Two Genetically Distinct Pathways for Transcriptional Regulation of Anaerobic Gene Expression in Salmonella typhimurium Expression of the tripeptide permease gene tppB is anaerobically induced .
This induction is independent of the fnr ( oxrA ) gene product , which is known to be required for the anaerobic induction of several respiratory enzymes .
We isolated , characterized , and mapped mutations in two genes , oxrC and tppR , which prevent the anaerobic induction of tppB expression .
Mutations in oxrC were highly pleiotropic , preventing the anaerobic expression of the formate dehydrogenase component of formate hydrogen lyase ( fhl ) , a tripeptidase ( pepT ) , and two of the three known hydrogenase isoenzymes ( hydrogenases 1 and 3 ) .
On the other hand , expression of nitrate reductase , fumarate reductase , and a number of other fnr ( oxrA ) - dependent enzymes was not affected by mutations in oxrC .
Thus , there appeared to be at least two distinct classes of anaerobically induced genes , those which required fnr for their expression and those which required oxrC .
It seems that fnr-dependent enzymes perform primarily respiratory functions , whereas oxrC-dependent enzymes served fermentative or biosynthetic roles .
We found the primary defect of oxrC mutants to be a deficiency in phosphoglucose isomerase activity , implying that a product of glycolysis functions as an anaerobic regulatory signal .
Mutations in tppR were specific for tppB and did not affect expression of other oxrC-dependent genes .
However , tppR did exhibit phenotypes other than the regulation of tppB .
Both oxrC and tppR mutants were hypersensitive to the toxic NAD analog 6-aminonicotinic acid .
This suggests that oxrC and tppR may play a role in the regulation of NAD biosynthesis or , alternatively , that NAD or a related nucleotide serves as the anaerobic signal for oxrC-dependent enzymes .
The enteric bacteria Escherichia coli and Salmonella typhimurium are facultative anaerobes .
When these bacteria are grown under anaerobic conditions , the synthesis of many proteins involved in aerobic respiration is repressed , whereas the synthesis of a specific class of approximately 50 proteins , including a number of respiratory enzymes , is specifically induced ( 5 , 33 ) .
The anaerobic induction of many genes encoding respiratory enzymes has been shown to be at the level of transcription ( 1 , 16 , 23 , 41 , 42 ) .
In addition , three genes whose products are related in function but are not directly involved in energy generation , the tripeptide permease gene tppB and two peptidase genes , pepT and pepN , are also anaerobically induced ( 8 , 14 , 36 ) .
The molecular mechanisms by which gene expression is regulated in response to anaerobiosis are poorly understood .
In E. coli , mutations in thefnr gene ( also variously called nirA or nirR ) prevent the anaerobic induction of several respiratory enzymes , including nitrate reductase , nitrite reductase , fumar-ate reductase , dimethyl sulfoxide reductase , and glycerol-3-phosphate dehydrogenase ( 2 , 16 , 17 , 21 , 34 ) .
The effect offnr mutations on these genes is at the transcriptional level .
The nucleotide sequence offnr shows that the Fnr protein shares considerable homology with the catabolite activator protein CAP .
This implies that the Fnr protein may be a DNA-binding protein and suggests the involvement of a nucleotide such as cyclic AMP in the regulation of anaerobic gene expression ( 31 ) .
In S. typhimurium , two genes designated oxrA and oxrB have been shown to be essential for the anaerobic induction of several respiratory enzymes ( 36 ) .
oxrA is identical the fnr of E. coli ( 14 , 36 ) .
The oxrB to gene characterized , although it is required gene is , as yet , poorly of subset of for the anaerobic expression the same genes as is oxrA ( fnr ) .
However , not all anaerobically induced genes are subject to fnr control .
We recently showed that transcription of tppB , which encodes the tripeptide permease , is specifically induced by anaerobiosis and that this induction is fnr independent ( 14 ) .
Similarly , the anaerobic induction of a tertiary amine oxidase ( torA ) , peptidase N ( pepN ) , and certain aspects of formate hydrogen lyase activity have been reported to be independent offnr ( 8 , 23 , 25 , 28 ) .
To investigate thisfnr-independent pathway , we isolated and characterized mutations which prevent the anaerobic induction of tppB .
Two distinct and unlinked regulatory genes were defined , designated oxrC and tppR .
The oxrC gene was found to play an important regulatory role in anaerobic gene expression .
Mutations in the oxrC gene were highly pleiotropic and affected the anaerobic synthesis of a number of enzymes whose expression is fnr independent .
Thus , oxrC and fnr mutations appeared to define two distinct pathways for the anaerobic induction of gene expression .
All strains used in this study tives of S. typhimurium LT2 unless otherwise indicated .
The genotypes and constructions of these strains are described in Table 1 .
were in LB Media and growth-conditions .
Cells grown on LB at with aeration , unless or 37 °C medium ( 20 ) agar stated .
strains were grown at 30 °C otherwise Mu-containing to prevent bacteriophage induction .
Anaerobic-growth was cells in completely filled and sealed achieved by growing vessels or by using Gas-Paks ( Oxoid Ltd. ) .
To ensure full aeration of aerobically grown cells , growth was in vigorously shaking conical flasks containing less than 1/20 the flask volume of medium .
LC medium is LB to which 2 mM CaC12 , 0.1 % glucose , and 0.001 % thymidine are added .
Minimal medium was based on the E medium of Vogel and Bonner ( 26 ) and was supplemented with 0.4 % glucose , fructose , or glycerol as the carbon source , as indicated .
Nutrient broth ( NB ) was obtained from Difco Laboratories .
MacConkey agar plates were prepared from MacConkey agar base ( Difco ) to which the appropriate sugar was added at 1 % .
MacConkey agar-nitrate medium is described by Stewart and MacGregor ( 35 ) , and glycerol-nitrate medium is described by Lambden and Guest ( 17 ) .
