Cobinamide in cobA Mutants of Salmonella typhimurium - Journal of ...

Vol. 175, No. 19

JOURNAL OF BACTERIOLOGY, Oct. 1993, p. 6328-6336

0021-9193/93/196328-09$02.00/0 Copyright © 1993, American Society for Microbiology

cobU-Dependent Assimilation of Nonadenosylated Cobinamide in cobA Mutants of Salmonella typhimurium GEORGE A. O'TOOLE AND JORGE C. ESCALANTE-SEMERENA* Department of Bacteriology, University of Wisconsin-Madison, 1550 Linden Drive, Madison, Wisconsin 53706 Received 20 May 1993/Accepted 22 July 1993

The cobA locus of Salmonella typhimurium is involved in the assimilation of nonadenosylated cobinamide, (CN)2CBI, into cobalamin (CBL) under aerobic and anaerobic growth conditions. Aerobically, cobA mutants are unable to assimilate (CN)2CBI into CBL. However, under anaerobic conditions, cobA mutants assimilate (CN)2CBI into CBL as efficiently as cobA+ strains. On the basis of this observation, we postulated the existence of a cobA-independent pathway for the assimilation of (CN)2CBI into CBL that is functional under anaerobic growth conditions (J. C. Escalante-Semerena, S.-J. Suh, and J. R. Roth, J. Bacteriol. 172:273-280, 1990). In this paper, we report the isolation and initial genetic characterization of derivatives of cobA mutants that are unable to assimilate (CN)2CBI into CBL during anaerobic growth. As demonstrated by complementation analysis, marker rescue, and DNA sequencing data, these mutations are alleles of cobU, a gene involved in the assembly of the nucleotide loop of CBL. We have shown that the block in CBL synthesis in these cobU cobA double mutant strains can be corrected by exogenous adenosyl-CBI. Our data indicate that this new class of cobU mutations blocks CBL biosynthesis but does not destroy the putative kinase-guanylyltransferase activities of the CobU protein. We propose that this new class of cobU mutations may affect an as yet unidentified ATP:corrinoid adenosyltransferase activity of the CobU protein. Alternatively, such mutations may alter the ability of CobU to use nonadenosylated CBI as a substrate.

anaerobic growth conditions (10, 24). A cobA mutant is unable to convert (CN)2CBI to CBL under aerobic growth conditions. However, under anaerobic conditions, this defect is no longer observed, suggesting that a cobA mutant is capable of assimilating (CN)2CBI into CBL in the absence of oxygen (see Table 2). On the basis of this observation, we postulated the existence of a function that could bypass a cobA defect under anaerobic growth conditions. We present the results of isolation and initial molecular genetic analysis of mutant derivatives of cobA strains that are unable to assimilate exogenous (CN)2CBI into CBL under anaerobic growth conditions. These derivatives of cobA strains carry mutations in the cobU gene. These lesions define a new class of cobU mutations that appear to block the synthesis of CBL from (CN)2CBI during anaerobic growth without eliminating the putative kinase-guanylyltransferase activities of the CobU protein. We propose that this new class of cobU mutations may affect an as yet unidentified ATP:corrinoid adenosyltransferase activity of the CobU protein. Alternatively, such mutations may alter the ability of CobU to use nonadenosylated CBI as a substrate.

Salmonella typhimurium synthesizes cobalamin (CBL) de novo under anaerobic growth conditions and assimilates exogenous incomplete nonadenosylated corrinoids into CBL both aerobically and anaerobically (10, 13). Incomplete corrinoids such as dicyano-cobinamide [(CN)2CBI] appear to require conversion to their adenosylated forms (e.g., Ado-CBI) before they can be joined to the lower base of the molecule, 5,6-dimethylbenzimidazole (DMB), in a series of reactions that assemble the nucleotide loop of CBL (Fig. 1)

(4, 5, 8, 19).

In S. typhimurium, the nucleotide-loop biosynthetic functions are encoded by the CobIII region of the CBL biosynthetic (cob) operon located at 41 min of the linkage map (13). Three genes, designated cobU, cobS, and cobT, have been proposed to encompass this region (21). Of direct relevance to this report is the cobU gene, whose gene product is approximately 40% identical at the amino acid sequence level to the CobP protein of Pseudomonas denitrificans (4, 6), with some domains of the proteins greater than 80% identical. In P. denitrificans, the cobP gene encodes a bifunctional enzyme with Ado-CBI kinase, GTP:Ado-CBI-P guanylyltransferase, activities that are needed for the synthesis of Ado-CBI-GDP from Ado-CBI (4, 6). The degree of homology between the S. typhimurium CobU and the P. denitrificans CobP makes it likely that the CobU protein carries out the same enzymatic activities (21). This idea has been strengthened by the isolation of cobU mutants of S. typhimunum whose growth phenotype can be rescued by exogenous (CN)2CBI-GDP (19). The cobA locus (34 min) of S. typhimunum has been implicated in the adenosylation of both complete and incomplete corrinoids. Furthermore, cobA has been shown to participate in CBL biosynthesis under both aerobic and *

MATERIALS AND METHODS

Bacteria, culture media, and growth conditions. All bacterial strains used were derivatives of S. typhimurium LT2, and their genotypes are listed in Table 1. Nutrient broth (NB [Difco, Detroit, Mich.]) (0.8%) containing 85 mM NaCl was used as complex medium, unless otherwise stated. E medium (26) supplemented with glucose (11 mM) was used as minimal medium. Nutritional supplements for auxotrophs were added at final concentrations reported elsewhere (7). The final concentrations of antibiotics in complex medium were (in micrograms per milliliter): tetracycline, 20; kanamycin, 50; and chloramphenicol, 20. Solid medium con-

Corresponding author. 6328

VOL. 175, 1993

ANAEROBIC ASSIMILATION OF COBINAMIDE BY S. TYPHIMURIUM

CN, OH, Hp

ICB

Assimilation of

CobA

denovo synthesis of

cobinamide V

non-adenosylated corrdnolds mplete

Ado

Ado

i

CobA

CcbU, CobS, CobT

I

Nudeotde-ioop

Assmilation of

non-adenosylated

assembly

incomplete corrinolds

OH, Hp0 ICN, I' |CBI I

CobU a

b

CobU

Synthesis of

non-adenosylated

[0

p

CobS,CobT Nucleotide-loop assembly

I?H.

