Evidence that a B₁₂-Adenosyl Transferase Is Encoded within the Ethanolamine Operon of Salmonella enterica

Abstract

Adenosylcobalamin (Ado-B₁₂) is both the cofactor and inducer of ethanolamine ammonia lyase (EA-lyase), a catabolic enzyme for ethanolamine. De novo synthesis of Ado-B₁₂ by Salmonella enterica occurs only under anaerobic conditions. Therefore, aerobic growth on ethanolamine requires import of Ado-B₁₂ or a precursor (CN-B₁₂ or OH-B₁₂) that can be adenosylated internally. Several known enzymes adenosylate corrinoids. The CobA enzyme transfers adenosine from ATP to a biosynthetic intermediate in de novo B₁₂ synthesis and to imported CN-B₁₂, OH-B₁₂, or Cbi (a B₁₂ precursor). The PduO adenosyl transferase is encoded in an operon (pdu) for cobalamin-dependent propanediol degradation and is induced by propanediol. Evidence is presented here that a third transferase (EutT) is encoded within the operon for ethanolamine utilization (eut). Surprisingly, these three transferases share no apparent sequence similarity. CobA produces sufficient Ado-B₁₂ to initiate eut operon induction and to serve as a cofactor for EA-lyase when B₁₂ levels are high. Once the eut operon is induced, the EutT transferase supplies more Ado-B₁₂ during the period of high demand. Another protein encoded in the operon (EutA) protects EA-lyase from inhibition by CN-B₁₂ but does so without adenosylation of this corrinoid.

The eut operon of Salmonella enterica encodes enzymes for use of ethanolamine as a carbon and nitrogen source (28, 29). Only 4 of its 17 genes (19, 33), eutD, eutE, eutB, and eutC, have been correlated directly with an enzymatic activity known to be required for ethanolamine utilization (9, 29). The operon and metabolic pathway are diagrammed in Fig. 1. A fifth gene, eutR, encodes a positive regulatory protein that activates transcription of the eut operon in response to the simultaneous presence of ethanolamine plus adenosylcobalamin (Ado-B₁₂) (27, 32). The eutT and eutA genes described here have not been associated with any specific enzymatic function.

FIG. 1.

Pathway and operon for ethanolamine degradation. Ado-B₁₂ serves as a coinducer of the operon and cofactor for EA-lyase. Degradation of ethanolamine provides both a carbon source and nitrogen source (30). Cbi (cobinamide) is a corrinoid lacking upper and lower ligands that can be adenosylated by CobA to produce Ado-Cbi, a biosynthetic precursor of B₁₂. All eut genes are indicated above the map with the two normal promoters: the major regulated promoter (P_I) and the minor constitutive eutR promoter (P_II) (19, 27-29). Transposon-associated introduced promoters are below the map.

Use of ethanolamine as a carbon or nitrogen source requires Ado-B₁₂, which is the cofactor of ethanolamine ammonia lyase (EA-lyase; EC 4.3.1.7) (2, 6, 8, 12, 28). Ado-B₁₂ also serves as a coinducer (with ethanolamine) of eut operon transcription (18, 27, 32). A rationale for using Ado-B₁₂ as a coinducer is the fact that operon expression is futile without Ado-B₁₂ to support EA-lyase and this cofactor (unlike most others) is not always present. This is the case because S. enterica synthesizes Ado-B₁₂ de novo only under anaerobic conditions (1, 15, 20, 30). Therefore under aerobic conditions, cells can grow on ethanolamine only if the environment can provide either Ado-B₁₂ or a suitable precursor. Since Ado-B₁₂ is subject to photolysis (4), exogenous precursors are likely to lack the upper (adenosyl) ligand and require internal adenosylation.

It is clear that S. enterica can add the upper adenosyl ligand to cobalamins since commercial CN-B₁₂, which lacks this ligand, allows aerobic growth on ethanolamine (28). The CobA enzyme, ATP:cob(I)alamine adenosyl transferase, adenosylates assimilated CN-B₁₂ and also contributes to de novo B₁₂ synthesis (anaerobic) by adenosylating a biosynthetic intermediate (11, 34). CobA also adenosylates the assimilated precursor cobinamide (Cbi) and converts it to the biosynthetic intermediate Ado-Cbi (11). Another cobalamin transferase (PduO) is encoded within the pdu operon and is induced only during growth on propanediol; this transferase is not involved in any of the metabolism described here (17).

Existence of a third adenosyl transferase was suggested by the fact that CN-B₁₂ allows a cobA mutant to use ethanolamine as a nitrogen source when glycerol is the carbon source (see Table 2), but not with glucose (11). This implied existence of another transferase, perhaps subject to catabolic repression. Evidence is presented that EutT is this transferase and contributes to ability of cells to grow on ethanolamine, especially with low levels of exogenous CN-B₁₂. Another transferase candidate (EutA) helps cells resist inhibitory effects of CN-B₁₂ on EA-lyase, but does so without converting CN-B₁₂ to the normal cofactor Ado-B₁₂.

TABLE 2.

Aerobic growth phenotypes of eut point and deletion mutants

Strain	Relevant genotype		Eut(N) growth phenotypea			No. of mutations examinedb
	eut	cobA	None	+CN−B12 (0.08 μM)	+Ado−B12 (0.08 μM)
TT20042	eut+	−	−	+	+
TT10674		+	−	+	+
TT20078	eutT86(Am)	−	−	−	+	9c
TT20079		+	−	+	+
TT20026	eutA111	−	−	−	+	2e
TT20025		+	−	(−)d	+
TT20138	eutPQTD333(Δ)f	−	−	−	+	1
TT20139		+	−	+	+
TT20136	eutDM302(Δ)f	−	−	+	+	1
TT20137		+	−	+	+
TT20064	eutD101(Am)	−	−	+	+	3g
TT17319		+	−	+	+
TT20024	eutE56	−	−	+	+	3h
TT17310		+	−	+	+

^aGrowth tests were done in aerobic liquid medium containing glycerol as a carbon source and ethanolamine as a nitrogen source with indicated additions; growth was scored after 2 days.

