INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bovine mastitis is the most costly infectious disease in animal agriculture. Of the multitude of microbial pathogens that can cause bovine mastitis, the coliform bacteria Escherichia coli and Klebsiella pneumoniae are pathogens commonly isolated from intramammary infections (IMIs) and are leading causes of clinical mastitis. To date, no universally acceptable plan for controlling coliform mastitis has been established, and the need continues to exist for recommendations on controlling mastitis based on empirical knowledge (20). Iron is essential for coliform bacteria to fulfill normal metabolic processes. However, the availability of free iron in bovine milk is severely restricted, because most iron is bound to citrate and, to a lesser degree, to lactoferrin, transferrin, xanthine oxidase, and some caseins (11). To overcome the iron-restricted condition in their mammalian hosts, coliform bacteria may utilize one or more iron assimilation systems to take up iron within a particular environmental niche in the host (3). Possession of iron uptake systems is known to be important in bacterial pathogenesis (22, 33). High-affinity iron uptake systems are widely utilized by coliform bacteria to take up iron in the host milieu (3). These involve the synthesis of a low-molecular-mass siderophore, the expression of iron-regulated outer membrane proteins (IROMPs) and enzymes to utilize the chelated iron (3). However, the bovine mammary gland is very suitable for coliform bacteria to utilize another strategy, a ferric citrate iron uptake system, to acquire iron because of the high concentration of citrate (7 mM) in normal milk (2, 3, 11).

Some coliform bacteria have developed the ability to take up iron directly from naturally occurring organic iron-binding acids, including citrate (3, 10). Although citric acid is not a siderophore, the uptake system has many properties of siderophore-mediated high-affinity systems (2, 3). The ferric citrate iron uptake system is repressed at high iron concentrations by Fur protein (2, 3). However, the induction of the citrate-mediated iron uptake system is distinct from that of normal high-affinity iron uptake systems (2, 3). The citrate iron uptake system requires ferric dicitrate for induction (10). More than 0.1 mM citrate is required for induction of this system. Such a high concentration of citrate is necessary to dissolve ferric ions in a low-molecular-mass form, most probably as ferric dicitrate (10). Citrate does not serve as a carbon or energy source, and it is not transported under aerobic growth conditions (10). Ferric dicitrate, the inducer, does not have to enter the cytoplasm to initiate transcription of the ferric citrate transport genes (10, 32). Ferric citrate receptor FecA is an 80.5-kDa IROMP that is responsible for the binding of ferric dicitrate (23). As a novel signal transduction model, the ferric citrate iron uptake system has been thoroughly investigated over the past 18 years (reviewed in references 2 and 3). However, no information exists concerning the role of the ferric citrate iron uptake system as a pathogenic mechanism in bacteria, because the concentration of citrate in the vertebrate host generally is too low to induce the bacterial ferric citrate system (29). However, bovine milk appears to provide an ideal environment for induction of the ferric citrate iron transport system. The average concentration of citrate in bovine milk is as high as 7 mM (11). In addition, the citrate/iron molar ratio in milk is in excess of 1,000 (11), which can easily result in the production of ferric dicitrate for induction. Furthermore, citrate in milk does not serve as a carbon or energy source for coliform bacteria under aerobic growth conditions, which are present in bovine milk in vivo (18). Therefore, based on the above information, we hypothesize that in the presence of a high concentration of citrate in bovine milk, the ferric citrate iron uptake system is induced in coliform bacteria.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. The coliform isolates tested were E. coli (n = 18) and K. pneumoniae (n = 17) from naturally occurring bovine IMIs in five herds. E. coli UT5600 [leu proC trpE rpsL entA (ompT-fepA)] was kindly provided by Dick van der Helm (Department of Chemistry and Biochemistry, The University of Oklahoma, Norman) E. coli ZI311/pSV66 [araD139 (argF-lac)U169 rpsL150 relA1 flbB5301 deoC1 ptsF25 rbsR aroB fecA? zag::Tn10/Cm^r fecIRA] was kindly provided by Volkmar Braun (Mikrobiologie, Universität Tübingen, Tübingen, Germany). All bacterial strains were stored on Trypticase soy agar slants at room temperature prior to use.

Construction of strain E. coli UT5600/pSV66. Plasmid pSV66 was isolated from E. coli ZI311/pSV66 and was transformed into E. coli UT5600 by following a standard protocol (26). Host strain E. coli UT5600 and transformant E. coli UT5600/pSV66 were grown in Trypticase soy broth (TSB) or TSB containing 200 µM -'-dipyridyl (Sigma Chemical Co., St. Louis, Mo.) and 1 mM citrate. Bacterial outer membranes were isolated and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described below.