CBL

CBI-GDP

FIG. 1. Proposed roles of CobU in the synthesis of CBL. Shown is a schematic view of de novo CBL biosynthesis and the assimilation of incomplete and complete corrinoids. CobA adenosylates an unidentified intermediate in the de novo synthesis of Ado-CBI and

adenosylates incomplete exogenous corrinoids prior to assembly of loop. CobA also participates in the adenosylation of complete corrinoids, such as CBL (10). CobU is proposed to act only on incomplete corrinoids, and it can only do so efficiently under anaerobic conditions. (a) In addition to performing its nucleotideloop assembly functions, CobU may adenosylate incomplete corrinoid intermediates. (b) CobU may be able to generate CBI-GDP in the absence of an adenosylated substrate. Single arrows may represent more than one biochemical reaction. The upper ligand on nonadenosylated complete and incomplete corrinoids in vivo has not been established and may vary. We speculate that, in vivo, the upper ligand of the nonadenosylated CBI-GDP and CBL may be a hydroxyl- or aquo- group (depending on the pH). Abbreviations: CN-, cyano; HO-, hydroxyl; H20-, aquo; DHSHC, dihydrosirohydrochlorin (a precursor of the corrin ring). the nucleotide

tained 15 g of BactoAgar (Difco) per liter. When added to the culture medium, corrinoids were present at a final concentration of 15 nM and DMB was present at a final concentration of 300 ,uM. The preparation of anoxic culture medium used in anaerobic growth studies has been described elsewhere (3). The medium was supplemented with 10 ml of trace minerals (2) per liter and 100 mM sodium bicarbonate, and the cell density of the cultures was monitored with a Spectronic 20D

spectrophotometer (Milton Roy Co., Rochester, N.Y.) at 650 nm. Aerobic cultures were grown as described previously (20). All cultures were grown at 37°C. The growth rate

(k) was

defined as 1/doubling time (in hours). Genetic techniques. (i) Transductions. All transductional crosses were performed with mutant phage P22 HT105 int-201 (22, 23), and all phage manipulations were performed as previously described (7). For transductional crosses that selected Kmr or Cmr transductants, cells and phage were mixed and preincubated nonselectively for 2 h at 30°C in liquid without shaking and then were plated on rich medium supplemented with kanamycin or chloramphenicol and 10 mM EGTA. The latter was added to minimize lysogen formation. (ii) Mutagenesis. P22 lysates were chemically mutagenized with hydroxylamine by the procedure of Hong and Ames (12) as described elsewhere (7). (iii) Construction of cobR4 cob-369 double mutant. A genetic test was used to determine whether the aerobic over-

6329

expression of cobU specifically, and not the cob genes in general, could rescue a CobA phenotype. The cobR4 mutation (located upstream of the CobI region) was used to increase aerobic cob expression. This mutation results in anaerobic levels of expression even in the presence of

oxygen but has no effect on the biochemical functions required for de novo synthesis of CBL (1). A lysate of P22 grown on strain JE1899 (cob-369 cob-109::Mu d11734, Kmr, 5-bromo-4-chloro-3-indolyl-P-Dgalactopyranoside+ [Fig. 2]) was used to transduce JE1490 (cobR4 cobA) to Kmr. The cob-109::Mu d11734 mutation is located in the CobII region on the cob genetic map (Fig. 2) and confers Kmr. cob-109::Mudll734 has no phenotype under these growth conditions, because DMB is provided. Kmr transductants were screened for loss of the ability to assimilate (CN)2CBI into CBL aerobically (i.e., inheritance of cobU369), on minimal E glucose agar plates supplemented with (CN)2CBI and DMB after a 48-h incubation period. Retention of the cobR4 mutation was assessed on nutrient agar plates supplemented with 5-bromo-4-chloro-3-indolyl3-D-galactopyranoside and kanamycin; Kmr colonies (i.e., those which carry the Mu d-lacZ fusion) were dark blue on this medium only when cobR4 was still present. In vivo assessment of CBL biosynthesis. CBL biosynthesis was assessed indirectly by the ability of a metE mutant strain to synthesize methionine on minimal medium. Strains carrying a metE mutation methylate homocysteine to methionine in a reaction catalyzed by the CBL-dependent MetH enzyme (25). Thus, metE mutants that carry mutations in the cob biosynthetic pathway are methionine auxotrophs. Our previous work indicates that, in S. typhimurium, the end product of the cob biosynthetic pathway is Ado-CBL (10); however, it is known that Ado-CBL is not the form of CBL used by the MetH enzyme (25). In experiments in which the Met phenotype of cob mutants was assessed, we refer to the product of the cob pathway as CBL. 13-Galactosidase assays. ,B-Galactosidase activity assays were performed by a modification of the method of Miller (17) as described elsewhere (9). Recombinant DNA techniques. Unless otherwise noted, all protocols for DNA manipulation were those of Maniatis et al. (15). Restriction enzymes, Klenow fragment of DNA polymerase I, and T4 DNA ligase were obtained from New England Biolabs (Beverly, Mass.), and S1 nuclease was obtained from Boehringer-Mannheim Biochemicals (Indianapolis, Ind.). All enzymes were used as per the manufacturers' recommendations. GeneClean (Bio 101, Inc., La Jolla, Calif.) was used for the isolation of DNA fragments from agarose gels. Plasmids were introduced into appropriate recipients by electroporation with a Bio-Rad Gene Pulser (Bio-Rad Laboratories, Richmond, Calif.) as previously described (19). Plasmids. All plasmids used in this study are listed in Table 1 and were derived from pJE2 (8). Plasmid pJO1 was generated by subcloning the approximately 3.4-kb S1 nuclease-blunted AatII fragment of pJE2 into the multiple cloning site of plasmid pSU21 (16), which had been previously digested with HincII. The construction of plasmid pJO7 (cobU+) was reported previously (19). This plasmid carries the complete 543-bp cobU coding sequence. The cobU+ gene is not under the control of the Plac promoter of pSU21. Plasmid pJO7 has approximately 220 bp of upstream sequence and 150 bp of downstream sequence flanking cobU (Fig. 3). Two subclones containing different fragments of the cobU gene were constructed for the marker rescue studies as

