(Received for publication, February 25, 1997, and in revised form, April 21, 1997)
From the Department of Bacteriology, University of Wisconsin-Madison, Madison, Wisconsin 53706
Thiamin pyrophosphate is an essential cofactor that is synthesized de novo by Salmonella typhimurium. In bacteria, the end product of the de novo biosynthetic pathway is thiamin monophosphate, which is then phosphorylated by thiamin-monophosphate kinase (EC 2.7.4.16) to form thiamin pyrophosphate. We have isolated and characterized the thiL gene of S. typhimurium and showed that thiL is a 978-base pair open reading frame encoding a 35-kDa protein with thiamin-monophosphate kinase activity. thiL was located in the 10-centisome region of the S. typhimurium chromosome. We demonstrated that altered thiamin-monophosphate kinase activity resulted in decreased repression of transcription of thiamin pyrophosphate-regulated thiamin biosynthetic genes. In contrast to other thi loci, thiL is not transcriptionally regulated by thiamin pyrophosphate. This result is consistent with a dual role for ThiL in de novo biosynthesis and in salvage of exogenous thiamin.
Thiamin pyrophosphate (TPP)1 is an
essential cofactor for a number of well characterized enzymes in the
cell (e.g. pyruvate dehydrogenase (EC 1.2.4.1),
-ketoglutarate dehydrogenase (EC 1.2.4.2), and acetolactate synthase
(EC 4.1.3.18). Despite the important role it plays in metabolism, the
TPP biosynthetic pathway and its regulation are not well understood in
any organism.
Thiamin consists of a 4-amino-5-hydroxymethylpyrimidine (HMP) moiety,
and a 4-methyl-5-(-hydroxyethyl)thiazole (THZ) moiety (see Fig. 1).
Phosphorylated derivatives of these two moieties (HMP(PP) and THZ(P))
are condensed to form thiamin monophosphate (TMP), the product of
de novo thiamin biosynthesis in both bacteria and yeast.
This condensation is catalyzed by thiamin-phosphate pyrophosphorylase
(EC 2.5.1.3), the product of the thiE gene in
Escherichia coli (1) and of the thi6 gene in
Sacchromyces cerevisiae (2). After the formation of TMP, the
strategies to generate TPP differ between enterics and yeast. In yeast,
TMP is first dephosphorylated to generate thiamin, and TPP is then formed by the addition of a pyrophosphate group from ATP in a step
catalyzed by thiamin pyrophosphokinase (EC 2.7.6.2), encoded by the
thi80 gene in S. cerevisiae (3). In contrast, in
both E. coli and Salmonella typhimurium, TMP is
directly phosphorylated to generate TPP by action of
thiamin-monophosphate kinase (EC 2.7.4.16) (4). Thus, unlike yeast,
E. coli and S. typhimurium require a salvage
enzyme (thiamin kinase (EC 2.7.1.89), encoded by thiK (4))
to incorporate exogenous thiamin into the TPP pools
(Fig. 1) (for review, see Ref. 5).
Previous work demonstrated that cell-free extracts of wild-type E. coli contained thiamin-monophosphate kinase activity, and mutants lacking this activity were subsequently isolated. These mutants required TPP for growth and defined the thiL locus (6). Hfr mapping experiments located thiL at centisome (Cs) 10 on the E. coli chromosome (4), and it was proposed that this locus encoded thiamin-monophosphate kinase.
During work on the regulation of thiamin biosynthesis in S. typhimurium, we isolated a point mutation (thiR927) that eliminated the ability of exogenous thiamin or TMP to repress transcription of the thi-operon located at 90 Cs (7). This mutation did not, however, affect the ability of TPP to repress transcription of this operon (7). Since both thiR927 and thiL mapped to the 10 Cs region, our results were consistent with thiR927 being an allele of thiL that resulted in altered enzyme activity. We suggested that thiR mutants had lower TPP pools resulting in constitutive expression of genes normally repressed by TPP. To test this model, characterization of thiL was initiated in S. typhimurium.
