INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

Salmonella typhimurium and Escherichia coli can use the nonfermentable amino alcohol ethanolamine as the sole carbon and/or nitrogen source (8, 21). The initial step in the catabolism of ethanolamine involves the cleavage of ethanolamine into acetaldehyde and ammonia by the adenosylcobalamin (AdoCbl)-dependent enzyme ethanolamine ammonia-lyase (5, 7, 8). In addition to the requirement of AdoCbl for the enzymatic degradation of ethanolamine, work with S. typhimurium has shown that AdoCbl is also required for the induction of the genetically defined eut operon (35, 36, 45). This operon encodes proteins involved in ethanolamine catabolism in this bacterium and E. coli (5, 6, 37, 48). The requirement of AdoCbl for both ethanolamine catabolism and eut operon expression presents a challenge to these organisms growing aerobically, since S. typhimurium can synthesize AdoCbl de novo only under anaerobic conditions and E. coli is unable to synthesize the complete coenzyme de novo (20, 24). Both organisms meet this challenge by using transport systems to acquire exogenous complete and incomplete corrinoids under aerobic conditions.

Transport of exogenous cobalamin (Cbl) and other corrinoids from the environment into the cytoplasm of S. typhimurium or E. coli requires two independently functioning transport systems; the first actively transports Cbl across the outer membrane, while the second transports Cbl across the cytoplasmic membrane (10). Transport across the outer membrane involves BtuB, a high-affinity outer membrane receptor for Cbl, and the TonB-dependent energy-transducing complex consisting of the cytoplasmic membrane proteins TonB, ExbB, ExbD, and other, yet to be identified proteins (4, 18, 32, 46). TonB is anchored in the cytoplasmic membrane and spans the periplasm to interact directly with a number of outer membrane receptors involved in Cbl or ferric siderophore transport (32). The TonB-dependent energy-transducing complex couples electrochemical potential from the cytoplasmic membrane to the active transport of Cbl and ferric siderophores across the outer membrane. In the absence of a functional transport system, aerobically growing cells become starved for iron and respond by hypersecreting siderophores in a futile attempt to access iron. More relevant to ethanolamine utilization, these cells cannot access exogenous Cbl unless Cbl is present in a concentration high enough to overcome the transport defect (4, 34). Transport across the cytoplasmic membrane is carried out by the ABC transport system of BtuB, BtuC, and BtuD and functions independently of the TonB-dependent system (10).

eutF mutants were originally identified by the inability to grow on ethanolamine as a sole source of carbon, and EutF was proposed to play a role in ethanolamine transport or regulation of an ethanolamine transporter (28). Since then, we have also observed other phenotypes associated with eutF mutations which included the inability to grow on 1,2-propanediol as a sole carbon source and reduced growth rates on the nonfermentable carbon sources propionate and succinate (30). Here we present evidence that these phenotypes are the result of partial-loss-of-function tonB alleles and are not due to a new gene locus.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Bacteria, media, and growth conditions. The bacterial strains and plasmids used in this study are listed in Table 1. Carbon source utilization was tested on no-carbon medium E (NCE) (9, 49) supplemented with 1 mM MgSO₄, 0.3 mM methionine, 0.1 mM tryptophan, 15 nM cyanocobalamin (CNCbl), and, as a carbon source, 30 mM ethanolamine, 1,2-propanediol, propionate, succinate, or glycerol. Concentrations were the same regardless of whether growth was tested on solid or liquid medium. The final concentrations of antibiotics in complex medium were 20 (tetracycline), 50 (kanamycin), 20 (chloramphenicol), and 100 (ampicillin) µg/ml. In NCE medium, the final concentrations used were 10, 100, 10, and 50 µg/ml, respectively. Growth of cultures at 37°C was monitored with a Spectronic 20D spectrophotometer (Milton Roy Co., Rochester, N.Y.) at 650 nm. CAS medium was a gift from Michelle R. Rondon. All experiments were done under aerobic conditions.

Genetic techniques. (i) Transductions. All transductional crosses were performed with mutant phage P22 HT105 int-201 (42, 43), and all phage manipulations were performed as previously described (9).