J. BACTERIOL.

O'TOOLE AND ESCALANTE-SEMERENA

6330

TABLE 1. Strains and plasmids used& Source

Genotype or characteristics

Strain or plasmid

TR6583 TR6583 derivatives JE50 JE212 JE588 JE761 JE1293 JE1445 JE1486 JE1490 JE1540 JE1897 JE1898 JE1899 JE2026 JE2027 JE2814 JE2948 JE2949 JE2950 JE2951

metE205 ara-9 derivative of strain LT2 (formerly SA2279)

K. Sanderson via J. R. Roth

cob-62::Mu d11734, located in CobII DE299 (hisG-cob)r cob-236::Tn1ODE16DE17(Tcr),b located in CobI zea-1872::Tn1ODE16DE17(Tcr), located outside CobIl

Laboratory collection Laboratory collection Laboratory collection Laboratory collection Laboratory collection Laboratory collection

cobA366::TnJODE16DE17(Cmr)c DE902, deletion of undetermined ends covering cobA and trp cobU369 derivative of JE588 zeb-1845::TnlO cobR4 derivative of JE1293 cobA366::TnJODE16DE17(Cmr) derivative of JE1486 cobU172 derivative of JE761 cobA366::TnJODEI6DEI7(Cmr) derivative of JE1897 cobU369 cob-109::Mu d11734 recAl DE902 derivative of JE1897 recAl DE902 derivative of JE1486 cobU369 cob-109::Mu dl1734 derivative of JE1490 pSU21 pJO7

cobA::Tnl0d(Tcr)/pSU21 cobA::Tnl0d(Tcr)/pJO7

Plasmids pJO7

19 8

Derivative of pJOl, cobU+, Cmr Derivative of pJE2, cobU+, Cmr Derivative of pJOl, Cmr Derivative of pJO7, Cm' Cloning vector, Cm'

pJOl pJO2d pJO15d pSU21

16

strains and plasmids, unless otherwise indicated, were constructed during the course of this work. aAll b Referred to in the text as TnlOd(Tcr) (27). c Referred to in the text as TnlOd(CmT). d See Materials and Methods for a complete description of these plasmids.

follows (Fig. 3). (i) Plasmid pJO2 was constructed by cloning the EcoRI-to-ClaI fragment of pJO1 into pSU21 previously digested with EcoRI and ClaI. Plasmid pJO2 contains the downstream 403 bp of the 543-bp cobU coding sequence as well as approximately 1.4 kb of downstream flanking sequence. (ii) Plasmid pJO15 was constructed by digesting pJO7 with HincII and ClaI, isolating the large fragment containing the vector, filling in the overhanging ends with the Klenow fragment of DNA polymerase I, and ligating. Plasmid pJO15 contains all of the cobU coding sequence, except for the last 54 bp, and approximately 220 bp of upstream flanking sequence.

Cobi

'>

I

cob-236::TnlOd(Tc)

X

| Donor

l

\

+543 I

Aatil

BstEll

EcoR1

Hincli

I

0o1

I

pJo7 pJ02

- cob-2::Mu dll734(Km)

l

cobU

+1

Cobil

Cobill l

Complementation and marker rescue. Complementation was assessed by the ability of a clone to compensate for a chromosomal defect in a recombination-deficient genetic background. Plasmids were moved into the desired genetic background by transduction.

|Recipient

cob-369

FIG. 2. Three-factor cross. The donor transducing fragment and recipient chromosome for the three-factor cross are shown. The top of the figure shows the three regions of the cob operon (CobI, CobIl, and CobIII). The donor strain (JE50) carries a Mu d1734 insertion (-) in the CobII region which confers Kmr. The recipient strain (JE1540) carries the cob-369 mutation (X) and TnlOd(TcT) (0) in the CobI region. JE1540 also carries cobA366::TnlOd(CmT) (not shown in this figure). The crossover events required to generate the donor class (thick lines) and rare class (thin lines) recombinants are shown.

Taql

Al

140

I

pJO15 489

FIG. 3. Clones used for complementation and marker rescue. A simplified restriction map of cobU and flanking regions is depicted at the top of the figure. A portion of the insert of plasmid pJOl, used to generate the subclone pJO7, is shown. Plasmid pJO7 (cobU+), used in the complementation experiments, is also shown. Plasmid pJO7 contains the full-length cobU coding sequence (543 bp) as well as approximately 220 bp of upstream and 150 bp of downstream flanking sequence. Plasmids pJO2 and pJO15 were used in the marker rescue studies. The indicated restriction sites were used to generate plasmid pJO7 and the marker rescue clones. The numbers in the figure refer to base pairs numbered relative to the + 1 position at the left-hand end of the cobU coding region. The cobU coding region is represented by thick lines and flanking sequence is represented by thin lines.

V 1ANAEROBIC ASSIMILATION OF COBINAMIDE BY S. TYPHIMURIUM VOL. 175, 1993

Marker rescue experiments were performed with clones pJO2 and pJO15. These plasmids did not complement any of the cobU mutations tested (data not shown). Clones were transduced into the desired genetic background (7), and the clone-containing strains were patched to NB plates. These plates were replica printed to E minimal medium containing glucose and (CN)2CBI and were incubated anaerobically. The appearance of small colonies within the patch of a mutant strain carrying a particular clone was scored as positive for marker rescue. All other phenotypes of the plasmid-containing strains were also verified. Mutant strains carrying the vector only (pSU21) were used as negative controls in both complementation and marker rescue experiments. DNA sequencing. The DNA sequences of wild-type and mutant cobU genes were determined in two steps. First, the region of the chromosome containing the mutations was amplified with a thermocycler (Perkin-Elmer Cetus, Norwalk, Conn.), with Vent DNA polymerase (New England Biolabs), deoxynucleotides (Promega, Madison, Wis.), and oligonucleotide primers (Genosys Biotech, Woodlands, Tex.) homologous to sequences flanking, or internal to, the cobU gene (14). The amplified DNA was gel purified (GeneClean) and sequenced in the thermocycler with the Sequi Therm cycle-sequencing kit (Epicentre Technologies, Madison, Wis.). Sequencing reactions were labeled with a-3SdATP (Dupont, Boston, Mass.) with a specific activity of 1,000 to 1,500 Ci/mmol. Primers were designed with the reported cobU sequence (21). Primers 1 and 3 were homologous to the coding strand, and primers 2 and 4 were homologous to the noncoding strand (see Fig. 4 for primer sequences). Primer sets 1/3 and 2/4 were used to amplify the upstream and downstream portions of cobU, respectively. Both strands of DNA from three independently amplified samples of the wild-type gene and each mutant gene were sequenced. Additionally, both the upstream and downstream portions of the wild-type gene and each mutant gene were sequenced to confirm the absence or presence of a mutation.

