Dermacentor andersoni transmission of Francisella tularensis subsp. novicida reflects bacterial colonization, dissemination, and replication coordinated with tick feeding.

Ticks serve as biological vectors for a wide variety of bacterial pathogens which must be able to efficiently colonize specific tick tissues prior to transmission. The bacterial determinants of tick colonization are largely unknown, a knowledge gap attributed in large part to the paucity of tools to genetically manipulate these pathogens. In this study, we demonstrated that Francisella tularensis subsp. novicida, for which a complete two-allele transposon mutant library has been constructed, initially infects the midguts of 100% of acquisition-fed Dermacentor andersoni nymphs, with stable colonization and replication during a subsequent molt. Increased dissemination to and marked replication within the salivary gland was closely linked to a second (transmission) feed and culminated in secretion of bacteria into the saliva and successful transmission. Simultaneous testing of multiple mutants resulted in total bacterial levels similar to those observed for single mutants. However, there was evidence of a bottleneck during colonization, resulting in a founder effect in which the most successful mutant varied when comparing individual ticks. Thus, it is essential to assess mutant success at the level of the tick population rather than in individual ticks. The ability of F. tularensis subsp. novicida to recapitulate the key physiological events by which bacteria colonize and are transmitted by ixodid ticks provides a new genome-wide approach to identify the required pathogen molecules and pathways. The identification of specific genes and, more importantly, conserved pathways that function at the tick-pathogen interface will accelerate the development of new methods to block transmission.

Salmonella pathogenicity island 2-dependent macrophage death is mediated in part by the host cysteine protease caspase-1.

Salmonella typhimurium invades host macrophages and can either induce a rapid cell death or establish an intracellular niche within the phagocytic vacuole. Rapid cell death requires the Salmonella pathogenicity island (SPI)1 and the host protein caspase-1, a member of the pro-apoptotic caspase family of proteases. Salmonella that do not cause this rapid cell death and instead reside in the phagocytic vacuole can trigger macrophage death at a later time point. We show here that the human pathogen Salmonella typhi also triggers both rapid, caspase-1-dependent and delayed cell death in human monocytes. The delayed cell death has previously been shown with S. typhimurium to be dependent on SPI2-encoded genes and ompR. Using caspase-1(-/-) bone marrow-derived macrophages and isogenic S. typhimurium mutant strains, we show that a large portion of the delayed, SPI2-dependent death is mediated by caspase-1. The two known substrates of activated caspase-1 are the pro-inflammatory cytokines interleukin-1beta (IL-1beta) and IL-18, which are cleaved to produce bioactive cytokines. We show here that IL-1beta is released during both SPI1- and SPI2-dependent macrophage killing. Using IL-1beta(-/-) bone marrow-derived macrophages and a neutralizing anti-IL-18 antibody, we show that neither IL-1beta nor IL-18 is required for rapid or delayed macrophage death. Thus, both rapid, SPI1-mediated killing and delayed, SPI2-mediated killing require caspase-1 and result in the secretion of IL-1beta, which promotes inflammation and may facilitate the spread of Salmonella beyond the gastrointestinal tract in systemic disease.

Actin-based motility is sufficient for bacterial membrane protrusion formation and host cell uptake.

Shigella flexneri replicates in the cytoplasm of host cells, where it nucleates host cell actin filaments at one pole of the bacterial cell to form a 'comet tail' that propels the bacterium through the host's cytoplasm. To determine whether the ability to move by actin-based motility is sufficient for subsequent formation of membrane-bound protrusions and intercellular spread, we conferred the ability to nucleate actin on a heterologous bacterium, Escherichia coli. Previous work has shown that IcsA (VirG), the molecule that is necessary and sufficient for actin nucleation and actin-based motility, is distributed in a unipolar fashion on the surface of S. flexneri. Maintenance of the unipolar distribution of IcsA depends on both the S. flexneri outer membrane protease IcsP (SopA) and the structure of the lipopolysaccharide (LPS) in the outer membrane. We co-expressed IcsA and IcsP in two strains of E. coli that differed in their LPS structures. The E. coli were engineered to invade host cells by expression of invasin from Yersinia pseudotuberculosis and to escape the phagosome by incubation in purified listeriolysin O (LLO) from Listeria monocytogenes. All E. coli strains expressing IcsA replicated in host cell cytoplasm and moved by actin-based motility. Actin-based motility alone was sufficient for the formation of membrane protrusions and uptake by recipient host cells. The presence of IcsP and an elaborate LPS structure combined to enhance the ability of E. coli to form protrusions at the same frequency as S. flexneri, quantitatively reconstituting this step in pathogen intercellular spread in a heterologous organism. The frequency of membrane protrusion formation across all strains tested correlates with the efficiency of unidirectional actin-based movement, but not with bacterial speed.