(ii) Isolation of allele 1235. The derivative of strain JE1418 capable of growing on ethanolamine as a carbon and energy source was isolated by plating JE1418 on NCE minimal medium supplemented with ethanolamine as a carbon source. Portions of 10 independent cultures (~10⁹ cells) were plated individually. After incubation for 3 days, we recovered a single colony with a reversion rate estimated to be 10⁹. The mutant strain JE4386 was reconstructed by growing phage P22 on the revertant, using the phage lysate as the donor to transduce strain JE1418 to Eut⁺, and then further characterized. A Tn10 1617 element (50) [referred to as Tn10d(Tc)] located near 1235 was isolated by genetic means as described elsewhere (12). This transposon [zde-6396::Tn10d(Tc)] was ca. 30% cotransducible with 1235 by phage P22.

Recombinant DNA techniques. (i) Plasmid constructions. To make pTONB1, pRZ526 was digested to completion with HpaI and the ~4.9-kb fragment containing the original insert was gel purified by the QIAquick gel extraction protocol as instructed by the manufacturer (Qiagen Inc., Valencia, Calif.). This fragment was digested to completion with BglII; the ca. 3,100-bp fragment containing tonB, yciC, yciB, and yciA was gel purified by QIAquick gel extraction and cloned into the BamHI/HincII site of vector pSU19, resulting in plasmid pTONB1.

(ii) PCR amplification of tonB from TR6583 and JE1291. tonB was PCR amplified from chromosomal DNA of boiled whole cells of TR6583 and JE1291. An 860-bp amplified product was generated by using the following primers in a standard PCR mixture: 5'tonB (5' TTCAGCTCTGGTTTTTCA 3', corresponding to bases 82 to 99 of the published tonB sequence [16]) and 3'tonB (5' TCCGACGGTAAACCTCGC 3'; corresponding to bases 941 to 924 of the published tonB sequence [16]). The amplification profile was as follows: 94°C for 5 min; 30 cycles of 94°C for 1 min, 50°C for 1 min, and 72°C for 1 min; 72°C for 10 min; 4°C for 12 h. All primers used in this study were obtained from Integrated DNA Technology, Inc. (Coralville, Iowa).

(iii) PCR amplification of tonB from strain JE1418 (903). tonB was amplified from strain JE1418 chromosomal DNA by using boiled whole cells as the template. The universal Tn10 primer (2) was used in conjunction with the 5'tonB primer described above to amplify the DNA between trpC3480::Tn10d(Cm^r) to the 5' end of tonB. The amplification profile was as follows: 94°C for 5 min; 30 cycles of 94°C for 1 min, 50°C for 1 min, and 72°C for 3 min; 72°C for 10 min; 4°C for 12 h.

(iv) PCR amplification of tonB from strain JE2123 (1235). tonB was amplified from JE2123 chromosomal DNA by using boiled whole cells as the template. The trpA primer 5' AGCTTAAAGAGTACCATGCC 3' (corresponding to bp 665 to 685 of the trpA sequence [26]) was used in amplifications with the 5'tonB primer described above to amplify from the trpA coding region, across the deletion, to the 5' end of tonB. The amplification protocol was the same as for the universal Tn10 and 5'tonB primers discussed above.

(v) DNA sequencing of PCR products. All PCR products used for sequencing were purified by using a QIAquick gel extraction kit as instructed by the manufacturer. All sequencing was done by nonradioactive sequencing at the Nucleic Acid and Protein Facility at the University of Wisconsin Biotechnology Center. Primers used for sequencing include those described above in addition to the primer tonB DEL (5' GCATCGGCGACCAGCAAG 3', corresponding to bases 538 to 555 of the S. typhimurium tonB sequence [16]).

Biochemical procedures. (i) -Galactosidase assays. -Galactosidase activity assays were performed by a modification of the method of Miller (25) as described elsewhere (14).