RESULTS Genetic analysis. (i) Isolation of mutants unable to assimilate (CN)2CBI into CBL under anaerobic conditions in a cobA background. Previous data indicated that CobA participated in CBL biosynthesis under both aerobic and anaerobic growth conditions (10). On the basis of this information, we predicted that any defects in the function(s) required for the CobA-independent assimilation of (CN)2CBI should be silent in a cobA+ genetic background. Furthermore, the aerobic overexpression of the cob operon could partially correct the phenotype of a cobA mutant (10). From this, we predicted that the function(s) responsible for the CobAindependent anaerobic assimilation of (CN)2CBI was located within the cob region. We performed localized mutagenesis of the cob region to isolate mutant derivatives of a cobA strain unable to synthesize CBL from (CN)2CBI under anaerobic growth conditions. Three classes of mutants were expected. Class i mutants are defective in nucleotide-loop biosynthesis (CobIII functions). Such mutants would be blocked in the assimilation of (CN)2CBI independent of cobA. Class ii mutants are defective in the CobA-independent assimilation of (CN)2CBI. Such mutants would be impaired for the assimilation of (CN)2CBI only in a cobA background. Class iii mutants simultaneously disrupt the function(s) required for

6331 50

1

+1

51

GATAAAATTTATATACCATCATGCAACAACATCAGGAGGAGCCGGTATGA

100

101

TGATTCTGGTGACGGGCGGGGCACGTAGTGGTAAAAGCCGTCATGCTGAA

150

151

GCCTTAATTGGCGATGCGCCGCAGGTACTGTATATCGCCACCTCGCAGAT

200

201

TCTTGATGACGAGATGGCGGCGAGAATTCAGCATCATAAAGATGGCAGGC

250

251

CGGCACACTGGCGGACCGCAGAATGCTGGCGGCATCTTGATACGTTGATT

300

3

301

ACCGCGGATCTTGCCCCTGACGACGCGATTTTGCTGGAATGTATTACCAC

350

-7 351

CATGGTGACGAATCTGCTGTTTGCGCTGGGAGGCGAGAACGATCCCGAAC N 2

400

401

AGTGGGATTACGCGGCGATGGAGCGCGCCATTGACGATGAAATTCAGATT

450

451

TTAATTGCAGCCTGCCAGCGCTGCCCGGCGAAAGTGGTACTGGTGACAAA

500

501

TGAGGTGGGAATGGGGATCGTCCCGGAAAACCGTCTGGCGCGCCATTTTC

550

551

600

601

GTCTGGCTGGTAGTCTCAGGTATTGGAGTCAAAATTAAATAATGAGTAAG 650

651

CAGTTTTGGGCCATGCTCGCTTTTATTAGCCGCTTGCCCGTCACGTCACG

700

701

CTGGTCGCAGGGACTGGATTTCGAGCAGTATTC N ~~~~~~4

750

FIG. 4. Sequence of the wild-type cobU gene, identification of mutations, and primers used in amplification and sequencing. The DNA sequence of cobU as derived from the chromosome is shown. The four primers used to amplify chromosomal DNA are also shown: primer pairs 1/2 and 3/4 allowed the amplification of the upstream and downstream portions of cobU, respectively. The sequences of the primers are underlined; the head of the arrow indicates the 3' end, and the tail indicates the 5' end of the oligonucleotides. The first codon (ATG) and stop codon (TGA) of cobU are shown in boldface; the position labeled +1 indicates the first base in the open reading frame. The mutation in JE1897 (cobU1172) is a G-to-A transition at position 88, and the mutation in JE1486 (cobU369) is a G-to-A transition at position 463. The mutations are indicated by the boldface G's and arrowheads.

the CobA-independent assimilation of (CN)2CBI and the nucleotide-loop biosynthetic function(s). The phenotype of these mutants would classify them as defective in nucleotide-loop assembly functions (class i). Hydroxylamine-mutagenized P22 lysates grown on strain JE588 [cob-236::Tn1Od(Tcr) or JE761 [zea-1872::Tn1Od(Tcr)] were used as donors to transduce strain JE1445 (DEL902 trp-cobA) to Tcr. The resulting Tcr transductants were screened for strains which could no longer synthesize CBL from (CN)2CBI under anaerobic growth conditions. To introduce a wild-type cobA locus into the putative mutants, all isolates were transduced to tryptophan prototrophy, thus repairing the deletion which spans trp-cobA. A P22 lysate grown on strain JE212 (DELhisG-cob) was chosen as donor of the cobA-trp region to avoid any possible recombination events involving wild-type sequences of the cob operon. The Cob phenotype of the resulting cobA + trp+ transductants was assessed under anaerobic conditions, and two classes of mutants were observed. One class of mutants was

6332

J. BACTERIOL.


TABLE 2. Supplementation of a metE mutant Aerobic growth' with:

Genotypea

cobA+ cobU+ cobA cobU+ cobA cobU1172

No addition'

(CN)2CBI

-

+

-

Anaerobic growth with:

Ado-CBI

+ +d +

No addition

(CN)2CBI

Ado-CBI

+

+ + -

+

-

+

+

a All strains carry mutation metE205. b Growth was on E minimal medium containing glucose with the indicated supplement. For anaerobic growth studies, the medium was supplemented with trace minerals. Concentrations and growth conditions are described in Materials and Methods. Growth was assessed after 15 h of incubation at 37'C. c De novo CBL biosynthesis in S. typhimurium occurs only under anaerobic growth conditions. d Limited growth is observed after extended incubation (>12 to 15 h). For a more detailed discussion, see Results. e Identical results were observed with cobU369.