The making of a gradient: IcsA (VirG) polarity in Shigella flexneri.

The generation and maintenance of subcellular organization in bacteria is critical for many cell processes and properties, including growth, structural integrity and, in pathogens, virulence. Here, we investigate the mechanisms by which the virulence protein IcsA (VirG) is distributed on the bacterial surface to promote efficient transmission of the bacterium Shigella flexneri from one host cell to another. The outer membrane protein IcsA recruits host factors that result in actin filament nucleation and, when concentrated at one bacterial pole, promote unidirectional actin-based motility of the pathogen. We show here that the focused polar gradient of IcsA is generated by its delivery exclusively to one pole followed by lateral diffusion through the outer membrane. The resulting gradient can be modified by altering the composition of the outer membrane either genetically or pharmacologically. The gradient can be reshaped further by the action of the protease IcsP (SopA), whose activity we show to be near uniform on the bacterial surface. Further, we report polar delivery of IcsA in Escherichia coli and Yersinia pseudotuberculosis, suggesting that the mechanism for polar delivery of some outer membrane proteins is conserved across species and that the virulence function of IcsA capitalizes on a more global mechanism for subcellular organization.

Salmonella-induced macrophage death: the role of caspase-1 in death and inflammation.

Salmonella typhimurium invades host macrophages and can induce either an almost immediate cell death or establish an intracellular niche within the phagocytic vacuole. Rapid cell death depends on the Salmonella pathogenicity island SPI1 and the host protein caspase-1, a member of the pro-apoptotic caspase family of proteases. Caspase-1-dependent cell death leads to the activation of the potent pro-inflammatory cytokines interleukin (IL)-1beta and IL-18 to produce bioactive cytokines. Animal studies indicate that the activation of these cytokines is necessary for efficient colonization of the mouse gastrointestinal tract. Salmonella that reside in the phagocytic vacuole do not cause this early cell death and can trigger a macrophage death at a much later time point. This late-phase cell death is dependent on SPI2-encoded genes and ompR.

Apoptosis as a common bacterial virulence strategy.

The comparison of common strategies used by bacterial pathogens to overcome host defenses provides us with the opportunity to analyze the biology of pathogenicity, as well as point out the unique interactions between a particular pathogen and its host. Here we compare and contrast apoptosis induced by three enteric pathogens, Salmonella, Shigella, and Yersinia. We point out that all three enteric pathogens induce apoptosis in macrophages in vitro, but the proposed mechanisms are quite different. Yersinia induces apoptosis by inhibiting the translocation of the transcriptional activator, NF-kappaB, into the nucleus, which results in the suppression of TNFalpha production; whereas Salmonella- and Shigella-induced apoptosis depend on the activation of caspase-1 (casp-1). The result of casp-1 activation is to induce apoptosis as well as to process the proinflammatory cytokines, pro-IL-1beta and pro-IL18 into their mature bioactive forms. Thus, in contrast to Yersinia, Salmonella and Shigella-induced apoptosis results in a proinflammatory cascade.

Salmonella exploits caspase-1 to colonize Peyer's patches in a murine typhoid model.

