two microliters of the concentrated linear DNA amplicons from above were electroporated into 40 µl of fresh electrocompetent cells and incubated at 3 7°C in 1 ml of SOC media for 1 hour. Five 100 µl aliquots were spread plated onto LB plates, containing either 50, 60, or 70 µglml ofkanamycin, and incubated at 37°C to cure any transfonnants of the temperature-sensitive plasmid The remainder of the electroporated culture was incubated overnight at room temperature and then spread plated as above. Transfonnants which successfitlly underwent homologous recombination ( Figure 13) were kanamycin-resistant and ampicillin-sensitive. Deletions in the maeB, sfrA or pta genes were confirmed by PCR utilizing flanking confirmation primers ( Table 2). These strains were designated S.
Exconjugants which successfully achieved homologous recombination with integration of the suicide vector pML55cradl-craz into its genome were selected on LB plates containing ampicillin and kanamycin (257). The integrants were confirmed on LB plates containing tetracycline, XLD plates, and LB plates containing approximately 70 µglml ofX-Gal (5-bromo-4-chloro-3-indolyl-J3-ngalactopyranoside; Sigma Chemical Co., St. Louis, MO) (302). The integrants are partial diploids, containing both a wild-type era gene and a era promoter-lacZ transcriptional fusion; these strains were therefore also confirmed for growth on M9 minimal agar plates separately supplemented with various carbon sources. The kanamycin resistance cassette, as well as the era promoter-lacZ transcriptional fusion (using era-COMP/forward and /acZ/reverse primers), were confirmed by PCR Complementation of the eraandjbp-genes. The S. typhimurium SR-11 Fad-strain was complemented (211) with the plasmid plliA8 containing the wild-type era gene (  (304). After isolating the modified plasmid DNA, it was electroporated into the SR-11 Fad-strain (Table 1 ).
The S. typhimurium SR-11 4fbpCam AX-3 segregant strain was complemented with a plasmid containing the wild-type fbp gene (211 ). Utilizing SR-11 wild-type genomic DNA as a template and the yijF/forward and yifG/reverse primers (Table 2) Statistics using the Student t distribution. To determine ifthe differences observed in pgalactosidase activities were significant, the student t distribution was calculated using a P value of 0.05. Significance or insignificance was reported with a 95% confidence limit (305).

52
Construction, confirmation, and characterization of the S. typhimurium SR-11 jbpmutant. A mutation in thejbp gene, which encodes the gluconeogenic enzyme fructose-1,6bisphosphatase was constructed in S. typhimurium SR-11 (Figure 1 ). A deletion of 878 bp from the wild-type jbp gene was made and replaced with a chloramphenicol resistance cassette. The resulting S.
The SR-11 jbp-mutant strain, like the SR-11 Fad-vaccine strain, was unable to utilize gluconeogenic substrates as sole carbon sources when streaked onto M9 minimal agar plates (Table 3).
With the expected exception of the SR-1 lfbpmutant's inability to metabolize glycerol as a sole carbon source, the phenotypes observed were identical to the SR-11 Fad-vaccine strain. Furthermore, complementation with a wild-type.fbp gene on the plasmid plliA:fbp+, restored the ability of the SR-11 jbpmutant to utilize gluconeogenic substrates as sole carbon sources on M9 minimal agar plates.
Virulence of the S. typhimurium SR-11 AjbpCam AX-3 segregant in BALB/c mice. The SR-11 jbp-mutant strain was administered perorally (inoculation of approximately 2.4 x 10 8 cfu) to 30 day-old female BALB/c mice and was found to be virulent ( Table 4). The cfu recovered from the livers and spleens ofthe mice fed the SR-1 lfbpmutant strain were similar to the numbers ofbacteria recovered from the mice fed the SR-11 wild-type control strain (Table 4). A mortality rate of 100% was observed in the SR-11 jbp-mutant at the same inoculum that is 100% avirulent in the SR-11 Fadvaccine strain.
P-galactosidase assays, for era promoter activity, in the SR-11 cradl craz integrant. The era promoter activity was highest in SR-11 cradl craz integrant when grown on gluconeogenic substrates ( Table 5). The highest era promoter activity was observed on succinate (3.3 units), followed by malate, pyruvate, citrate, oleate, acetate, and phospho{enol)pyruvate (1.7 units). The J3galactosidase assays in SR-11 cradl craz integrant were lower when grown on glycolytic substrates versus gluconeogenic substrates (Table 5). For glycolytic substrates, the highest era promoter activity was observed on glucose (1.3 units), followed by gluconate (1.0 units), and lowest on fructose (0.9 units). As expected, the ~-galactosidase assays revealed no activity in the SR-11 wild-type control grown on glucose (0.0 units).
The ratio of the ~-galactosidase units for each substrate assayed relative to the ~-galactosidase units for fructose, the lowest era promoter activity observed (0.9 units), was calculated. The highest ratio, at 3.6, was for the SR-11 cradl craz integrant when grown on succinate (Table 6, Figure 15). This translates to 3.6 times more era promoter activity observed when grown on succinate versus fructose. The average increase in era promoter activity, for all the gluconeogenic substrates tested, was 2.5-fold above the era promoter activity for growth on fructose.
The generation times of SR-11 cradl craz integrant when grown on gluconeogenic substrates were longer than when grown on glycolytic substrates ( Table 5). The generation times on gluconeogenic substrates were greater 100 minutes, whereas on glycolytic substrates, the generation times were less than 91 minutes. The fastest growth rate for the SR-11 cradl craz integrant was recorded in M9 minimal broth supplemented with fructose (73 minutes).