still unable to convert (CN)2CBI to CBL in a cobA+ background. We concluded that these mutants were defective in the nucleotide-loop assembly (CobIII) functions. The second class of mutants regained the ability to synthesize CBL from (CN)2CBI under anaerobic growth conditions. We concluded that these strains were defective in the CobA-independent pathway for the assimilation of (CN)2CBI (Table 2, line 3). Two independent isolates of this class were obtained and designated cob-369 and cob-1172. (ii) Genetic mapping. Three-factor cross analysis was performed to confirm that cob-369, mapped within the cob operon (Fig. 2). Strain JE50, which carries insertion cob62::Mu dl1734(Kmr), served as donor in a transductional cross with JE1540 [cob-369 cob-236::TniOd(Tcr) cobA366:: Tnl0d(Cmr)] as recipient. The cob-369 mutation was 97% cotransducible with the cob-236: :Tnl0d(Tcr) marker, an insertion known to map in the downstream part of the CobI region of the cob operon (8). A total of 999 Kmr transductants were screened aerobically for Tcr and anaerobically for their ability to synthesize CBL from (CN)2CBI and DMB (i.e., loss of cob-369). The results of this analysis are summarized below. The genotypes and drug resistances of each class observed are given, followed by the number of recombinants and their relative frequency: class i, cob', Tc' Km', 983, 98.4% (donor-type recombinant); class ii, cob', Tcr Kmr, 12, 1.2%; class iii, cob-369, Tcr KMr, 3, 0.3%; and class iv, cob-369 Tcs Kmr, 1, 0.1% (rare-type recombinant). If cob-369 was located between cob-236::TnlOd(Tcr) and cob62::Mu dl1734(Kmr), the donor-type transductant (cob', Tcs KMr) would comprise the most frequently isolated class and the strain requiring double recombinational exchanges (cob-369, Tc' Km') would be the least frequently isolated class. The results of the three-factor cross, therefore, placed cob-369 between markers cob-236::TnlOd(Tcr) and cob-62::Mu dl1734(Kmr). Genetic mapping with deletions spanning the upstream portion of the CobIII region was consistent with the threefactor analysis (data not shown) and showed that cob-369 was located either in the downstream region of CobI or the upstream region of CobIII. (iii) Complementation analysis. cob-369 and cob-1172 appear to be alleles of cobU. To determine the gene affected by the cob-369 and cob-1172 mutations, previously constructed clones containing single genes of the CobIII region (19) were tested for their ability to complement these mutations. Plasmid pJO7 (cobU+ [Fig. 3]) restored the ability of strains JE2026 (cob-1172 cobA recAl) and JE2027 (cob-369 cobA recAl) to assimilate (CN)2CBI into CBL under anaerobic growth conditions. Other single-gene clones of the CobIII genes did not complement the defects in strains JE2026 and JE2027 (data not shown). These complementation results

suggested that both cob-369 and cob-1172 were alleles of cobU. (iv) Marker rescue. cob-369 and cob-1172 are rescued by clones which carry fragments of cobU. The phenotype of a strain carrying cob-1172 in a cobA background was recombinationally repaired by pJO15 but not by pJO2 (Fig. 3). The phenotype of a strain carrying cob-369 in a cobA background was recombinationally repaired by pJO2 but not by pJO15. These results allow the localization of the two alleles within cobU; allele cobU1172 was localized to the upstream portion, and allele cobU369 was localized to the downstream portion of cobU. (v) Identification of mutations. The sequencing of these mutant alleles showed that cob-369 and cob-1172 are singlebase changes in cobU (Fig. 4). Allele cob-1172 is a G-to-A transition at base number 88 of the cobU gene. This mutation results in a single amino acid change at position 30 from alanine to threonine. The mutated alanine is conserved in both S. typhimurium and P. denitrificans (Fig. 5) and is located near a putative phosphate-binding domain (11, 18). It is unclear whether this phosphate-binding domain is part of a motif required for ATP or GTP binding, because the portion of the motif required for adenine versus guanine specificity is not present in CobU. Allele cob-369 is a G-to-A transition at base 463. This results in a single amino acid change at position 155 from glycine to serine. This mutated glycine is located within a region at the C terminus of the protein that is conserved in both S. typhimurium and P. denitrificans and includes a portion of a consensus GTP-binding domain (28). The portion of a GTP-binding domain present in the CobU protein (D-X2-G) imparts specificity for the guanine moiety (28), but the phosphate-binding portion of this domain is not present. The positions of the mutations cob-369 and cob-1172 are consistent with the marker rescue studies described above. The mutations are both G-to-A transitions, as expected for hydroxylamine-induced changes. The genetic mapping, complementation, marker rescue, and sequencing experiments described above demonstrate that cob-369 and cob-1172 mutations are alleles of cobU, and they will hereafter be referred to as cobU369 and cobU172. Physiological role of the cobU gene product. (i) CobA function is not required under anaerobic growth conditions for the assimilation of (CN)2CBI in a cobU+ background. To quantitate the effect of a lack of a functional CobA during anaerobic growth on the ability of the cell to assimilate (CN)2CBI into CBL, we compared the growth rate (k) and final cell density (A650) of a culture of strain TR6583 (cobU+ cobA+, k = 0.56 h-1, A650 = 0.84) with those of strain JE1293 (cobU+ cobA, k = 0.57 h-1,A650 = 0.84). These data showed no difference in the anaerobic growth behavior of

VOL. 175, 1993


6333

S

VPZNRLARHFRD.IA

H2N-

I _1

II

I

-COO H

VT j G AR S alIL,R H A Z A L I GD A P Q V L Y I AT S T

FIG. 5. The CobU protein, homology to CobP, and locations of amino acid changes of mutant CobU proteins. A schematic of the CobU protein is shown. Solid blocks indicate regions of amino acid identity shared by the S. typhimurium CobU and P. denitrificans CobP proteins. Amino acids composing and surrounding the putative nucleotide-binding domains are shown. Amino acids conserved in S. typhimurium and P. denitnificans are in boldface, and residues that are part of a binding domain are underlined. The amino acid changes for JE1897 (A-)T) and JE1486 (G-+S) are also shown. The new amino acids which result from mutations are in boldface and are shown below or above the mutated residue. Near the N terminus is a region of homology which includes a phosphate-binding domain (G-X3-G-K-T-or-S-X6-I-or-V) but lacks the portion of the domain which is thought to confer nucleotide specificity (11, 18). Near the C terminus of the protein is the part of a GTP-binding domain (D-X2-G) which imparts specificity for the guanine moiety (28); the phosphate-binding portion of this domain is not present.