When necessary , minimal-medium was supplemented as follows : amino-acids , 0.4 mM ; ampicillin , 50 jig ml - ' or 25 pug ml - ' in rich and minimal media , respectively ; tetracycline , 20 , ug ml - ' and 10 , ug ml - ' in rich and minimal media , respectively ; kanamycin , 25 , ug ml - ' ; chloramphenicol , 25 , ug ml - ' ; streptomycin , 150 , ug ml - ' ; and 5-bromo-4-chloro-3-indolyl-f-D-galactoside ( X-gal ) , 20 , ug ml - ' .
When exogenous electron-acceptors were added , they were used at the following concentrations : sodium formate , 0.5 g liter - ; KNO3 , 10 g liter - ' ; and sodium fumarate , 5 g liter-1 Genetic techniques .
Transductions were performed by using a high-transducing derivative of phage P22 int4 as described by Roth ( 26 ) .
Because of its greater transducing capacity , the phage derivative P1 Tn9 clr-100 was occasionally used ( 20 , 32 ) .
As S. typhimurium is normally resistant to P1 infection , galE derivatives were used for P1 transduction .
Donors and recipients for P1 transduction were grown in LC medium .
Conjugations for HFr matings or F ' transfer were performed as described previously ( 9 , 20 ) .
Random chromosomal insertions of TnS were obtained by transduction of the appropriate recipient to Kanr with a P22 lysate of strain TT3416 as the donor , as described previously ( 4 ) .
Mu dl-8 ( Ampr lac ) insertions were obtained by transduction of strain TT7610 ( supD ) to Ampr by using TT7674 as the donor ( 13 ) .
Random chromosomal insertions of the mini-TnJOA16AJ 7 ( Cmlr ) element were obtained by transduction of strain TT10427 ( which carries the transposase helper plasmid pNK972 [ 39 ] ) to chloramphenicol resistance with a P22 lysate of strain TT10605 ( 39 ) .
After transductions involving either a TnS or Mu phage insertion , the correct location of the insertion and the presence of just a single copy of the transposon in the transductant were ascertained by marker rescue .
Mutants resistant to the toxic peptide alafosfalin were selected by plating washed cells on a minimal glucose plate containing 80 , ug of alafosfalin ml - ' ( 9 ) .
Screening for sensitivity or resistance to alafosfalin was by radial streaking on an MG plate around a filter disk containing 250 , ug of the antibiotic ( 9 ) .
Sensitivity to the toxic NAD analog 6-aminonicotinic acid ( 6-AMN ) was similarly determined by using 10 , ug of the analog per disk .
Wild-type strains gave a zone of killing with a diameter of about 10 mm , whereas hypersensitive strains showed a 30-mm zone of killing .
, B-Galactosidase activity was determined as described by Miller , by using the sodium dodecyl sulfatechloroform permeabilization method ( 20 ) .
Phosphoglucose isomerase ( PGI ) was assayed as described by Fraenkel and Horecker ( 7 ) .
Nitrate reductase activity was detected by the overlay technique described by Sawers et al. ( 29 ) .
RESULTS Isolation of Mu dl-8 ( Ampr lac ) operon fusions to tppB .
To isolate mutants defective in the anaerobic induction of tppB , we took advantage of the relatively simple phenotypic plate screens for reduced expression of P-galactosidase from tppB-lacZ-fusions .
We previously isolated operon fusions between tppB and lacZ by using the bacteriophage derivative Mu dl ( Ampr lac ) ( 14 ) .
However , Mu dl-mediated lacZ-fusions are relatively unstable , the phage transposing to other sites on the chromosome at a significant frequency .
We therefore constructed tppB-lacZ-fusions by using the recently described phage Mu dl-8 ( Ampr lac ) ( 13 ) .
This phage contains an amber mutation in the transposase gene and is consequently stable in strains that do not harbor an appropriate suppressor .
A collection of 10,000 random Mu dl-8 insertions into the chromosome of strain TT7610 ( supD ) was made as described in Materials and Methods .
From this collection , Mu dl-8 insertions in tppB were selected by their resistance to alafosfalin , and the resulting fusions were mapped and characterized as described previously ( 9 , 14 ) .
One tppB : : Mu dl-8 fusion was stabilized by transduction into a wild-type ( supD + ) strain , and this derivative ( CH776 ) was used for all further experiments .
CH776 was shown to harbor just a single Mu dl-8 insertion , and this insertion was shown by marker rescue to be responsible for the TppB phenotypes .
Regulation of 0-galactosidase expression from this fusion was similar to that found previously for tppB : : Mu dl ( Ampr lac ) fusions ( 14 ) ( Table 2 ) .
Isolation of anaerobic regulatory mutations .
Mutants defective in the anaerobic induction of tppB expression were identified by color changes on MacConkey agar-lactose plates .
Preliminary experiments showed that color changes on anaerobically incubated MacConkey agar-lactose plates were unreliable due to a general increase in acid production under such conditions .
However , when incubated aerobically , colonies of tppB-lacZ fusion strains gave a characteristic fish-eye appearance ; , B-galactosidase was expressed only in the center of a colony which had become anaerobic , whereas the aerobic perimeter of the colony remained white .
To ensure complete inactivation of any regulatory gene and to facilitate its characterization , the transposon TnS was used as a mutagen .
A random collection of 12,000 independent Tn5 insertions in strain CH776 ( tppB84 : : Mu dl-8 ) was pooled , washed twice in minimal-medium , and plated on MacConkey agar-lactose plates at a density of about 500 cells per plate .
Any colonies which were white or less red than CH776 , i.e. , having lost the fish-eye appearance , were picked , purified , and characterized further .
Those which despite giving an altered colony color on MacConkey agarlactose plates showed unaltered levels of P-galactosidase activity were discarded as mutations causing a general defect in acid production .
Derivatives in which the TnS insertion was essentially 100 % linked by cotransduction to the Mu dl-8 insertion were assumed to have insertions in the lacZ or lac Y genes of the Mu derivative and were discarded .