^bSeveral different mutations in each gene were tested; all mutations in the same gene exhibited the same phenotype. Full genotypes are in Table 1.

^cAll known eutT mutants are nonsense mutations (19). The alleles tested were eutT67 [gln229(Am)], eutT74 [gln529(Am)], eutT75 [gln379(Oc)]eutT77[gln181(Am)], eutT78 [gln402(Op)], eutT86 [pro227(Leu)] [gln229(Am)], eutT88[gln184(Am)], eutT90[gln184(Am)], and eutT104[trp374(Op)].

^dThis strain has a Eut(N⁺) phenotype at 0.02 μM CN-B₁₂ but fails to grow at 0.08 μM CN-B₁₂; this inhibition phenotype will be described later.

^eThe alleles tested were eutA111 and eutA114.

^fIn-frame nonpolar deletion mutations (19).

^gThe alleles tested were eutD53[gln103(Oc)], eutD101[gln4656(Am)] and eutD124[arg518(Op)].

^hThe alleles tested were eutE56, eutE65, and eutE79.

MATERIALS AND METHODS

Bacterial strains and transposons.

Strains were derived from S. enterica (serovar Typhimurium) LT2 (Table 1). Normal and introduced promoters are diagrammed in Fig. 1. The cobA366::Tn10d(Cm) mutation maps far from the eut operon and eliminates the general B₁₂ adenosyl transferase (11, 34). An inserted element referred to as eutA::P_int throughout this paper is the eutA208::Tn10d(Tc) insertion (28), carrying an outward-directed constitutive promoter, P_int (32), created by a point mutation within the Tn10d(Tc) element (37). This strong promoter expresses eut genes downstream of eutA, including EutBC (EA-lyase) and EutR (32). Another Tn10 derivative (T-POP) (25) inserted in the eutJ gene (or in eutP) allowed tetracycline-inducible expression of genes downstream of its insertion site. Nonpolar deletion mutants eutPQTD333 and eutDM302 were isolated as tetracycline-sensitive Eut⁺ revertants of T-POP insertions in the eutQ or eutD genes (19) and a set of constructed single-gene, in-frame deletions will be described (23). The eut-38::MudA insertion lies just within the distal end of the eut transcript but outside of all reading frames; it reports operon transcription without impairing any eut function (19, 27). Other eut mutations have been described previously (19, 27).

TABLE 1.

Bacterial strains used in this study

Strain	Genotype
BE119	E. coli ompT hsdS (rB− mB−) dcm+ Tetrgal λ(DES) endA DE3[argU ileY leuW Camr]/pTA925 (T7 expression vector without insert)
BE205	E. coli ompT hsdS (rB− mB−) dcm+ Tetrgal λ(DES) endA DE3[argU ileY leuW Camr]/pTA1019 (T7 expression vector with eutT insert)
TT10000	LT2 wild type
TT10674	eut-38::MudA
TT17300	eutA208::Tn10d(Tc)Pint eutR156::MudJ
TT17306	eutA208::Tn10d(Tc)Pint
TT17310	eutE56 eut-38::MudA
TT17312	eutA114 eut-38::MudA
TT17319	eutD101 eut-38::MudA
TT17320	eutQ18::Mu dJ eutA208::Tn10d(Tc) Pint
TT20024	eutE56 eut-38::MudA cobA366::Tn10d(Cm)
TT20025	eutA111 eut-38::MudA
TT20026	eutA111 eut-38::MudA cobA366::Tn10d(Cm)
TT20033	eutQ18::MudJ eutA208::Tn10d(Tc)Pint cobA366::Tn10d(Cm)
TT20039	eutE163::MudJ eutA208::Tn10d(Tc)Pint
TT20041	eutE163::MudJ eutA208::Tn10d(Tc)Pint cobA366::Tn10d(Cm)
TT20042	eut-38::MudA cobA366::Tn10d(Cm)
TT20043	eutT11::MudA eutA208::Tn10d(Tc)Pint
TT20044	eutT11::MudA eutA208::Tn10d(Tc)Pint cobA366::Tn10d(Cm)
TT20045	eutP171::MudJ eutA208::Tn10d(Tc)Pint cobA366::Tn10d(Cm)
TT20046	eutP171::MudJ eutA208::Tn10d(Tc)Pint
TT20047	eutG3::MudA eutA208::Tn10d(Tc)Pint
TT20054	eutE10::Mud eutA208::Tn10d(Tc)Pint cobA366::Tn10d(Cm)
TT20069	eutE6::MudA eutA208::Tn10d(Tc)Pint
TT20071	eutE10::MudA eutA208::Tn10d(Tc)Pint
TT20078	eutT86 eut-38::MudA cobA366::Tn10d(Cm)
TT20079	eutT86 eut-38::MudA
TT20114	cobA366::Tn10d(Cm) cobR4 cob24::MudJ metE205 ara-9 zeb-1845::Tn10
TT20124	eutQ18::MudJ eutJ269::Tn10(Tc)T-POP
TT20125	eutQ18::MudJ eutJ269::Tn10(Tc)T-POP cobA366::Tn10d(Cm)
TT20126	eutT11::MudA eutJ269::Tn10(Tc)T-POP
TT20127	eutT11::MudA eutJ269::Tn10(Tc)T-POP cobA366::Tn10d(Cm)
TT20128	eutE10::MudA eutJ269::Tn10(Tc)T-POP
TT20129	eutE10::MudA eutJ269::Tn10(Tc)T-POP cobA366::Tn10d(Cm)
TT20136	eutDM302 eut-38::MudA
TT20137	eutDM302 eut-38::MudA cobA366::Tn10d(Cm)
TT20138	eutPQTD333 eut-38::MudA
TT20139	eutPQTD333 eut-38::MudA cobA366::Tn10d(Cm)
TT24803	eutT371::FRT(Δ)
TT25127	cobA366::Tn10d(Cm)
TT25128	cobA366::Tn10d(Cm)eutT371::FRT

Media.