cobA+ and cobA strains, indicating that cobU can compensate for a defective CobA protein under these growth conditions. We also found that in a cobA+ background, the cobUll72 mutation had only a minor effect on the final cell density of the culture (k = 0.52 h-, A650 = 0.69), which resulted in no discernible phenotype when the strain was grown on solid medium. This result showed that the cobU172 mutation has virtually no effect on anaerobic growth in a cobA+ background under the conditions tested. In contrast, the previously described cobU mutants were defective for CBL synthesis independent of cobA (19). Furthermore, this result indicated that cobA participates in the assimilation of (CN)2CBI into CBL anaerobically, in addition to its role in the de novo synthesis of CBL (10). The double mutant strain JE1898 (cobU1172 cobA) could not grow on (CN)2CBI anaerobically (k = no growth), indicating that the conversion of (CN)2CBI into CBL was blocked in this strain (Table 2, line 3). The same results were observed in all of the experiments described above with strains carrying allele cobU369 (data not shown). (ii) The cobU gene is responsible for the limited ability of a cobA mutant to assimilate (CN)2CBI during aerobic growth. cobA mutants are severely impaired for the aerobic assimilation of (CN)2CBI into CBL (10). The cobU1l72 cobA double mutant is completely defective for (CN)2CBI assimilation under aerobic conditions (Table 2), even after extended periods of incubation (>50 h). These data indicated that the residual aerobic growth of a cobA mutant was dependent on a functional cobU locus. Additionally, the CobU-dependent assimilation of (CN)2CBI is inefficient under aerobic growth conditions. (iii) Overexpression of cobU under aerobic conditions partially corrects the effect of a cobA defect. The inefficient assimilation of (CN)2CBI by a cobU+ cobA strain growing under aerobic conditions could be explained by the low level of cobU expression in the presence of oxygen. We used the cobR4 mutation to increase aerobic expression of cobU to a level observed under anaerobic growth (1). A cobR4 cobA double mutant (JE1490) grows substantially better than a cobA mutant (JE1293) under aerobic growth conditions. The cobR4 mutation had the effect of nroving the growth rate of JE1490 (cobR4 cobA, k = 0.18 h-, final cell density = 180 Klett units [KU]) over that of

JE1293 (cobA, k = 0.05 h-1, final cell density = 55 KU). The final cell density of the JE1490 cultures was identical to that of the wild type. In control strains carrying the cobR4 mutation in a cobA+ background, there was no discernible effect on the growth rate (data not shown). Despite the increased cob expression, the growth rate of JE1490 was still threefold lower than that of the wild type (k = 0.59 h-1, final cell density = 183 KU). Additionally, JE1490 still exhibited a lag time 5- to 10-fold greater than that of the wild type before the onset of exponential growth. Although the cobR4 mutation results in increased expression of the entire cob operon, the data presented below suggest that it is the overexpression of cobU alone that improves the aerobic growth of a cobA mutant. Strain JE1490 (cobR4 cobA) can grow when provided with (CN)2CBI (k = 0.18 h-1, final cell density = 180 KU). Strain JE2814 (cobR4 cobU369 cobA mutant [see Methods and Materials for construction of this strain]) is unable to grow when provided with (CN)2CBI, suggesting that cobU369 renders JE2814 unable to assimilate (CN)2CBI into CBL. The cobU369 mutation eliminates the improved growth of a cobR4 cobA strain relative to a cobA strain. This result suggested that the aerobic overexpression of cobU was responsible for the improved growth of a cobA mutant. Providing cobU on a multicopy plasmid also improves the aerobic growth of a cobA mutant. Strain JE2951 [cobAl pJO7(pcobU+), k = 0.20 h-1, final cell density = 190 KU] grows significantly better than a cobA mutant carrying only the vector pSU21 (JE2950, k = 0.01 h-1, final cell density = 30 KU). Additionally, JE2951 (cobA/pcobU+) has a lag time before the onset of exponential growth which is one-half that of the 14 h observed for strain JE2950 (cobAIpSU21). However, the growth of the cobA mutant carrying pcobU (JE2951, k = 0.20 h-', final cell density = 190 KU) is still threefold slower than that of the wild-type control strain JE2949 (cobA+/pcobU+, k = 0.63 h-1, final cell density = 240 KU). The lag time before the onset of exponential growth for JE2949 (cobA+IpcobU+) is 4 h. The vector plasmid pSU21 had no effect on the growth of a cobA + strain (JE2948, k = 0.56 h-1, final cell density = 240 KU, lag time = 4 h). These data demonstrate that cobU overexpression is sufficient for the improved aerobic growth of a cobA mutant. (iv) Oxygen is a negative effector of cobU fimction. The results presented above suggested that other factors besides


6334

J. BACrERIOL.

1.0

-0.6

OX S

. * , . , . , . , .

0

-0.4 0

~ ~~~~~~~~~~~C 2-~

previously demonstrated that the defect in aerobic CBL synthesis of a cobA mutant can be corrected by exogenous Ado-CBI (10). The defect in anaerobic CBL synthesis of the cobUl172 cobA double mutant strain (JE1898) could also be corrected by exogenous Ado-CBI (Table 2, columns 3 and 6). Identical results were observed with strains carrying allele cobU369 (data not shown). This result demonstrates that the phenotype of cobUJ172 and cobU369 mutants (in a cobA background) can be rescued by the addition of an adenosylated corrinoid. Furthermore, because CBL synthesis occurs when Ado-CBI is provided, this result shows that the functions of CobU required for the synthesis of CBI-GDP are not eliminated by cobUl172 or cobU369.