It was also anticipated that the screening procedure used would identify TnS insertions in the tppA ( ompR ) gene .
tppA is a positive regulator of tppB expression and has recently been shown to be identical with the ompR gene ( 9 ; M. M. Gibson and C. F. Higgins , submitted for publication ) .
tppA ( ompR is not involved in the anaerobic regulation of tppB ( 14 ) .
To identify and eliminate tppA ( ompR ) mutations , the transductional linkage between each regulatory TnS and an ompR : : Mu dl insertion ( strain CH656 ) was determined .
All strains in which the TnS was closely linked to ompR were presumed to be insertions in ompR and were not studied further .
Three strains containing putative regulatory TnS insertions remained , CH804 , CH805 , and CH878 .
The tppB : : Mu dl-8 fusion from these strains was transduced into LT2 , and regulation of P-galactosidase was shown to be normal ( i , e. , identical to that of the parental strain CH776 ) .
Thus , the Mu dl-8 had not mutated or transposed .
Similarly , the TnS insertions were transduced into unmutagenized CH776 to confirm that the strains contained only a single TnS insertion and that reduction in P-galactosidase expression from the tppB : : Mu dl-8 fusion was due to the TnS insertion and not to an incidental point mutation .
P-Galactosidase assays of these strains showed that the three regulatory mutations fell into two classes , designated oxrC and tppR .
Genetic mapping and further phenotypic characterization showed that the TnS insertions in strains CH804 and CH805 were indistinguishable .
Thus , all further characterization of the oxrC locus was performed by using CH805 ( oxrCJ02 : : Tn5 ) .
Mutations in oxrC ( oxygen regulation ) prevented the anaerobic induction of tppB expression but had no effect on induction by leucine ( Table 2 ) .
We previously presented evidence that the anaerobic and leucine-dependent inductions of tppB expression are mediated independently ( 14 ) .
The isolation of regulatory mutants which prevent only the anaerobic induction substantiate this view and , in addition , show that anaerobic induction is not simply a consequence of increased intracellular leucine pools .
Mutations in tppR ( the regulatory locus for tppB ; as shown below , tppR mutations did not have a pleiotropic affect on anaerobic enzymes ) were found to confer partial auxotrophy ( see below ) .
Thus , tppR mutants were unable to grow in minimal-medium , and all assays had to be performed on cells grown in LB ( which contains leucine ) .
Cells grown aerobically in LB showed a basal level of tppB expression due to induction by leucine ( Table 2 ) .
HIowever , this expression was not increased by growing the cells anaerobically , showing that , like oxrC mutations , mutations in tppR prevent the anaerobic induction of tppB expression .
oxrC and tppR are not alleles of fnr .
The only gene so far which is known to regulate anaerobic gene expression is fnr ( oxrA ) ( also called nirA and nirR [ 17 , 21 , 31 ] ) .
We previously obtained evidence that anaerobic expression of tppB is independent of fnr ( 14 ) .
Although the phenotypes of oxrC and tppR mnutations were very different from those of fnr mutations ( see below ) , it remained a possibility that different alleles of fnr exhibit different phenotypes .
To demonstrate that oxrC is not an allele Qffnr , plasmid pCH21 ( containing the fnr gene [ 14 ] ) was introduced into strains CH804 and CH805 and 3-galactosidase activity was assayed anaerobically .
The cloned fnr gene did not complement oxrC ( Table 2 ) .
In addition , we also showed that oxrC and tppR map to very different chromosomal locations , both from each other and from fnr ( see below ) .
oxrC is a pleiotropic regulatory gene .
A number of respiratory enzymes are known to be induced by anaerobiosis .
Although some of these are fnr dependent , others are unaffected by fnr mutations .
To determine whether oxrC mutations are specific to tppB or whether oxrC defines a pleiotropic anaerobic regulatory locus , the effect of oxrC mutations on the expression of a variety of anaerobically induced genes was examined .
Operon fusions between lacZ and three anaerobically induced genes , fl , hyd , and pepT , were recently described in S. typhimurium ( 1 , 36 ) .
An oxrC : : TnS mutation was introduced into these fusion strains by transduction to Kanr .
The TnS was shown by marker rescue to have remained in oxrC and not to have transposed .
Mutations infnr ( oxrA ) were also transduced into thefhl and pepT fusion strains , taking advantage of the TnS insertion in CH602 which is-70 % linked to fnr ( oxrA ) .
Derivatives were checked for coinheritance of fnr and the Tn5 insertion by screening for the formation of red colonies on anaerobic MacConkey agar-nitrate plates and by the failure of the derivatives to grow on anaerobic glycerol-nitrate plates .
The J7fl locus to which the fusion was made probably encodes the formate dehydrogenase component of formate hydrogen lyase ( FDH-BV ) ( 1 , 15 , 24 ) .
fhl expression is induced anaerobically , and this induction was further enhanced by exogenous formate ( Table 3 ) .
A mutation in oxrC strongly reduced the anaerobic induction offrl but had little effect on induction by formate .
In addition , the effects of the oxrC mutation could not be suppressed by supplying exogenous formate .
Thus , it seems clear that the formate and anaerobic induction offrl expression were mediated by independent processes ; oxrC mutations affected only the anaerobic induction .
In contrast to the effects of oxrC mutations , fnr mutations only reduced anaerobic Al expres LT2 5.03 10.9 5.94 12.49 CH1021 ( oxrC ) 0.19 0.46 0.13 0.64 JF165 ( pasA ) ND ND 0.18 0.25 a PGI activity was determined as described in Materials and Methods .
Cells were grown aerobically or anaerobically in LB or NB medium .
sion about twofold and expression was fully restored to wild-type levels by the addition of formate .
Thus , the effect of fnr mutations on fhi expression appears to be indirect , probably the result of decreased formate production in fnr strains ; it has been suggested that pyruvate-formate lyase activity isfnr dependent ( 28 ) .
Like tppB , expression offhl is oxrC dependent and fnr independent .