Rich medium (NB) was nutrient broth (0.8%; Difco Laboratories) supplemented with NaCl (5 g/liter). Minimal media were variants of E medium, which contains citrate. NCE medium lacks citrate (26), and NCN medium lacks both citrate and a nitrogen source (5). Ethanolamine hydrochloride (20.5 mM; Aldrich) was used as a carbon source in NCE medium, as a nitrogen source in NCN medium with glycerol (20 mM), or as a carbon and nitrogen source in NCN medium. CN-B₁₂ (80 nM; Sigma Chemical Co.) was the usual exogenous vitamin B₁₂ source. Amino acids and purines were added to minimal medium at the concentrations previously recommended (10). NB agar media contained antibiotics at the following concentrations: ampicillin, 30 μg/ml; chloramphenicol, 20 μg/ml; kanamycin, 50 μg/ml; and tetracycline, 20 μg/ml. Solid media were prepared by adding agar (1.5%, Difco) to E, NCN, NCE, or NB medium.

Assay of β-galactosidase.

Strains were grown in NCE or NCN medium containing glycerol (0.2%) as a carbon source and ethanolamine (20.5 mM) plus CN-B₁₂ (80 nM) as inducers. Following growth to mid-log phase at 37°C on a New Brunswick Shaker, model 50, cells were held at 4°C until assayed. Enzyme activity was determined in chloroform-sodium dodecyl sulfate-permeabilized cells (21). Activity was expressed as nanomoles of nitrophenol produced per minute per optical density at 650 nm (OD₆₅₀) of cell culture turbidity. All presented values are the average of at least four determinations on two independent cultures.

Cloning the eutT gene for expression.

The eutT coding sequence was amplified by PCR using forward primer GCCGCCAGATCTGATGAACGATTTCATCACCGAAACGTGG and reverse primer GCCGCCAAGCTTTCATGGCTTCTCTCCCAACCGTTG (17). Template DNA from S. enterica was prepared with a Bio-Rad Aquapure genomic DNA isolation kit. Polymerase for PCR was a mixture of Pfu and Taq polymerases (7:1 [unit/unit]). The PCR products were cloned into the T7 expression plasmid pTA925 with BglIII and HinDIII (17). The DNA sequences of selected clones were verified. Plasmids pTA925 (no eutT insert) and pTA1019(eutT insert) were introduced into expression strain Escherichia coli BL21 DE3RIL (Stratagene, La Jolla, Calif.), to form strains BE119 and BE205, which were used for EutT production. Adenosyl transferase activity was assayed as described previously (17).

Overexpression of EutT and preparation of cell extracts.

EutT expression strains were grown at 30^oC with shaking in 1,000-ml baffled Erlenmeyer flasks containing 500 ml of Luria broth supplemented with 25-μg/ml kanamycin. Cells were grown to an OD₆₀₀ of 0.6 to 0.8. At that density, protein expression was induced by the addition of isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. Cultures were shaken for an additional 3 h at 30°C, placed on ice for 5 min, and collected by centrifugation for 5 min at 4,650 × g (maximum) using a Beckman J2-HS centrifuge and a Beckman JLA10.500 rotor. The cell pellet was frozen at −80°C until used. Cell pellets were resuspended in 50 mM potassium phosphate buffer at pH 7, containing 50 mM NaCl, and cells were disrupted in a 5-ml French pressure cell (SLM Aminco, Urbana, Ill.). Cell debris and unbroken cells were pelleted by centrifugation at 31,000 × g (maximum) with a Beckman JA20 rotor. The supernatant served as the soluble fraction for enzyme assays and polyacrylamide gel electrophoresis. The pellet obtained by centrifugation was extracted with Bacterial Protein Extraction Reagent II (B-PERII; Pierce, Rockford, Ill.) according to the manufacturer's directions, but with the following modifications. The B-PERII solution used for the first extraction was supplemented with the protease inhibitor phenylmethylsulfonyl fluoride at a concentration of 100 μg/ml. The B-PERII solution used for the second extraction was supplemented with DNase at a concentration of 20 μg/ml. Inclusion bodies were washed twice with 10 ml of a 1/20 dilution of B-PERII and then resuspended in 50 mM potassium phosphate pH 7.0 containing 50 mM NaCl; this suspension was used for assay of inclusion bodies (17).

Growth inhibition by CN-B₁₂.

A eutA mutant fails to grow on ethanolamine as a carbon source, regardless of the cobalamin form provided. It also fails to use ethanolamine as a nitrogen source when CN-B₁₂ is provided at a high level (80 nM) but does grow if Ado-B₁₂ is given. The Ado-B₁₂-stimulated growth is inhibited by high CN-B₁₂ levels (29). We show here that growth on ethanolamine as a nitrogen source is allowed (inhibition is avoided) if the level of CN-B₁₂ is reduced to 20 nM. To measure this inhibition, overnight nutrient broth cultures were diluted 1/100 into 200 μl of NCN medium containing glycerol (0.2%), ethanolamine (20.5 mM), and CN-B₁₂ at 0, 20, 80, or 140 nM. Cultures were incubated, with aeration, at 37°C in a Bio-Tec automated plate reader. Three parallel cultures were grown at each concentration of CN-B₁₂, and OD₆₅₀ turbidity was measured every hour.

Effect of cobA and eutT mutations on growth rates.

