1Department of Pathology, Division of Clinical Microbiology, University of Texas Medical Branch at Galveston, 301 University Blvd., Galveston, TX 77555-0740, USA.
Involvement of host cell tyrosine phosphorylation in the invasion of HEp-2 cells by Bartonella bacilliformis
Article first published online: 17 JAN 2006
FEMS Microbiology Letters
Volume 171, Issue 2, pages 191–201, February 1999
How to Cite
Williams-Bouyer, N. M. and Hill, E.McGinnis. (1999), Involvement of host cell tyrosine phosphorylation in the invasion of HEp-2 cells by Bartonella bacilliformis. FEMS Microbiology Letters, 171: 191–201. doi: 10.1111/j.1574-6968.1999.tb13432.x
- Issue published online: 17 JAN 2006
- Article first published online: 17 JAN 2006
- Received 1 September 1998, Revised 6 November 1998, Accepted 24 November 1998
- Bartonella bacilliformis;
- Tyrosine phosphorylation;
- Protein kinase inhibitor;
- Epithelial cell;
- Endothelial cell
We have provided evidence that exposure of human cells to protein kinase inhibitors results in decreased invasion of these cells by Bartonella bacilliformis in a dose-dependent manner. Preincubation of human laryngeal epithelial cells in the presence of genistein, a tyrosine protein kinase inhibitor, decreased the invasion of these cells by B. bacilliformis significantly. Further, exposure of normal human umbilical vein endothelial cells to staurosporine, a potent inhibitor of protein kinase C and some tyrosine protein kinases, resulted in a considerable reduction in the number of organisms internalized by these cells. Moreover, Bartonella infection of HEp-2 cells induced tyrosine phosphorylation of several Triton X-100 soluble proteins with approximate molecular masses of 243, 215 179, 172 (doublet), 160, 145 and 110 kDa that were absent or reduced in the presence of genistein in cells after 1 h of infection. Exposure of HEp-2 cell monolayers to anti-α5 and anti-β1 chain integrin monoclonal antibodies resulted in a moderate decrease in the invasion of these cells, suggesting a possible role of α5β1 integrins in the uptake of Bartonella into nucleated cells.
Bartonella bacilliformis is the causative agent of Carrion's disease and is transmitted to humans by the phlebotomine sandfly found in the Andean valleys of Ecuador, Colombia and Peru. Its predilection for erythrocytes and endothelial cells is manifested in two distinct clinical syndromes known as the Oroya fever syndrome and verruga peruana. The Oroya fever syndrome is characterized by the presence of bartonellae in nearly all of the red blood cells and a resulting febrile anemia, which if not treated with antibiotics, can be fatal. The tissue phase of the disease, verruga peruana, develops in some individuals following resolution of the hematic phase and is characterized by the formation of hemangioma-like lesions in the facial area and lower extremities .
B. bacilliformis invades a variety of different human cell types in vitro including skin fibroblasts, umbilical vein endothelial cells, HeLa cells and laryngeal epithelial cells; but little is known about mechanisms utilized by this hemotrophic bacterium for the invasion process. The inhibition of invasion by cytochalasin D treatment of HUVECs and HEp-2 cells suggests an active role of host cells in the uptake of Bartonella and that actin polymerization and cytoskeletal rearrangements are probably involved in the process . Recently, a two-gene locus was characterized from B. bacilliformis that is associated with the ability to invade human erythrocytes . Further, the polar flagella of Bartonella and deformin  have also been implicated in the invasion of erythrocytes.
The ability of microorganisms to be internalized by host cells is an important virulence attribute of several important pathogens such as Salmonella, Shigella, Listeria and enteropathogenic Escherichia coli. However, the mechanisms describing the entry of Yersinia pseudotuberculosis into mammalian cells are perhaps the most well characterized and include at least three independent pathways that are utilized by Yersinia[6–9]. The inv-mediated pathway is well-characterized and encodes a 103-kDa outer membrane protein, invasin, which enables the organism to interact with multiple β1 integrin receptors that are cell adhesion molecules connected to the cytoskeleton . How the interactions between the integrin receptor and the cytoskeleton generate a signal that results in the invasion of host cells is an important question and the subject of intense investigation [11–13]. The possible involvement of tyrosine phosphorylation in transducing signals from integrins to the cytoskeleton was suggested by the fact that tyrosine protein kinase inhibitors blocked invasin-promoted uptake of E. coli by epithelial cells . Moreover, the induction of tyrosine phosphorylation of host cell proteins by several pathogens has been reported [15–20] as well as actin polymerization and cytoskeletal rearrangements [20–24]. Since protein kinases play a major role in eukaryotic cells in transducing extracellular signals, of interest is the involvement of these enzymes and protein phosphorylation in the invasion of host cells by B. bacilliformis. Thus, this study examined the hypothesis that Bartonella invasion of nucleated cells triggers the generation of phosphorylation signals that are involved in the induction of cytoskeletal rearrangements that promote the entry of Bartonella into nucleated host cells.