Salmonella typhimurium invades host macrophages and induces apoptosis and the release of mature proinflammatory cytokines. SipB, a protein translocated by Salmonella into the cytoplasm of macrophages, is required for activation of Caspase-1 (Casp-1, an interleukin [IL]-1beta-converting enzyme), which is a member of a family of cysteine proteases that induce apoptosis in mammalian cells. Casp-1 is unique among caspases because it also directly cleaves the proinflammatory cytokines IL-1beta and IL-18 to produce bioactive cytokines. We show here that mice lacking Casp-1 (casp-1(-/)- mice) had an oral S. typhimurium 50% lethal dose (LD(50)) that was 1,000-fold higher than that of wild-type mice. Salmonella breached the M cell barrier of casp-1(-/)- mice efficiently; however, there was a decrease in the number of apoptotic cells, intracellular bacteria, and the recruitment of polymorphonuclear lymphocytes in the Peyer's patches (PP) as compared with wild-type mice. Furthermore, Salmonella did not disseminate systemically in the majority of casp-1(-/)- mice, as demonstrated by significantly less colonization in the PP, mesenteric lymph nodes, and spleens of casp-1(-/)- mice after an oral dose of S. typhimurium that was 100-fold higher than the LD(50). The increased resistance in casp-1(-/)- animals appears specific for Salmonella infection since these mice were susceptible to colonization by another enteric pathogen, Yersinia pseudotuberculosis, which normally invades the PP. These results show that Casp-1, which is both proapoptotic and proinflammatory, is essential for S. typhimurium to efficiently colonize the cecum and PP and subsequently cause systemic typhoid-like disease in mice.

The Salmonella invasin SipB induces macrophage apoptosis by binding to caspase-1.

Recently, Salmonella spp. were shown to induce apoptosis in infected macrophages. The mechanism responsible for this process is unknown. In this report, we establish that the Inv-Spa type III secretion apparatus target invasin SipB is necessary and sufficient for the induction of apoptosis. Purified SipB microinjected into macrophages led to cell death. Binding studies show that SipB associates with the proapoptotic protease caspase-1. This interaction results in the activation of caspase-1, as seen in its proteolytic maturation and the processing of its substrate interleukin-1beta. Caspase-1 activity is essential for the cytotoxicity. Functional inhibition of caspase-1 activity by acetyl-Tyr-Val-Ala-Asp-chloromethyl ketone blocks macrophage cytotoxicity, and macrophages lacking caspase-1 are not susceptible to Salmonella-induced apoptosis. Taken together, the data demonstrate that SipB functions as an analog of the Shigella invasin IpaB.

Yersinia-induced apoptosis in vivo aids in the establishment of a systemic infection of mice.

Pathogenic Yersinia cause a systemic infection in mice that is dependent on the presence of a large plasmid encoding a number of secreted virulence proteins called Yops. We previously demonstrated that a plasmid-encoded Yop, YopJ, was essential for inducing apoptosis in cultured macrophages. Here we report that YopJ is a virulence factor in mice and is important for the establishment of a systemic infection. The oral LD50 for a yopJ mutant Yersinia pseudotuberculosis increases 64-fold compared with wild-type. Although the yopJ mutant strain is able to reach the spleen of infected mice, the mutant strain seldom reaches the same high bacterial load that is seen with wild-type Yersinia strain and begins to be cleared from infected spleens on day 4 after infection. Furthermore, when in competition with wild-type Yersinia in a mixed infection, the yopJ mutant strain is deficient for spread from the Peyer's patches to other lymphoid tissue. We also show that wild-type Yersinia induces apoptosis in vivo of Mac-1(+) cells from infected mesenteric lymph nodes or spleens, as measured by quantitative flow cytometry of TUNEL (Tdt-mediated dUTP-biotin nick-end labeling)-positive cells. The levels of Mac-1(+), TUNEL+ cells from tissue infected with the yopJ mutant strain were equivalent to the levels detected in cells from uninfected tissue. YopJ is necessary for the suppression of TNF-alpha production seen in macrophages infected with wild-type Yersinia, based on previous in vitro studies (Palmer, L.E., S. Hobbie, J.E. Galan, and J.B. Bliska. 1998. Mol. Microbiol. 27:953-965). We conclude here that YopJ plays a role in the establishment of a systemic infection by inducing apoptosis and that this is consistent with the ability to suppress the production of the proinflammatory cytokine tumor necrosis factor alpha.

Macrophage-dependent induction of the Salmonella pathogenicity island 2 type III secretion system and its role in intracellular survival.

Salmonella pathogenicity island 2 (SPI-2) encodes a putative type III secretion system necessary for systemic infection in animals. We have investigated the transcriptional organization and regulation of SPI-2 by creating gfp fusions throughout the entire gene cluster. These gfp fusions demonstrated that SPI-2 genes encoding structural, regulatory and previously uncharacterized putative secreted proteins are preferentially expressed in the intracellular environment of the host macrophage. Furthermore, the transcription of these genes within host cells was dependent on the two-component regulatory system SsrA/SsrB and an acidic phagosomal environment. Most SPI-2 mutants failed to replicate to the same level as wild-type strains in murine macrophages and human epithelial cells. In orally infected mice, SPI-2 mutants colonized the Peyer's patches but did not progress to the mesenteric lymph nodes. We conclude that SPI-2 genes are specifically expressed upon entry into mammalian cells and are required for intracellular growth in host cells in vivo and in vitro.