Three independent NB cultures of each strain were inoculated from single colonies grown on NB plates. After overnight growth, cells were pelleted, washed in minimal medium (pH 7.0), and used (35 μl of final suspension) to inoculate 6-ml cultures in minimal medium. Liquid minimal medium was a mixture of 5 mM KH₂PO₄, 5 mM NaNH₄HPO₄, and 1 mM MgSO₄ buffered at pH 7.0 by 50 mM MOPS (morpholinepropanesulfonic acid). The carbon source was 41 mM ethanolamine hydrochloride (0.04%). Cyanocobalamin concentrations are indicated later. Minimal medium also contained biotin; Ca-d-pantothenic acid, nicotinamide, and pyridoxine HCl at 4 × 10⁻⁴% (wt/vol); and thiamine and riboflavin at 2 × 10⁻⁵% (wt/vol) and trace metals as previously described (24). Cultures were aerated slightly by shaking at 240 rpm in tubes standing upright. Growth was monitored by observing OD₆₅₀ on a Spectronic 20D+ spectrophotometer.

RESULTS

Evidence for a CobA-independent route of cobalamin adenosylation.

Mutants of S. enterica lacking the CobA adenosyl transferase cannot use ethanolamine as a nitrogen source when provided with CN-B₁₂ and glucose as a carbon source (11). This failure is due to inability to convert CN-B₁₂ to the required EA-lyase cofactor, Ado-B₁₂; growth is restored if Ado-B₁₂ replaces CN-B₁₂. We noted that CN-B₁₂ allowed cobA mutants to use ethanolamine as a nitrogen source when glycerol rather than glucose provided carbon (Table 2, top row) or when ethanolamine provided both nitrogen and carbon. This implied that an adenosyl transferase (other than CobA) was expressed during growth on poor carbon sources. Conditions under which this additional activity was inferred were ones that induce the eut operon (27, 32), suggesting that a gene in the eut operon might encode the inferred transferase.

Preliminary evidence for two eut genes that might provide Ado-B₁₂.

Several eut mutations were combined with a cobA mutation and tested for their effect on the inferred ability to adenosylate cobalamin. When CN-B₁₂ was provided aerobically, only eutT and eutA point mutants failed to use ethanolamine as a nitrogen source on glycerol, Eut(N⁻). Below we characterize these eutT mutations and then return to eutA.

In an otherwise wild-type strain, available eutT mutations (all nonsense types) eliminated growth on ethanolamine as the sole carbon source even when Ado-B₁₂ is provided (19). This is due to their polar effect on expression of the downstream genes for EA-lyase (eutB and eutC), which is essential for use of ethanolamine. However, these simple eutT mutants can use ethanolamine as a nitrogen source, demonstrating that their level of EA-lyase is sufficient for the less-demanding task of supplying nitrogen (Table 2, rows 1 to 4). In strains lacking cobA, eutT mutations eliminate use of ethanolamine as a nitrogen source with provided CN-B₁₂, a Eut(N⁻) phenotype. This defect is corrected by providing Ado-B₁₂ (Table 2). The Eut(N⁻) phenotype on CN-B₁₂ was caused by nonpolar deletion mutations that remove eutT (eutPQTD333) but was not seen in eutT⁺ strains carrying a nonpolar deletion (eutDM302) or in strains with a polar point mutation downstream of the eutT gene (e.g., eutD101). Since the Eut(N⁻) phenotype of eutT mutants on CN-B₁₂ was corrected by providing either Ado-B₁₂ or a functional CobA transferase activity, EutT was a prime candidate for the eut-specific adenosyl transferase.

Genetic evidence that EutT contributes to cobalamin adenosylation in vivo.

The cofactor Ado-B₁₂ is required to induce the eut operon and to serve as cofactor for EA-lyase. Thus, if eutT mutants fail to make Ado-B₁₂, both induction and enzyme activity should be impaired. These processes were tested independently in strains with insertion eutA::P_int (32), which constitutively expresses the EA-lyase genes (eutBC). Fusions of lac to the eut operon upstream of eutA reported operon induction (Table 3). Operon induction was tested on glycerol-ammonia medium, where no EA-lyase activity is required. All cobA⁺ strains showed induction. However, in cobA mutants, no induction was seen for Mud insertions that eliminated EutT activity either directly or by polarity (eutP, eutQ, or eutT; Table 3, rows 3 to 6). Inducibility was normal in strains with a eutE::Mud fusion, distal to the eutT gene (Table 3, rows 7 and 8). The induction defect in strains lacking EutT (and CobA) was corrected by Ado-B₁₂ (Table 3, right column). Thus, EutT allows CN-B₁₂ to induce the operon without CobA and is therefore inferred to contribute to conversion of CN-B₁₂ to the inducer Ado-B₁₂.

TABLE 3.

Two tests of EutT-mediated Ado-B₁₂ production, induction and activity

Strain	Genotypea		Eut(N) growth phenotypeb			eut operon induction (β-galactosidase activity)c
	eut	cobA	None	+CN-B12	+Ado-B12	None	+EA	+EA +CN-B12	Fold increase by CN-B12d	Effect of CobA+e	+EA +Ado-B12
TT20045	eutP171::lac (EutT−)	−	−	−	+	2.7	3.9	6.2	1.6	21	130
TT20046		+	−	+	+	1.6	4.6	130	28		110
TT20033	eutQ18::lac (EutT−)	−	−	−	+	9.8	32	46.	1.4	28	1350
TT17320		+	−	+	+	21	56	1300	23		960
TT20044	eutT11::lac (EutT−)f	−	−	−	+	5.5	18	27.	1.5	34	590
TT20043		+	−	+	+	6.0	20	910	46		590
TT20041	eutE163::lac (EutT+)g	−	−	+	+	5.8	29	960	33	0.9	740
TT20039		+	−	+	+	5.1	30	850	29		760

^aAll strains contain the eutA::P_int mutation that eliminates EutA protein and highly expresses the EutBC (lyase) and EutR (transcription activator) proteins (Fig. 2A). Complete genotypes are in Table 1.

^bGrowth phenotypes were scored + or − by replica-printing cell patches from an NB plate to an NCN plate with glycerol as a carbon source and ethanolamine as a nitrogen source.

^cCells were grown on NCE medium containing glycerol as a carbon source and ammonia with additions as shown. β-Galactosidase activity was determined as described in Materials and Methods.

^dInduction caused by CN-B₁₂.