The data from Jl1 operon fusions are fully substantiated by direct assay for enzyme activity ( 15 ) .
Thus , mutations in oxrC , but not those in fnr , specifically reduced FDH-BV activity .
The hyd locus to which the lacZ fusion was made has been mapped to 59 min on the chromosome .
The lesiop is pleiotropic , lacking all three hydrogenase isoenzyme activities ( 15 ) and , therefore , seems unlikely to be a hydrogenase structural gene .
Several hyd genes with pleiotropic phenotypes have been mapped to this 59-min locus , one of which can be phenotypically restored to Hyd + by growth in the presence of nickel ( 38 ) ; the hyd-lacZ fusion is not nickel suppressible .
oxrC mutations also prevented the anaerobic induction of expression of this hyd locus ( Table 3 ) .
pepT encodes an anaerobically inducible tripeptidase whose expression is known to depend on fnr function ( 36 ) .
oxrC was also required for pepT expression from pepT-lacZ operon fusions ( Table 3 ) .
Thus , unlike tppB , hyd , and fhl , pepT was both oxrC and fnr dependent .
Because lacZ-fusions to other anaerobically induced genes have not been isolated in S. typhimurium , the effects of oxrC and fnr on other respiratory enzymes had to be determined by direct enzyme assay ( 15 ) .
Neither nitrate reductase nor fumarate reductase activities , both of which are fnr dependent ( 17 , 21 ) , were affected by oxrC mutations ( 15 ) .
Similarly , activity of the respiration-linked hydrogenase ( hydrog-oxrCJ02 : : Tn5 EB137 Jlzl : : Mu dl CH951 jhl : : Mu dl oxrCJ02 : : Tn5 TN2021 pep T7 : : Mu dl CH940 pepl7 : : Mu dl oxrCJ02 : : TnS a Cells were grown anaerobically in the medium indicated , and 1Bgalactosidase was assayed as described in Materials and Methods .
b Units are as described by Miller ( 20 ) .
wtroncwoitntdohtions U of 13-galactosidaseb ffroom sttran ain : Leucine 02 Acetate CH776 CH805 - + -80 40 -- 442 60 + 316 191 - + -- + 527 276 + + -162 ND + -- 495 ND + + + 191 ND + - + 501 ND a Strain CH776 ( tppB84 : Mu dl-8 ) or CH805 ( tppB84 : : Mu dl-8 oxrCJ02 : : TnS ) was grown in minimal glucose medium with leucine , oxygen , and acetate ( 1 % ) .
b Units are as described by Miller ( 20 ) .
enase 2 ) was fnr dependent and oxrC independent .
On the other hand , the formate hydrogenlyase-associated hydrogenase ( hydrogenase 3 ) was oxrC dependent and fnr independent , whereas hydrogenase 1 , which seems to be associated with both hydrogen uptake and formate hydrogenlyase activities , was both oxrC and fnr dependent .
It seems that the effects of oxrC were specific to anaerobically induced genes ; no effect on the expression of lafZ fusions to a variety of oxygen-independent genes was foupd .
It therefore seems clear that oxrC mutations defined a pleiotropic regulatory locus required for the expression of several but not all anaerobically induced enzymes .
At least for some and probably for all these genes , regulation was at the level of transcription .
Effects of tppR mutations on other anaerobically induced genes .
tppR was very much more specific than was oxrC ( Table 3 ) .
Thus , tppR mutations had no effect on the expression of lacZ-fusions to fhl , hyd , or pepT and did not affect the activity of the following enzymes : nitrate reductase , fumarate reductase , FDH-BV , respiratory form-ate dehydrogenase , or any of the three hydrogenase isoenzymes ( 15 , 29 ) .
Effect of oxrC and tppR mutations on sugar fermentation .
During the characterization of oxrC mutants , it was noticed that strains harboring an oxrC mutation grew as white colonies on green plates .
These plates are essentially pH indicators , and this observation therefore implies a defect in fermentation ( 18 ) .
When streaked on MacConkey agarglucose plates , oxrC strains grew as pale pink colonies NAD pool levelsb fhil ( 93 min ; FDH-BV ) Hydrogenase 3 Nitrate reductase Hydrogenase 2 Fumarate reductase Respiratory formate dehydrogenase pepT ( tripeptidase ) -- + Hydrogenase 1 -- + a Data concerning the regulation of hydrogenase isoenzyme activity and the activities of a number of respiratory enzymes by oxrC , tppR , and fnr are described by Jamieson et al. ( 15 ) .
+ , Function is present in strains carrying the mutation ; - , function is reduced or absent in strains carrying the mutation .
When a gene is indicated , we demonstrated that the effect of oxr mutations is at the transcriptional level .
When only the enzyme is listed , we assayed activity and did not assay transcription directly ( see text ) .
b As indicated by 6-AMN hypersensitivity .
compared with the dark red color of oxrC + strains .
It therefore seemed likely that oxrC mutants were defective in either glucose transport or metabolism .
As the major route for glucose uptake is via the phosphotransferase system ( PTS ) , the activities of the PTS enzymes were assayed .
No significant differences in enzyme I activity or in-vitro phosphorylation were found between OxrC + and OxrC-strains ( P. W. Postma , personal communication ) .
In addition , oxrC mutants were found to ferment mannitol on MacConkey agar-mannitol plates .
As mannitol is transported only via the PTS , it seemed unlikely that the effects of oxrC on carbohydrate fermentation were at the level of transport ; it seems instead that there was a defect in the production of acetate , formate , or both from glucose .
To identify the defect in glycolysis , oxrC mutants were tested for their ability to ferment various sugars .
oxrC mutants were found to ferment fructose , galactose , and arabinose normally and to grow on glycerol as the sole carbon source , yet they were defective in both glucose and maltose fermentation .
This suggests a defect in PGI activity .
On assay , it was found that PGI activity increased twofold in response to anaerobiosis .
This effect , while not major , is in agreement with data presented previously ( 30 ) .
