Regulatory Gene Integrates Control of Vitamin Synthesis and Propanediol Degradation The cob operon of Salmonella typhimurium encodes enzymes required for synthesis of adenosyl-cobalamin ( vitamin-B12 ) .
The pdu operon encodes enzymes needed for use of propanediol as a carbon source , including an adenosyl-cobalamin-dependent enzyme , propanediol dehydratase .
These two operons both map near min 41 of the S. typhimurium linkage map and are transcribed divergently .
Here we report that the cob and pdu operons form a single regulon .
Transcription of this regulon is induced by either glycerol or propanediol .
The metabolism of these compounds is not required for induction .
Propanediol induces the regulon either aerobically or anaerobically during-growth on poor carbon sources .
Aerobically glycerol induces only if its metabolism is prevented by a mutational block such as a glpK mutation .
Under anaerobic conditions , glycerol induces in both glpK + and glpK mutant strains during-growth on poor carbon sources .
A new class of mutations , pocR , prevents induction of the coblpdu regulon by either propanediol or glycerol and causes a Cob-Pdu-phenotype .
The pocR gene is located between the cob and pdu operons and appears to encode a trans-acting protein that acts as a positive regulator of both operons .
Transcription of the pocR regulatory gene is induced , even without the PocR protein , during aerobic-growth on poor carbon sources and during anaerobic respiration .
With the functional PocR protein , transcription of the pocR gene is autoinduced by propanediol but not by glycerol .
The growth-conditions that increase pocR gene expression correlate with growth-conditions that allow high induction of the coblpdu regulon .
A model for control of this regulon suggests that the PocR protein is a transcriptional activator of both the cob and pdu operons and that both glycerol and propanediol can individually serve as effectors of the PocR protein .
We suggest that global control mechanisms cause variation in the level of the PocR protein ; an increased level of the PocR protein permits higher induction by propanediol or glycerol .
Salmonella typhimurium synthesizes the vitamin-B12 co-enzyme adenosyl-cobalamin ( Ado-Cbl ) de novo ( 24 ) .
Twen-ty-five genes that play a role in Ado-Cbl biosynthesis are known ( 16 , 30 ) .
The majority of these genes lie in a single large operon , the cob operon , that includes three contiguous clusters of functionally related genes ( see below and reference 14 ) .
Genes of the CobI and CoblI clusters are required for biosynthesis of adenosyl-cobinamide ( Ado-Cbi ) and di-methylbenzimidazole ( DMB ) , respectively .
CobIII gene products mediate the covalent joining of Ado-Cbi and DMB to form Ado-Cbl ( 25 ) .
In S. typhimurium , de novo vitamin-B12 synthesis occurs only under anaerobic-growth-conditions ; however , exogenous corrinoids can be taken up and converted to vitamin-B12 coenzymes both aerobically and anaerobically ( 18 ) .
Previous work suggested that the cob operon is induced only anaerobically ( redox control ) and is stimulated by the complex of the cyclic-AMP-receptor-protein ( CRP ) with cyclic AMP ( cAMP ) ( 3 , 17 ) .
Rondon and Escalante-Semerena recently found that the cob operon can also be induced under aerobic conditions by propanediol ( 29 ) .
Propanediol utilization is one of the four metabolic functions known require vitamin-B12 in S. typhito coenzymes munium ( 23 ) .
Ado-Cbl is a cofactor of propanediol dehydratase , which catalyzes dehydration of propanediol to propionaldehyde , the first step of propanediol utilization ( 23 , 32 ) .
Propanediol as a sole carbon and energy can serve source for S. typhimunium only aerobically , but only if an exogenous corrinoid is provided , since Ado-Cbl is not synthesized de novo under aerobic conditions ( 24 ) .
Mutations causing a defect in utilization of propanediol map in thepdu operon , which is adjacent to the cob operon .
While thepdu operon encodes an Ado-Cbl-dependent propanediol dehydratase , other genes of this operon have not been characterized .
Induction of the pdu operon by propanediol occurs during aerobic-growth on a poor carbon source , such as succinate ( 23 ) .
Here we investigate the system that coregulates the cob and pdu operons .
Both operons appear to be induced by propanediol and by glycerol .
In addition , both are subject to global regulation .
We propose a model for induction and discuss the paradoxical aspects of the regulatory behavior of this system .
MATERUILS AND METHODS Bacterial strains and transposons .
Strains used in this study were derivatives of S. typhimurium LT2 ( Table 1 ) .
MudA refers to a conditionally transposition-defective derivative of the phage Mu dl ( 7 ) that was constructed by Hughes and Roth ( 21 ) .
MudJ refers to phage Mu dl-1734 constructed by Castilho et al. ( 8 ) .
MudA and MudJ elements confer Apr and Knr , respectively .
When any of these transposable Mud elements inserts into a chromosomal operon in the proper orientation , transcription of that operon extends into the element .
the lacZ gene of the inserted By assaying resulting I-galactosidase one can estimate the activity of the chromosomal promoter .
These lac operon fusions were used to study transcriptional control of the cob and pdu operons .
TnlOdTc and TnlOdCm are transposition-defective derivatives of transposon TnlO that confer Tcr and Cmr , respectively ( 15 , 34 ) .
Several pdu insertion mutants were kindly provided by Randall Jeter .
The rich-medium used was nutrient broth ( 0.8 % ; Difco ) supplemented with 85 mM NaCl .
The minimal me-dium was either E medium supplemented-with-glucose ( 33 ) or NCE ( E medium lacking citrate ) ( 5 ) supplemented with one or more of the following compounds : D-glucose , 0.2 % for aerobically grown cultures and 0.4 % for anaerobically 0.2 % ; Na2 succinate , 1.0 % ; Na grown cultures ; glycerol , pyruvate , 0.44 % ; Na2 fumarate , 0.32 % ; and DL-1,2-propane-diol , 0.2 % .
Glucose , fumarate , and succinate were from Sigma Chemical Co. ( St. Louis , Mo. ) Propanediol and pyruvate were from Aldrich Chemical Co. ( Milwaukee , Wis. ) , and glycerol was from Em Scientific ( Cherry Hill , N.J. ) .
Antibiotics were added to nutrient agar as follows : Na ampicillin , 30 , ug/ml ; chloramphenicol , 20 , g/ml ; kanamycin monosulfate , 50 FLg/ml ; and tetracycline hydrochloride , 20 All antibiotics were from Sigma Chemical .
acids , when required , were added to minimal-medium at the concentrations recommended previously ( 13 ) .
Transductional crosses were performed with the high-frequency generalized transducing phage mutant P22 HT105/1 int-201 ( 31 ) .
Cells ( 2 x 108 ) were mixed with phage ( 107 to 108 ) and plated directly on selective medium .
When drug-resistant transductants ( Tcr Knr Apr Cmr ) were selected , the cell culture and phage were incu-in for 15 to 45 min bated together nonselectively liquid before plating on solid selective medium .