^eThis is the ratio (cobA⁺/cobA mutant) of operon induction in strains isogenic except for the cobA mutation.

^fEssentially the same results were seen with eutT17::Mud-lac.

^gEssentially the same results were seen with all other Mud-lac insertions located promoter distal to eutT (Fig. 2A). MutA insertion alleles tested were eut-6 (between eutN and eutE), eutE12, eutE10, eutE1, eutJ26, and eutH25. The MudJ alleles tested were eut-168 (between eutD and eutM) eutE163, and eutE181.

Further evidence for a role of EutT in Ado-B₁₂ production was its effect on EA-lyase-dependent growth. In the strains used above, EA-lyase (EutBC) was expressed constitutively from a strong promoter (P_int) within the inserted eutA::P_int element. Since EA-lyase is the only enzyme required to produce ammonia from ethanolamine, these strains can use ethanolamine with or without operon induction but do so only if cofactor Ado-B₁₂ is available (32). Mud insertions that eliminate EutT expression (those in the eutP, eutQ, or eutT genes) caused a Eut(N⁻) phenotype in the absence of CobA (Table 3, first six rows), whereas a eutE::Mud insertion (distal to eutT) is Eut(N⁺) (Table 3, last two rows). The eutT growth defect on CN-B₁₂ was corrected by Ado-B₁₂ or by restoring a cobA⁺ allele (Table 3). Similar results were seen for point mutations in genes eutT, eutD, and eutE (data not shown). Thus, EutT served to provide Ado-B₁₂ by the criterion of ethanolamine growth, as well as operon induction. (Effects of a simple in-frame eutT deletion on growth will be shown later.)

Assays of EutT adenosyl transferase activity.

The genetic evidence above suggested that EutT was a B₁₂ adenosyl transferase. This conclusion was confirmed by transferase assays. Since assay of the CobA and PduO adenosyl transferases requires overproduction of the protein (34), the EutT protein was produced in a T7 expression system (17). A strain (BE205) with eutT inserted in the expression plasmid produced relatively large amounts of a 30.2-kDa protein in both the soluble and inclusion body fractions; this protein was not found in a strain (BE119) whose expression plasmid lacked eutT. Extracts of strains expressing eutT (BE205) showed 50- to 100-fold more transferase activity than the control strain (BE119) (Table 4). These assays were done exactly as described previously and used OH-B₁₂ (reduced to the CobI form) as an adenosyl-accepting substrate (17). It should be noted that a substantial fraction of adenosyl transferase activity sedimented with cell debris and may be present as inclusion bodies or associated with membranes. These results demonstrate that EutT protein has ATP:cob(I)alamin adenosyl transferase activity; the phenotypes above show that this activity is relevant in vivo.

TABLE 4.

ATP:cob(I)alamin adenosyl transferase activity

Straina	Plasmidb	Gene expressed	Fractionc	Sp act (nmol/ min/mg)d
BE119	pTA925	Vector only	Soluble	0.3
BE119	pTA925	Vector only	Inclusion bodies	0.4
BE205	pTA1019	EutT	Soluble	33.2
BE205	pTA1019	EutT	Inclusion bodies	21.1

^aComplete genotypes are in Table 1.

^bPlasmid construction is described in Materials and Methods.

^cPreparation of soluble and inclusion body fractions for enzyme assay is described in Materials and Methods.

^dATP:cob(I)alamin specific activities were determined as described in Materials and Methods. The values shown are the averages of three trials.

Phenotypes of eutA mutations.

Like eutT mutants, eutA mutants failed to use ethanolamine as either a carbon or nitrogen source when standard levels of CN-B₁₂ (80 nM) were provided. It was shown earlier that growth on ethanolamine as a nitrogen source was restored when Ado-B₁₂ replaces CN-B₁₂, but not when Ado-B₁₂ and high CN-B₁₂ were supplied together (29), suggesting an inhibitory effect of CN-B₁₂. A lower CN-B₁₂ concentration (20 nM) allowed eutA mutants to use ethanolamine as a nitrogen source. The inhibition of growth by high levels of CN-B₁₂ agrees with the previous in vitro demonstration that purified EA-lyase (EutBC) is inhibited by CN-B₁₂ (3, 16, 31). Thus EutA seems to protect against inhibition by CN-B₁₂, reminiscent of reactivating factors that remove inappropriate cofactors from other B₁₂-dependent enzymes (22, 35, 36). Since EutA might protect EA-lyase by adenosylating excess CN-B₁₂, it was tested further as a candidate for an adenosyl transferase.

EutA protein does not provide an adenosylation function.

The above tests of EutT function were done in strains that carried the eutA::P_int insertion and thus lacked EutA function, showing that EutA protein is not required for EutT transferase activity. To test whether EutA can provide adenosylation independent of EutT, the eutA gene was expressed from a tetracycline-inducible promoter (T-POP) inserted in the eutJ gene; this promoter also expressed EA-lyase (eutBC) genes needed for growth on ethanolamine. In this strain, ability to convert CN-B₁₂ to Ado-B₁₂ could be assessed as ability to grow on ethanolamine as a nitrogen source (which is independent of operon induction). Alternatively, Ado-B₁₂ production can be scored by operon induction (which does not require EA-lyase activity). Promoter proximal insertions of Mud-lac in the eutQ, eutT, or eutE genes were added to the eutA::P_int strain in order to block expression of upstream genes and as reporters of operon induction. The eutQ and eutT insertions eliminate EutT expression, while the eutE::MudJ insertion leaves eutT expression intact. As seen in Table 5, high-level expression of the eutA gene (from the eutJ::T-POP promoter) does not provide Ado-B₁₂ in strains that lack EutT and CobA. That is, EutA could not substitute for the EutT or CobA adenosyltransferases in supporting either induction or EA-lyase-dependent growth. We conclude that EutA does not provide an adenosyl transferase activity and must provide resistance to the inhibitory effect of high CN-B₁₂ in some other way. Additional tests of this inhibition are below.

TABLE 5.