Construction, confirmation, and characterization of the S. typhimurium SR-11 maeK
and SR-11 sfcA-mutants. Mutations in the maeB gene (encoding the gluconeogenic NADPdependent malic enzyme) and the sfeA gene (encoding the gluconeogenic NAO-dependent malic enzyme) were constructed in S. typhimurium SR-11 ( Figure 1 ). A deletion of 2191 bp from the wildtype maeB gene and a separate deletion of 1612 bp from the wild-type sfeA gene were made; both deletions were replaced with a kanamycin resistance cassette. The resulting S. typhimurium SR-11 mae/J" and S typhimurium SR-11 sfi::Amutant strains were confirmed by PCR (Figure 16, 17, 18).
The S typhimurium SR-11 maeE and sfi::Amutants grew as well as the SR-11 wild-type on M9 minimal agar plates supplemented with glucose or fructose, but grew more slowly, versus the wildtype, on M9 minimal agar plates supplemented with malate or succinate. Noticeably longer generation times for the SR-11 maeff and sfi::Amutants, versus the SR-11 wild-type, were observed in both mutants when grown on malate or succinate in M9 minimal broth ( Table 7). The differences in growth rates of all three strains were minor when utilizing glycolytic substrates as sole carbon sources (Table   7).
f3-galactosidase assays, for era promoter activity, in the S. typhimurium SR-11 maeH cradl craz integrant and the SR-11 sf cA-cradl craz integrant The era promoter activity in the SRl l maeff mutant, measured by f3-galactosidase assays, was slightly down-regulated (l.1-to 1.4-fold) on the two gluconeogenic substrates and one glycolytic substrate, versus the SR-11 cradl craz integrant control strain (Table 8). The greatest down-regulation was observed for growth on succinate (1.4fold), followed by malate (1.1-fold), and glucose (1.1-fold). The era promoter activity in the SR-11 maeff mutant was essentially the same as the control strain when grown on fructose ( Table 8).
Virulence of the S. typhimurium SR-11 maeH and SR-11 sf cAmutants in BALB/c mice.
The S. typhimurium SR-11 maelf and SR-11 sfcA-mutant strains were virulent when administered perorally (inoculation of approximately 1.3 x 10 8 cfu) to 33 day-old female BALB/c mice. A mortality rate of 100% was observed in BALB/c mice fed either the SR-11 maeB-mutant strain or the SR-11 sfeA-mutant strain ( Table 9). The day of expiration post-peroral inoculation for the BALB/c mice fed the SR-11 maeB-and the SR-11 sfcA-strains was similar to age-matched BALB/c mice fed the SR-11 wild-type control ( Table 9). The cfu recovered from the livers and spleens of the mice fed the SR-11 maeB-mutant strain (4.8 x 10 7 cfu and 3.9 x 10 7 cfu respectively) were approximately 1/2 ofa log below the numbers of bacteria recovered from the mice fed the SR-11 wild-type control strain ( Table   9). The cfu from the livers and spleens of the mice fed the SR-11 sfeA-mutant strain (1.5 x 10 7 cfu and 7.9 x 10 7 cfu respectively) were approximately 1 log below the cfu recovered from the SR-11 wildtype control strain ( Table 9).
The S. typhimurium SR-11 sfeA-mutant strain was also fed (inoculation of approximately 2.1 x 10 8 cfu) to 50 day-old female BALB/c mice. Again, the SR-11 sfeA-mutant strain was found to be virulent with a 100% mortality rate (Table 10). The day of expiration post-peroral inoculation for the BALB/c mice fed the sfcA-mutant strain was similar to the age-matched BALB/c mice fed the SR-11 wild-type control ( Table 10). The cfu recovered from the livers and spleens ofthe mice fed the SR-11 sfcAmutant strain were slightly elevated versus the mnnbers of bacteria recovered from the mice fed the SR-11 wild-type control strain (Table 10).
Construction, confirmation, and characterization of the S. typhimurium SR-11 ptamutant. To further examine the regulation of the global regulatory era gene, a mutation in the pta gene, which encodes the enzyme phosphotransacetylase was constructed in S. typhimurium SR-11 ( Figure 1 ). A deletion of 1911 bp from the wild-type pta gene was made and replaced with a kanamycin resistance cassette. The resulting S. typhimurium SR-11 ptd mutant strain was confirmed by PCR ( Figure 19).
The SR-11 ptd mutant strain was also unable to utilize acetate as a sole carbon sol.ll'ce in both M9 minimal agar plates and M9 minimal broth supplemented with potasshnn acetate. The generation times ofthe SR-11 ptacradl craz integrant were 19 minutes (22%) longer on glucose and 31 minutes (31%) longer on fructose, versus the SR-11 cradl craz integrant control, in M9 minimal broth (Table   11). The differences in the generation times of the SR-11 ptd cradl craz integrant and the SR-11 cradl craz integrant control were minor for growth on succinate and malate (Table 11 ).
Virulence of the S. typhimurium SR-11 ptamutant in BALB/c mice. The S. typhimurium SR-11 ptamutant strain was administered perorally (inoculation of approximately 1.6 x 10 8 cfu) to 33 day-old female BALB/c mice and was found to be avirulent (Table 13). However, the mice showed signs of disease: ruffled fur, eye infections, and lethargy. Two mice recovered by Day 14 of the experiment and the remaining two mice were also recovering. The intensity of the disease was not as proooWlced as the mice fed the SR-11 wild-type strain. The cfu recovered from the livers and spleens of the mice fed the SR-11 ptamutant strain ( 6.9 x 10 6 cfu and 6. 7 x 10 6 cfu respectively), sacrificed on Day 15 , were 1.3 logs less than mnnbers of bacteria recovered from the mice fed the SR-11 wild-type control strain (Table 13) The availability of µM concentrations of glucose or another glycolytic substrate(s) would allow the biosynthesis of essential metabolites such as ribose-5-phosphate for ribonucleotides and deoxynucleotides, erythrose-4-phosphate for aromatic amino acids, and glucose-6-phosphate and pentose-5-phosphate for LPS (69,76). Without co-metabolism from a separate glycolytic substrate such as glucose, these metabolites would not be biosynthesized in an SR-11 jbp-mutant, since gluconeogenesis is blocked at :fructose-1,6-bisphosphate and never reaches glucose-6-phosphate and the pentose phosphate pathway (Figure 1 ). Furthermore, it appears that in vivo, mM concentrations of the era effector :fructose-1,6-bisphosphate did not accumulate and. inactivate the Cra DNA binding protein ( Figure 1 ). The SR-11 jbp-mutant, in vivo, appears to possess a functioning Cra protein, unlike the SR-11 Fad-vaccine strain which is avirulent in BALB/c mice (Table 4) (33,210).