-0.2

2

4

6

8

10

12

[Oxygen] kPa FIG. 6. Effects of initial oxygen concentration on the growth rate and lag time of JE1293 (cobU+ cobA). The lag time of the culture before exponential growth (0) and the growth rate (k) expressed as 1/doubling time (0) are plotted against the initial oxygen concentration of the cultures. The maximum concentration of oxygen used in this experiment (11 kPa) is equal to approximately one-half of the level of oxygen in air. Cultures were incubated at 37°C in sealed Balch tubes containing 5 ml of anoxic E minimal medium supplemented with glucose, trace minerals, (CN)2CBI, and sodium bicarbonate. Sterile air was added to the headspace of the cultures with a syringe to give the final concentration of oxygen indicated. An equivalent amount of gas was withdrawn from the headspace before addition of the sterile air. The data plotted represent the average of three experiments. The lag time of the cobA+ strain under these conditions is typically less than 1 h.

transcriptional or translation control, such as growth rate of the cultures, redox potential, and/or effects of oxygen, may impact CobU function. If CobU activity is sensitive to oxygen, then increasing the oxygen concentration would be expected to slow the growth of cells dependent on CobU. We investigated the effects of varying the initial oxygen levels on the growth of strain JE1293 (cobU+ cobA) and expected that increasing oxygen concentrations would impair CobU function and thereby slow the growth of these cells. In the range of 0 to 2.8 kPa, oxygen had little effect on the growth rate (k = 0.58 h-1), but it did have a marked influence on the lag time before the onset of exponential growth (Fig. 6). At oxygen concentrations above 5.7 kPa (approximately one-fourth the amount of oxygen in air), the lag time continued to increase linearly, but the growth rate of the cells also decreased by half (k = 0.25 h-'). The decrease in growth rate as the oxygen level increased was expected in light of the results of the aerobic growth studies presented above. The linear relationship between oxygen concentrations up to 5.7 kPa and the onset of exponential growth suggests that the dissolved oxygen level may have to be reduced before the wild-type CobU protein is able to efficiently assimilate nonadenosylated CBI in a cobA strain. These data are consistent with the idea that CobU function is impaired by oxygen.

The phenotype of cobA strains carrying allele cobU369 or cobU1172 is corrected by exogenous Ado-CBI. It has been

DISCUSSION We have isolated derivatives of cobA mutants that can no longer assimilate (CN)2CBI into CBL under anaerobic growth conditions. The evidence presented here demonstrates that the lesions responsible for this phenotype are alleles of the cobU gene, the promoter-proximal gene of the CoblIl region. Specifically, the CobU protein is proposed to catalyze the synthesis of Ado-CBI-GDP from Ado-CBI (4, 6, 19, 21). The two mutations analyzed in this study comprise a new class of cobU mutations. Strains carrying the cobU369 or cobUl172 mutations in a cobA background fail to assimilate nonadenosylated CBI into CBL, but can synthesize CBL if provided with Ado-CBI. The previously reported cobU mutations (19) blocked CBL synthesis regardless of the substrate provided (data not shown). In a cobA+ background, cobU369 and cobUl172 are phenotypically silent. The phenotype of the previously described cobU mutants was independent of a functional cobA locus (19). The data presented here also strongly suggest that the kinase and guanylyltransferase activities of the CobU protein are not eliminated by the cobU369 or cobUl172 mutation. The DNA sequence demonstrates that mutations cobU369 and cobU1172 are single-base-pair changes in cobU. The amino acids affected are in regions conserved in both the S.

typhimurium CobU and the P. denitrificans CobP proteins,

are within spans of the CobU protein that have high amino acid identity with CobP, and are located in or near putative substrate-binding domains. It is unclear how the wild-type CobU protein can bypass the need for a functional CobA protein under anaerobic growth conditions, especially in light of the proposed activities of CobU and CobA. The CobU protein has been proposed to have Ado-CBI kinase and GTP:Ado-CBI-P guanylyltransferase activities (4, 6, 19, 21). CobA is likely to be an ATP:corrinoid adenosyltransferase (24). We offer two models to explain the data presented. Two important points to consider about both of these models are that (i) the activities of the CobU protein required to convert Ado-CBI to CBL should still be present in strains carrying the cobU369 or cobUl172 allele and (ii) the wild-type CobU protein can efficiently bypass a cobA defect only in a low-oxygen environment. The CobU protein may have an as yet undescribed ATP: CBI adenosyltransferase activity that is active only under low oxygen tensions. Consistent with this model, a cobA cobU+ strain can efficiently assimilate (CN)2CBI into CBL only under anaerobic growth conditions. Additionally, AdoCBI corrected the growth phenotype of the cobU1172 cobA double mutant, suggesting a block in the synthesis of this


VOL. 175, 1993

compound (Table 2). Should this putative ATP:CBI activity be associated with the CobU protein, it would presumably be in a domain of the protein different from those responsible for the kinase and guanylyltransferase activities. Alternatively, the cobU1l72 and cobU369 mutations may affect the ability of the CobU protein to utilize a nonadenosylated substrate. That is, CobU can utilize nonadenosylated CBI as a substrate under anoxic conditions but requires an adenosylated substrate when functioning in the presence of oxygen. This model is based on the observation that strains with a wild-type CobU protein (in a CobA-deficient background) can efficiently use nonadenosylated CBI as a substrate only under anaerobic growth conditions. The activity of the wild-type CobU protein may decrease in the presence of oxygen such that an adenosylated substrate is required for the aerobic synthesis of CBI-GDP. Consistent with this model are the results of experiments which implicate oxygen as a negative effector of the wild-type CobU. The negative affect of oxygen on wild-type CobU function is of interest because there are no obvious motifs in the amino acid sequence that would suggest the presence of oxygen-labile elements, such as Fe-S centers. According to this alternative model, strains carrying allele cobU369 or cobU1l72 retain the kinase and guanylyltransferase functions of the CobU protein. The amino acid changes conferred by cobU369 and cobU1172 may result in a mutant CobU protein whose kinase and guanylyltransferase activities are dependent on an adenosylated substrate under all growth conditions. In summary, the cobU369 and cobU1l72 mutations may either affect a previously undescribed oxygen-sensitive ATP:CBI adenosyltransferase activity or alter the ability of the CobU protein to utilize a nonadenosylated substrate under anaerobic growth conditions. A careful biochemical study of the wild-type and mutant CobU proteins under both aerated and anoxic conditions, including examining kinetic properties and binding constants of substrates and determining whether the CobU protein encodes an ATP:CBI adenosyltransferase, will be required to distinguish between the two models presented. ACKNOWLEDGMENTS This work was supported by Public Health Service grant GM40313 from the National Institute of General Medical Sciences to J.C.E-S. G. A. O'Toole was supported by the NIH Biotechnology training grant GM08349. We thank S.-J. Suh for input and many helpful discussions and B. M. Cali for critical reading of the manuscript. We also thank J. R. Roth for providing us with cob nucleotide sequence data prior to

publication. REFERENCES 1. Andersson, D. I., and J. R. Roth. 1989. Redox regulation of the genes for cobinamide biosynthesis in Salmonella typhimurium. J. Bacteriol. 171:6734-6739. 2. Balch, W. E., G. E. Fox, L. J. Magrum, C. R. Woese, and R. S. Wolfe. 1979. Methanogens: reevaluation of a unique biological group. Microbiol. Rev. 43:260-296. 3. Balch, W. E., and R. S. Wolfe. 1976. New approach to the cultivation of methanogenic bacteria: 2-mercaptoethanesulfonic acid (HS-CoM)-dependent growth of Methanobacterium ruminantium in a pressurized atmosphere. Appl. Environ. Microbiol. 32:781-791. 4. Blanche, F., L. Debussche, A. Famechon, D. Thibaut, B. Cameron, and J. Crouzet. 1991. A bifunctional protein from Pseudomonas denitnificans carries cobinamide kinase and cobinamide

5.