The EutA protein cannot provide adenosyl transferase activity

Strain	Genotypea		Eut(N) growth phenotypeb			Operon induction (β-galactosidase activity)c
	eut insertion	cobA	None	+CN-B12	+Ado-B12	None	+CN-B12	Fold increase by CN-B12d	Effect of CobAe	+Ado-B12	Fold increase by Ado-B12f
TT20125	eutQ18::lac (EutT−A+)	−	−	−	+	5.7	5.8	1.0	240	1,200	220
TT20124		+	−	+	+	5.6	1,400	250		1,300	230
TT20127	eutT11::lac (EutT− A+)	−	−	−	+	4.2	3.8	0.9	190	640	150
TT20126		+	−	+	+	5.3	710	135		670	130
TT20129	eutE163::lac (EutT+ A+)	−	−	+	+	4.2	680	160	1.2	870	210
TT20128		+	−	+	+	6.3	900	140		920	150

^aAll strains carry (distal to the lac fusion) a eutJ::T-POP insertion, which expresses the eutGHABCR genes in the presence of tetracycline. Complete genotypes are in Table 1.

^bAll strains were tested for growth on NCN glycerol medium with 20.5 mM ethanolamine, 0.08 μM CN-B₁₂ or Ado-B₁₂, and 2-μg/ml tetracycline.

^cAll strains were grown on NCE glycerol NH₃ medium with 20.5 mM ethanolamine, 0.08 μM CN-B₁₂ or Ado-B₁₂, and 2-μg/ml tetracycline. β-Galactosidase activity was determined as described in Materials and Methods.

^dInduction caused by CN-B₁₂; large induction is due to high expression of EutR from T-POP.

^eThis is the ratio (cobA⁺/cobA mutant) of operon induction in strains isogenic except for the cobA mutation.

^fInduction caused by Ado-B₁₂.

Inhibition of eutA mutants by CN-B₁₂.

The inhibitory effect of CN-B₁₂ (80 nM) on eutA mutants noted above was relieved at lower concentrations of CN-B₁₂ (20 nM), which allowed these mutants to grow on glycerol using ethanolamine as a nitrogen source. The growth inhibition by high CN-B₁₂ is not due to a failure to express the operon, since normal induction was seen in eutA mutant strains by 80 nM CN-B₁₂ (in combination with ethanolamine). We presume that CN-B₁₂ inhibits growth by inhibiting EA-lyase directly as seen previously for the pure enzyme (2). However this sensitivity seemed to require other genes in the eut operon.

A role of other genes in sensitivity to CN-B₁₂ was demonstrated with the eutA::P_int insertion (described above), which disrupts eutA and expresses EA-lyase (EutBC) from a constitutive promoter. A strain with this insertion grew on ethanolamine as a nitrogen source (with 20 nM CN-B₁₂) but was inhibited by a higher concentration (80 nM) of CN-B₁₂. The inhibition by high CN-B₁₂ was abolished by an insertion in the promoter proximal eutQ gene, but not by an insertion in the eutG gene. This suggested that one or more genes upstream of eutG (eutS, -P, -Q, -D, -T, -M, -N, -E, or -J) must be needed for sensitivity. Consistent with this, a eutR mutation, which prevents expression of these genes, eliminated sensitivity to high CN-B₁₂. (As described above, the strains used have no EutA function and express EA-lyase [EutBC] constitutively from the P_int promoter.)

To identify the gene or genes necessary for inhibition by CN-B₁₂, a series of in-frame deletions of individual eut genes were added to the eutA::P_int mutant. The sensitivity of the parent strain to high CN-B₁₂ (Fig. 2A) appears as delayed growth of liquid cultures. This sensitivity was not relieved by deletion of eutT (adenosyl transferase) or the eutD, eutE, or eutJ genes (encoding metabolic enzymes acting after acetaldehyde). Growth of the eutT and eutE mutants is shown in Fig. 2B and C. Inhibition by high CN-B₁₂ was eliminated by removal of any one of the eutM, -N, -P, -Q, or -S genes. Growth of the eutM deletion mutant was typical and is shown in Fig. 2D. The EutM, -N, and -S proteins are homologues of carboxysome shell proteins; the functions of EutP and -Q are unknown. These effects are discussed below. It should be noted that delayed growth caused by high CN-B₁₂ occurs in tests reported in Table 3, but growth is scored after 48 h in these tests so the inhibition does not interfere.

FIG. 2.

Genes required for inhibition of eutA mutants by CN-B₁₂. Mutants lacking EutA function can grow aerobically on ethanolamine as a nitrogen source when supplied with a low level of CN-B₁₂ but are inhibited by higher concentrations of CN-B₁₂ (A). This sensitivity requires expression of some eut genes upstream of the eutA gene (eutS, -P, -Q, -M, and -N) but not others (eutT, -D, -E, -J, -G, and -H). All strains carry the eutA::P_int insertion and therefore express EA-lyase constitutively. Added mutations are constructed nonpolar deletion mutations. (B) ΔeutE. (C) ΔeutT. (D) ΔeutM. The responses shown are typical of mutations in the two groups. Growth rates were determined with a Biotec automated plate reader.

Testing autogenous induction of the eut operon by EutT and EutR.

Evidence is described above that EutT and CobA (but not EutA) contribute to formation of Ado-B₁₂, the inducer and required cofactor for ethanolamine degradative enzymes. This raises the question of the relative roles of EutT and CobA during normal growth on ethanolamine. The positive regulatory protein EutR is encoded within the eut operon (Fig. 1) and thus serves to activate its own production (autogeneous regulation) (27, 32). This autoinduction circuit is required for maximal operon expression (27, 32). The EutT protein also contributes to its own expression by producing Ado-B₁₂, a coinducer of the operon. Thus a strain with a cobA mutation and an uninduced eutT in the operon should not convert CN-B₁₂ to Ado-B₁₂ and therefore should show no induction of its eut operon. However, this situation might be unstable if a momentary stochastic expression of the operon provided enough EutT adenosyl transferase to initiate an autocatalytic induction (i.e., if a tiny bit of EutT produces enough Ado-B₁₂ to supply more EutT and launch a full induction).