Isolation of outer membranes. Outer membranes were isolated by the method of Todhunter et al. (28) with slight modification. Bacteria were grown in TSB or TSB containing 200 µM -'-dipyridyl and 1 mM citrate. Cultures were incubated at 37°C for 18 h on a rotary shaker (200 rpm). Following incubation, bacteria were harvested by centrifugation at 2,500 × g for 30 min at 4°C and washed three times in 0.15 M NaCl. Cells were resuspended in deionized, distilled water and disrupted by sonication for 10 min. Sonicated bacteria were centrifuged at 5,000 × g for 10 min at 4°C. N-Lauroylsarcosine sodium salt (Sigma Chemical Co.) was added to the supernatant at a final concentration of 2% and incubated for 30 min at room temperature. Outer membranes were collected by centrifugation at 50,000 × g for 60 min at 4°C, washed twice in deionized, distilled water, and stored at 70°C. Total protein was determined with the bicinchonic acid protein assay reagent (Pierce Chemical Co., Rockford, Ill.).

Rabbit anti-FecA serum. FecA protein was derived from E. coli UT5600/pSV66, which was cultured in TSB containing 200 µM -'-dipyridyl and 1 mM citrate. FecA immunogen was prepared by SDS-PAGE, and the FecA band was excised and fragmented as described in reference 7. New Zealand White rabbits were immunized four times over an 8-week period with prepared FecA immunogen. FecA immunogen (50 µg) in Freund's complete adjuvant was injected subcutaneously at six sites along the back of each rabbit. Booster immunizations in Freund's incomplete adjuvant were given 2, 5, and 8 weeks after the primary immunization. The rabbits were bled from the marginal ear vein 10 days after each immunization, and sera were tested by immunoblotting as described below. The reactivity of anti-FecA serum to purified ferric enterobactin receptor FepA (kindly provided by Dick van der Helm, Department of Chemistry and Biochemistry, The University of Oklahoma, Norman) was also determined.

Electrophoresis. Outer membrane proteins were separated by SDS-PAGE (12.5% polyacrylamide) by utilizing the discontinuous buffer system of Laemmli (12). Outer membrane protein samples were prepared for SDS-PAGE by being heated at 100°C for 5 min in 0.0625 M Tris (pH 6.8), 2% SDS, 5% -mercaptoethanol, 10% glycerol, and 0.00125% bromphenol blue. Proteins were detected with Commassie brilliant blue R-250 (28).

Immunoblots. Western immunoblots were performed as described previously (14) with slight modification. Proteins were transferred to nitrocellulose sheets and blocked in phosphate-buffered saline (PBS) plus 5% instant nonfat dry milk for 1 h at room temperature. The nitrocellulose sheets were washed in PBS plus 0.05% Tween 20 (PBS-Tween) and then incubated for 1 h at room temperature in diluted rabbit anti-FecA serum. Nitrocellulose sheets were rinsed three times in PBS-Tween and incubated in a 1:16,000 (vol/vol) dilution of horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (whole molecule) for 1 h at room temperature. Nitrocellulose sheets were washed as described previously and incubated with a substrate solution containing 50 mg of 3, 3'-diaminobenzidine dissolved in 100 ml of PBS and 0.1 ml of 30% (vol/vol) H₂O₂. The reaction was stopped after 10 min by washing the nitrocellulose with distilled water. All antibodies were diluted in PBS plus 5% instant nonfat dry milk.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of transformant E. coli UT5600/pSV66. Both E. coli UT5600 and transformant E. coli UT5600/pSV66 induced FecA with a molecular mass of approximately 80.5 kDa when grown in TSB supplemented with 200 µM -'-dipyridyl and 1 mM citrate (Fig. 1). The FecA was not observed in bacteria that were grown in TSB, which is an iron-sufficient medium. Transformant E. coli UT5600/pSV66 produced much more abundant quantities of FecA proteins in the presence of citrate than did the host strain E. coli UT5600 (Fig. 1). Therefore, E. coli UT5600/pSV66 was chosen as an ideal source of ferric citrate receptor FecA. Another advantage of purification of FecA from E. coli UT5600/pSV66 is that any possible contamination of 81-kDa ferric enterobactin receptor FepA, which has a molecular mass similar to that of FecA, is eliminated because FepA protein is not produced by E. coli UT5600 (FepA) under iron-restricted conditions.

View larger version (59K): [in this window] [in a new window]   FIG. 1.   Outer membrane protein profiles of E. coli UT5600 (A) and UT5600/pSV66 (B) separated by SDS-PAGE and stained with Coomassie blue. The amount of protein used per lane was 20 µg. Bacteria were grown in TSB (lane 1) or TSB plus 200 µM -'-dipyridyl and 1 mM citrate (lane 2). The left lane (M) contains molecular mass (103) standards. The approximate position of ferric citrate receptor FecA is indicated.