(ii) Immunoblot analysis of TonB. Immunoblot analysis of TonB was done as previously described (23), with minor modifications. Briefly, cells were grown in NCE medium supplemented with glycerol (30 mM), MgSO₄ (1 mM), methionine (0.3 mM), and tryptophan 0.1 mM, in the presence or absence of FeSO₄ (45 µM). Cells were grown to A₆₅₀ of 0.5 in 5 ml of NCE medium. A 1-ml sample of culture was removed, diluted with 0.5 ml of 15% trichloroacetic acid, and incubated on ice for 30 min. The acid-precipitated material was pelleted, washed once with 1.0 M Tris-Cl (pH 8.0) at 25°C, resuspended in 50 µl of 2× sample buffer, and boiled for 5 min. Samples were then resolved by sodium dodecyl sulfate-12% polyacrylamide gel electrophoresis (SDS-PAGE) (22) and transferred to an Immobilon-P membrane (Millipore, Bedford, Mass.) by means of a Mini Trans-Blot apparatus (Bio-Rad, Richmond, Calif.) at 100 V for 1 h according to the manufacturer's specifications. Immunoblot analyses were performed according to the Phototope-horseradish peroxidase Western blot detection kit (New England Biolabs, Inc., Beverly, Mass.) protocol, with minor modifications. Membranes were blocked at 25°C overnight, and 1× Tris-buffered saline-0.1% Tween 20-5% dry milk was used for all incubations. Mouse monoclonal antibody 4H4 (a gift from Kathleen Postle) was used at a 1:5,000 dilution. The secondary antibody was horseradish peroxidase-conjugated donkey anti-mouse immunoglobulin G (a gift from Heidi Goodrich-Blair) diluted 1:20,000. X-ray film (XAR-50; Kodak) was used to detect signal.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

eutF mutants were identified by the inability to grow on ethanolamine as source of carbon (28). Two of the mutants, one putative point mutant (strain JE1690) and a deletion mutant (strain JE1418), are characterized in this study.

The strain carrying allele 903 (JE1418) was reported to be able to use ethanolamine as a sole nitrogen source but not as a sole carbon source, suggesting the eutF locus was only partially affected by this mutation. Characterization of strain JE1418 revealed that the eutF locus mapped between the trp operon and tonB. Strain JE1418 was found to be a tryptophan auxotroph, thus defining one end of the deletion somewhere within the trp operon. The other end of the deletion was determined based on the following tests for TonB protein function. First, strain JE1418 was sensitive to infection by bacteriophage ES18, which requires a functional TonB protein for infection (47). Second, strain JE1418 could use CNCbl or AdoCbl for the synthesis of methionine at a concentration of 15 nM. tonB mutants require micromolar concentrations of Cbl to overcome the transport defect (4). The putative point mutant showed TonB function as defined by these two criteria (data not shown). Further phenotypic analysis of the eutF locus determined that eutF mutants were unable to grow on 1,2-propanediol as the sole carbon source and had reduced growth rates on the nonfermentable carbon sources propionate and succinate (30).

Complementation of eutF mutants. Attempts to isolate a complementing clone from an S. typhimurium library were unsuccessful. However, complementation of the EutF phenotypes was achieved with a plasmid carrying an ~4,900-bp fragment containing the E. coli tonB locus and surrounding loci. This plasmid, pRZ526, also contained a fragment of bacteriophage 80 DNA (33). Plasmid pRZ526 complemented strains JE1418 and JE1690 for growth on ethanolamine, suggesting that pRZ526 contained the eutF locus (Fig. 1). This finding suggested that the eutF locus was complemented by one (or more) of five possible E. coli genes or possibly by a 80 gene. To narrow the possibilities, complementation was tested with plasmid pRZ531 (Fig. 1). pRZ531 complemented both JE1418 and JE1690, suggesting eutF was not yciA, yciC, or 80 DNA. Previous minicell analysis of protein expression from pRZ531 determined that only two proteins were synthesized from this cloned region of E. coli DNA (33). The apparent molecular weights of the products, based on SDS-PAGE analysis, suggested these proteins were TonB and YciB. To address whether yciB was eutF, yciB and yciC were subcloned from pRZ526 resulting in plasmid pYCIBC (Fig. 1). Plasmid pYCIBC failed to complement strain JE1418 or JE1690, suggesting that yciB was not eutF. These results supported the previous finding that yciC was not involved in the complementation of EutF phenotypes.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Gene organization of plasmids tested for complementation. All represented reading frames are based on the E. coli sequence (6). Complementation is scored as growth (+) or no growth () on NCE plates supplemented with ethanolamine (30 mM), Met (0.3 mM), Trp (0.1 mM), MgSO4 (1 mM), and CNCbl (15 nM).

eutF alleles result in a TonB phenotype. To show that strains JE1418 and JE1690 contained lesions that affected tonB function, we tested both strains for the hypersecretion of siderophores. Disruption of TonB function causes a hypersecretion of siderophores due to iron starvation, and the presence of the secreted siderophores can be indirectly detected by the use of CAS medium (44). Both strains (JE1418 and JE1690) grown on CAS medium resulted in large orange-yellow zones around the colonies, diagnostic of siderophore hypersecretion. tonB⁺ control strains TR6583 and JE1291 showed no significant zones (data not shown).