Transductants were purified and made phage free by streaking for single colonies on nonselective green indicator plates ( 9 ) .
Transductants were tested for phage sensitivity by cross-streaking with the clear-plaque P22 mutant H5 .
The CobIII + phenotype was scored ( in strains carrying a metE mutation ) as the ability to grow without methionine on E-glucose medium containing 0.2 , ug of cobinamide dicyanide ( Sigma ) per ml and 50 , ug of 5,6-DMB ( Aldrich ) per ml .
This method is based on the fact that S. typhimurium has alternative enzymes ( encoded by metE and metH ) capable of catalyzing the last step in methionine synthesis .
The metH enzyme depends on vitamin-B12 ; therefore , mutants lacking metE function can grow on minimal-medium only if they have vitamin-B12 .
Strains that are phenotypically CobIII + possess the enzymes needed to produce vitamin-B12 from cobinamide and DMB ; these enzymes are expressed both aerobically and anaerobically .
The Cob ' phenotype was scored ( in strains with a metE mutation ) as the ability to grow anaerobically on E-glucose medium with or without 5 , uM CoCl2 ( Mallinckrodt Chemicals , St. Louis , Mo. ) .
The growth-conditions necessary for de novo synthesis of vitamin-B12 were provided by an anaerobic chamber ( Forma Scientific model 1024 ) containing an atmosphere of N2-CO2-H2 ( 89:5:6 ) The pdu phenotype was scored on MacConkey indicator plates ( Difco , Inc. ) containing 1 % propanediol and 0.2 , ug of vitamin-B12 per ml .
On these plates , Pdu + strains form red colonies and Pdu-colonies are white .
The,-galactosidase activity in strains grown under specified conditions was assayed as described by Miller ( 27 ) .
For each table , the data presented are derived from a single experiment ; each entry is based on a single assay , but the experiments were repeated several times with essentially the same results as those presented .
For aerobic-growth of cells , an overnight nutrient broth culture , inoculated from a frozen strain , was prepared .
This broth culture ( 0.1 ml ) was used to inoculate 2 ml of the desired minimal-medium .
Cultures were grown to the log phase , and 0.1 ml was used as the inoculum for a second 2-ml culture with similar minimal-medium .
When the second culture reached log phase , cells were harvested and assayed for P-galactosidase activity .
Cultures were grown with shaking in 2 ml of medium in a test tube ( 15 by 125 mm ) .
For anaerobic-growth of cells , 0.1 ml of an overnight nutrient broth culture was used to inoculate 2 ml of minimal-medium .
This culture was grown aerobically to log phase , brought inside an anaerobic chamber , and used ( 0.2 ml ) to inoculate 5 ml of similar minimal-medium in an anaerobic aluminum crimp sealable culture tube ( 18 by 150 mm ) ( Bellco Biotechnology , Vineland , N.J. ) .
This medium was previously made anaerobic by incubation in an anaerobic chamber for at least 10 h for gas exchange .
After inoculation , tubes were fitted with black rubber stoppers ( Bellco ) , removed from the anaerobic chamber , and sealed .
The headspace gas of the tubes was replaced with N2 by three cycles of evacuation and pressurization ( 4 ) .
Cells were grown to log phase at 37 °C with shaking under 2 atm ( 202 kPa ) of N2 and then assayed for I-galactosidase activity .
If N2 pressure was lost during incubation , assays were repeated .
When strains with duplications were grown for assay of f-galactosidase , loss of duplications was counterselected by supplementing all media with-ampicillin ; resistance to this antibiotic is encoded by the transposon located between copies of the duplicated segments .
Ampicillin was provided at 30 , ug/ml in rich-medium , 15 Rg/ml in minimal-medium when cultures were incubated aerobically , and 2 , ug/ml in minimal-medium when cultures were incubated anaerobically .
Cultures of strains with duplications were routinely tested for the presence of haploid segregants by plating for single colonies and replica printing to score the loss of the Apr marker between copies of the duplicated segment ; reported 0-galactosidase activities were obtained from cultures with < 5 % Aps segregants .
Transposon mutagenesis of the coblpdu region .
For TnlOdTc mutagenesis , a pool of approximately 100,000 independent TnlOdTc insertion mutants was prepared by transducing a TnlOdTc insertion from an Escherichia coli F ' lac strain into a recipient strain with a plasmid-encoded TnlO transposase ( 34 ) .
The pool of Tcr transductant clones was used as a donor in transductional crosses with strain TT17073 , which contained a CobIII : : MudJ insertion .
Transductants ( CobIII + ) were selected and screened for Tcr recombinants that had inherited a TnlOdTc linked to the CoblIl region .
To obtain MudJ insertion mutations in the cob/pdu region , a pool of about 100,000 independent MudJ insertions was prepared by the cis complementation method of Hughes and Roth ( 22 ) and was used as donor in transductional crosses with a recipient strain which contained a CobIII : : TnlO insertion ( TT9814 ) .
CobIII + transductants were selected and scored for coinheritance of Knr , encoded by MudJ .
The resulting CobIII + clones were screened for their Pdu phenotype and their ability to synthesize vitamin-B12 de novo under anaerobic conditions .
Orienting pocR : : MudJ insertions .
The orientation of pocR : : MudJ ( Kn ) insertions was determined by transductional crosses that scored recombination between a pocR : : MudJ ( Kn ) element and one of two donor CobII : : MudA ( Ap ) insertions of known orientation , cob-61 ( clockwise , TT10364 ) and cob-62 ( counterclockwise , TT10365 ) .
Recombinants inheriting the donor Apr were scored for loss of the recipient Knr and for acquisition of the CoblIl-phenotype expected for a deletion extending from the pocR gene to the CoblI region .
The orientation of the pocR : : MudJ element was determined by which of the two donor cob insertions could recombine with the recipient pocR : : MudJ element to generate a deletion ( 20 ) .
The same method used for determining the orientation of Mud elements ( see above ) was used to construct a set of deletions each extending rightward from an insertion in the his operon to the sites of similarly oriented MudJ insertions in pdu , pocR , or the silent zeb region ( 20 ) .
Similarly , we constructed deletions extending leftward from the CobIl region to the same MudJ insertions in pocR and the zeb region .
These Mud-directed deletions were used in constructing the deletion map .
Deletions from his to pocR or to the zeb region were constructed by transducing the recipient hisD9952 : : MudA strain ( TT7691 ) with phage grown on strains carrying apocR or zeb : : MudJ insertion mutation .
Deletions from his to pdu were constructed by transducing recipients carrying a pdu : : MudA ( Ap ) insertion with phage grown on an insertion mutant hsiCJ264 : : MudJ ( Kn ) strain ( TT17144 ) .
Deletions from pdu to CobII were constructed by transducing the recipient cob-61 : : MudA strain ( TT10364 ) with phage grown on strains carrying a pdu : : MudJ insertion .
In all of these constructions , donor strains carried a MudJ ( Kn ) element and the recipients carried a MudA ( Ap ) element .