We report here the identification of the thiL gene at Cs 9.5 on the S. typhimurium chromosome. We show herein that the thiL gene encodes a 35-kDa protein with thiamin-monophosphate kinase activity. Further, we demonstrated that the thiR927 mutation is an allele of thiL resulting in a glycine to aspartate change at position 132 of the predicted amino acid sequence of the protein, and thus we re-designated this allele thiL927. This work supports the hypothesis that TPP is the regulatory molecule for thiamin synthesis in S. typhimurium and predicts the existence of a sensor/regulatory protein.
Bacterial Strains
All strains used in this study are derivatives of S. typhimurium LT2, unless noted, and are listed in
Table I. MudJ is used throughout the paper to refer to
the MudI1734 transposon, which has been described (8), and
Tn10d(Tc) refers to the transposition defective
mini-Tn10(Tn1016
17) (9).
|
Culture Media and Biochemicals
No-carbon source E medium supplemented with 1 mM
MgSO4 and 11 mM glucose was used as minimal
media (10, 11). Difco nutrient broth (8 g/liter) with NaCl (5 g/liter)
added was used as rich medium. Luria broth was used for experiments
involving plasmid manipulation. Difco BiTek agar (15 g/liter) was added
for solid medium. Thiamin and its phosphoesters were added, where
indicated, to a final concentration of 100 nM. Antibiotics
were added as needed to the following concentrations in rich and
minimal media, respectively: kanamycin (50, 125 mg/ml), ampicillin (30, 15 mg/ml), tetracycline (20, 10 mg/ml), and chloramphenicol (20, 4 mg/ml). All chemicals were purchased from Sigma except
isopropyl-1-thio--D-galactopyranoside (IPTG), which was
purchased from Fisher Biotech (Chicago, IL).
Genetic Methods
Transduction MethodsAll transductional crosses were performed by using the high frequency transducing bacteriophage P22 mutant HT 105/1 int-201 (12) as described (13). Transductants were purified and identified as phage-free by cross-streaking on green plates (14).
Mutant IsolationStrains auxotrophic for thiamin pyrophosphate (TPP) were isolated by insertional mutagenesis with one of two defective transposons; Tn10d(Tc) or MudJ elements. In the case of the Tn10d(Tc) elements, a pool of cells containing >80,000 independent insertions was generated as described elsewhere (15, 16). A P22 lysate grown on these cells was used to transduce DM460 (thiH910::MudJ) to tetracycline resistance (Tcr). The Tcr transductants were screened for those that were corrected by TPP but not thiamin. In the case of the MudJ elements, kanamycin-resistant transductants were generated in DM2275 (thiH942::Tn10d(Tc)) by independent transposition events allowed by cis-complementation, as has been described (17). In each case, putative mutants were verified by transducing the respective insertion into a wild-type genetic background (LT2) and confirming the nutritional requirement for TPP. TPP was present in all media except that used for screening the thiamin requirement.
Molecular Biological Techniques
Recombinant DNA TechniquesPlasmid DNA was isolated by either alkaline lysis or a QIAprep Spin Plasmid kit (Qiagen, Chatworth, CA), and was electroporated into different backgrounds using the E. coli Pulser (Bio-Rad). Standard methods were used for DNA restriction and ligation. Restriction endonucleases and ligase were purchased from Promega (Madison, WI).
Cloning of thiLDNA from the 10 Cs region of the S. typhimurium chromosome was isolated by inducing a
proC693::MudP located at 8.4 Cs and isolating the
phage DNA, as described elsewhere (18, 19). The isolated DNA was
digested with EcoRI, ligated into similarly cut pSU19 (20),
and electroporated into a thiL mutant strain (DM1683). The
resulting chloramphenicol-resistant electroporants were selected in the
presence of TPP and then assessed for growth on minimal
chloramphenicol, which would require complementation of the mutant
defect. From this screen, a plasmid containing an ~7.2-kb insert was
isolated and designated pThiL7.2. Subsequent subcloning steps, shown in
Fig. 2, using SalI and Sau3A resulted in a
complementing clone, pThiL1.8, that contained an ~1.8-kb insert (350 bp of which is noncontigous due to the Sau3A digest dropping
out a 1.2-kb fragment).