2Materials and methods
2.1Growth of B. bacilliformis
B. bacilliformis, strain KC584 (American Type Culture Collection, Rockville, MD) was grown according to methods described in . Cells were plated on brain heart infusion agar plates containing a final concentration of 0.5% bovine serum albumin (Sigma Chemical Co., St. Louis, MO), 2.0% casamino acids (Difco), a 0.2% glucose and 10% (v/v) human blood erythrocyte lysate, which was prepared by the addition of 3:1 (v/v) sterile distilled water to packed outdated erythrocytes (American Red Cross, Nashville, TN). Cells were incubated at 26°C in a water-saturated atmosphere.
2.2Radiolabeling of cells
B. bacilliformis was radiolabeled with 200 μCi trans-[35S]methionine per ml, specific activity, 1000 Ci mmol−1 (ICN Biochemicals, Irvine, CA) in brain heart infusion broth supplemented with 1.0% human erythrocyte lysate and 10% fetal bovine serum for 1 h. Bartonella incorporated an average of approximately 10−4 cpm per organism.
2.3Tissue culture cells
Human laryngeal epithelial cells (HEp-2, ATCC CCL 23) were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum. Human umbilical vein endothelial cells (HUVECs) were provided by Dr. Richard Hoover, Department of Pathology, Vanderbilt University, Nashville, TN, and were isolated from umbilical veins by collagenase digestion as described by Jaffee et al. . They were identified as endothelial cells on the basis of their cobblestone morphology, uptake of acetylated low-density lipoprotein and positive staining for factor VIII antigen. They were maintained in M199 growth medium supplemented with 10% neonatal calf serum, heparin, and endothelial cell growth supplement. Cells were grown in 60 mm diameter tissue culture dishes at 37°C in the presence of 5% CO2 and 95% air in a humidified atmosphere.
2.4Radioactive invasion assay
The number of intracellular radiolabeled bartonellae was determined according to procedures described in . Monolayers were infected with 108 cells and incubated 1 or 3 h at 37°C. Bacterial cell counts were determined using a Petroff-Hausser counting chamber. At the end of the incubation period, monolayers were washed 5–6 times with PBS to remove nonadherent parasites and incubated in the presence of 0.08% trypsin for 20 min at 37°C to remove externally bound organisms. The detached monolayer and released extracellular parasites were subjected to differential centrifugation in an IEC-HN-SII bench top centrifuge at 1000×g for 5 min. Pelleted HEp-2 cells containing intracellular bartonellae were washed and trichloroacetic acid-precipitable radioactivity was measured. The percent Bartonella invasion represents percentages of added counts per minute invading monolayers, calculated as recovered cpm/added cpm×100. The percent inhibition is calculated as (treated % invasion/untreated % invasion×100−(100).
2.5Treatment of cells with inhibitors
To determine the effects of inhibitors on the invasion of cells by Bartonella, subconfluent monolayers of target cells on 60 mm diameter tissue culture dishes were preincubated in the presence of increasing concentrations of genistein (50–500 μM over 30 min), tyrphostin 25 (100–1000 μM over 1 h) or staurosporine (0.2–1.0 μM over 1 h), prior to infection with radiolabeled bartonellae at a multiplicity of infection (MOI) of 100:1. Genistein inhibits the binding of ATP to tyrosine protein kinases while tyrphostin competes with the tyrosine binding to the enzyme. Staurosporine is a potent inhibitor of protein kinase C but also inhibits some tyrosine protein kinases such as pp60c-src or serine/threonine protein kinases [14, 26–28]. Inhibitors were purchased from Upstate Biotechnology, Lake Placid, NY, and stock solutions dissolved in dimethyl sulfoxide and stored at −20°C. Invasion was carried out in the continued presence of the drugs for about 3 h and the percent radioactivity internalized by host cells determined as described above. To determine whether anti-integrin antibodies exert an inhibitory effect on the invasion of HEp-2 cells, monolayers were preincubated on ice for 60 min in the presence of 5 μg anti-β1 integrin monoclonal antibody (anti-CD29) or anti-α5 integrin monoclonal antibody (anti-CD49e) (Pharmingen) per ml of serum-free RPMI 1640 medium containing 0.2% BSA. Monolayers were also exposed to an irrelevant mAb as a control. After two washes, monolayers were then infected with 35S-radiolabeled Bartonella for 1 h to allow uptake. The percent radioactivity invading monolayers was determined as described above.