Transductants ( Knr ) were selected and screened for those that had lost Apr and acquired the directed deletion .
Deletions from the pocR or zeb region to CobIl were constructed by transducing recipients containing a pocR or zeb : : MudJ ( Kn ) insertion mutation with phage grown on a cob-62 : : MudA ( Ap ) insertion mutant ( TT10365 ) .
In these crosses , Apr recombinants were selected and screened for the loss of the recipient Knr phenotype and acquisition of the CobIlI phenotype characteristic of the constructed deletion .
Deletion map of the control region .
All mapping involved transductional crosses between recipient strains carrying one of the deletions described above ( with either a Knr or an Apr phenotype ) and donor strains carrying a simple insertion of TnlOdTc ( with a Tcr phenotype ) or TnlOdCm ( with a Cmr phenotype ) .
Transductants resistant to tetracycline ( or chloramphenicol ) were selected and replica printed to me-dium with-tetracycline ( or chloramphenicol ) and either ampicillin or kanamycin , depending on whether the deletion recipient had a MudA or a MudJ at the join point .
( We use `` join point '' to designate the site or region between the two copies of the duplicated segment .
) Formation of transductants showing both Tcr ( or Cmr ) and either Apr or Knr phenotypes indicated recombination between the recipient deletion and the donor insertion .
A lack ( or paucity ) of such recombinants indicated that the donor insertion mapped inside the recipient deletion .
Crosses were scored as positive for recombination when greater than 5 % of the Tcr transductants also showed the recipient drug resistance .
Crosse were scored as negative if fewer than 1 % of the Tcr transductants showed the recipient drug resistance .
No pair of strains showed intermediate values .
The rationale for this mapping method is described in the Appendix .
Construction of tandem chromosomal duplications .
A modification of the method of Hughes and Roth ( 20 ) was used to construct duplications shown in Fig. 2 ( see below ) .
Duplications with the basic structure of duplication 1 were constructed in several steps by P22 transduction .
A hisD 9952 : : MudA ( Ap ) insertion ( FT7691 ) was crossed into a strain with a cob-24 : : MudJ ( Kn ) element ( 1T10852 ) , selecting Apr ( Fig. 1 ) .
Sequences near the donor insertion can recombine with the chromosome at one site ( the his operon ) , and sequences within the donor insertion can recombine with the chromosome at the distant site of the recipient Mud insertion within the cob operon .
Such events can construct a duplication .
The phenotype of 8 of 30 Apr transductants was Apr Knr His ' , indicative of a duplication from the hisD9952 site to the cob-24 site with a MudA ( Ap ) element at the join point and a MudJ ( Kn ) element at the cob-24 outside end of the duplication ( Fig. 1 ) .
This duplication was separated from the outside MudJ ( Kn ) insertion by transduction of the join point into a new recipient ( TR6583 ) , selecting Apr. .
The phenotype of six of six transductants was Apr Kn ' His ' , indicating that these strains inherited the his-cob duplication with the MudA ( Ap ) at the join point but lacked the outside MudJ ( Kn ) element .
This duplication had the structure of duplication 1 with pocR + alleles at both pocR loci ( Fig. 2 ) .
To construct duplication 1 with various combinations of pocR alleles , a strain with this duplication was used as recipient in a cross with donor strain pocR8 : : TnlOdTc ( TT17085 ) , selecting Tcr Apr transductants .
All transduc-his pdu pocR ' his ' pdu pocR ' cob cobil FIG. 2 .
Duplications used in complementation tests .
Merodip-loids were constructed by recombination between the inserted elements diagrammed in the top section .
The structures of the three duplication types are diagrammed .
Various mutant alleles were transduced into these strains to construct the strains described in Tables 5 and 6 .
tants tested ( 30 of 30 ) had the phenotype expected of a his-cob duplication with an added pocR : : TnlOdTc insertion ( Tcr Apr His ' Pdu + ) .
To determine whether the TnlOdTc element was located cis or trans to the MudA element at the duplication join point , some of these strains were used as transductional donors in crosses with strain TR6583 .
Inheritance of the marker ( Apr ) at the duplication join point was selected , and approximately 200 transductants were scored for cotransduction of Tcr .
In about 1 of 15 of the strains with a single pocR : : TnlOdTc element , the TnlO element showed about 70 % linkage to the MudA element at the duplication join point ; in about 14 of 15 of the strains with a duplication , the TnlOdTc element was unlinked to the MudA element at the join point .
We inferred that the linked and unlinked TnlOdTc elements were located cis and trans to the MudA element , respectively .
To construct a duplication with a pocR8 : : TnlOdTc insertion at both pocR loci , the transductional cross described above was repeated at a high multiplicity of infection ( approximately 5 to 10 ) and transduction mixtures were plated on MacConkey indicator plates with propanediol , vitamin-B12 , tetracycline , and ampicillin .
The Tcr transductants could thus be directly screened for their Pdu phenotypes .
About 1 of 500 transductants resulting from this procedure was Apr Tcr Pdu - .
To verify that these transductants contained the recipient his-cob duplication and a TnlOdTc element at bothpocR loci , strains were grown nonselectively overnight in nutrient broth .
Dilutions of these cultures were plated on nutrient agar , and colonies that developed were printed to rich-medium with and without-ampicillin or tetracycline .
As expected , the MudA element ( Apr ) marker was unstable and the Tcr phenotype was stable .
In order to block vitamin-B12 synthesis and preven repression of the cob operon by endogenous cobalamin ( 17 ) , insertion cob-364 : : TnlOdCm was introduced into the strains with the duplications described above .
This insertion maps outside the duplicated his-cob region and thus prevents vitamin-B12 synthesis .
It was introduced by selecting for resistance to chloramphenicol .
Duplication 2 ( Fig. 2 ) was constructed by a procedure analogous to that for construction of duplication 1 .
The initial duplication was made by transducing a cob-61 : : MudA ( Ap ) insertion ( Tf10364 ) into a strain ( FT17131 ) with insertion pdu-12 : : MudJ ( Kn ) .
The phenotype of 9 of 16 transductants was that expected for a strain with a cob-pdu duplication generated by recombination between the Mud elements ( Apr Knr CobII + ) .
This duplication was crossed into TR6583 to eliminate the MudJ element located at the end point of the duplication .
A strain carrying duplication 2 was identified as a transductant having the phenotype Apr Kn ' Pdu + .
ApocRl : : TnlOdTc element was then transduced into the cis copy , the trans copy , or both copies of the pocR gene present in duplication 2 .
The cis and trans TnlOdTc elements showed 75 and 4 % linkage , respectively , to the MudA ( Ap ) element at the join point of the duplication .
Duplication 3 was constructed by transducing the hisD 9952 : : MudA element from strain TF7691 into a recipient carrying insertion pocR12 : : MudJ ( FT17104 ) , selecting Apr. .
Among the transductants , 16 of 30 showed the Apr Knr His ' Pdu-phenotype indicative of a his-pocR duplication with a pocRl2 : : MudJ insertion at the outside end of the duplicated segment .