DNA Sequencing
Complete DNA sequence of pThiL1.8 was generated in part by using the dideoxy method (21) with the Sequenase 2.0 kit (U. S. Biochemical Corp.) and by nonradioactive sequencing at the University of Wisconsin Biotechnology Center-Nucleic Acid and Protein Facility, and the sequence presented is the result of both strands being sequenced more than three times in their entirety. DNA sequence was analyzed using the Genetics Computer Group (GCG) (Madison, WI) DNA sequence analysis program, BLAST (22), and SeqEd (Applied Biosystems Inc., Foster City, CA).
The thiL927 mutation was sequenced as follows. thiL was amplified from the chromosome of a strain containing the thiL927 mutation (strain DM946) using primers ThiL-48-2 and ThiL-40-2 under conditions described below for PCR amplification. This amplification produced a 1.2-kb fragment containing thiL that could be visualized on a 1% agarose gel. This fragment was purified using Qiaquick Gel Extraction Kit (Qiagen) and was entirely sequenced on both strands at least 3 times.
Mapping Insertions by PCRInsertions were mapped by
modifying a PCR mapping protocol, described elsewhere (23).
Amplification between two primers was performed using Vent
(exo) polymerase (New England Biolabs, Inc., Beverly, MA)
in a Thermolyne Temp-Tronic Thermocycler. Reaction conditions were as
follows: 95 °C denaturation for 1 min, 55 °C annealing for 1 min,
and 72 °C extension for 1 min. Primers used were; Tn10-I
(5
-GACAAGATGTGTATCCACCTTAAC-3
), which hybridizes to the 66-bp
inverted repeat Tn10 sequence; MuL (5
-ATCCCGAATAATCCAATGTCC-3
), which hybridizes to the left end of the
MudJ insertion; MuR (5
-GAAACGCTTTCGCGTTTTTCGTGC-3
), which hybridizes
to the right side of the MudJ; and ThiL-2-40
(5
-GAGCAGGTGCCACGGATT-3
) and ThiL-2-48
(5
-GTGATTCGTCCAACAAAAGTG-3
), which annealed downstream and upstream
of thiL, respectively, and whose relative locations are
indicated in Fig. 2. Size of amplified fragments was determined by
agarose gel electrophoresis with standard size markers.
-Galactosidase Assays
-Galactosidase assays were performed using the Miller method
(24) as described previously (25).
Overexpression and Visualization of ThiL
The complete insert from pThiL1.8 was ligated into T7
overexpression vectors pT7-5 and pT7-6 (26) using an
EcoRI-PstI double digest (sites shown in Fig. 2).
The resulting plasmids, pThiL-5 and pThiL-6, were electroporated into
E. coli strain BL21/DE3, generating strains DM2571 and
DM2572, respectively. These strains contain the T7 RNA polymerase in a
-lysogen, under the control of an IPTG-inducible promoter. The
induction protocol was as follows. Strains were grown shaking at
37 °C to 60 Klett units, IPTG was added to 400 µM, and
incubation was continued at 37 °C for ~3.5 h. After this time,
cells were pelleted, resuspended in 2 ml of 50 mM Tris-HCl
(pH ~7.5), and sonicated for 40 s (2 × 20) at 50% duty
cycle using a model 550 sonic dismembrator (Fisher, Itasca, IL).
Membrane and soluble fractions were separated by centrifugation at
40,000 × g for 1 h. Proteins from both
supernatant and pellet fraction were separated by 12.5% SDS-PAGE and
visualized by Coomassie Blue staining.
Thiamin-monophosphate Kinase Assay
Thiamin-monophosphate kinase was assayed by a modification of a previously described protocol (6). The assay mixture contained 1 mM TMP, 5 mM ATP, 50 mM Tris-HCl (pH ~7.5), 5 mM MgCl2, and 0.35 M KCl. The reaction was started by the addition of ~0.16 mg of protein to a final volume of 200 µl in 1.5-ml Eppendorf tubes. The reaction mixture was incubated for 30 min at 37 °C. To precipitate the protein after the incubation, 10 µl of 1 M HCl was added, and the reaction mixture was incubated at 100 °C for 15 min. Addition of 10 µl of 1 M KOH neutralized the sample, and centrifugation removed the precipitated proteins. Thiamin and its phosphoesters in the supernatant were derivatized to the corresponding thiochromes by mixing 150 µl of the assay supernatant with 250 µl of 2.65% KFe(CN)6 in 0.01 N NaOH. 1 µl of this mixture was analyzed via HPLC as described below.