To test this, cobA strains with three different preinduced levels of EutT were examined for operon induction by CN-B₁₂ as the Ado-B₁₂ precursor (Table 6). The first two rows describe a normal eut operon with lac genes that were fused at a point within the operon distal to all structural genes (eut-38::lac), leaving EutT expression dependent on induction. With CobA (row 2), normal induction is seen, but without CobA, uninduced EutT allows very little induction (1.8-fold) by CN-B₁₂ ethanolamine on NH₃-glycerol-ethanolamine medium. A slightly greater induction by CN-B₁₂ (fivefold) was seen when ethanolamine must provide the nitrogen source (Table 6, row 1, numbers in parentheses). Previous work has revealed no effect of nitrogen limitation on eut operon transcription (27, 32). We suggest that when ethanolamine must provide a nitrogen source, the only cells that grow are those that have stochastically achieved a slightly higher operon expression level, which can be maintained by the produced EutT enzyme.

TABLE 6.

Initiating operon induction with EutT as the sole Ado-B₁₂ source

Strain	Genotypea		eut operon induction (β-galactosidase activity)b
	eut	cobA	None	+EA	Fold increase by EAc	+EA + CN-B12	Fold increase by CN-B12d	+EA + Ado-B12	Fold increase by Ado-B12e
TT20042	eut-38:Mudlac (Eut+)f	−	NA	6.0		11 (34)	1.8 (5.6)	110 (74)	19 (12)
TT10674		+	NA	5.8		240 (164)	42 (19)	150 (110)	25 (19)
TT20129	eutE10:Mudlac eutJ269::T-POP	−	NA	4.2		680	160	870	210
TT20128		+	NA	6.3		900	140	920	150
TT20127	eutT11::Mudlac eutJ269::T-POP	−	NA	4.2		3.8	0.9	640	150
TT20126		+	NA	5.3		710	130	670	130
TT20054	eutE10:Mudlac eutA::Pint	−	3.6	36	10.0	1,300	37	690	19
TT20071		+	3.5	27	7.7	920	34	590	22
TT20044	eutT11::Mudlac eutA::Pint	−	5.5	18	3.3	27	1.5	590	33
TT20043		+	6.0	20	3.3	910	46	590	30

^aComplete genotypes are in Table 1.

^bStrains were grown on NCE glycerol medium; 20.5 mM ethanolamine, and either 80 nM CN-B₁₂ or 80nM Ado-B₁₂. Strains with a T-POP insertion were grown with 2-μg/ml tetracycline. NA, not assayed. Numbers in parentheses are for strains grown without NH₃; ethanolamine must serve as a nitrogen source.

^cInduction by ethanolamine.

^dInduction by CN-B₁₂ plus ethanolamine over that seen for ethanolamine alone.

^eInduction by Ado-B₁₂ plus ethanolamine over that seen for ethanolamine alone.

^fThis strain has a functional eut operon with a lac operon fusion to the transcribed region distal to all genes.

To test the effect of a higher basal level of EutT on operon inducibility, a eutJ::T-POP insertion was added to strains carrying various eut-lac reporter fusions; the T-POP insertion allows tetracycline-inducible expression of downstream genes including eutR. Increasing the level of EutR, is known to allow modest operon expression without regulatory effectors (27, 32). In eutT⁺ strains expressing eutR from the eutJ::T-POP (Table 6, rows 3 and 4), ethanolamine alone did not induce operon induction with or without CobA, but ethanolamine plus CN-B₁₂ caused full operon induction independent of CobA (compare rows 3 and 4). The full induction depended on the eutT⁺ gene in strains lacking cobA: that is, when the reporter disrupted eutT, no induction was seen without CobA (rows 5 and 6). Thus a very slight increase in basal operon expression provided enough EutT (and adenosylated CN-B₁₂) for full operon induction by CN-B₁₂ independent of CobA.

A still higher basal (uninduced) level of operon expression is seen in strains that express eutR from the stronger P_int promoter (Table 6, rows 7 to 10). When such strains had a functional eutT gene (rows 7 and 8), ethanolamine alone caused a slight induction but ethanolamine plus CN-B₁₂ caused full induction independent of CobA (compare rows 7 and 8). This CobA-independent induction depended on EutT (compare rows 9 and 10 to 7 and 8). In all of the strains above, Ado-B₁₂ induced the operon even when both CobA and EutT are limiting (see right column).

Physiological roles of EutT and CobA enzymes in ethanolamine metabolism.

The above conclusions raise the question of the relative importance of EutT and CobA to growth on ethanolamine. To test this, an in-frame (nonpolar) deletion of the eutT gene was constructed and placed in strains with and without a cobA mutation. These strains were compared for their ability to grow on ethanolamine (Fig. 3). Lack of EutT caused little impairment of growth on high levels of exogenous B₁₂, but reduced growth rate at low B₁₂ levels. Lack of CobA alone delayed growth initiation regardless of B₁₂ level, presumably by making it difficult to induce the operon, as described above.

FIG. 3.

Effects of cobA and eutT mutations on use of ethanolamine as a carbon source. Growth and media are described in Materials and Methods. In panel A, CN-B₁₂ was added at 10 nM. In panel B, CN-B₁₂ was added at 100 nM. All cultures were grown under low aerobic conditions at 30°C. The isogenic strains used were TT10000 (wild type; squares), TT25127 (cobA; upward triangles), TT24803 (eutT; downward triangles), and TT25128 (cobA eutT; circles).

Evidence that EutT adenosyl transferase does not contribute to de novo B₁₂ synthesis or assimilation of cobinamide.