The join point of this duplication was crossed into strain TR6583 .
Duplication 3 was carried by an Apr Kn ' His + Pdu + transductant obtained from this cross .
Three derivatives of this parental duplication-carrying strain were constructed ; each received one of the following mutations .
The pocR8 : : TnlOdTc andpdu-207 : : TnlOdTc elements were each introduced by selecting Tcr Apr recombinants .
The mutation glpK7 was introduced into the parent strain by using a linked TnlO element as a selective marker ( Tcr ) .
RESULTS Coinduction of the cob and pdu operons by propanediol .
Transcription of both operons was induced by propanediol during either aerobic or anaerobic-growth of cells on poor carbon sources ( Table 2 ) .
Induction was determined by measuring the 13-galactosidase activity in strains with cob-lac and pdu-lac operon fusions formed by Mud ( Lac ) insertions .
Neither operon was induced during aerobic-growth-on-glucose or glycerol .
During fermentation of glucose , pro-panediol induced both operons but the maximum level was low .
The highest induced level was observed under conditions of anaerobic respiration of pyruvate-fumarate .
The inductive effect of propanediol did not require propanediol metabolism ; strains with mutations causing a propanediol dehydratase deficiency showed normal induction by pro-panediol ( data not shown ) .
The aerobic inductive effect of propanediol on the pdu operon has been reported previously ( 23 ) .
Previous work suggested that the cob operon was not expressed aerobically ( 3 ) ; the induction of the cob operon by propanediol under aerobic conditions was surprising .
The parallel behavior of the two operons in these experiments and those described below suggest that the cob and pdu operons constitute a single regulon with a single control mechanism .
Coinduction of the cob and pdu operons by glycerol .
Transcription of both the cob and pdu operons was induced by glycerol as well as propanediol ( Table 3 ) .
The aerobic induction by glycerol was observed only in strains unable to metabolize glycerol .
In a glpK mutant , which lacks glycerol kinase , glycerol induced both the cob andpdu operons during aerobic-growth-on-succinate and during anaerobic respiration of pyruvatefumarate .
Since glycerol kinase is required for both aerobic and anaerobic metabolism of glycerol , these results indicated that unmetabolized glycerol stimulated aerobic and anaerobic transcription of both the cob and pdu operons .
In glpK + strains , aerobic induction of the cob and pdu operons by glycerol did not occur .
This indicated that aerobic glycerol metabolism blocked induction .
However , glycerol metabolism did not prevent induction of either operon under anaerobic conditions .
Induction was maximal during anaerobic respiration of glycerol-fumarate .
Under these conditions , addition of propanediol caused no further induction ( data not shown ) .
Previously it had been seen that the cob operon is maximally induced during anaerobic-growth-on-glycerol-fumarate ( 17 ) , but this was attributed to metabolic consequences of glycerol respiration ( 3 ) .
The finding that glycerol per se induces this regulon is surprising .
Insertion mutagenesis of the coblpdu region .
The region between the divergently transcribed cob and pdu operons was mutagenized with TnlOdTc and MudJ elements ( see Materials and Methods ) .
Four classes of mutants were obtained .
One class showed a simple Cob-phenotype , and another showed a simple Pdu-phenotype .
These mutations are presumed to affect the two operons individually .
A third class of mutants ( zeb ) had a normal growth phenotype ( Cob ' Pdu + ) but showed slight alterations in expression of the cob andpdu operons ( see below ) .
The fourth class of insertions ( pocR ) showed a pleiotropic Cob-Pdu-phenotype .
Both the Cob ' Pdu + ( zeb ) and the Cob-Pdu - ( pocR ) mutations mapped between the cob and pdu operons ( see below ) .
A likely reason for the Cob-Pdu-phenotype of pocR mutants was that they failed to induce both the cob and pdu operons .
Qualitative tests employing MacConkey agar-lac tose indicator plates suggested that all 11 of the pocR : : TnJOdTc insertion mutants we isolated were unable to induce the cob and pdu operons .
Four different pocR : : TnlOdTc mutants were assayed to quantitate the effect of the insertion on expression of the cob and pdu operons .
The results for a representative pocR : : TnlOdTc insertion are presented in Table 4 .
During aerobic-growth-on-succinate minimal-medium or during anaerobic respiration , the pocR mutant showed no induction of either operon by propanediol or glycerol ( Table 4 , lines 2 and 6 ) .
The failure of glycerol to induce the cob and pdu operons in a pocR mutant was also seen in strains with glycerol metabolism blocked by a glpK mutation ( data not shown ) .
It should be noted that under anaerobic-growth-conditions , all pocR insertion mutants showed a leaky Cob-phenotype that was corrected by the addition of CoCl2 to the growth medium .
This phenotype is probably due to the fact that pocR insertion mutations reduce cob operon expression to a level at which Co2 + transport , thought to be encoded by the cbiO gene within the cob operon ( 30 ) , limits vitamin-B12 synthesis on medium with no deliberately added cobalt .
The pocR gene is not directly involved in Co2 + transport , since pocR mutants have a Cob + phenotype if one provides for expression of the cob operon ( 1 , 2 ) .
The Cob + Pdu + class of insertions will be referred to as zeb , by the convention for naming insertions by their chromosomal map position ( 11 ) .
The cob-lac and pdu-lac fusion strains with a zeb-3721 : : TnlOdTc insertion showed about twofold less induction of the coblpdu regulon by propanediol during aerobic-growth-on-succinate ( Table 4 , lines 3 and 7 ) .
The same zeb insertion had little or no effect on induction during anaerobic-growth-on-glycerol-fumarate medium .
Similar results were obtained with three other zeb : : TnlOdTc insertions ( data not shown ) .
It is conceivable that pocR mutants might fail to express one operon as a secondary consequence of their failure to express the other .
To test this , four simple pdu : : TnlOdTc insertions were tested ; none affected regulation of a cob-lac fusion .
The results obtained with a representative insertion are shown in Table 4 ( strain TT17123 ) .
Conversely , deletion mutations eliminating the entire CobI region retain a Pdu + phenotype .
Since neither operon is required for expression of the other , we conclude that the Cob-Pdu-phenotype of pocR mutations is due to impaired expression of both the cob and pdu operons .
Genetic map of the region between the pdu and cob operons .
By deletion mapping , the following map order was determined : pdu , zeb , pocR , and cob .
For each Mud ( lac ) insertion mutation type ( pdu , zeb , and pocR ) , two deletions were constructed , one extending leftward to the his operon and the other extending rightward to the CobII region ( see Fig. 3 ) .
Deletion mutants were used as recipients in crosses with donor TnJOdTc ( or TnlOdCm ) insertion mutations in each of the critical regions .
The importance of the deletion pairs is explained in the Appendix .
Doubly mutant recombinants that carry both the Tcr ( or Cmr ) phenotype of the donor insertion and the Kr or Apr phenotype of the recipient deletion were selected .