The thiochromes in the derivatized assay supernatant were separated utilizing normal phase HPLC with a Lichrosorb-NH2 10 µm column (Alltech, Deerfield, IL), as has been described (3, 27). Conditions for the separation employed an acetonitrile, 90 mM potassium phosphate (pH 8.4) mix at a ratio of 60:40 using a flow rate of 1.4 ml/min. Under these conditions, thiochrome standards of thiamin, TMP, and TPP eluted with retention times of 2.5, 5.3, and 7.1 min, respectively. Elution of thiochromes was monitored with a Waters 990 spectrofluorimeter detector set at 375 nm (excitation) and 432 nm (emission). The area under the eluted peak was determined with Waters Millenium 2,000 software and quantitated utilizing a standard curve generated with known TPP concentrations.
Quantitation of TMP and TPP
TMP and TPP concentrations in strains DM3286 (zaj-8048:: Tn10d(Tc)) and DM3287 (zaj-8048::Tn10d(Tc)thiL927) were determined via a modification of CNBr thiochrome derivatization, (27).2 Briefly the protocol is as follows. Cells were grown in minimal medium to 100 Klett units and pelleted, resuspended in 1 ml of dH20, and divided into two aliquots. One aliquot was used to determine dry weight, and the other was adjusted to pH 2.0 with 1 M HCl and boiled for 20 min. The sample was then centrifuged at 30,000 × g for 20 min to remove cellular debris. The supernatant was used for quantitation via derivatization to TMP/TPP thiochromes with 0.3 M CNBr. The thiochromes formed were analyzed with the same HPLC protocol described above.
Six independent insertions (2 Tn10d (Tc) and 4 MudJ), which caused a TPP auxotrophy, were identified by screening 10,000 independent insertion mutants. Transductional analysis found that all six insertions were ~50% co-transducable by P22 with an insertion in thiI, a locus known to be located at 9.6 Cs (28).
Cloning of thiLTo confirm that the above insertions were in a single gene that was thiL, a complementing clone was isolated using MudP/Q technology (see "Experimental Procedures"). Plasmid pThiL, contained a 7.2-kb fragment and complemented DM1683 (thiL933::Tn10d(Tc)). This plasmid was sequentially subcloned to more precisely localize thiL (Fig. 2). Digestion of pThiL with SalI resulted in a 3-kb fragment that was isolated and ligated into pSU19 to produce pThiL3.0. This smaller plasmid retained the ability to complement DM1683. Plasmid pThiL3.0 was subjected to a partial digestion with Sau3A. The digestion mixture was ligated and electroporated into DM1683 selecting chloramphenicol-resistant and assessing complementation. This screen identified a plasmid, pThiL1.8, that contained an 1.8-kb insert fragment that was released from pSU19 by a SalI digestion. Because this plasmid complemented DM1683, it contained the thiL gene, but as shown in Fig. 2, included some non-contiguous DNA resulting from the drop-out of a 1.2kb-Sau3A fragment. Further analysis of this plasmid focused only on the contiguous sequences.
Identification of thiL ORFThe insert of pThiL1.8 was sequenced completely, and a 1173-bp contiguous sequence between primers ThiL-2-40 and ThiL-2-48 has been submitted to GenBankTM (accession number [GenBank]). BLASTN and BLASTX (22) data base searches were used to identified homologies to the 1.2-kb thiL fragment. BLASTN searches determined that the ends of the insert were homologous to E. coli genes, pgpA and nusB, respectively (shown in Fig. 2). Both of these loci mapped to 9.5 Cs on the E. coli chromosome, and were separated by approximately 1 kb of unsequenced DNA. The only sequence homologous to the entire 1-kb internal fragment was from Hemaphilus influenzae and corresponded to a region of unknown function (HIU32756). The sequence comparison was consistent with the P22 mapping results and definitively placed thiL as a new gene at 9.5 Cs in S. typhimurium.