A subsequent publication revealed the existence of a second :fructose-1,6-bisphosphatase in E.
coli encoded by the glpX gene (306). S. typhimurium also contains this redlllldant :fructose-1,6bisphosphatase enzyme (307). After this research was completed, an SR-11 jbp-glpX double mutant, also tmable to utilize gluconeogenic substrates as sole carbon sources in vitro, was constructed and proved to be virulent in BALB/c mice (308).
The SR-11 Fad-, llllable to utilize gluconeogenic substrates as sole carbon sources in vitro (Table 3), was able to invade and was viable in enclosed vacuoles within M cells of the Peyer's patches in BALB/c mice (33). Additionally, when SR-11 Fad-crd is fed to BALB/c mice, lowmnnbers of the strain (approximately 10 3 cfu) were recovered from both the liver and spleen (Table 4). 1bis suggests that the SR-11 Fad-and the SR-1 ljbpmutant both have access in vivo to a glycolytic substrate in BALB/c mice.
Clearly, the SR-11 jbpmutant's interruption "high" in the gluconeogenic pathway does not renders. typhimurium SR-11 avirulent in BALB/c mice (Figure 1, Table 4). Either co-metabolism utilizing both gluconeogenic and glycolytic substrates occurs in vivo, or the Cra protein is involved in the regulation of some other cellular function related to virulence. Perhaps the TCA cycle must generate a sufficient concentration of the precursor metabolite phospho(enol)pyruvate via malate or oxaloacetate for virulence ( Figure 1 ). PTS carbohydrate phosphorylation is coupled to carbohydrate translocation across the cytoplasmic membrane; the energy for these processes being provided by the precursor metabolite phospho(enol)pyruvate (91). The general PTS proteins, EI and HPr, transfer phosphoryl groups from phospho(enol)pyruvate to the sugar-specific transporters (EIISllfPI) in the PTS system (91,207). A strain of S. typhimurium, with a mutation in the ptsH-encoded EI enzyme of the PJ'S system, has been reported as attenuated for virulence in BALB/c mice (207). The PTS system is also an integral part of cAMP-CRP regulation ( catabolite repression) and the genes of its regulon (183)(184)(185). The live oral S. typhimurium ~cya (adenylate cyclase) ~crp (cAMP receptor protein) vaccine is based on deletions in the cAMP-CRP global regulatory system (197). The above mutant strains, both attenuated for virulence in BALB/c mice,-reveal the importance for a fully functional PTS system driven by the precursor metabolite phospho( enol)pyruvate; the SR-11 jbpmutant, although blocked in the gluconoegenic pathway at fructose-1,6-bisphosphate, may have generated a sufficient concentration of phospho( enol )pyruvate (Figure 1 ). A S. typhimurium sfcAmutant, deficient in the NAD-dependent malic enzyme, was reported to be avirulent in BALB/c mice (20,80). Furthermore, four-carbon gluconeogenic substrates such as malate and/or oxaloacetate were noted as clearly important for growth in the murine host (20). Unlike the SR-11 fbp-mutant, where gluconeogenesis is interrupted up "high" in the gluconeogenic pathway, the reported sfcAmutant interrupts gluconeogenesis directly off the TCA cycle (Figure 1). The difference in the location of the interruption in the gluconeogenic pathway may account for the virulence observed in the SR-11 jbp-mutant.
The era promoter activity in the SR-11 cradl craz integrant To further investigate the link between the era gene, gluconeogenesis, and virulence in BALB/c mice, the era promoter activity in the SR-11 cradl craz integrant (containing a era promoter-lacZ transcriptional fusion) was determined for growth on various glycolytic and gluconeogenic substrates {104, 232, 233). Overall, the era promoter activity was up-regulated (1.6-to 3.1-fold) when grown on gluconeogenic substrates versus glycolytic substrates ( Table 5). The average increase in era promoter activity, for all the gluconeogenic substrates tested, was 2.5-fold above the era promoter activity for growth on fructose (Table 6, Figure 15). Except for pyruvate, the era promoter activities were highest for growth on substrates within the TCA cycle. These data are consistent with the fimction of the Cra protein: up- The era promoter activity of SR-11 cradl craz integrant during growth on malate was upregulated by 2.1-fold higher versus growth on glucose (Table 5). S. typhimurium contains two redundant malic enzymes: an NAD-dependent malic enzyme (encoded by the sfcA gene) and NADPdependent malic enzyme (encoded by the maeB gene) (123,125,126). Both enzymes are malate oxidoreductases which convert malate, normally present in the TCA cycle, to pyruvate ( Figure 1) (125,126,309,310). Both biochemical reactions, studied more extensively in E. coli, are essentially irreversible (81,(309)(310)(311). In E.coli, it has been suggested that the ifcA-encoded NAD-dependent malic enzyme is involved in gluconeogenesis and that the maeB-encoded NADP-dependent malic enzyme supplies the cell with NADPH when growing on C 4 carbon sources (312). Another study in E.
coli suggested that the NADP-dependent malic enzyme is also involved in the supply of acetyl-CoA from malate, which is utilized for the biosynthesis of lipids as well as for the maintenance of TCA cycle intennediates (310).