This novel mapping method ( described in more detail in the Appendix ) was required because there was no phenotype to permit selection for wild-type recombinants .
The resulting genetic map is in Fig. 3 .
Each of the four phenotypic classes of TnlOdTc insertions mapped as a contiguous group , demonstrating a physical clustering of mutations that confer the same phenotype .
Since all of the zeb and pocR insertions mapped outside the longest of the deletions extending from his to pdu , these regions were placed to the right of thepdu operon , as shown in Fig. 3 .
The zeb insertions mapped outside the deletions from pocR to CobII ; pocR insertions mapped inside these deletions .
Therefore , the pocR region lies to the right of the zeb region and immediately to the left of the cob operon .
Using deletions extending from severalpocR insertions to CobII or to his , we found that the pocR : : TnlOdTc insertions mapped in two distinct deletion intervals within the pocR gene .
The placement of these insertions within the pocR gene was confirmed by studies of polarity effects of the TnlOdTc insertions on expression of the pocR : : lac fusions .
That is , eachpocR : : TnlOdTc insertion reduces expression of pocR : : lac fusions mapping to its right ( Fig. 3 ) .
All four zeb : : TnlOdTc insertions mapped into a single deletion interval .
The region between the pdu and cob operons thus includes the zeb region and the pocR gene in that order ; TnlOdTc and MudJ insertions map at several points in the pocR gene ( Fig. 3 ) .
The fact that pocR : : TnlO insertions block transcription of Mud ( lac ) fusions mapping to their right in the pocR gene suggests that the pocR gene is transcribed from left to right as the map appears in Fig. 3 .
This conclusion was confirmed by orienting the pocR : : Mud insertions ( see Materials and Methods ) ; it was found that in these insertions the lacZ gene is oriented such that expression requires transcription from left to right in the map in Fig. 3 .
Orienting transcription of the pocR gene .
The direction of transcription of the cob operon was previously reported to be counterclockwise and that of the pdu operon was reported to be clockwise ( 23 , 25 ) .
Here we report that the direction of transcription of the pocR gene , like the cob operon , is counterclockwise , or left to right as the map is drawn in Fig. 3 .
In transductional crosses , transposons at separated chromosomal sites can recombine to construct a deletion if the two elements have the same orientation in the chromosome ( 12 , 20 ) .
Crosses ( described in Materials and Methods ) demonstrated that the cob-62 : : MudA element ( transcribed counterclockwise ) recombined with the pocRl2 : : MudJ element to form a deletion ; 28 of 30 transductants inheriting the donor MudA ( Ap ) element acquired a deletion of the region between the pocR and CobIl Mud elements .
The formation of deletions indicated that these transposons had the same orientation in the chromosome .
Since the lacZ genes of both insertions are expressed , both target genes must be transcribed in the same direction .
The CobII region is known to be transcribed counterclockwise ( 25 ) .
Thus , the pocR gene must also be transcribed counterclockwise ( left to right in Fig. 3 ) .
Counterclockwise transcription of the pocR gene was supported by orientation of two additional Lac ' pocR : : MudJ insertions ; both proved to be in the same orientation as pocRl2 : : MudJ .
Orientation of all three pocR : : MudJ elements was confirmed by individually crossing strains containing these elements with a strain carrying insertion cob-61 : : MudA ( Lac - ) , oriented clockwise .
In each cross , no transductants ( O of 30 ) inheriting the donor MudA had the phenotype predicted for a deletion formed by recombination between the Mud elements .
Mutations in the pocR region can be complemented in trans .
Tandem chromosomal duplications were used to perform complementation tests demonstrating that the regulatory phenotypes of a pocR mutation could be corrected by a wild-typepocR gene provided in trans .
Duplication 1 ( Fig. 2 ) was used to test regulation of the cob operon .
It has a cob : : lac fusion at the join point ( the point or region between the two copies of the duplicated material ) .
The duplication provides twopocR genes , one located immediately cis to the cob : : lac fusion and the other located far from the cob : : lac fusion , essentially in trans .
For the complementation tests , pocR : : TnlOdTc alleles were introduced into strains carrying a pocR duplication to construct one derivative that had its sole functional pocR allele cis to the cob : : lac fusion and second that had the functional pocR allele trans to the fusion .
A third derivative that had two pocR : : TnlOdTc alleles was constructed .
Each duplication-carrying strain contained the insertion cob-364 : : TnlOdCm to prevent endog-enous synthesis of vitamin-B12 , which could have repressed the cob operon ( 17 ) .
Regulation of cob operon expression in these strains is shown in Table 5 , lines 2 to 5 .
The regulatory behavior of a haploid strain ( TT10327 ) is also shown by the diploid strains that have at least one pocR + allele , regardless of whether the functional pocR allele is located cis or trans to the assayed cob-lac fusion .
The strain with no functional pocR allele showed little or no induction ( Table 5 , TT17177 ) , similar to the behavior of a haploid strain with apocR mutation ( Table 4 , TT17119 ) .
Duplication 2 ( in Fig. 2 ) was used to test the effects of pocR on expression of thepdu operon .
This duplication has a pdu-lac fusion at its join point and includes two copies of the pocR gene , one copy immediately cis to the join point pdu-lac fusion being assayed and the other effectively trans to the fusion .
Strains were constructed with this duplication and apocR : : TnlO insertion in cis or in trans to the fusion ; a third derivative hadpocR : : TnlO at both pocR loci .
Since the pdu operon is not subject to repression by vitamin-B12 ( 1 ) , no cob insertion was added to these strains .
Regulation of thepdu operon in these strains is presented in Table 5 , lines 6 to 10 .
During aerobic-growth-on-succinate plus propanediol and during anaerobic-growth-on-glycerolfumarate or pyruvate-fumarate plus propanediol , the pdu operon showed normal induction in all strains having at least onepocR + allele , regardless of the location of the functional copy with respect to the pdu : : lac fusion .
No induction was seen in strains lacking a functional pocR allele .
Throughout Table 5 , it should be noted that all strains with a duplication of the control region showed a maximal induced level of the cob/pdu regulon that is 30 to 85 % of that seen in a haploid strain ; this is even true for strains with two wild-type PocR + alleles .
This applies to both the cob and pdu operons , to aerobic and anaerobic conditions , and to strains with either duplication 1 or duplication 2 .
The basis of this effect is not yet understood .
The above data also demonstrate that there is little read-through transcription from the pocR gene into the cob operon .
The induced level of the cob operon is essentially the same regardless of whether PocR + function is provided by a pocR allele located in cis or in trans to the assayed operon .
The presence of a TnlO insertion in the proximal pocR gene does not reduce cob operon expression .
Since TnlO insertions are strongly polar ( 26 ) , this is evidence that transcription from the pocR gene promoter ( which is considerable , as seen below ) does not contribute to expression of the adjacent cob operon under the conditions tested here .