However , crude extracts of oxrC mutants were shown to have a 37-fold reduction in PGI activity , although activity was not totally abolished ( Table 4 ) .
This finding was somewhat surprising , as PGI is not thought to be a major regulatory enzyme in glycolysis .
In contrast to oxrC mutants , tppR mutants ferment glucose normally on MacConkey agar-glucose plates and are not defective in PGI activity ( data not shown ) .
tppR mutations result in auxotrophy .
Although tppR mutants fermented glucose normally , they were found to grow extremely slowly on minimal glucose medium , suggesting an auxotrophic requirement .
tppR mutants grew normally in LB , and growth-on-minimal-medium could be restored by the addition of 0.25 % LB or 0.25 % Casamino Acids ( Difco ) .
The addition of aspartate or methionine , but not of any other amino-acid , also stimulated growth of oxrD mutants , although it did not completely restore growth to wild-type levels .
The precise nature of the requirement for multiple amino-acids remains unclear .
Suppression of the oxrC phenotype .
During characterization of the oxrC mutation , we noticed that the effects of oxrC on the anaerobic induction of tppB observed in minimal-medium were completely suppressed by growth in LB .
Thus , when grown in LB , oxrC derivatives ofpepT , tppB , or fhl operon fusions showed normal anaerobic induction of P-galactosidase expression ( Table 5 ) .
NB , on the other hand , did not suppress the oxrC lesion .
This implies that a component present in LB but not in NB is responsible for the phenotypic suppression .
However , although growth in LB suppressed the effect of oxrC mutants on tppB , fll , and pepT expression , the defect in PGI activity remained ( Table 5 ) .
This implies that either the oxrC effects on tppB and PGI are mediated by different routes or , alternatively , that oxrC causes a defect in PGI synthesis and , as a consequence of this defect , the expression of other genes is altered .
This latter view was shown to be correct .
Thus , fructose or any other sugar which entered the glycolytic pathway below PGI all suppressed the OxrC phenotypes except the loss of PGI activity ( Table 6 ; unpublished data ) .
Presumably , fructose or other sugars present in LB but not in NB also mediate suppression by this medium .
Further evidence that the primary defect in oxrC mutants is loss of PGI activity comes from the map location of this gene ( see below ) .
The observation that the primary defect in oxrC mutants is the absence of PGI activity implies that a product of glycolysis which is synthesized in altered amounts during anaerobic-growth plays a role in the anaerobic induction of oxrC-dependent enzymes .
Two possible candidates for such a signaling-molecule are acetate and formate .
Formate can not play this role , as exogenous formate did not suppress the effects of oxrC mutations on tppB or fhl expression ( Table 2 ) .
Acetate induced tppB expression aerobically and stimulated anaerobic induction in an oxrC strain ( Table 6 ) .
However , as tppB could be induced independently by either leucine or anaerobiosis , it seems possible that acetate simply mimics the effects of exogenous leucine and does not affect anaerobic expression directly .
The data in Table 6 support this view .
Thus , neither acetate nor formate appear to be the mediators of oxrC-dependent anaerobic gene expression .
Effect of nucleotide analogs .
During the mapping of oxrC and tppR mutations ( see below ) , both were found to confer hypersensitivity to the toxic nucleotide analog 6-AMN .
This implies that the mutations cause a defect in nucleotide biosynthesis .
6-AMN hypersensitivity is a result of reduced NAD pools ( 6 , 12 ) .
The hypersensitivity of oxrC mutants to 6-AMN was suppressed by fructose , showing it to be a direct consequence of the defect in PGI , whereas fructose had no effect on the hypersensitivity of tppR mutations .
As the nadA and nadB genes are anaerobically inducible ( 3 , 11 ) , it may be that the oxrC and tppR mutations interfere directly with the regulation of NAD synthesis .
Effect of electron-acceptors on tppB expression .
Expression of tppB is induced anaerobically , and this induction was prevented by the oxrC mutation .
The addition of potential electron-acceptors other than oxygen ( e.g. , nitrate and fumarate ) to anaerobic cultures reduced the anaerobic induction to some extent but by no means completely repressed oxrC function ( Table 2 ) .
Chromosomal locations of the oxrC and tppR genes .
The approximate location of the oxrC gene on the S. typhimurium chromosome was determined by introducing the mutation into various HFr strains .
These derivatives were used as conjugation donors with a series of auxotrophi strains as recipients , selecting for prototrophic transconjugants .
The prototrophic colonies were screened for kanamy-cin resistance to determine the percentage of coinheritance of oxrC with each auxotrophic marker ( data not shown ) .
These data showed oxrC to be located in the 89.5-to 96-min region of the S. typhimurium chromosome flanked by metA and purA .
To more precisely locate oxrC on the chromosome , the P22-mediated cotransductional linkage of oxrCJ02 : : TnS with markers in this region of the chromosome was determined ( Fig. 1 ) .
oxrC was found to be 3 % linked to a malB point mutation and 11 % linked to a TnJO insertion ( ib-1708 : : TnJO ) in this region of the chromosome .
The well-characterized zja-861 : : TnS insertion is located between metA and malB , 2 to 5 % linked to malB .
However , no linkage ( less than 1 % ) between zja-861 : : TnS and oxrC or zjb-1708 : : TnJO could be detected .
It therefore seems clear that oxrC must lie on the mel side of malB .
Because the size of the TnJO and TnS transposons is relatively large compared with the transducing capacity of P22 , the precise location of oxrC with respect to zjb-J 708 : : TnJO was determined by P1-mediated transduction .
Three-point crosses were performed with strain CH1027 ( mal + oxrCJ02 : : TnS ) as the recipient and strain CH1080 ( mal zjb-1708 : : Tn1O ) as the donor .
Transductions were carried out selecting for Tetr recombinants .
A total of 100 recombinants were subsequently screened for coinheritance of oxrC : : TnS ( Kan ) and the Mal-phenotype .