Antisera against FecA. Serum from an FecA-immunized rabbit strongly reacted with FecA protein induced by E. coli UT5600/pSV66 in the presence of citrate at a 1:2,000 dilution (Fig. 2). In iron-replete medium TSB, E. coli UT5600/pSV66 produced a very small amount of FecA that reacted with postimmunization serum (Fig. 2). No anti-FecA antibodies were detected in preimmunization serum. In addition, anti-FecA serum did not react with purified FepA (lane 2 in Fig. 2).

View larger version (35K): [in this window] [in a new window]   FIG. 2.   Immunoblots of separated outer membrane proteins of E. coli UT5600/pSV66 (lanes 1 and 3) and purified FepA reacted with preimmunization (PRE) and postimmunization (POST) rabbit anti-FecA sera. Both sera were diluted 1:2,000 (vol/vol). E. coli UT5600/pSV66 cells were grown in TSB (lane 1) or TSB plus 200 µM -'-dipyridyl and 1 mM citrate (lane 3). The left lane (M) contains molecular mass (103) standards. The approximate position of ferric citrate receptor FecA is indicated.

Distribution of the ferric citrate system among coliforms isolated from bovine IMIs. The isolates tested were E. coli (n = 19) and K. pneumoniae (n = 17) from naturally occurring bovine IMIs in five herds. The expression of FecA can be taken as an indicator of induction of the transport system in cases in which transport cannot be measured (10). Figure 3 shows the typical outer membrane protein profiles of E. coli and K. pneumoniae isolates separated by SDS-PAGE. The induction of FecA protein was observed in each isolate that was grown in the presence of citrate, while no FecA was observed in each isolate that was grown in iron-rich medium TSB. In the presence of citrate, all isolates induced large quantities of FecA that reacted with rabbit anti-FecA serum under iron-restricted conditions. Figure 4 shows the immunoblot of typical coliform isolates. All isolates that were grown in an iron-replete culture expressed FecA at a very low level (lane 1 in Fig. 4) compared with large quantities of FecA in iron-restricted medium containing citrate (lane 2 in Fig. 4). The molecular mass of serum-reactive FecA is approximately 80.5 kDa.

View larger version (33K): [in this window] [in a new window]   FIG. 3.   Typical outer membrane protein profiles of E. coli and K. pneumoniae isolates separated by SDS-PAGE and stained with Coomassie blue. The amount of protein used per lane was 20 µg. Bacteria were grown in TSB (lane 1) or TSB plus 200 µM -'-dipyridyl and 1 mM citrate (lane 2). The left lane (M) contains molecular mass (103) standards. (A) E. coli 17. (B) E. coli 414. (C) E. coli 471. (D) K. pneumoniae 531. (E) K. pneumoniae 564. (F) K. pneumoniae 32. The approximate position of FecA is indicated.

View larger version (24K): [in this window] [in a new window]   FIG. 4.   Immunoblots of the separated IROMPs of typical isolates of E. coli and K. pneumoniae reacted with the rabbit anti-FecA serum. The amount of protein used per lane was 20 µg. Bacteria were grown in TSB (lane 1) or TSB plus 200 µM -'-dipyridyl and 1 mM citrate (lane 2). Rabbit anti-FecA serum was diluted 1:2,000 (vol/vol). (A) E. coli 17. (B) E. coli 414. (C) E. coli 471. (D) K. pneumoniae 531. (E) K. pneumoniae 564. (F) K. pneumoniae 32. The approximate position of FecA is indicated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The purposes of the current study were (i) to determine the distribution of ferric citrate iron-uptake systems in coliform bacteria, (ii) to initiate research on the role of ferric citrate receptor FecA in the pathogenesis of coliform bacteria, and (iii) to develop a possible approach to the prevention and control of coliform mastitis. Mastitis control is achieved either by decreasing teat-end exposure to pathogens or by increasing the resistance of cows to IMI (27). Control of coliform mastitis historically relied on reduced exposure to the coliform bacteria in the environment of the dairy cows (27). However, teat-end exposure to coliform bacteria can occur anytime, because the primary reservoir of coliform bacteria is the dairy cows' environment (27). Elimination of coliform bacteria from the environment is not economically feasible; therefore, increasing the resistance of cows against coliform bacteria would be a logical method to reduce coliform mastitis. The area of coliform mastitis control with the greatest advances in recent years is vaccination (34). However, the commercial sale of E. coli J5 (O111:B4) vaccines does not prevent bovine IMI, and influx of protective antibodies into the gland begins 12 to 24 h after bacterial populations have begun to increase (8, 9, 30). The need still exists for an effective vaccine to prevent IMI and control the growth of bacteria in the bovine mammary gland.