6.

7. 8.

9. 10.

11.

12.

13. 14. 15. 16.

17. 18. 19.

20.

21.

22.

23.

6335

phosphate guanylyltransferase activities. J. Bacteriol. 173:60526057. Cameron, B., F. Blanche, M.-C. Rouyez, D. Bisch, A. Famechon, M. Couder, L. Cauchois, D. Thibaut, L. Debussche, and J. Crouzet. 1991. Genetic analysis, nucleotide sequence, and products of two Pseudomonas denitrificans cob genes encoding nicotinate-nucleotide:dimethylbenzimidazole phosphoribosyltransferase and cobalamin (5'-phosphate) synthase. J. Bacteriol. 173:6066-6073. Crouzet, J., S. Levy-Schil, B. Cameron, L. Cauchois, S. Rigault, M.-C. Rouyez, F. Blanche, L. Debussche, and D. Thibaut. 1991. Nucleotide sequence and genetic analysis of a 13.1-kilobase-pair Pseudomonas denitrificans DNA fragment containing five cob genes and identification of structural genes encoding cob(I) alamin adenosyltransferase, cobyric acid synthase, and bifunctional cobinamide kinase-cobinamide phosphate guanylyltransferase. J. Bacteriol. 173:6074-6087. Davis, R. W., D. Botstein, and J. R. Roth. 1980. A manual for genetic engineering: advanced bacterial genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Escalante-Semerena, J. C., M. G. Johnson, and J. R. Roth. 1992. The CobII and CobIII regions of the cobalamin (vitamin B12) biosynthetic operon of Salmonella typhimurium. J. Bacteriol. 174:24-29. Escalante-Semerena, J. C., and J. R. Roth. 1987. Regulation of cobalamin biosynthetic operons in Salmonella typhimunum. J. Bacteriol. 169:2251-2258. Escalante-Semerena, J. C., S.-J. Suh, and J. R. Roth. 1990. cobA function is required for both de novo cobalamin biosynthesis and assimilation of exogenous corrinoids in Salmonella typhimurium. J. Bacteriol. 172:273-280. Fry, D. C., S. A. Kuby, and A. S. Mildvan. 1986. ATP-binding site of adenylate kinase: mechanistic implications of its homology with ras-encoded p21, F1-ATPase, and other nucleotidebinding proteins. Proc. Natl. Acad. Sci. USA 83:907-911. Hong, J.-S., and B. N. Ames. 1971. Localized mutagenesis of any specific small region of the bacterial chromosome. Proc. Natl. Acad. Sci. USA 68:3158-3162. Jeter, R. M., B. M. Olivera, and J. R. Roth. 1984. Salmonella typhimurium synthesizes cobalamin (vitamin B12) de novo under anaerobic growth conditions. J. Bacteriol. 159:206-213. Joshi, A. K., V. Baichwal, and G. Ferro-Luzzi Ames. 1991. Rapid polymerase chain reaction amplification using intact bacterial cells. BioTechniques 10:42-44. Maniatis, T., E. F. Fritsch, and J. Sambrook 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Martinez, E., B. Bartolome, and F. de la Cruz. 1988. pACY184derived cloning vectors containing the multiple cloning site and lacZa reporter gene of pUC8/9 and pUC18/19 plasmids. Gene 68:159-162. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Moller, W., and R. Amons. 1985. Phosphate-binding sequences in nucleotide-binding proteins. FEBS Lett. 186:1-7. O'Toole, G. A., M. R. Rondon, and J. C. Escalante-Semerena. 1993. Analysis of mutants of Salmonella typhimurium defective in the synthesis of the nucleotide loop of cobalamin. J. Bacteriol. 175:3317-3326. Rondon, M. R., and J. C. Escalante-Semerena. 1992. The poc locus is required for 1,2-propanediol-dependent transcription of the cobalamin biosynthetic (cob) and propanediol utilization (pdu) genes of Salmonella typhimurium. J. Bacteriol. 174:22672272. Roth, J. R., J. G. Lawrence, M. Rubenfield, S. Kieffer-Higgins, and G. M. Church. 1993. Characterization of the cobalamin (vitamin B12) biosynthetic genes of Salmonella typhimurium. J. Bacteriol. 175:3303-3316. Schmieger, H. 1971. A method for detection of phage mutants with altered transduction ability. Mol. Gen. Genet. 100:378-381. Schmieger, H., and H. Bakhaus. 1973. The origin of DNA in transducing particles of P22 mutants with increased transduction frequencies (HT-mutants). Mol. Gen. Genet. 120:181-190.

6336


24. Suh, S.-J., and J. C. Escalante-Semerena. 1993. Cloning, sequencing, and expression of cob which encodes ATP:corrinoid adenosyltransferase activity in Salmonella typhimurium. Gene 129:93-97. 25. Taylor, R. T., and H. Weissbach. 1973. N5-Methylenetetrahydrofolate-homocysteine methyltransferases, p. 121-165. In P. D. Boyer (ed.), The enzymes. Academic Press, Inc., New York. 26. Vogel, H. J., and D. M. Bonner. 1956. Acetylornithase of

J. BACTERIOL.

Escherichia coli: partial purification, and some properties. J. Biol. Chem. 218:97-106. 27. Way, J. C., M. A. Davis, D. Morisato, D. E. Roberts, and N. Kleckner. 1984. New TnlO derivatives for transposon mutagenesis and for construction of lacZ operon fusions by transposition. Gene 32:369-379. 28. Woolley, P., and B. F. C. Clark. 1989. Homologies in the structures of G-binding proteins: an analysis based on elongation factor EF-Tu. Bio/Technology 7:913-920.