Yersinia signals macrophages to undergo apoptosis and YopJ is necessary for this cell death.

Pathogenic Yersinia spp. carry a large common plasmid that encodes a number of essential virulence determinants. Included in these factors are the Yersinia-secreted proteins called Yops. We analyzed the consequences of wild-type and mutant strains of Yersinia pseudotuberculosis interactions with the macrophage cell line RAW264. 7 and murine bone marrow-derived macrophages. Wild-type Y. pseudotuberculosis kills approximately 70% of infected RAW264.7 macrophages and marrow-derived macrophages after an 8-h infection. We show that the cell death mediated by Y. pseudotuberculosis is apoptosis. Mutant Y. pseudotuberculosis that do not make any Yop proteins no longer cause host cell death. Attachment to host cells via invasin or YadA is necessary for the cell death phenotype. Several Yop mutant strains that fail to express one or more Yop proteins were engineered and then characterized for their ability to cause host cell death. A mutant with a polar insertion in YpkA Ser/Thr kinase that does not express YpkA or YopJ is no longer able to cause apoptosis. In contrast, a mutant no longer making YopE or YopH (a tyrosine phosphatase) induces apoptosis in macrophages similar to wild type. When yopJ is added in trans to the ypkAyopJ mutant, the ability of this strain to signal programmed cell death in macrophages is restored. Thus, YopJ is necessary for inducing apoptosis. The ability of Y. pseudotuberculosis to promote apoptosis of macrophages in cell culture suggests that this process is important for the establishment of infection in the host and for evasion of the host immune response.

Functional analysis of ssaJ and the ssaK/U operon, 13 genes encoding components of the type III secretion apparatus of Salmonella Pathogenicity Island 2.

We have investigated the structure and transcriptional organization of 13 genes of Salmonella Pathogenicity Island 2 (SPI2) that encode components of the second type III secretion apparatus of Salmonella typhimurium. ssaK, L, M, V, N, O, P, Q, R, S, T, U constitute one operon of 10 kb. ssaJ lles upstream of ssaK and is the terminal gene of another operon. The deduced products of ssaJ, ssaK, ssaV, ssaN, ssaO, ssaQ, ssaR, ssaS, ssaT, and ssaU show greatest similarity to the Yersinia spp. genes yscJ, yscL, lcrD, yscN, yscO, yscQ, yscR, yscS, yscT, and yscU, respectively. The products of the ssaL, ssaM and ssaP genes do not have significant similarity to products of other type III secretion systems, and might be important for the specific function of the SPI2 type III secretion system. Bacterial strains carrying different ssa mutations display minor alterations in terms of serum sensitivity when compared with the wild-type strain, but none are defective in replication within macrophage-like RAW 264.7 cells. However, some of the ssa mutant strains invade HEp2 cells less efficiently and are less cytotoxic to RAW 264.7 macrophages than the wild-type strain. We show that the invasion defect is correlated with a lack of SipC in culture supernatants of these mutant strains. SipC is a product of the SPI1 type III secretion system of S. typhimurium, and is important for epithelial cell invasion. Therefore, mutations in SPI2 can affect the SPI1 secretion system, which raises the possibility of an interaction between the two type III secretion systems.

Salmonella typhimurium invasion induces apoptosis in infected macrophages.