The recombinant phenotypes were as follows : Kanr Mal ' , 7 % ; Kanr Mal - , 0 % ; Kans Mal ' , 83 % ; and Kans Mal - , 10 % .
The three-point crosses indicated that oxrC is located between malB and zjb-1708 : : TnJO .
The presumed structural gene for PGI has been mapped approximately to this region of the chromosome ( 27 ) .
As oxrC mutants are deficient in PGI activity and because suppression by fructose indicates that the PGI defect is the primary cause of all oxrC phenotypes , it seems probable that the two mutations are at the same locus .
A mutation which confers 6-AMN hypersensitivity ( pasA [ 6 ] ) is also located in this region of the chromosome .
We showed that the pasA mutation is located between malB and zjb : : TnJO and is linked to these markers to about the same extent as is oxrC ( Fig. 1 ) , taking into account the differences in cotransduction frequencies , which result when mapping point ( pasA ) and insertion ( oxrC : : TnS ) mutations .
As oxrC is also hypersensitive to 6-AMN , it seems likely that the two mutations are in the same gene .
We therefore tested a pasA mutant and showed it to be defective in glucose but not fructose fermentation on MacConkey agar plates , indicating a defect in PGI activity .
Subsequent assays for PGI activity confirmed this defect ( Table 5 ) .
Thus , oxrC , pgi , and pasA are almost certainly alleles of the same locus , the primary defect being a deficiency in PGI .
It seems likely that this locus encodes the structural gene for PGI , although it remains a possibility that it encodes a positive regulator of pgi expression .
Because tppR mutations confer a general auxotrophy , mapping was facilitated by isolating a mini-TnJO insertion closely linked to the tppR : : TnS mutation .
This was achieved by transducing strain CH878 to Cmlr with a P22 lysate grown on a collection of random mini-TnJO ( CmV ) insertions in the S. typhimurium chromosome .
This collection of insertions was prepared as described in Materials and Methods .
The Cmlr derivatives were screened for those which had simultaneously become Kans. .
The cotransductional linkage between one such mini-TnJO insertion ( zae-1709 : : TnJOA16A17 ) and the tppR : : TnJO insertion was found to be greater than 95 % .
This mini-TnJO insertion did not confer the tppR phenotypes .
The approximate map location of the tppR-linked mini-TnJO insertion was determined to be between leu ( 2 map units ) and pro ( 7 map units ) by HFr mapping , as described for oxrC above .
P22 cotransduction showed both the tppR : : TnS and the tppR-linked mini-TnJO insertions to be 66 % linked to a panC : : TnJO insertion ( TT421 ) and 78 % linked to a panC point mutation ( SA2628 ) , both at 3 min on the S. typhimurium chromosome .
tppR is not , as far as we have been able to determine , an allele of any of the known genes in this region of the chromosome .
DISCUSSION The transcription of a number of bacterial genes is coordinately induced by anaerobiosis .
This global regulatory system has been the subject of much interest , yet little is known of the mechanisms by which the anaerobiosis is sensed or how the anaerobic switch is controlled .
The only gene known to play a key role in the control of anaerobic gene expression is fnr ( oxrA ) , a pleiotropic regulatory gene required for the anaerobic induction of a number of genes including nar , frd , glp , and pepT ( 17 , 21 , 36 ) .
However , not all anaerobically induced genes are fnr dependent .
We previously showed that the anaerobic induction of the tripeptide permease gene tppB is independent of the fnr gene product ( 14 ) .
In this study , we identified two genetic loci , oxrC and tppR , which were required for the anaerobic induction of tppB .
Both were highly pleiotropic but exhibited different phenotypes .
Our results imply the existence of at least two distinct classes of anaerobically induced genes .
These two classes of genes seemed to respond to different regulatory signals , and regulation was apparently mediated by entirely different mechanisms .
The first class of anaerobically induced genes were fnr dependent and oxrC independent .
In contrast , the second class of genes , which included tppB and Jhl ( the structural gene for FDH-BV ) , were fnr independent and were defined by their dependence on the normal function of the pleiotropic regulatory locus oxrC .
All the anaerobically induced genes we examined fell into these two classes , with two exceptions , pepT and hydrogenase 1 , whose expression required the function of both oxrC andfnr .
Those genes identified as belonging to each regulatory pathway are shown in Table 7 .
Interestingly , there seemed to be a functional distinction between the two classes of anaerobically induced genes ; fnr-dependent enzymes served primarily respiratory roles , whereas oxrC-dependent enzymes served fermentative or biosynthetic roles .
This observation is substantiated ( 15 ) in an accompanying paper in which the effects offnr and oxrC on respiratory and fermentative hydrogen metabolism are examined .
Particularly significant is the finding that hydrogenase 1 was linked to both respiratory and fermentative hydrogen metabolism , compatible with the dual regulation of this isoenzyme by both fnr and oxrC .
Surprisingly , oxrC mutants were found to be defective in sugar fermentation , and this was shown to be due to a defect in PGI activity .
We showed that the absence of this enzyme was the primary defect of oxrC mutants and that all other phenotypes , including the defects in anaerobic induction of gene expression , were a direct consequence of the loss of PGI activity .
Thus , none of the phenotypes of oxrC mutants , except the deficiency in PGI , were observed when cells were grown on fructose or on other sugars which enter the glycolytic pathway below PGI .
Evidence that oxrC maps to approximately the same chromosomal location as does pgi implies that the two genes are identical ( i.e. , that oxrC is the structural gene for PGI ) .
However , we can not rule out th possibility that oxrC encodes a positive regulator of pgi .
Indeed , this latter possibility is suggested by the observation that , despite the oxrC lesion being due to a TnS insertion , residual PGI activity is still detectable and this residual activity shows two-to threefold anaerobic induction .
Thus , if oxrC is the structural gene for PGI , there must be at least one additional isoenzyme .
In addition , the oxrC gene in S. typhimurium is located on the opposite side of malB than pgi is located in E. coli .