Altered TonB proteins in eutF mutants. To analyze directly whether TonB was altered in eutF mutants, immunoblotting using an anti-TonB monoclonal antibody was performed to detect chromosomally expressed TonB protein in wild-type and mutant strains (Fig. 2).

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Immunoblot analysis of TonB protein in strains TR6583 (tonB+), JE1418 (903), JE1291 (tonB+), and JE1690 (eutF1115). The amount of material loaded in each lane was the equivalent of 0.1 A650 of whole cells. Proteins were separated by SDS-PAGE. Samples were grown in the presence or absence of 45 µM FeSO4. Deg., degradation.

Iron-dependent regulation of tonB expression in eutF mutants. TonB protein levels in S. typhimurium and E. coli are reduced in the presence of high concentrations of iron due to Fur regulation of tonB expression (31, 51). To ensure that iron regulation of tonB was still functional, we grew strains JE1418 and JE1690, along with the wild-type controls, in the presence or absence of 45 µM FeSO₄ and detected TonB by immunoblotting as discussed above. Since we routinely use NCE medium without iron supplementation, all strains were expected to express TonB at some increased level to scavenge residual iron in the medium, as seen in studies of tonB regulation in E. coli (51). Upon addition of FeSO₄ to the medium, and assuming that the promoter region of tonB is intact, there should be a reduction in the level of TonB. All strains showed iron regulation of TonB levels (Fig. 2), consistent with the promoter region of tonB being unaffected in both eutF mutants.

Physical characterization of tonB in strains JE1418 and JE1690. To show that tonB was eutF, we attempted amplification and sequencing of chromosomal tonB alleles in strains JE1418 and 1690. Primers flanking tonB allowed amplification and sequencing of the complete coding region of tonB. For both of the tonB⁺ strains TR6583 and JE1291, an 860-bp product was amplified as expected from the published sequence (16) (data not shown). The product from TR6583 was sequenced, and the results were consistent with the amplified product being tonB (data not shown). In a control experiment, the cobU gene of S. typhimurium was amplified in a parallel reaction and always resulted in successful amplification from both wild-type and mutant DNAs (data not shown).

(i) DNA sequence analysis of the tonB allele in strain JE1418. Reactions using DNA from strain JE1418 failed to produce an amplified tonB product. These results supported the hypothesis that tonB was affected by the 903 allele in this strain. To characterize the 903 allele in strain JE1418, the transposition-deficient element trpC3480::Tn10d(Cm) (>90% cotransducible with allele 903) was recombined into the chromosome of JE1418, with retention of the 903 mutation (strain JE4163). The DNA between the insertion and the 5' end of tonB was amplified by using a primer specific for the end of the insertion and the 5'tonB primer. The amplification product (ca. 3,000 bp) was sequenced by using the transposon primer. DNA sequence information determined the trpC sequence on one side of the PCR product, while the 5'tonB primer determined the tonB sequence as expected (data not shown). Primer walking from the tonB side of the PCR product identified the deletion site of strain JE4163 (Fig. 3A). Sequence data demonstrated that the fusion of tonB with trpA resulted in the removal of the last nine codons of tonB. In-frame translation across the tonB-trpA junction predicted the removal of the final 9 amino acids of TonB and the addition of 45 amino acids after amino acid 231 of wild-type TonB (Fig. 3B). These data confirmed that tonB was disrupted in strains carrying the 903 allele and that the orientation of tonB transcription was toward the trp operon as originally proposed (19).

(ii) DNA sequence analysis of the tonB allele in strain JE1690. Surprisingly, tonB could not be amplified from the putative point mutant, strain JE1690, which suggested that the inability to amplify tonB from JE1690 was probably due to the deletion of one of the tonB primer sites. We hypothesized that the 3' end of tonB in this strain was affected since the 5'tonB primer was designed to hybridize immediately downstream of the 10 region, and hence a deletion that removed this site would abolish tonB expression.