Invasive Salmonella typhimurium induces dramatic cytoskeletal changes on the membrane surface of mammalian epithelial cells and RAW264.7 macrophages as part of its entry mechanism. Noninvasive S. typhimurium strains are unable to induce this membrane ruffling. Invasive S. typhimurium strains invade RAW264.7 macrophages in 2 h with 7- to 10-fold higher levels than noninvasive strains. Invasive S. typhimurium and Salmonella typhi, independent of their ability to replicate intracellularly, are cytotoxic to RAW264.7 macrophages and, to a greater degree, to murine bone marrow-derived macrophages. Here, we show that the macrophage cytotoxicity mediated by invasive Salmonella is apoptosis, as shown by nuclear morphology, cytoplasmic vacuolization, and host cell DNA fragmentation. S. typhimurium that enter cells causing ruffles but are mutant for subsequent intracellular replication also initiate host cell apoptosis. Mutant S. typhimurium that are incapable of inducing host cell membrane ruffling fail to induce apoptosis. The activation state of the macrophage plays a significant role in the response of macrophages to Salmonella invasion, perhaps indicating that the signal or receptor for initiating programmed cell death is upregulated in activated macrophages. The ability of Salmonella to promote apoptosis may be important for the initiation of infection, bacterial survival, and escape of the host immune response.

Applications for green fluorescent protein (GFP) in the study of host-pathogen interactions.

The green fluorescent protein (GFP) from Aequorea victoria is a novel fluorescent marker that has potential use in the study of bacterial pathogenicity. To explore some of the potential applications of GFP to the study of host-parasite interactions, we constructed two GFP expression vectors suitable for different facultative intracellular bacterial pathogens. The first expression vector was tested in the enteric pathogens, Salmonella typhimurium and Yersinia pseudotuberculosis, and the second vector tested in Mycobacterium marinum (Mm). Both expression vectors were found to be stable and to direct high levels of GFP synthesis. Standard epifluorescence microscopy was used to detect all three bacterial pathogenic species during the early and late stages of infection of live mammalian cells. Mm expressing gfp was also visualized in infected animal tissues. gfp expression did not adversely affect bacterial survival, nor did it compromise entry into mammalian cells or their survival within macrophages. In addition, all three gfp-expressing bacterial pathogens could be detected and sorted in a flow cytometer, either alone or in association with epithelial cells or macrophages. Therefore, GFP not only provides a convenient tool to image pathogenic bacteria, but allows the quantitative measurement of bacterial association with mammalian cells.

Host restriction phenotypes of Salmonella typhi and Salmonella gallinarum.

Salmonella typhi and Salmonella gallinarum phenotypes correlated with mouse host restriction have been identified by using in vitro and in vivo systems. S. typhi is capable of entering the murine intestinal epithelium via M cells, as is Salmonella typhimurium, which causes systemic infection in the mouse. But, unlike S. typhimurium, S. typhi does not destroy the epithelium and is cleared from the Peyer's patches soon after M-cell entry. S. gallinarum appears to be incapable of entering the murine Peyer's patch epithelium. Our in vitro evidence suggests that S. gallinarum is taken up in murine phagocytic cells by a mechanism different from that of S. typhimurium. S. typhimurium is taken up at a higher frequency and is maintained at higher viable counts throughout a 24-h time course in a murine macrophage-like cell line than are S. gallinarum and S. typhi.

Medicaid and managed care: opportunities for innovative service delivery to vulnerable populations.

The crisis in state "safety-net" funding for high-need, high-risk, high-cost vulnerable populations has led to the development of creative strategies and financing mechanisms by the states for delivering essential community-based services. This is also an opportunity for private-sector providers who can recognize the special needs of these populations, form links with other key constituencies and apply managed care concepts to long-term health and human service delivery.

Cloning of Bordetella bronchiseptica urease genes and analysis of colonization by a urease-negative mutant strain in a guinea-pig model.

The genes encoding urease were cloned from Bordetella bronchiseptica and the 5.2 kb of DNA essential for expression analysed in a T7 RNA polymerase transcription-translation system. At least four polypeptides with predicted molecular weights of 69,000, 26,000, 12,200 and 11,000 were found. Partial DNA sequence of the gene encoding the 69,000 Da polypeptide revealed high amino acid identity to the alpha-subunit of Proteus mirabilis urease, UreC and jack bean urease. A stable, unmarked deletion was constructed in this gene to create a urease-negative mutant of B. bronchiseptica. To assess colonization in a guinea-pig model, the urease-negative strain was inoculated with the urease-positive parental strain in a mixed infection. The urease-negative strain out competed the urease-positive strain in the trachea, lungs and caecum. We demonstrate that urease is not essential for B. bronchiseptica colonization of the guinea-pig respiratory and digestive tracts.

The bvgAS locus negatively controls motility and synthesis of flagella in Bordetella bronchiseptica.