2.6Preparation of whole cell lysates from HEp-2 cells
Lysates of infected and uninfected HEp-2 cells were prepared by lysis in 1% Triton X-100 containing 20 mM Tris-HCl, pH 7.2, 150 mM NaCl and 1 mM sodium orthovanadate (Sigma) for 20 min at 4°C. Centrifugation was carried out for 20 min at 14 000×g in a microcentrifuge at 4°C.
2.7SDS-PAGE and Western blotting
To identify proteins phosphorylated on tyrosine residues, aliquots of cell lysates (approx. 25 μg of protein) were separated on 7% SDS polyacrylamide gels and transferred to nitrocellulose using a Bio-Rad Mini Trans Blot Cell and the Tris-glycine buffer system of Towbin et al. . Blots were probed with anti-phosphotyrosine monoclonal antibody, 4G10 (Upstate Biotechnology Inc., Lake Placid, NY) and a 1:3000 dilution of goat anti-mouse alkaline phosphatase conjugated IgG as secondary antibody. Development of blots was with an alkaline phosphatase conjugate substrate development kit containing nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3 indoylphosphate (BCIP) color developing reagents (Bio-Rad, Hercules, CA).
2.8Preparation of Bartonella antiserum
Antisera against whole B. bacilliformis was prepared by injecting New Zealand white female rabbits with formalin-killed organisms. Bartonella cells were collected from blood agar plates, washed in PBS and centrifuged. The cell pellet was resuspended in 0.1% formalin for 4 h at 22°C to achieve killing. A mixture of formalin-killed organisms was emulsified 1:1 in complete Freund's adjuvant for the initial subcutaneous injections in the thigh, back and neck. Two subsequent injections at two-week intervals consisted of a 1:1 emulsion of formalin-killed organisms and incomplete Freund's adjuvant.
Immunofluorescence was used to determine whether treatment of cells with tyrosine protein kinase inhibitors inhibited the binding of Bartonella to cell monolayers . Monolayers of HEp-2 cells were preincubated in the presence of the inhibitors as already described prior to and during infection with Bartonella for 1 h at 0°C. Following the washing of monolayers to remove nonadherent organisms, anti-Bartonella immune serum (1:100) dilution was added to monolayers for 1 h at 4°C. Following two washes with PBS, a fluorescein-conjugated goat anti-rabbit immunoglobulin G (Gibco-BRL) was added and incubation carried out for 30 min at room temperature. Cells were visualized with an Olympus BX-60 fluorescence microscope.
The statistical significance of the data was determined by Student's t-test. A P value of less than 0.05 was taken as significant.
3.1Inhibitory effects of protein kinase inhibitors on B. bacilliformis invasion of host cells
We examined the effects of three protein kinase inhibitors on the entry of B. bacilliformis into nucleated cells. Following a preincubation of monolayers in the presence of each inhibitor, 35S-radiolabeled organisms (108) were added and incubation carried out for 3 h in the continued presence of the drug. As shown in Fig. 1, exposure of HEp-2 cell monolayers to genistein had an inhibitory effect on the ability of these cells to internalize Bartonella in a dose-dependent manner. Invasion of cells was decreased by 60% in the presence of 250 μM genistein (P<0.01). Moreover, in the presence of 1 mM tyrphostin 25, results indicated a 40.2±7.26% decrease in invasion (data not shown, P<0.02). Conversely, exposure of HUVECs to 250 μM genistein resulted in only a 27.25±3.96% inhibition of invasion of these cells (data not shown). There was no significant difference between control untreated and treated HUVECs in ability to internalize Bartonella (P > 0.05). Interestingly, however, exposure of HUVECs to increasing concentrations of staurosporine resulted in a dose-dependent decrease in the invasion of these cells by Bartonella. The data in Fig. 2 show an approximate 75% inhibition of invasion into HUVECs in the presence of 1 μM staurosporine (P<0.001) compared to an approximate 25.92±2.82% inhibition in HEp-2 cells (data not shown). There was no significant difference between control untreated and treated HEp-2 cells in ability to internalize Bartonella (P > 0.05). Moreover, results from indirect immunofluorescence microscopy shown in Fig. 3 suggested no apparent effect of genistein on the binding of organisms to HEp-2 cells previously exposed to the drug. Similar results were obtained for staurosporine (data not shown). There were no apparent toxic effects of the drugs on cell monolayers as indicated by viability measurements using trypan blue exclusion. Viability of bacteria was also not affected and when the effects of these drugs on invasion were tested by removing drugs from cells prior to Bartonella infection, there was still a decrease in invasion, indicating that the effect was on host cells and not on the bacteria.