During growth on gluconeogenic substrates such as malate (succinate, citrate, pyruvate, acetate, or oleate) the pyruvate pool is divided between conversion to phospho(enol)pyruvate (up the gluconeogenic pathway) and conversion to acetyl-CoA to fuel the TCA cycle; both routes are required for the biosynthesis of precursor metabolites for growth (Figure 1 ). Because of the high demand for acetyl-CoA to fuel the TCA cycle, it is tmlikely that much of the acetyl-CoA is converted to acetyl phosphate ( Figure 1 ).
The era promoter activity in SR-11 cradl craz integrant grown on the glycolytic substrate glucose was 2.1-fold lower than when grown on malate (Table 5). In these non-limiting glucose growth conditions, an incomplete branched, biosynthetic form of the TCA cycle is utilized ( Figure 2) (74,78). The full TCA cycle is not required under these conditions because the bulk of energy is derived from glycolysis (74,78). In this glucose-rich environment, cells produce excess acetyl-CoA which drains through acetyl phosphate, to further produce ATP with the associated secretion of acetate ( Figure 2) (78,79). As cell density increases, the build-up of acetate also serves to increase the size of the acetyl phosphate pool (78).
The era promoter activity for SR-11 cradl craz integrant cells grown on acetate was 1.4-fold higher than glucose-grown cells (Table 5). S. typhimurium contiins two pathways for the utilization of acetate: the first pathway is a direct route to acetyl-CoA catalyzed by the acs gene product acetyl-CoA synthetase; the second pathway forms the intermediate metabolite, acetyl phosphate, catalyzed by the ackA-encoded acetate kinase A enzyme and then acetyl-CoA is formed by the pta-encoded phosphotransacetylase enzyme (Figure 1) (279,280,(313)(314)(315)(316). Biochemical studies in E. coli on the acs pathway and the pta-ackA pathway, which both lead to acetyl-CoA, may have clarified their function ( Figure 1) (317)(318)(319). In these biochemical studies, the pta and ackA genes were both downregulated by approximately 2-fold while the acs gene was induced by greater than 8-fold in a global expression profiling experiment of acetate-grown E. coli using glucose grown cultures as a reference (317). In another experiment, investigating the global regulation of the main metabolic pathways of E.
l . 1-.~,,,,.A on 2-dimensional electrophoresis, similar results were observed: the pta-encoded and aekAco z~ encoded enzymes were down-regulated (1.3-fold and 1.7-fold respectively) for aerobic growth on acetate versus the glucose-grown control (318). These combined data for both experiments suggest that the acs gene pathway is mainly responsible for acetate uptake, whereas the pta-ackA pathway is utilized for acetate excretion (Figure 1) (317)(318)(319). Consequently, a limited pool of acetyl phosphate would be expected for aerobic growth in acetate.
Overall, the predicted size of the acetyl phosphate pool, the biochemical pathways utilized, and the data from the f3-galactosidase assays, implicate acetyl phosphate in the regulation of the era gene; acetyl phosphate appears to function, indirectly or directly, as a repressor of the era gene.
Furthennore, acetyl phosphate is a logical candidate since it is a known global regulator of chemotaxis, the phosphate-specific transport system, and nitrogen regulation (24,79,189,190). Acetyl phosphate also has been proposed to play a role in outer membrane proteins involved in osmo-regulation and the regulation of enzymes in the TCA cycle and glyoxylate shunt (24,79,189,190).
The era promoter activity of the S. typhimu.tium SR-11 ma.eH, SR-11 sfcA-, and SR-11 pta· mutants. AS. typhimurium sfeA· mutant was reported to be avirulent in BALB/c mice and the importance of four carbon gluconeogenic substrates such malate and/or oxaloacetate for growth in the murine host was implied (20,80). This warranted further study and also suggested that the sfcA gene may be involved in the regulation of the era gene or the Cra protein. To test this hypothesis and the above implication that acetyl phosphate appears to function as a repressor of the era gene, deletions in the maeB, sfeA, and pta genes were constructed in S. typhimurium SR-11 wild-type.
The generation times of the SR-11 maeE, and SR-11 sfeA· mutants were essentially the same as the SR-11 wild-type for growth in M9 minimal broth supplemented with glucose or fructose (Table   7). However, considerably longer generation times (versus the SR-11 wild-type) for both the SR-11 maeB" and SR-11 sfeAmutants were observed for growth in malate or succinate (Table 7). This suggests that both malic enzymes are required for maximum growth rates when utilizing gluconeogenic substrates. Conversely, both malic enzymes are not required for maximum growth rates when utilizing glycolytic substrates.

62
Of interest, the SR-11 maeE mutant's growth rate was 12 minutes (11%) slower on succinate and 41 minutes (22%) slower on malate than the SR-11 sfcA-mutant grown on these gluconeogenic substrates (Table 7). These differences in growth rates may be due to the reported functions of the malic enzymes: the sfcA-encoded NAD-dependent malic enzyme is involved in gluconeogenesis whereas the maeB-encoded NADP-dependent malic enzyme supplies the cell with NADPH when growing on C 4 carbon sources and is also involved in the supply of acetyl-CoA from malate, which is utilized for the biosynthesis of lipids as well as for the maintenance of TCA cycle intermediates (310,312).
Overall, considering the long generation times for growth in gluconeogenic substrates, a deletion in the pta gene in the SR-11 ptacradl craz integrant strain exhibited minor pleiotropic effects for growth utilizing gluconeogenic substrates (Table 11 ). Conversely, the growth rate of the SR-11 ptd cradl craz integrant was 19 minutes (22%) slower on glucose and 31 minutes (31%) slower on fructose versus the SR-11 cradl craz integrant control (Table 11 ). Excess acetyl-CoA cannot drain through acetyl phosphate, to further produce A 1P with the associated secretion of acetate in the SR-11 ptd cradl craz integrant ( Figure 2). This implies that the pta gene, and the ability to produce acetyl phosphate from acetyl-CoA, appear to be required for maximum growth rates when utilizing glycolytic substrates (Table 11 ). A similar conclusion was reported for the growth of E. coli in glucose: the ability to produce acetyl phosphate influences (depresses) the ~ growth rate (79).