Transcription of the pocR gene was examined in a haploid strain containing a pocR-lac fusion and in strains containing a duplication with apocR-lac fusion in one copy of the pocR locus and either a pocR + or a pocR8 : : TnlOdTc allele at the secondpocR locus ( Table 6 ) .
The structure of the strains used is shown in Fig. 2 ( duplication 3 ) .
Slight induction of pocR transcription occurred during aerobic-growth on poor carbon sources ( Table 6 , aerobic conditions ) .
Transcription of the pocR gene increased threefold during aerobic-growth-on-succinate compared with that seen during aerobic-growth-on-glucose .
This effect did not require propanediol , glycerol , or a PocR + allele .
Anaerobiosis also increasedpocR transcription ( Table 6 , strain TT17166 ) .
Expression of apocR-lac fusion was 3-fold higher during fermentation of glucose than during aerobic-growth-on-glucose , and maximal induction occurred during anaerobic respiration of glycerol-fumarate ( 17-fold higher than during aerobic-growth-on-glucose ) .
Anaerobic induction did not depend on a PocR + allele ( Table 6 , strains TT17166 and TT17168 ) .
Transcription of the pocR gene shows autoinduction ( about threefold ) by propanediol during aerobic-growth-on-succinate and during anaerobic-growth on pyruvate-fuma-rate ( Table 6 ) .
This is termed autoinduction because it requires a pocR + allele ( provided by a genetic duplication which contained both the pocR-lac fusion and a pocR + allele ) .
No propanediol induction was seen in a haploid pocR : : lac fusion strain ( Table 6 , strain TT17166 ) or in a strain carrying a duplication with a TnlO insertion in the secondpocR allele ( Table 6 , strain TT17168 ) .
Propanediol metabolism was not required for autoinduction of the pocR gene .
A strain with apdu : : ThlO insertion in both copies of the duplicated material lacked the ` ability to metabolize propanediol but still showed propanediol induction of pocR gene transcription ( Table 6 , strain TT17169 ) In contrast to propanediol , glycerol did not mediate induction of pocR gene transcription even in a strain which provided a pocR + allele ( Table 6 ) .
Similar results were obtained with a glpK derivative of the pocR + strain .
This indicated that the failure of glycerol to induce pocR transcription was not due to aerobic metabolism of glycerol ( which prevented aerobic induction of the cob : : lac or pdu : : lac fusion as described above ) .
Although glycerol did not mediate aerobic autoinduction of pocR transcription , anaerobic expression of a pocR-lac fusion was maximal during-growth-on-glycerol-fumarate , suggesting that anaerobic metabolism of glycerol resulted in strong induction of pocR gene transcription .
This effect is independent of the PocR function and may be mediated by a global redox control system as suggested previously ( 3 ) .
Induction of CobIIacZ and CobII4cZ fusions by pro-panediol and glycerol .
The regulatory effects described above have been observed with a lac operon transcribed by the promoter for the CobI region of the cob gene cluster .
The same effects apply to lac fusions to more distal sites within the Coblll and CobIl regions of the gene cluster ( Table 7 ) .
These results suggest that the CobI , CobIII , and CobIl genes constitute a single operon regulated as described above .
Transcription of the CoblIl and CobII regions was induced by the same growth-conditions that induced expression of a CobI fusion , that is , during aerobic-growth-on-succinate plus propanediol and during anaerobic-growth-on-glycerol-fumarate or on pyruvate-fumarate plus propanediol .
Note , however , that the uninduced transcription levels of CobIII and CobIl fusions are higher than that of the CobI fusion and that the magnitude of the inductive effects is smaller ( Table 7 , strains TT10857 and TT10858 ) .
These results are attributed to weak promoters for the CobIII and CobII genes located within the operon and to a terminator that stops a significant fraction of transcripts originating at the regulated operon promoter at the far left end of the cluster .
To determine whether the internal promoters were regulated by propanediol , we constructed strains containing a CobI , a CobIll , or a CobIl MudJ insertion and an upstream CobI : : TnlOdTc insertion to block transcription from the left ( as seen in Fig. 3 ) .
The TnJO insertion completely blocks transcription , since it prevents expression of a CobI : : lac fusion located immediately downstream of the TnlO insertion site ( Table 7 , strain TT17160 ) .
Any expression of the distal CobIII and CobII genes in strains with this TnlO insertion must be due to the internal promoters .
The lac operon fusions in the CobIII and CobII regions are weakly expressed in strains with the CobI : : TnlOdTc insertion , but the residual expression is not regulated ( Table 7 , strains FT17195 and FT17194 ) .
This weak expression from internal promoters is sufficient to account for the CobIII + CobII + phenotype of CobI insertion mutants .
We infer that the entire cob gene cluster is a single operon controlled by the regulatory mechanisms discussed here , which vary transcription of the entire gene cluster by acting at the far left end of the cob region as the map is presented in Fig. 3 .
On the basis of sequence data , this single operon includes about 25 genes ( 10 , 30 ) .
propose a cob/pdu regulon ( Fig. 4 ) .
The key of that model aspects are listed below .
( i ) The pocR protein acts in trans as an activator of the cob and pdu operons and also activates transcription of its own gene .
Insertion mutations in thepocR gene prevent induction of the cob and pdu operons .
Such mutations are complemented equally well by a pocR + allele provided either in cis or in trans .
Read-through from the pocR promoter is not a major contributor to cob operon expression , since a cob : : lac fusion shows full induction in a strain that contains apocR + allele in trans and a strongly polar pocR insertion cis to the cob : : lac fusion .
Autoinduction ofpocR transcription is supported by the fact that propanediol induces transcription of a pocR-lac fusion in a PocR + merodiploid but not in an isogenic PocR-strain .
Since the pdu operon and the pocR gene are transcribed divergently and cob operon expression is independent of the pocR promoter , we infer that the pdu , pocR , and cob regions must be expressed from three distinct promoters .
All three are regulated by the PocR protein .
Sequencing data are consistent with the above interpretation .
The base sequence of the region between the pdu and cob operons reveals a reading frame ( read left to right as the map in Fig. 3 is drawn ) with an inferred protein product that is highly homologous to the araC family of transcriptional regulatory proteins ( 10 , 30 ) .
Furthermore , Richter-Dahlfors and Andersson have characterized a regulated promoter lying between the pocR and cob regions ( 28 ) .
( ii ) Both propanediol and glycerol can serve individually as effectors of the pocR protein .
The cob and pdu operons are most markedly induced in the presence of either propanediol or glycerol .
The metabolism of these inducers is not a requirement for induction .
Autoinduction ofpocR is caused only by propanediol , and this effect is also independent of propanediol metabolism .
It is surprising that either propanediol or glycerol can serve as effectors of PocR for induction of the pdu and cob operons , yet only propanediol is able to induce pocR transcription .
This suggests that glycerol and propanediol may stabilize different conformations of the PocR protein or possibly that additional components , such as the mediators of global control , are involved in autoregulation .
An alternative interpretation of the data is that a presented complex of propanediol and PocR protein activates pocR gene expression .