3.2B. bacilliformis infection of HEp-2 cells induces tyrosine phosphorylation of host proteins
Based on the inhibitory effects of genistein on Bartonella invasion of epithelial cells, tyrosine protein kinases and tyrosine phosphorylation are implicated as mediating the invasion process. Thus, we examined whether Bartonella-infected cells had elevated levels of proteins phosphorylated on tyrosine. Monolayers of HEp-2 cells were infected with Bartonella for increasing amounts of time and cell lysates prepared as described in Section 2 and analyzed by 7% SDS-PAGE. Phosphotyrosine proteins were identified by Western blotting using an antiphosphotyrosine monoclonal antibody, 4G10. As shown in Fig. 4, lane 7, after 1 h of infection, there was an increase in the intensity of three phosphotyrosine protein bands that are 243, 215, 160 and 145 kDa in size and the appearance of a new band approximately 110 kDa in size and a doublet (179 and 172 kDa). However, there was a decrease in the intensity of all protein bands in infected cells exposed to genistein (lane 8) and in cells after 2 h of infection (lane 9). To determine if phosphotyrosine proteins were host or bacterial, Bartonella cell extract was included in lane 2 of the immunoblot. There were no apparent protein bands similar to bands seen in infected host cells, indicating the absence of Bartonella proteins phosphorylated on tyrosine. The data implicate tyrosine kinases and or tyrosine phosphorylation as possible signals that are involved in the internalization of Bartonella into HEp-2 cells.
3.3Anti-integrin monoclonal antibodies exert inhibitory effects on the invasion of HEp-2 cells by B. bacilliformis
The fact that tyrosine kinases and/or tyrosine phosphorylation which are known to transduce a variety of cell signals may be involved in the invasion of host cells suggested that invasion of nucleated cells by Bartonella is perhaps mediated by host cell integrins in a manner similar to Y. pseudotuberculosis. Thus, whether anti-integrin antibodies exert an inhibitory effect on the invasion of HEp-2 cells by Bartonella was next examined. Monolayers of HEp-2 cells were incubated on ice in the presence of anti-β1 integrin monoclonal antibody (anti-CD29) and anti-α5 integrin monoclonal antibody (anti-CD49e), respectively, prior to infection with radiolabeled Bartonella. As shown in Table 1, the α5 and β1 integrin monoclonal antibodies resulted in 37 and 35% inhibition, respectively, in invasion of cells by radiolabeled Bartonella. The simultaneous addition of both antibodies to monolayers failed to result in additional decreases in the number of organisms internalized. An irrelevant monoclonal antibody had no affect on the ability of cells to internalize Bartonella. These results indicate a moderate level of involvement of the α5β1 integrins in promoting the uptake of Bartonella into HEp-2 cells.
|Number of experiment||Integrin||% Invasionb,c||% Inhibitiond|
|irrelevant mAb||11.22±3.27 (NSe)||0|
In this study, inhibitors of protein kinase activity were used to show that Bartonella invasion of host cells involves tyrosine phosphorylation. Genistein and tyrphostin treatment of HEp-2 cells resulted in significant decreases in the uptake of Bartonella into these cells (60 and 40%, respectively) but had a lesser effect on the ability of HUVECs to internalize this organism. Conversely, staurosporine, a potent inhibitor of protein kinase C and some tyrosine protein kinases, had a significant inhibitory effect on the ability of HUVECs to internalize Bartonella compared to a lesser effect on HEp-2 cells. These results indicate the significance of tyrosine kinases and or tyrosine phosphorylation in the internalization of Bartonella by these cells. The basis for the differential effects of inhibitors on the ability of host cells to internalize Bartonella is not clearly understood but may relate to the fact that different uptake signals and strategies for bacterial uptake may be employed by different cell types. Moreover, the fact that exposure to these drugs failed to completely block internalization suggests that multiple receptor/bacterial ligand mediated pathways of invasion probably exist in these cells. In addition, an earlier report also demonstrated that both genistein and tyrphostin blocked the internalization of E. coli carrying the cloned Yersinia enterocolitica inv-mediated invasion pathway . All three drugs also resulted in a decrease in the invasion of human monocytes by Legionella pneumophila.