Either the pgi structural gene is located at somewhat different chromosomal positions in the two species or , alternatively , oxrC encodes a positive regulator of pgi .
In this context it is worth noting that S. typhimurium lacks an XylE function ( P. J. Henderson , personal communication ) ; xylE in E. coli has been mapped close to pgi .
Interestingly , a mutation with the properties of a pgi lesion that maps to a similar chromosomal location has been shown to prevent the anaerobic induction of threonine dehydratase in E. coli ( 19 ) .
A mutation which confers 6-AMN hypersensitivity ( pasA ) is also located in this region of the chromosome ( 6 ) .
As oxrC mutants were also supersensitive to this analog , we tested the pasA mutation and found it to be deficient in PGI activity and to be genetically inseparable from oxrC .
Thus , it seems that mutations at a single locus , whose primary defect was the loss of PGI activity and which was probably the structural gene for the enzyme , could give rise to each of these phenotypes .
How does loss of PGI activity prevent anaerobic enzyme induction ?
Because oxrC mutations had no effect on anaerobic metabolism when fructose was used as a carbon source , it seems clear that it is not the PGI protein itself which plays a regulatory role but that the anaerobic induction of oxrC-dependent enzymes must require a normal flow of carbon down the glycolytic pathway .
Presumably , the defect in glycolysis prevents the synthesis of a metabolic intermediate which functions as a regulatory signal .
The nature of this compound remains a matter for speculation .
Two possible candidates are acetate or formate , which are produced during fermentation and which accumulate during the switch from respiratory to fermentative energy generation .
However , we showed that neither of these compounds is the anaerobic signaling-molecule , although each does play a role in the regulation of specific genes ( e.g. , formate increases the anaerobic induction of hil , and acetate affects intracellular tppB expression indirectly , possibly by altering leucine pools ) .
An alternative possibility is that oxrC mutations lead to an imbalance in NAD biosynthesis and that an NAD-related nucleotide or , alternatively , the NAD/NADH ratio serves as an indicator of anaerobiosis .
This hypothesis is suggested by the observed hypersensitivity of oxrC mutants to 6-AMN , which implies a decreased intracellular pool of NAD ( 6 , 12 ) .
Under anaerobic conditions , the NAD/NADH ratios are known to alter ( 40 ) .
It is easy to envision a mechanism by which a defect in glycolysis ( the oxrC mutants ) could mimic this effect as dihydroxyacetone phosphate ( a glycolytic intermediate ) is required for NAD biosynthesis ( 12 ) .
An alternative explanation for the 6-AMN hypersensitivity , that oxrC and tppR prevent the anaerobic induction of the nadA and nadB genes , has been shown to be incorrect ( E. Ellis and C.F.H. , unpublished data ) .
This question requires further analysis .
The tppR mutation defined a second locus which affected the expression of tppB .
tppR was located at 3 min on the S. Unlike oxrC , this mutation did typhimurium chromosome .
not prevent the anaerobic induction of Jhl , pepT , or any other known anaerobically induced gene which we tested .
However , tppR was pleiotropic in that it caused a deficiency in NAD biosynthesis and resulted in a complex auxotrophic requirement .
The exact nature of the defect was unclear .
It is noteworthy that the NAD and amino-acid biosynthetic defects of tppR mutants were apparent during both aerobic and anaerobic-growth .
In addition , the poor growth ( auxo-trophic requirement ) of tppR mutants in minimal media could be suppressed by growth in rich media ( LB ) , whereas the anaerobic induction of tppB is not restored by growth in LB .
Thus , unlike oxrC , the failure of tppR mutants to induce tppB expression anaerobically did not seem to be a second-ary consequence of a defect in metabolism .
tppR specifically regulated tppB expression , rather than defining a component of a global anaerobic regulatory pathway .
We have shown here that the anaerobic induction of several genes encoding products involved in carbohydrate and amino-acid metabolism ( oxrC dependent ) was regulated by a more or less distinct mechanism from that controlling the induction of respiration-linked enzymes ( fnr dependent ) .
oxrC-dependent genes required the normal function of glycolysis , whereas the fnr-dependent enzymes are apparently unaffected by altered patterns of fermentation .
Clearly the mechanisms by which genes are regulated by anaerobiosis are varied and complex .
We defined two apparently independent pathways by which anaerobic gene expression is controlled .
To learn whether additional pathways exist and to identify the various components of each pathway will require further genetic analysis .
We are grateful to E. L. Barrett , D. H. Boxer , C. G. Miller , P. W. discussions and to E. L. Postma , and R. G. Sawers for helpful C. G. J. R. and K. E. Sanderson Barrett , D. Fraenkel , Miller , Roth , for providing bacterial strains .
We thank P. W. Postma for performing the PTS assays on our oxr mutants .
This work was supported by a Medical Research Council grant to C.F.H. and a Science and Engineering Research Council student-ship to D.J.J. C.F.H. is a Lister Institute Research Fellow .
Barrett , E. L. , H. S. Kwan , and J. Macy .
Anaerobiosis , formate , nitrate , and pyrA are involved in the regulation of formate hydrogenlyase in Salmonella typhimurium .
BDous , P. T. , and J. H. Weiner .
Dimethyl sulfoxide reductase activity by anaerobically grown Escherichia coli HB101 .
Bussey , L. B. , and J. L. Ingraham .
A regulatory gene ( use ) affecting the expression of pyrA and certain other pyrimidine genes .
Cairney , J. , C. F. Higgins , and I. R. Booth .
Proline uptake through the major transport system of Salmonella typhimurium is coupled to sodium ions .
The number of anaerobically regulated genes in Escherichia coli .
Foster , J. W. , E. A. Holley , and S. Mya .
NAD metabolism in Salmonella typhimurium : isolation of pyridine analogue supersensitive ( pas ) and pas suppressor mutants .
Fraenkel , D. G. , and B. L. Horecker .
Pathways of D-glucose metabolism in Salmonella typhimurium .
Gharbi , S. , A. Belaich , M. Murgier , and A. Lazdunski .