Further analysis of the internal sequence of pThiL1.8 identified two
putative ORFs initiating at nucleotide 1 or 388 and sharing a
termination codon at nucleotide 978 (Fig. 3). Results
described below eliminated nucleotide 388 as the start of
thiL and supported designation of the 978-bp ORF as
thiL encoding a 35-kDa protein (Fig. 3).
Recently, a motif found near the C terminus of many phosphate-binding enzymes has been identified as a putative phosphate-binding domain (29). Consistent with the expectation that ThiL binds both ATP and TMP, this putative phosphate-binding motif was found in the C-terminal region of the thiL ORF and is underlined in Fig. 3.
Mapping Insertions to thiLTo confirm that the insertion mutations that caused a TPP auxotrophy were in the 978-bp thiL ORF, all six insertions were mapped by a PCR-based protocol. Strains containing each of the insertions were subjected to PCR amplification using one primer specific to the insertion and the other upstream of thiL (Fig. 2). Each amplification resulted in a product <700 bp in length, indicating that all of the insertions were in the 978-bp ORF.
Overexpression of ThiLTo confirm our assignment of the
978-bp ORF to the thiL gene, we cloned the entire insert
from pThiL1.8 into the T7 overexpression vectors, pT7-5
and pT7-6, generating pThiL-5 and pThiL-6, respectively. Plasmid
pThiL-6 contained the insert in the orientation expected to express the
978-bp ORF from the T7 promoter. These plasmids were moved into
E. coli strain BL21/DE3, which generated strains DM2571
(pThiL-5) and DM2572 (pThiL-6) that were then subjected to a protocol
inducing T7-specific expression. Following the induction, proteins in
the crude extract were resolved by SDS-PAGE and visualized by Coomassie
Blue staining. As shown in Fig. 4, extracts from the
pThiL-6 containing strain (lane B) contained an extra
protein band of approximately 35 kDa compared with the strain
containing the oppositely oriented insert (lane A). The
molecular mass of the overexpressed protein correlated well with the
predicted molecular mass of 35 kDa for ThiL.
ThiL Has Thiamin-monophosphate Kinase Activity
To show that
thiL encoded thiamin-monophosphate kinase, activity assays
were performed on crude extracts of DM2571 and DM2572 after induction
of T7 polymerase. The assays were performed by providing TMP and ATP as
substrates and measuring TPP accumulation after 30 min of incubation
(see "Experimental Procedures"). Crude extract from strain DM2572
had a specific activity of 9.01 ± 0.90 nmol TPP formed/mg of
protein/min, whereas strain DM2571 had 1.93 ± 0.80 nmol TPP
formed/mg of protein/min, when data from three independently performed
experiments were analyzed. Representative HPLC tracings detecting the
products of this assay are shown in Fig. 5. A small peak
with the retention time of TPP is visible in the control tracing in
Fig. 5. Control experiments determined that this peak reflected a
contaminant in the TMP purchased from Sigma.
thiR927 Is an Allele of thiL
Results presented above confirmed that thiL encoded thiamin-monophosphate kinase, and we sought to determine if the point mutation previously designated thiR927 was an allele of thiL, as we had proposed. The thiL gene from DM946 (thiL927) was amplified from the chromosome and sequenced. A comparison of the resulting sequence to wild-type thiL using SeqEd identified thiL927 as a missense mutation in thiL at residue 132, causing a glycine (GGT) to an aspartate (GAT) change in the predicted amino acid sequence. The affected codon is underlined in Fig. 3. The A to G transition was consistent with the use of hydroxylamine, a mutagen, to generate this mutation (7).