The era promoter activity in the· SR-11 maeE cradl craz integrant was slightly downregulated (1.1-to 1.4-fold) when grown on two gluconeogenic substrates and one glycolytic substrate, versus the SR-11 cradl craz integrant control strain (Table 8). In the SR-11 maeE cradl craz integrant, an increase in the concentration of its redwidant sfcA-encoded NAD-dependent malic enzyme should be expected; mutational analysis has demonstrated that the malic enzymes can compensate for each other in acetate-grown cultures (317). It is possible that the same occurs in glucose-grown, succinategrown, and malate-grown cultures. This minor down-regulation in the era promoter activity of the SR-11 maeB" cradl craz integrant may be due to an increased concentration of the efcA-encoded malic enzyme.
The era promoter activity was significantly (P = 0.05) up-regulated (2.3-to 2.8-fold) in the SR-l l sfeAcradl craz integrant grown on both gluconeogenic and glycolytic substrates, versus the SR-11 cradl craz integrant control strain (Table 8). This significant up-regulation becomes more striking when focused at the extremely high J3-galactosidase units observed: 7.0 units for growth on succinate and 8.2 units for growth on malate (Table 8). The efeA-encoded NAD-dependent malic enzyme appears to function as a transcriptional repressor, or play a role in the transcriptional repression, of the global regulatory era gene.
Similar to the SR-11 sfeAcradl craz integrant, the era promoter activity was significantly considerably up-regulated (3.0to 3.4-fold) in the SR-11 ptd cradl craz integrant on both gluconeogenic and glycolytic substrates, versus the control strain (Table 12). Although the J3galactosidase activities were slightly lower in the SR-11 cradl craz integrant control strain for this set of assays, the ratio of units of J3-galactosidase for each substrate, relative to units of J3-galactosidase for the SR-11 cradl craz integrant control strain, compensates for this and reveals a clearer view of the era promoter activity (Table 12). It appears that a repressor of the era gene ceased to function in the SR-11 ptacradl craz integrant: this implicates acetyl phosphate, indirectly or directly, in the regulation of the era gene.
Virulence of the S. typhimu.rium SR-11 maeK, SR-11 sfcA-, and SR-11 ptamutants in BALB/c mice. Both the SR-11 maeR and SR-11 sft:Amutants were virulent in BALB/c mice (Table   9). Since an efeAmutant strain of S. typhimurium, deficient in the NAD-dependent malic enzyme, was reported to be avirulent in mice, a second virulence experiment was performed utilizing more mature mice (20,80). The SR-11 sfeAmutant strain proved to be virulent in more mature BALB/c mice (Table 10). It was later determined that the sfeAmutant strain reported to be avirulent, was submitted to the American Type Culture Collection (A TCC) as a strain of S. typhimurium, but the ATCC characterized this strain as S. eholeraesuis (a swine pathogen) (320). The pathogenicity of a S.
choleraesuis sfcAmutant strain may deviate from the pathogenicity of a S. typhimurium sfeAmutant strain in the mouse model.
Tue SR-11 ptd mutant strain was found to be avirulent in BALB/c mice (Table 13). The mice exhibited symptoms of murine typhoid fever, but the intensity of the disease was not as prooounced as the mice fed the SR-11 wild-type strain. Two mice recovered by Day14 ofthe experiment and the remaining two mice were also recovering. The cfu recovered from the livers and spleens of the mice fed the SR-11 ptd mutant strain (6.9 x 10 6 cfu and 6.7 x 10 6 cfu respectively), sacrificed on Day 15 , were 1.3 logs less than numbers of bacteria recovered from the mice fed the SR-11 wild-type control strain (Table 13). Since the SR-11 ptamutant strain cannot synthesize acetyl phosphate from acetyl-CoA and acetyl phosphate is a known global regulator, it appears that an essential virulence factor was not induced (24,79).
Mouse typhoid infections are directly dependent on the toxicity of lipid A (9). A S.
typmmurium waaN, lacking a single acyl chain on its lipid A domain of LPS molecules, was 90% avirulent in BALB/c mice (9). Extremely high counts of bacteria (10 9 cfu) were recovered from the livers and spleens of the mice (9). Most of the mice carrying the high bacterial loads slowly cleared the Salmonella from their organs (9). Perhaps acetyl phosphate, regulating over fifty different twocomponent signal transduction systems in E. coli, is involved in the regulation of lipid A biosynthesis (24). The SR-11 ptd mutant strain may also have an altered lipid A component, since it cannot synthesize acetyl phosphate from acetyl-CoA.
Another possibility of an essential virulence factor not induced in the SR-11 ptamutant strain is the SPI-2 type III secretion system (22,56). The two-component signal transduction system, SsrA-SsrB, is required for SPI-2 gene expression of the SPI-2 type Ill secretion system as well as its translocated effectors (22,56). SsrA-SsrB is positively regulated by the OmpR-EnvZ two-component system (22,56). Acetyl phosphate is a major secondary source of phosphoryl groups for response regulators of two-component signal transduction pathways; in E. coli, acetyl phosphate has been shown to phosphorylate OmpR in vivo (24,25,79,190). Acetyl phosphate is not synthesized in the SR-11 ptamutant strain.