Multiple controls exerted on in-vivo expression of the pepN gene in Escherichia coli : studies with pepN-lacZ operon and protein fusion strains .
Gibson , M. M. , M. Price , and C. F. Higgins .
Genetic characterization and molecular cloning of the tripeptide permease ( tpp ) genes of Salmonella typhimurium .
Higgins , C. F. , M. M. Hardie , D. Jamieson , and L. M. Powell .
Genetic map of the opp ( oligopeptide permease ) locus of Salmonella typhimurium .
Holley , E. A. , M. P. Spector , and J. W. Foster .
Regulation of NAD biosynthesis in Salmonella typhimurium : expression of nad-lac fusions and identification of a nad regulation locus .
Hughes , K. T. , B. T. Cookson , D. Ladika , B. M. Olivera , and J. R. Roth .
6-Aminonicotinamide-resistant mutants of Salmonella typhimurium .
Hughes , K. T. , and J. R. Roth .
Conditionally transposi-tion-defective derivative of Mu dl ( Amp Lac ) .
Jamieson , D. J. , and C. F. Higgins .
Anaerobic regulation of a peptide transport gene in Salmonella typhimurium .
Jamieson , D. J. , R. G. Sawers , P. A. Rugman , D. H. Boxer , and C. F. Higgins .
Effects of anaerobic regulatory mutations and catabolite-repression on regulation of hydrogen metabolism and hydrogenase isoenzyme composition in Salmonella typhimurium .
Kuritzkes , D. R. , X.-Y .
Zhang , and E. C. C. Lin .
Use of 4 ( glp-lac ) in studies of respiratory regulation of the Escherichia coli anaerobic sn-glycerol-3-phosphate dehydrogenase genes ( glpAB ) .
Lambden , P. R. , and J. R. Guest .
Mutants of Escherichia coli K12 unable to use fumarate as an anaerobic electron-acceptor .
Levine , M. , and R. Curtiss .
Genetic fine structure of the c region and the linkage map of phage P22 .
Merberg , D. , and P. Datta .
Altered expression of biodegradative threonine dehydratase in Escherichia coli mutants .
Experiments in molecular genetics .
Cold Spring Harbor Laboratory , Cold Spring Harbor , N.Y. 21 .
Newman , B. M. , and J. A. Cole .
The chromosomal location and pleiotropic effects of mutations of the nirA gene of Escherichia coli K12 : the essential role of nirA + in nitrite reduction and in other anaerobic redox reactions .
Palva , E. T. , P. Liljestrom , and S. Harayama .
Cosmid cloning and transposon mutagenesis in Salmonella typhimurium using phage X vectors .
Burini , and M. Chippaux .
Regulation of the trimethylamine N-oxide ( TMAO ) reductase in Esche-richia coli : analysis of tor : : Mudl operon fusions .
Pecher , A. , F. Zinoni , and A. Bock .
The seleno-polypeptide of formic dehydrogenase ( formate hydrogen lyase linked ) from E. coli : genetic analysis .
Pecher , A. , F. Zinoni , R. C. Jatisatien , R. Wirth , H. Hennecke , and A. Bock .
On the redox control of synthesis of anaerobically induced enzymes in Enterobacteriaceae .
Genetic techniques in studies of bacterial metabolism .
Sanderson , K. E. , and J. R. Roth .
Linkage map of Salmonella typhimurium , edition VI .
Sawers , R. G. , S. P. Ballantine , and D. H. Boxer .
Differential expression of hydrogenase isoenzymes in Escherichia coli K-12 : evidence for a third isoenzyme .
Sawers , R. G. , D. J. Jamieson , C. F. Higgins , and D. H. Boxer .
Characterization and physiological roles of membrane-bound hydrogenase isoenzymes from Salmonella typhimurium .
Schreyer , R. , and A. Bock .
Phosphoglucose isomerase from Escherichia coli K1O : purification , properties and formation under aerobic and anaerobic conditions .
D. D. W. and R. Guest .
Shaw , J. , Rice , J. 1983 .
Homology between CAP and a of anaerobic respiration Fnr , regulator in Escherichia coli .
Silhavy , T. J. , M. L. Berman , and L. W. Enquist .
Experiments with gene fusions .
Cold Spring Harbor Laboratory , Cold Spring Harbor , N.Y. 33 .
Smith , M. W. , and F. C. Neidhardt .
Proteins induced by anaerobiosis in Escherichia coli .
Requirement of Fnr and NarL functions for nitrate reductase expression in Escherichia coli K-12 .
Stewart , V. , and C. H. MacGregor .
Nitrate reductase in Escherichia coli K-12 : involvement of chlC , chlE , and chlG loci .
Strauch , K. L. , J. B. Lenk , B. L. Gamble , and C. G. Miller .
Oxygen regulation in Salmonella typhimurium .
Thomas , A. D. , H. W. Doelle , A. W. Westwood , and G. L. Gordon .
Effect of oxygen on several enzymes involved in aerobic and anaerobic utilization of glucose in Escherichia coli .
Waugh , R. , and D. H. Boxer .
Pleiotropic hydrogenase mutants of Escherichia coli K12 : growth in the presence of nickel can restore hydrogenase activity .
Way , J. C. , M. A. Davis , D. Morisato , D. E. Roberts , and N. Kleckner .
New TnlO derivatives for transposon mutagenesis and for construction of lacZ operon fusions by transposition .
Wimpenny , J. W. T. , and A. Firth .
Levels of nicotinamide adenine dinucleotide in facultative bacteria and the effect of oxygen .
Yerkes , J. H. , L. P. Casson , A. K. Honkanen , and G. C. Walker .
Anaerobiosis induces expression of ant , a new Esche-richia coli locus with a role in anaerobic electron transport .
Zinoni , F. , A. Beier , A. Pecher , R. Wirth , and A. Bock .
Regulation of the synthesis of hydrogenase ( formate hydrogen-lyase linked ) of E. coli .