Glycerol would then serve as a PocR effector for induction of the pdu and cob operons when a sufficient level of PocR protein has been achieved .
Both models suggest that the PocR protein binds glycerol and propanediol ; these interactions have not yet been directly demonstrated .
( iii ) Expression of the pocR gene varies in response to global controls as well as autoregulation .
In strains with functional PocR protein , propanediol inducespocR gene transcription .
This autoinduction is seen only during-growth on poor carbon sources , suggesting that global controls as well as PocR protein are involved in autoinduction .
In the absence of PocR protein , the level of pocR gene transcription increases during-growth on poor carbon sources or under anaerobic conditions but propanediol has no inductive effect .
Compared with aerobic-growth of cells on glucose , aerobic-growth-on-succinate causes a 3-fold increase in pocR transcription and anaerobic respiration of glycerol-fumarate causes a 17-fold increase .
We suggest that these increases are due to global regulatory mechanisms that act on pocR gene transcription independently of the PocRmediated propanediol induction .
The global control mechanisms at work here have not yet been investigated .
Previous studies indicated that both ca-tabolite repression and redox control influence cob operon transcription ( 3 , 17 ) .
We present evidence that expression of both the cob and pdu operons requires the PocR protein ; induction of this regulon is improved under conditions that increase pocR gene expression ( see below ) .
Therefore , we propose that CRP-cAMP and redox control are two indepen dent global control mechanisms that regulate pocR gene transcription .
However , neither the direct involvement of CRP-cAMP nor the existence of the redox regulatory protein depicted in Fig. 4 has yet been demonstrated .
( iv ) Global control of the PocR protein level mediates global control of the coblpdu regulon .
There is a correlation between growth-conditions under which pocR gene transcription is high and growth-conditions under which propanediol and glycerol cause maximal induction of the coblpdu regulon ( Tables 2 , 3 , and 6 ) .
Moreover , during aerobic-growth of glpK strains on succinate , the coblpdu regulon is induced to higher levels by propanediol than by glycerol .
This coincides with the fact that propanediol but not glycerol autoinduces transcription of the pocR gene .
These correlations are consistent with the idea that global control signals increase expression of the pocR gene and thereby enhance induction of the cob/pdu regulon by propanediol and glycerol .
This idea explains why glycerol catabolism prevents aerobic induction of the coblpdu regulon ; decreased cAMP levels may prevent adequate expression of the pocR gene .
General points regarding the model .
Anaerobic induction of the cob operon by glycerol per se seems contradictory to the findings of Andersson and Roth ( 3 ) .
They observed maximum operon expression during anaerobic respiration of glycerol-fumarate and concluded that a high level of expression was due not to induction by glycerol but to a highly reduced cell interior achieved by rapid oxidation of glycerol and slow transport of electrons to fumarate .
They shifted glpK + and glpK mutant strains from glucose to glycerol anaerobically and saw partial induction of a cob-lac fusion only in the glpK + strains .
From these results the authors suggested that glycerol metabolism was needed for the anaerobic induction of the cob operon and that glycerol per se was not an inducer of the cob operon .
In this report , we suggest that glycerol is an inducer of the coblpdu regulon but that a high level of PocR protein is also required for induction .
We show that anaerobic respiration of glycerol or pyruvate allows a high level of pocR gene transcription .
Under our experimental conditions , glycerol induction of the cob and pdu operons was demonstrated in glpK cells that were actively engaged in anaerobic respiration of pyruvate-fumarate , a growth-condition that leads to relatively high levels ofpocR gene transcription .
We suggest that Andersson and Roth , in shifting a glpK mutant from glucose to anaerobic glycerol ( 3 ) , put cells under starvation conditions that did not allow induction ofpocR gene expression ; therefore , their experiments did not reveal the inductive effect of glycerol .
Previous studies indicated that transcription of the cob operon was induced only in the absence-of-oxygen ( 17 , 24 ) .
Subsequently , evidence was presented that a highly reduced cell interior ( rather than absence-of-oxygen per se ) was needed for induction ( 3 ) .
In this report , we show that the cob operon ( and the adjacent pdu operon ) can also be induced aerobically , but only if cells are grown on a poor carbon source and propanediol ( or glycerol ) is provided as an inducer .
We suggest that CRP-cAMP is a second global regulatory system ( in addition to redox control ) that can stimulatepocR gene transcription and thereby permit induction of the coblpdu regulon .
The cob region is a single regulated operon .
Previously it was concluded that the cob genes were arranged as three independent operons , each including genes of related function ( 17 ) .
This was based on the phenotypes of mutants .
CobI insertion mutants are blocked in formation of the intermediate Ado-Cbi but still express the promoter-distal CobIlI and CobII gene clusters which are responsible for synthesis of DMB ( CobII ) and assembly of Ado-Cbi and DMB into the cofactor-Ado-Cbl ( CoblIl ) ( 18 , 25 ) .
It is now clear that the entire Cob gene cluster ( about 25 genes ) is a single operon , regulated as a unit , with internal weak constitutive promoters that allow expression of distal genes arranged as logical groups within the operon .
This is dem-onstrated by the data in Table 7 and by additional unpublished data ( 14 ) .
The paradox of the coblpdu regulatory pattern .
At first glance , the regulatory behavior of the cob/pdu regulon makes physiological sense .
Since Ado-Cbl is required as a cofactor for degradation of propanediol , it seems logical for propanediol to serve as an inducer of the vitamin-B12 synthetic enzymes as well as the enzymes for propanediol degradation .
However , on closer inspection , some problems are presented both by the global regulatory pattern and by the ability of glycerol to serve as an inducer .
In fact , all of the described conditions that induce this regulon are apparently futile in that no condition allows cells to use propanediol as a carbon and energy source by means of endogenously synthesized vitamin-B12 .
S. typhimunum can use propanediol aerobically as a ( very poor ) source of carbon and energy but can not synthesize vitamin-B12 under these conditions , even when the cob operon is fully induced ( 2 , 24 ) .
It appears that some vitamin-B12-synthetic step is oxygen sensitive or that some critical synthetic enzyme ( outside the operon ) is not induced aerobically .
Therefore , aerobic induction of the coblpdu regulon provides enzymes for the degradation of propanediol but does not provide the needed cofactor .
Anaerobic induction of the regulon also appears futile .
Under these conditions , vitamin-B12 is synthesized and the propanediol-degradative genes are induced , but cells are unable to utilize propanediol as a sole carbon and energy source without oxygen to serve as an electron-acceptor .
Salmonellae can not ferment propanediol , as do klebsiellae ( 19 ) , nor can salmonellae use propanediol as a carbon source during anaerobic respiration with alternative electron-acceptors such as nitrate or fumarate ( 6 ) .
This apparently paradoxical regulatory pattern may make sense if salmonellae usually grow with oxygen levels insufficient to inhibit vitamin-B12 synthesis but sufficient to permit respiration of propanediol ( and other carbon sources ) .