TMP and TPP Quantitation in thiL927A simple explanation for
the regulatory phenotype caused by thiL927 would be that it
resulted in an enzyme unable to maintain optimal TPP pools. To test
this prediction, an isogenic pair of strains, DM3286
(zaj-8048::Tn10d(Tc)) and DM3287
(zaj-8048::Tn10d(Tc)thiL927) were
constructed. The intracellular TMP and TPP pool sizes were determined
in each strain after growth in minimal medium. These results showed
that while the TPP pools were similar in the two strains, significantly
increased TMP pools were accumulated in the thiL927 mutant
strain. Strain DM3287 had 45.23 ± 7.2 pmol of TMP/mg dry weight
and 25.45 ± 3.2 pmol of TPP/mg dry weight, while DM3286 had
5.05 ± 0.34 pmol of TMP/mg dry weight and 20.85 ± 2.3 pmol
of TPP/mg dry weight. These data reflect TMP/TPP ratios of 1.77 ± 0.36 and 0.242 ± 0.31 for DM3287 and DM3286, respectively, and
are shown in Fig. 6 as representative HPLC tracings.
Transcriptional Regulation
The transcription of
thiL was investigated since some genes involved in thiamin
biosynthesis are transcriptionally regulated in S. typhimurium (7).3 Regulation of
thiL was tested in a strain containing insertion thiL953::MudJ that had been shown by PCR
amplification to be in the correct orientation for transcription from
the thiL promoter. To account for the possibility that ThiL
may be involved in its own expression, transcriptional analysis was
performed in a strain that also contained pThiL1.8. Strain DM3154
(thiL953::MudJ/pThiL1.8) was grown in minimal
medium in the presence and absence of exogenous thiamin, TMP, or TPP
(100 µM), and -galactosidase activity was measured. In
all media, a basal level of ~50 Miller units was obtained, indicating
that thiL was in a class of thiamin biosynthetic genes not
transcriptionally regulated by TPP.
Results presented herein showed that thiamin-monophosphate kinase in S. typhimurium is encoded by thiL, a 978-bp ORF that produces a 35-kDa protein. The chromosomal location of the thiL gene in S. typhimurium was determined genetically and physically to be 9.5 Cs.
This work also showed that altered ThiL activity can cause a regulatory phenotype. The increased TMP/TPP ratio found in the thiL927 containing strain was consistent with the mutant enzyme being unable to increase TPP pools to repressing levels, even when supplied with exogenous thiamin. Such a scenario explains the lack of thi gene repression found in strains carrying this lesion when grown in excess thiamin. The similar TPP pools maintained by the mutant and wild-type strains after growth on minimal medium is consistent both with the lack of thi derepression in the mutant strain carrying allele thiL927 and the lack of an obvious growth defect. Based on the measured pool sizes and phenotypic considerations, the thiL927 mutation appears to result in a protein with either an altered equilibrium or subject to increased allosteric inhibition by TPP.
Interestingly, thiL was not transcriptionally regulated by TPP, making it the second of four characterized thi loci that appear to be constitutively transcribed. Identification of previously isolated mutations that cause regulatory effects as alleles of thiL predicts that there is a regulatory protein involved in the transcription of thiamin genes. Such a protein remains to be identified.
At this time, thiamin biosynthesis is not well understood in any organism. Although in both yeast and enteric bacteria TMP is the product of de novo synthesis, the conversion of TMP to TPP differs in these organisms. Unlike E. coli and S. typhimurium, the yeast fail to directly phosphorylate TMP to form TPP, but rather dephosphorylate it to form thiamin prior to adding a pyrophosphate group to generate TPP. This reaction in S. cerevisiae is catalyzed by thiamin pyrophosphokinase and is encoded by the TPP repressible gene thi80 (3). In contrast, we have shown that in S. typhimurium the terminal step in TPP synthesis, encoded by thiL, is not transcriptionally regulated by thiamin or its phosphoesters. These data emphasize two important differences between the terminal reactions in TPP biosynthesis in S. typhimurium and S. cerevisiae. Not only do the two organisms utilize different substrates (TMP and thiamin, respectively), they regulate the step in distinct ways. The physiological relevance of these differences remains to be clarified with additional work on thiamin biosynthesis in both organisms.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U74758[GenBank].