In view of these findings, tyrosine kinases and or tyrosine phosphorylation are implicated as possible cell signals involved in the internalization of Bartonella into host cells. An increase in the intensity of the 243-, 215-, 160- and 145-kDa phosphotyrosine protein bands and the appearance of the 179-, 172- (doublet) and 110-kDa phosphotyrosine proteins in infected cells after 1 h, but not in infected cells in the presence of genistein, provide further support for this notion. Moreover, the apparent increase in the intensity of phosphotyrosine protein bands after 1 h of infection correlates with the observations from previous findings which show intracellular Bartonella in human diploid fibroblasts after 1 h of infection and the absence of organisms in cells infected for only 30 min . The dramatic decrease in the intensity of all phosphotyrosine protein bands after 2 h is interesting and perhaps indicates that these proteins are involved during the early stages of the invasion process. Interestingly, the fact that protein bands in lanes 4 and 6 are still visible in the presence of genistein may represent a basal level of tyrosine phosphorylation in uninfected cells and cells infected for 30 min which is less sensitive to genistein.
Less than clear, however, is whether or not Bartonella-induced tyrosine activity is needed for invasion of HEp-2 cells or whether tyrosine phosphorylation of the identified proteins or some other proteins are required in the invasion process. The fact that these proteins are Triton X-100 soluble suggests that they may be membrane proteins and possibly receptors. Studies are under way to determine the identity of the phosphotyrosine proteins reported here and whether any Triton X-100 insoluble proteins, which would include cytoskeletal proteins, are also being phosphorylated on tyrosine. In view of these results and the fact that tyrosine phosphorylation has been correlated with the stimulation of integrin receptors, it is reasonable to suggest that host cell integrins may be involved in the uptake of Bartonella into these cells. Consistent with the idea of integrin involvement during the invasion of cells by Bartonella is the fact that exposure of HEp-2 cells to monoclonal antibodies against α5β1 integrins resulted in a moderate inhibition of the internalization of Bartonella into these cells. It was reported that monoclonal antibodies to several β1 integrins caused a reduction in the uptake of Y. pseudotuberculosis by host cells . Studies are now under way to determine if monoclonal antibodies against other β1 integrins also exert inhibitory effects on Bartonella invasion. Further, integrins appear to be common targets exploited by many other pathogens including Bordetella pertussis, Shigella flexneri, L. pneumophila, E. coli type I fimbriae, Borrelia burgdorferi and Leishmania. Moreover, the cytoplasmic domains of integrins associate with cytoskeletal proteins and can affect cytoskeletal rearrangements via changes in actin polymerization . Thus, it is interesting to speculate that perhaps Bartonella infection of HEp-2 cells induces a tyrosine protein kinase inhibitor-sensitive uptake signal which is perhaps tyrosine kinase activity and or tyrosine phosphorylation. The resulting phosphorylation of receptors as well as cytoskeletal proteins could mediate the rearrangement of the cytoskeleton of the host cell around the bacteria as they enter host cells. In HUVECs, perhaps Bartonella binding to its receptor involves the induction of protein kinase C activity which in turn would be involved in the tyrosine phosphorylation of various cytoskeletal proteins.
We thank Dr. Richard Hoover and Rhoda Jones for supplying human umbilical vein endothelial cells. This research was supported by Grants SO6 GM8037 and HRD 9550643 from the National Institutes of Health and the National Science Foundation, respectively. N.M. Williams-Bouyer was supported by a Patricia Roberts Harris Fellowship and NSF Grant RII8714805.
- 31995) Characterization of a two-gene locus from Bartonella bacilliformis associated with the ability to invade human erythrocytes. Infect. Immun. 3, 1552–1562.and Minnick, M.F. (
- 181995) Invasion of epithelial cells by Shigella flexneri induces tyrosine phosphorylation of cortactin by a pp60c-src-mediated signaling pathway. EMBO J. 11, 2471–2482., Prevost, M. and Sansonetti, P.J. (
- 251973) Culture of human endothelial cells derived from umbilical cord veins. Identification by morphologic and immunologic criteria. J. Clin. Invest. 52, 2745–2756., Nachman, R.L., Becker, C.G. and Minich, C.R. (