The CobA adenosyl transferase acts on several substrates, including CN-B₁₂ or the incomplete corrinoid cobinamide (Cbi; the corrinoid ring lacking both the upper and lower ligands); CobA also acts on some biosynthetic intermediate (perhaps cobyric acid) leading to production of Ado-Cbi, the precursor of Ado-B₁₂ (11). The experiments above demonstrated that EutT adenosylates assimilated CN-B₁₂ but do not test its ability to use Cbi or a biosynthetic intermediate.

EutT, unlike CobA, does not adenosylate either Cbi or the biosynthetic intermediate. Strains for these tests expressed eutT from the tetracycline-inducible promoter in a eutP::TPOP insertion (19). Production of B₁₂ in vivo was monitored by using a metE mutant, whose synthesis of methionine depends on the alternative B₁₂-dependent enzyme MetH (reviewed in reference 30). Results are outlined below.

First, a metE mutant can grow aerobically without methionine if Cbi is provided as a B₁₂ precursor. This growth occurs because CobA enzyme adenosylates Cbi to Ado-Cbi, which can be converted to B₁₂ (11). A metE cobA double mutant lacks the CobA adenosyl transferase and cannot produce B₁₂ from Cbi; therefore, it cannot grow on minimal medium without added methionine (or B₁₂). This defect was not corrected by inducing EutT, suggesting that (unlike CobA) EutT cannot adenosylate assimilated Cbi.

Second, a metE mutant can grow anaerobically on minimal medium because B₁₂ is synthesized de novo under these conditions. The de novo B₁₂ pathway depends on adenosylation of an intermediate by CobA (11). A metE cobA double mutant cannot synthesize B₁₂ de novo and fails to grow anaerobically on minimal medium. This defect was not corrected by expressing EutT, suggesting the EutT cannot adenosylate the intermediate in the normal biosynthetic pathway.

DISCUSSION

Ethanolamine metabolism requires Ado-B₁₂ as a coinducer (with ethanolamine) of operon transcription and as a cofactor for the first enzyme (EA-lyase). Under natural conditions, Ado-B₁₂ is not always available because it is synthesized only anaerobically and adenosylated corrinoids may not be present in all aerobic environments (30). By requiring Ado-B₁₂ (plus ethanolamine) for induction of the eut operon, cells ensure that this gene system (17 proteins) is expressed only if both its substrate and required enzyme cofactor are present.

Induction of eut operon transcription is autocatalytic in the sense that the positive regulatory protein EutR is encoded within the operon that it controls (27). This arrangement is thought to be needed because the EutBC (EA-lyase) and EutR (regulator) proteins compete for binding Ado-B₁₂ (32). By producing these proteins at a constant ratio (from genes in the same operon), the cell can respond to a wide range of Ado-B₁₂ levels and remain sensitive to induction at all levels of operon expression; this can be achieved with very little investment in EutR protein prior to induction. The EutT adenosyl transferase adds a second autocatalytic element to this regulatory circuit by contributing to the level of a coinducer of the operon (Ado-B₁₂). Because of this circuit, cells need maintain only a minimal pool of Ado-B₁₂ (made by CobA)—just sufficient to initiate eut operon induction. As the operon (and its eutT gene) is induced, EutT may supplement the Ado-B₁₂ level during high-demand growth on ethanolamine.

The three cobalamin adenosyl transferases, CobA, PduO, and EutT, play distinct biological roles. The CobA enzyme may be the basal housekeeping activity. It supports de novo B₁₂ synthesis (anaerobically) and can adenosylate the assimilated B₁₂ precursor cobinamide, allowing its conversion to Ado-B₁₂. This basal Ado-B₁₂ level initiates eut operon induction. The EutT enzyme adds Ado-B₁₂ to support full induction and allows full growth ability during the period of high demand. The PduO enzyme is thought to serve an analogous function during growth on propanediol, in which CobA seems to play a very minor aerobic role. The pdu operon is induced by propanediol alone (with no need for cofactor), and mutants lacking CobA or PduO singly show only slightly impaired growth on propanediol using CN-B₁₂, while the cobA, pduO double mutant fails to grow (17). It is not known whether PduO can substitute for CobA in de novo B₁₂ synthesis or in assimilation of cobinamide.

It is surprising that the three cobalamin adenosyl transferases of S. enterica, EutT, PduO, and CobA, show no obvious amino acid sequence similarity (17). This could either reflect independent evolutionary origins or such extensive divergence from a common ancestor that homology is not apparent. If the three adenosyl transferases are derived from a common ancestor, their divergence may have been driven by the need (for PduO and EutT) to form close interactions with distinct groups of proteins contributing to the two pathways. The function of both these pathways is associated with formation of carboxysomes (7, 13, 14, 23).

The EutA protein appears to protect EA-lyase from inhibition by excessive CN-B₁₂ (29). Evidence is provided that this protection does not involve conversion of inhibitory CN-B₁₂ to Ado-B₁₂. It seems likely that EutA serves as a reactivating factor that removes damaged or inappropriate cofactor from the enzyme, as has been shown for several B₁₂-dependent enzymes (22, 35, 36). It is curious that while purified EA-lyase is strongly inhibited by CN-B₁₂ (2) in vitro, this inhibition depends in vivo on the EutM, -N, -P, -Q, and -S proteins. While the functions of EutP and -Q are not known, the EutM, -N, and -S proteins resemble structural proteins or the carboxysome, an organelle thought to contain enzymes of the ethanolamine pathway. Thus sensitivity of eutA mutants to CN-B₁₂ may require an intact carboxysome. We suggest that adenosylation (by both CobA and EutT) may occur primarily outside this compartment, and CN-B₁₂ that enters the compartment may escape adenosylation and inhibit EA-lyase. Disruption of the carboxysome may make it impossible for excessive CN-B₁₂ to escape adenosylation. Alternatively, these proteins may form complexes with EA-lyase that increase its sensitivity to CN-B₁₂.

ADDENDUM IN PROOF

Evidence that EutA removes CN-B₁₂ from lyase was published recently (K. Mori, R. Bando, N. Hieda, and T. Turaya, J. Bacteriol. 186:6845-6854, 2004).