The products of the bvgAS locus coordinately regulate the expression of Bordetella virulence factors in response to environmental conditions. We have identified a phenotype in Bordetella bronchiseptica that is negatively controlled by bvg. Environmental signals which decrease (modulate) the expression of bvg-activated genes lead to flagellum production and motility in B. bronchiseptica. Wild-type (Bvg+) strains are motile and produce peritrichous flagella only in the presence of modulating signals, whereas Bvg- (delta bvgAS or delta bvgS) strains are motile in the absence of modulators. The bvgS-C3 mutation, which confers signal insensitivity and constitutive activation of positively controlled loci, eliminates the induction of motility and production of flagellar organelles. The response to environmental signals is conserved in a diverse set of clinical isolates of both B. bronchiseptica and B. avium, another motile Bordetella species; however, nicotinic acid induced motility only in B. bronchiseptica. Purification of flagellar filaments from B. bronchiseptica strains by differential centrifugation followed by CsCl equilibrium density gradient centrifugation revealed two classes of flagellins of Mr 35,000 and 40,000. A survey of clinical isolates identified only these two flagellin isotypes, and coexpression of the two forms was not detected in any strain. All B. avium strains tested expressed a 42,000-Mr flagellin. Amino acid sequence analysis of the two B. bronchiseptica flagellins revealed 100% identity in the N-terminal region and 80% identity with Salmonella typhimurium flagellin. Monoclonal antibody 15D8, which recognizes a conserved epitope in flagellins in members of the family Enterobacteriaceae, cross-reacted with flagellins from B. bronchiseptica and B. avium. Our results highlight the biphasic nature of the B. bronchiseptica bvg regulon and provide a preliminary characterization of the Bvg-regulated motility phenotype.

Structural and genetic analysis of the bvg locus in Bordetella species.

The bvg locus contains two genes, bvgA and bvgS, which control the expression of the virulence-associated genes in Bordetella species by a system similar to the two-component systems used by a variety of bacterial species to respond to environmental stimuli. We determined the nucleotide sequence of the bvg loci of Bordetella parapertussis and Bordetella bronchiseptica and compared them with the previously determined sequence of Bordetella pertussis. The nucleotide and amino acid sequences of the bvg loci of these species are well conserved in those regions coding for the protein domains which have putative kinase and DNA-binding activities. In marked contrast, the region of BvgS that codes for the protein domain with putative sensor activity shows a high degree of variability. In total, we find 198 base-pair changes in the bvg loci of B. parapertussis and B. bronchiseptica relative to the bvg locus of B. pertussis. One hundred and seventy-three of these base-pair changes are identical in B. parapertussis and B. bronchiseptica. This confirms our previous observation that B. parapertussis and B. bronchiseptica are more related to each other than to B. pertussis. We have mapped the mutations that cause phase changes in B. bronchiseptica and we have found that in three cases these are due to spontaneous deletions in the bvgS gene. The wild-type bvg locus present on a multicopy plasmid cannot complement avirulent derivatives of B. bronchiseptica to wild-type levels, but it can do so when the bvgA gene on the plasmid is inactivated. This suggests that hyperexpression of bvgA down-regulates the bvg system.

Phase variants of Bordetella bronchiseptica arise by spontaneous deletions in the vir locus.

Bordetella bronchiseptica is a common respiratory tract pathogen of many mammalian species. Nucleotide sequences from the locus involved in coordinate regulation of B. pertussis virulence factors, vir, were shown to have a high degree of homology to chromosomal DNA from virulent (Vir+) and avirulent (Vir-) strains of B. bronchiseptica. Small deletions, 50 bp to 500 bp, within the vir locus were found in some of the Vir- phase variants. The vir locus and the adjacent 5' portion of the fhaB structural gene were cloned from the parental Vir+ B. bronchiseptica strain on a 23.5 kb BamHI fragment. Restriction enzyme mapping of the cloned B. bronchiseptica vir locus revealed similarities with and differences from the previously cloned B. pertussis vir locus. The cloned B. bronchiseptica vir locus complemented spontaneous Vir- variants of Bordetella pertussis and B. bronchiseptica as well as vir::Tn5 mutants of B. pertussis. Comparison of various functions of the vir loci of B. bronchiseptica and B. pertussis revealed some interesting differences in the coordinate regulation of virulence factors.