Iron is an essential growth factor for survival and multiplication of coliform bacteria that commonly cause bovine IMI. Genera classified as coliform bacteria include Escherichia, Klebsiella, and Enterobacter, which are gram-negative bacteria belonging to the family Enterobacteriaceae (5). However, because of the low solubility of ferric iron (21) and the need to avoid its participation in potentially damaging Harber-Weiss-Fenton chemistry (6), higher organisms have evolved mechanisms for lowering the levels of free iron to well below those required for the growth of gram-negative bacteria (17). Free iron may be regulated in vertebrates as a defense mechanism against bacteria (17). Most iron is bound intracellularly to proteins such as ferritin, hemoglobin, and myoglobin and extracellularly to high-affinity iron-binding proteins such as transferrin and lactoferrin in serum and mucosal secretions (17). To overcome such iron limitation, coliform bacteria have developed different strategies to acquire iron in an iron-restricted environment (3, 17). The ferric enterobactin system, a siderophore-mediated iron uptake system, was suggested to be widely distributed among coliform bacteria isolated from naturally occurring bovine IMI (13, 14). Our findings (13-16) indicate that the ferric enterobactin system is a dominant iron acquisition system utilized by coliform bacteria during the nonlactation period, when most iron is bound to lactoferrin in the bovine mammary gland. However, the role of ferric enterobactin in iron acquisition appears to diminish in the lactating gland as the concentration of lactoferrin decreases and the concentration of citrate increases (1). The ferric citrate iron uptake system that we investigated in the current study may be critically important for the pathogenesis of coliform bacteria during lactation due to high concentration of citrate in bovine milk. Additionally, the finding that the presence of citrate in the growth medium represses the ferric enterobactin iron uptake system in coliform bacteria (25) suggests that the ferric citrate iron uptake system may play a dominant role for coliform bacteria during the lactation period.

The current study demonstrates that the ferric citrate iron uptake system is widely present among coliform bacteria associated with bovine mastitis. We successfully made constructs which provide an ideal source for FecA purification and induced high titers of anti-FecA serum. The expression of FecA is an indicator of induction of the ferric citrate iron uptake transport system (10). Therefore, polyclonal anti-FecA serum was used to determine the distribution of ferric citrate system and the degree of antigenic homology of FecA among coliform bacteria from naturally occurring bovine IMI. The rabbit anti-FecA serum used in this study is an accurate and effective probe for detection of the expression of FecA by coliform bacteria. Although the molecular mass and many properties of physical chemistry between FecA and ferric enterobactin receptor FepA are similar (31), these two IROMPs have low homology in amino acid sequence (database accession code: FecA, ae000499; FepA, A25953). The current study showed that anti-FecA serum did not react with purified FepA. In addition, anti-FepA monoclonal antibodies did not react with FecA (data not shown).

The ability of coliform bacteria to colonize and proliferate within a particular environmental niche in the host is essential for initiation of an infection. The IROMPs of coliform pathogens are often suggested as vaccine candidates for immunoprophylactic therapy because they are surface exposed and antigenic and may induce antibodies that block the essential iron uptake of the bacteria (22, 33). Antibodies directed against IROMPs, some of which are siderophore receptors, can block iron uptake and inhibit bacterial growth in vitro (14, 16, 19, 24). Antibodies in bovine mammary secretions are derived from blood or produced locally by plasma cells in the subepithelial connective tissue (4). Therefore, if cows immunized with IROMPs produce antibodies directed against the epitopes responsible for iron uptake, these antibodies will transfer from blood or local connective tissue to the mammary gland, thereafter blocking the essential iron uptake function and inhibiting the growth of coliforms in the mammary gland. We suggested that coliform bacteria may primarily use the ferric enterobactin iron uptake system to assimilate iron during the nonlactation period due to the high concentration of lactoferrin (14-16). Data from our laboratory have demonstrated the ability of anti-FepA antibodies to inhibit coliform bacterial growth in media containing lactoferrin or in involuted mammary secretion (14, 16). Ferric enterobactin receptor (FepA) may be an attractive vaccine component with which to control coliform mastitis during the nonlactation period. Our findings in the present study strongly suggest that FecA is widely expressed and highly conserved among coliform bacteria isolated from naturally occurring bovine IMI. Consequently, if FecA-immunized cows produce antibodies directed against the epitopes at the ligand binding site of FecA, such antibodies will be transferred into milk and may block the essential iron uptake function and inhibit the growth of coliforms during lactation. The next logical step in developing knowledge concerning the interactions among the FecA receptor and the host defense during the lactation period is to determine if humoral immunity to the molecule will alter bacterial growth in biological secretion.