The efcA-encoded NAD-dependent malic enzyme and acetyl phosphate may act in concert as a repressor of the era gene. The SR-11 sfcAand SR-11 ptamutants both exhibited significant (P = 0.05) up-regulation of era promoter activity for growth on both gluconeogenic and glycolytic substrates. This significant up-regulation, combined with their virulence da~ suggest that the sfeA-encoded NAD-dependent malic enzyme and acetyl phosphate may act in concert as corepressors of the era gene. Acetyl phosphate may phosphorylate or bind to the tetrameric NADdependent malic enzyme causing a conformational change in the enzyme (24,79,81,321) To date, the NAD-dependent malic enzyme has only been crystallized from the human mitochondrion, which is 57% similar to the efcA-encoded NAD-dependent malic enzyme in S.
typhimurium (277,278,322). X-ray diffraction reveals a helix-tum-helix motif which is a recognition motif common to many proteins that bind DNA (322,323). However, this helix-tum-helix motif may not exist in the sfeA-encoded NAD-dependent malic enzyme. The.protein sequence of the NADdependent malic enzyme from S. typhimurium does not contain the DNA-binding helix-tum-helix consensus sequence of the Lacl-GalR family of bacterial transcription regulatory factors (321,324).
The Lacl-GalR family of bacterial transcription regulatory factors contains over 25 regulatory DNAbinding proteins including the CRP and Cra proteins (324,325). The NAD-dependent malic enzyme may bind DNA by an alternative mechanism from the Lacl-GalR family of bacterial transcription regulatory factors.
Does the data fully support the hypothesis? One component of the data does not: the units of ~-galactosidase in the SR-11 sfeAmutant grown on glucose and fructose (3.8 units and 2.6 units respectively) are not at the same levels observed for growth on succinate and malate (7.0 units and 8.2 units respectively) (Table 8). If the efcA-encoded NAD-dependent malic enzyme is an integral component of a repressor of the era gene, it seems logical to expect similar J3-galactosidase activity for growth in both glycolytic and gluconeogenic substrates; the era gene should be derepressed in a SR-11 sfcAmutant. The ratio of units of J3-galactosidase, for each substrate, relative to writs of J3galactosidase for the SR-11 cradl craz integrant control strain are similarly increased, but the actual units are not (Table 8). It appears that the transcription of the era gene, like many other genes and operons, is subject to regulation by multiple regulatory proteins or multiple forms of regulation.
The orientation of the era promoter-/acZ transcriptional fusion versus the wild-type era gene in the SR-11 cradl craz strains. A second possibility, although tmlikely, to explain the significant differences in the era promoter activities of the SR-11 maeR cradl craz integrant, SR-11 sfcAcradl craz integrant, the SR-11 ptacradl craz integrant, and the SR-11 cradl craz integrant control strain, involves the orientation of the era promoter-lacZ transcriptional fusion versus the wildtype era gene. If the orientation of the era promoter-/acZtranscriptional fusion and the wild-type era gene, in the SR-11 maeR cradl craz integrant and the SR-11 cradl craz integrant control strain, is inverted versus the other two strains, then there is a slight possibility of an transcriptional regulatory effect occurring in a distant location upstream of the era promoter-/acZtranscriptional fusion ( Figure   21). This seems highly improbable since approximately 947 bp ofDNA upstream of the era promoter was engineered into these constructs to retain any possible regulatory sequences ( Figure 5) (256,259).
Reported regulation of the era gene by catabolite repression. It has been suggested, that FruR (encoded by the fruR gene; homologous to the Cra protein in S. typhimurium) synthesis in E. coli is subject to control by the cyclic AMP receptor protein (CRP) (213). Catabolite repression involves the cytoplasmic sensor of carbon and energy concentrations, cAMP, and the CRP protein; cAMP binds to CRP at specific DNA sequences in promoters, induces bends in the DNA, and interacts with RNA polymerase to promote transcriptional initiation (184,185). However, the mechanism of cAMP-CRP regulation varies; CRP can function not only as an activator, but also as a repressor depending where it binds relative to the promoter ( 184,185). Adenylate cyclase, which catalyzes the formation of cAMP from ATP, is more active when cellular concentrations of catabolites are low and less active when catabolite concen~ations are high ( 186). The synthesis of cAMP is regulated by a protein phosphorylation mechanism that is catalyzed by the PTS system; exogenous PTS sugars (i.e. glucose, fructose, mannitol) inhibit the synthesis of cAMP (91 , 94, 184).
Catabolite repression of the era gene would explain the lower f3-galactosidase activity obserVed in SR-11 sfcAcradl craz integrant and the SR-11 cradl craz integrant control strain grown in glycolytic substrates versus gluconeogenic substrates ( Table 6, 8). For transcriptional activation, the CRP protein generally binds at two sites upstream of the promoter: -45 to - 49 and -70 to -74 bp upstream of the transcriptional start site (326). In order to determine ifthe era gene is subjected to catabolite repression, the DNA binding consensus sequence of the CRP protein ] was compared to the DNA sequence of the era promoter of S. typhimurium (71,323). Absolutely no significant homology was detected within 200 bp upstream and downstream of the + 1 transcription start site; the CRP does not appear to bind to the era promoter and is not subjected to catabolite repression (183).
Putative phosphorylated-ArcA DNA-binding sequence in the promoter of the era gene. of the era promoter is within the DNA sequence 5'-ATGTIAACGATIITAA-3' and the tmderlined bases match the first six bases of the ArcA-P DNA binding site consensus sequence (33,193). A gap of five bp follows and then the remaining four bases (italicized) of the ArcA-P DNA binding site consensus sequence continues. It is possible that the ArcA-P response regulator can bind and repress the transcription of the era gene at this ArcA-P consensus sequence; considerable diversity in the consensus sequence, specifically in the last four bases, occurs in many operons regulated by the ArcA-ArcB two component regulatory system (193). For example, diversity in base substitution of the ArcA-P binding site consensus sequence is encountered in gltA (citrate synthase) and the sdhCDAB (succinate dehydrogenase) promoters (193). At least eleven other promoters, shown to bind the ArcA-P response regulator, contain base substitutions (including C and G base substitutions) in the last four base sequences of the ArcA-P binding site consensus sequence (193).