Isolation and characterization of a mutation that restores growth of strain JE1418 on ethanolamine. Before determining that tonB was disrupted in strain JE1418, we isolated a derivative of it capable of growing on ethanolamine as a sole carbon source. This revertant strain displayed eut operon expression to ca. 60% of the level measured in a wild-type strain in the presence of ethanolamine and Cbl (Table 2). A Tn10d(Tc) insertion linked to the mutation causing this reversion was isolated and found to be ca. 30% cotransducible with the 903 allele by phage P22. This result raised the possibility that the mutation causing the reversion was in tonB.

View this table: [in this window] [in a new window]   TABLE 2.   Effects of allele 1235 on eut operon expression and growth on ethanolamine

View larger version (15K): [in this window] [in a new window]   FIG. 3.   Partial nucleotide and predicted amino acid sequences of tonB genes from TR6583, JE4163, and JE2123. (A) Nucleotide sequence flanking the tonB-trpA fusion site. The TR6583 sequence corresponds to bases 790 to 837 of wild-type tonB (16). Boldface bases correspond to bases from trpA; boldface underlined bases correspond to bases deleted in a strain carrying allele 1235. (B) Predicted amino acid sequence of TonB. The starting proline residue corresponds to amino acid 221 of the published sequence (16). Boldface residues are predicted amino acids from in-frame translation of tonB from the 903 allele; boldface underlined residues are predicted amino acids from in-frame translation of the 1235 allele. DEL903 and DEL1235 are referred to as 903 and 1235, respectively, throughout the text.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

Data presented herein show that the mutations initially defining the eutF locus are partial-loss-of-function tonB alleles. This conclusion is based on the following findings: (i) mutations in the eutF locus can be complemented by a plasmid containing only the wild-type allele of the tonB gene from E. coli; (ii) eutF mutant strains hypersecrete siderophores, consistent with lesions in tonB; (iii) immunoblot analyses of TonB from both mutant strains suggest the synthesis of an unstable TonB fusion protein; (iv) sequencing of tonB from a eutF mutant, JE1418 (903), showed that the mutation resulted in the replacement of the final 9 amino acids of TonB with 45 amino acids from a tonB-trpA gene fusion; and (v) a 2-bp deletion (1235) in the coding region of the tonB-trpA gene fusion of strain JE1418 removed the bulky addition to the carboxy terminus of TonB, resulting in a partially functional TonB protein.

Why were eutF mutants not identified as mutants partially defective in tonB function? The initial characterization of eutF mutants originally eliminated tonB from consideration for two reasons. First, eutF strains were still sensitive to infection by bacteriophage ES18, which requires a functional tonB locus for infection (47). Second, both strains were able to grow in the presence of CNCbl or AdoCbl at a concentration of 15 nM. tonB mutants require micromolar concentrations of Cbl to overcome the TonB transport mutation (4). We now believe that the original phenotypic screen for eutF alleles led to the isolation of partial-loss-of-function tonB alleles. Strain JE1418 was isolated in a screen for deletions of cobA, a gene encoding the adenosyltransferase enzyme involved in AdoCbl biosynthesis. cobA mutants do not grow on ethanolamine as a sole carbon source because (i) ethanolamine ammonia-lyase requires AdoCbl as a coenzyme and (ii) cobA mutants cannot synthesize AdoCbl (15). Strain JE1418 did not grow on ethanolamine but was found to be wild type for cobA, which suggested that an alternative locus was responsible for the phenotype; this locus was named eutF (28).

Explaining EutF phenotypes in terms of partially functional TonB proteins. Based on the conclusion that eutF mutants were actually partial-loss-of-function tonB alleles, the various carbon source utilization phenotypes can be explained in the following way. As described above, ethanolamine utilization requires Cbl for breakdown of ethanolamine and the induction of the genetically defined eut operon (35, 36, 45). Recent molecular characterization of a portion of the putative eut operon identified a proposed ethanolamine permease (48). Based on the findings presented here, the expression of this permease is likely to be dependent on the genetically defined regulator eutR (35). Therefore, a strain unable to access exogenous Cbl would not fully express the ethanolamine permease and would exhibit decreased ethanolamine transport. The combination of reduced Cbl acquisition and decreased ethanolamine transport would result in decreased expression of eutR-controlled genes and reduced growth rates on ethanolamine. This phenotype is exactly what was observed in eutF mutants (28).