68
Summary. Mutations in thejbp, maeB, or the sfrA genes of S. typhimurium SR-11 remained virulent in BALB/c mice when administered peroraly. Conversely, a mutation in the pta gene attenuated the S. typhimurium SR-11 strain, and avirulence was observed in BALB/c mice. The era promoter activity was higher in the S. typhimurium SR-11 cradl craz integrant control strain when grown on gluconeogenic substrates versus glycolytic substrates. A mutation in the maeB gene slightly down-regulated the era promoter activity when grown on either gluconeogenic or glycolytic substrates.
In contrast, mutations in either the sfrA or the pta gene significantly (P = 0.05) up-regulated the era promoter activity for growth on both gluconeogenic and glycolytic substrates. This significant upregulation, combined with their virulence in BALB/c mice, suggest that the sfrA-encoded NADdependent malic enzyme and acetyl phosphate act in concert as a repressor of the era gene. It also appears that the era gene, like many other genes and operons, may be subject to regulation by multiple regulatory proteins or multiple forms of regulation.      (Table 1 ). DEL denotes a primer used for deleting, and confirming the the deletion, of a gene. Flank denotes a primer flanking the gene of interest used for confirming the wild-type and deleted genes.   (Table 1 ). DEL denotes a primer used for deleting, and confirming the deletion, of a gene. Flank denotes a primer flanking the gene of interest used for confirming the wild-type and deleted genes. COMP denotes a primer utilized to clone, and confirm the cloning, of a gene. A cloned gene, amplified with a COMP primer, was intended to complement a deleted gene. The /acZ/reverse primer was utilized to confirm the cra-lacZ transcriptional fusion in S. typhimurium SR-11 cradl craz integrant strains.

.---
The M9 minimal agar plates, supplemented with various carbon sources, were incubated @ 3 7°C for 45 hours. A ( +) denotes growth to at least the third quadrant on the plate. A (-) denotes no growth.  • assayed @ OD 500 of~ 1.0 • ~-galactosidase assays were performed~ triplicate for each culture.
• assayed @ OD 500 of::::: 1.0 . • (3-galactosidase assays were performed in triplicate for each culture.  Pathway, the pentose phosphate cycle, the tricarboxylic acid cycle (TCA), and the Entner-Doudoroff Pathway are shown above and represent the fueling pathways in central metabolism (72,73). The glyoxylate shllllt, when induced, bypasses several reactions in the TCA cycle (74). Two steps in the reversible Emlxlen-Meyerhofpathway are replaced in the gluconeogenic pathway by the pps-encoded enzyme phosphoenolpyruvate synthetase and thejbp-encoded enzyme fructoste-1 ,6-bisphosphatase (73,82,120). The pathways for acetate metabolism, fructose metabolism, and fatty acid oxidation ( ~leate) are also shown above ( 172,173,(177)(178)(179). P represents phosphate in the diagram. Diagram highly modified from references (73,212). utilizing a non-limiting glucose source. The full TCA cycle is not required under these conditions because the bulk of energy is derived from glycolysis (74,78). In this rich environment, cells produce excess acetyl-CoA which drains through acetyl phosphate, to further produce A 1P with the associated secretion of acetate (78,79). As cell density increases, the build-up of acetate also serves to increase the size of the acetyl phosphate pool (78). P represents phosphate in the diagram. Diagram highly modified from references (69,73,212).  Figure 3. Overview of the transcriptional regulatory effects of the Cra protein on key enzymes in the Emlxlen-Meyerhof pathway, the TCA cycle, the glyoxylate shtmt, the gluconeogenic pathway, and the Entner-Doudoroffpathway in Salmonella typhimurium; (+)denotes genes positively regulated by Cra~ ( ~) denotes genes negatively regulated by Cra. P represents phosphate in the diagram. Diagram highly modified from reference (69, 73 , 212).   Figure 5. A control strain, S. typhimurium SR-11 cradl craz integrant, containing a era promoter-laeZ transcriptional fusion, was obtained from Mary P. Leatham (256). The strain was constructed by conjugative-dependent allele replacement utilizing the suicide vector pML55cradl-craz (256,257) The control strain carries a partial upstream ilvl gene, an upstream RvH gene, the upstream intergenic sequence, and the era promoter ligated to the promoterless laeZ gene from pCB267 (256,258). Approximately 947 bp of DNA upstream of the era promoter were engineered into this construct to retain any possible regulatory sequences (256,259). The 3' end of the lacZ reporter gene is ligated to approximately 200 bp of the remaining truncated era gene, followed by approximately 534 bp of downstream DNA (a partialyabB gene) (256). The SR-11 maeR cradl craz integrant, SR-11 sfeAcradl craz integrant, the SR-11 ptd cradl craz integrant were also constructed by this technique employing the suicide vector pML55cradl-craz. Xho I. The plasmid contains the entire jbp gene with partial upstream and downstream yifG and yifF genes. The amplicon was ligated into the Barn HI/Not I sites of the phagemid cloning vector pBluescript II SK ( +) ( Table 1 ). The munbers in parentheses, which denote base pairs, are approximate.   Table 2). The amplicon effectively deleted 878 bp from the fbp gene. The PCR prcxluct was then digested with Barn Ill and ligated to itself The munbers in parentheses, which denote base pairs, are approximate. ampicillin resistance Figure 8. The plasmid p~jbp was mapped with the restriction endonucleases Not I, Pst I, and Xho I. The plasmid contains a deleted jbp gene with partial upstream and downstream yifG and yifF genes in the cloning vector pBluescript II SK ( +) ( Table I). A Bam Ill site was engineered adjacent to the deleted fbp gene for insertion of a chloramphenicol resistance cassette. The numbers in parentheses, which denote base pairs, are approximate.

Afbp
A(bp  ' BamHl (5229) chlorarnphenic ol resistance cassette.,..,.... Pst I, and Xho I. The plasmid contains a chloramphenicol resistance cassette adjacent to the deleted fbp gene and is flanked by partial upstream and downstream yijG and yifF genes. This insert is contained within the Bam IIl!Not I sites of the cloning vector pBluescript II SK(+) ( Table 1). The numbers in parentheses, which denote base pairs, are approximate. . c::::Jtl Kan.R Figure 13. The linear DNA amplicon containing the kanamycin resistance cassette and homologous flanking sequences to the maeB gene was electroporated into competent SR-11 wild-type cells containing the plasmid pKD46 (Table 1) (281 ). The gene encoding A. Red recombinase, which greatly enhances the rate of homologous recombination, is located the plasmid pKD46 and is controlled by an arabinose-inducible promoter. The wild-type gene is replaced by a deleted version and an antibiotic resistance marker in one step (281 ). This method was also used to make permanent deletions in the sfcA and pta genes. The SR-11 wildtype yields a 3.1 kb band which represents the wild-typejbp gene (994 bp) plus approximately 1 kb upstream (a partialyifG gene) and approximately 1 kb downstream (a partialyifF gene). Lane 3: The SR-11 f!..jbj£,am 1-3 integrant exhibits two bands: one corresponding to the wild-type band and a 3.5 kb band representing the deleted.fbp gene plus the chloramphenicol resistance cassette. Lane 4: The SR-11 4/b/£,am AX-3 segregant yields one 3.5 kb band denoting the deletedfbp gene plus the chloramphenicol resistance cassette. The PCR reactions utilized the yi.jF/forward and yi.jG/reverse Primers (  sue CD -- Figure 15. The ratio of (3-galactosidase units for each substrate assayed relative to the (3-galactosidase units for fructose. (3-galactosidase assays, for era promoter activity, were performed in SR-11 cradl craz integrant grown in M9 minimal broth supplemented separately with D-fructose, D-glucose, Dgluconic acid potassium salt, potassium acetate, sodium citrate, L(-)malic acid, phospho( enol)pyruvate monosodium salt, pyruvic acid sodium salt, oleic acid sodium salt, and succinic acid dissodilllll salt (232,233). A(+) denotes genes positively regulated by the Cra protein; a (-) denotes genes negatively regulated by the Cra protein. P represents phosphate in the diagram. Diagram highly modified from references (69,73,212).   (sfcA-DEL!furward & reverse primers) inserted in the deleted sfcA gene. Lane 6: A 3. 7 kb band (sfcA-Flank!forward & reverse primers) corresponding to the wild-type sfcA gene (1697 bp) plus the entire upstream rps V gene, partial upstream yddX gene, and partial downstream adhP gene. Lane 7: A (-) DNA control. Lanes 2-5 utilized SR-11 sfi:Agenomic template DNA and Lane 6 utilized SR-11 wildtype genomic template DNA. The PCR reactions contained 1.8 mM MgCli and subjected to an initial denaturation of94°C for 5 minutes followed by 32 cycles of: 94°C for 1 minute, 57°C for 1.5 minutes, and 72°C for 3 minutes. A final extension of 72°C for 7 minutes completed the PCR . PCR amplification of the pta gene in S. typhimurium SR-11 ptd and SR-11 wild-type. Lane 1: 1 kb DNA ladder. Lane 2: A (-) DNA control. Lane 3: A 2.0 kb band (pta-Flanklforward & pta-DEL/reverse primers) corresponds to the kanamycin cassette, deleted pta gene, and the partial upstream ackA gene. Lane 4: A 1.9 kb band (pta-Flankl reverse & pta-DEL!forward primers) reveals the kanamycin cassette, deleted pta gene, and downstream sequences. Lane 5: The 2.3 kb band (pta-Flank /forward & reverse primers) represents the partial upstream ackA gene, the deleted pta gene, the kanamycin cassette, plus downstream sequences. Lane 6: The 1.5 kb kanamycin resistance cassette (pta-DEL!forward & reverse primers) inserted in the deletedpta gene. Lane 7: A 2.6 kb band (pta-Flank:/forward & reverse primers) corresponding to the partial upstream ackA gene, the wild-type pta gene (2155 bp), plus downstream sequences. Lanes 3-6 utilized SR-11 ptagenomic template DNA and Lane 7 utilized SR-11 wild-type genomic template DNA. The PCR reactions contained 1.5 mM MgCl2 and were subjected to an initial denaturation of94°C for 5 minutes followed by 32 cycles of: 94°C for 1 minute, 54°C for 1.5 minutes, and 72°C for 2.5 minutes.  Figure 21. A second possibility, although unlikely, to explain the significant differences in the era promoter activities of the SR-11 maeE cradl craz integrant, SR-11 sfeA-cradl craz integrant, the SR-11 ptacradl craz integrant, and the SR-11 cradl craz integrant control strain, involves the orientation of the era promoter-/aeZtranscriptional fusion versus the wild-type era gene. If the orientation of the era promoter-/acZ transcriptional fusion and the wild-type era gene, in the SR-11 cradl craz integrant control strain and the SR-11 maeE cradl craz integrant strain, is inverted versus the other two strains, then there is a slight possibility of an transcriptional regulatory effect occurring in a distant location upstream of the era promoter-/aeZtranscriptional fusion. This seems highly improbable since approximately 947 bp of DNA upstream of the era promoter was engineered into these constructs to retain any possible regulatory sequences (256,259). 3 1. Uzzau, S., L. Bossi, and N. Figueroa-Bossi. 2002