Induction of the coblpdu regulon by glycerol is also hard to rationalize .
S. typhimurium shows no known vitamin-B12-dependent metabolism of glycerol , either aerobically or anaerobically .
Related enteric bacteria can ferment both glycerol and propanediol by vitamin-B12-dependent pathways .
In some organisms , a single diol dehydratase is able to perform the dehydration of both propanediol and glycerol ( 19 ) .
The ability of glycerol to induce the coblpdu regulon of S. typhimurium may be simply due to the structural similarity of glycerol to propanediol .
We suspect that the cob/pdu regulatory system makes physiological sense to salmonellae and that some surprises will be forthcoming before the logic is apparent to us .
APPENDIX Deletion mapping by dominant marker phenotypes .
The genetic map of the cob/pdu control region presented in this paper was constructed by a method that allows deletion mapping of any chromosomal region , even regions with mutations that cause no counterselectable phenotype .
This method could be applied to noncoding regions of the chromosome This method was employed here because mutations in the cob/pdu control region have no phenotypes to allow selection of wild-type recombinants ; therefore , we could not apply the standard deletion mapping method .
Mutant phenotypes for pdu are scored on MacConkey plates ; Pdu ' strains grow so poorly on minimal propanediol medium with vitamin-B12 that selection of recombinants is difficult .
The zeb : : TnlO insertions have no easily observable phenotypic defect .
Mutations in the pocR gene have a leaky Cob-phenotype .
The new mapping method employs as selective markers the drug resistance determinants associated with the donor and recipient insertion mutations .
We select for inheritance of two dominant mutant alleles rather than the standard counterselection of recessive mutant alleles .
A standard deletion mapping cross is diagrammed in Fig .
AlA ; both donor and recipient mutations cause a recessive counterselect-able phenotype , and one selects wild-type recombinants that lack both parental lesions .
If the donor point mutation lies within the region missing from the recipient deletion , no recombinants can form because both parents lack essential information corresponding to the site of the donor mutation .
Since the deletion can not revert , there is no background of revertants and sensitive crosses can be performed to distinguish between mutations lying within the deleted region ( no recombinants ) and mutations lying outside the deletion ( some recombinants ) .
This standard method could not be applied to the cob/pdu region .
Instead , we used insertion mutations which add drug resistance determinants to the chromosome .
These dominant phenotypes were used as selective markers in the mapping crosses .
Deletion mutations were created by recombination between two Mud insertions ; the deletion is inseparable from the drug resistance encoded by the Mud element ( Apr or Kn ' ) .
Donor mutations are insertions of TnlO derivatives and therefore provide a drug resistance phenotype ( Tcr or Cmr ) that is inseparable from the insertion mutation .
In scoring recombination between the donor point mutation ( insertion ) and the recipient deletion , we select for recombinants that inherit both the donor and recipient drug resistance markers .
This cross is diagrammed in Fig .
If the donor insertion lies within the recipient deletion , one might expect that the two mutations would be mutually exclusive and double mutants could not form , making this an effective mapping method .
However , notice that regardless of the position of the donor mutation , all of the genetic information needed for resistance to both drugs is provided in the cross .
Therefore , even if the insertion lies inside the deletion , recombinants will be observed if some genetic event permits addition of the donor element to the recipient chromosome without eliminating the recipient deletion and its associated drug resistance .
In practice , such events are observed at a low frequency even when the donor point mutation lies within the recipient deletion .
Therefore , all crosses lead to some frequency of recombinants .
Two mechanisms for addition of overlapping markers are diagrammed in Fig .
If the donor cell population includes , at low frequency , cells with a duplication whose join point is cotransducible with the donor selective recombination can add the donor marker to the marker , without chromosome removing the deletion which overlaps the site of the donor insertion ( Fig .
This event seems to explain the recombinants seen for deletions extending leftward from the pdu , zeb , orpocR locus to the his operon .
The few recombinants scored in these cases are His-and show both the donor and recipient drug resistances , but the donor drug resistance is lost with high frequency by segregation of the duplication ( exchanges between the repeated cd sequences noted in Fig .
The recipient drug resistance is never lost , because the deletion removes homology ab to the left of the donor insertion .
If the recipient strain carries a duplication that includes the site of the deletion mutation , then the donor insertion can replace the deletion in one copy of the duplication while the second copy of the deletion and its drug resistance are retained .
This is diagrammed in Fig .
These events appear to explain the few recombinants between donor insertions within the deleted region and deletions extending rightward from the pdu , zeb or pocR locus to CobIl .
These recombinants are Cob ' and show both donor and recipient drug resistances , but they frequently lose either one drug resistance marker or the other .
Which resistance marker is lost depends on which repeated sequences ( ab or cd in Fig .
A2 ) recombine to cause segregation .
Recombinants involving the his-cob deletions as recipients and donor insertions within the deleted region did not arise b this mechanism because these leftward deletions are too large to be repaired by a single transduced fragment .
Because of the events described above , recombinants are observed in all deletion mapping crosses , regardless of the position of the donor insertion .
However , mapping is still possible if one uses pairs of recipient deletions .
For each of the critical Mud insertions in the coblpdu region , we constructed a deletion extending to the right ( to the CobIl region ) and another extending to the left ( to the his operon ) .
The donor mutation must lie within one or the other of these deletions .
The recombination events that require involvement of a duplication are understandably rare , since duplications of any particular chromosomal site are typically carried in populations at a frequency of about 10 ' to 10 ' .
All of the donor insertion mutations mapped showed 100-to 1,000-fold more recombinants with one deletion than with the other .
The donor mutation was assigned to a location within the deletion whose cross gave fewer recombinants .
From all of the crosses yielding few recombinants , the recombinants found invariably showed instability of one of the sorts described above .
In addition , such recombinants showed cotransductional linkage between the Mud and TnlO elements that was reduced in comparison to the linkage seen in standard haploid strains ; this is consistent with the presence of duplications in such recombinants .
This work was supported by a grant from the National Institute of General Medical Sciences ( ROl GM34804 ) .
M.A. was supported in part by the Howard Hughes Undergraduate Initiative Program .
Some of the initial phases of this project were inspired and aided by three high school students ( all now undergraduates ) , Kathleen Treseder ( University of Utah ) , Jeffrey Cook ( Occidental College ) , and Sohail Tavazoie ( University of California , Berkeley ) .
The first pocR mutants were isolated by Kathleen Treseder .
Andersson , D. I. , and J. R. Roth .
Mutations affecting regulation of cobinamide biosynthesis in Salmonella typhimu-num .
Andersson , D. I. , and J. R. Roth .
Redox regulation of the genes for cobinamide biosynthesis in Salmonella typhimunum .
Balch , W. E. , G. E. Fox , L. J. Magrum , C. R. Woese , and R. S. Wolfe .
Methanogens : reevaluation of a unique biological group .
Berkowitz , D. , J. M. Hushon , H. J. Whitfield , Jr. , J. R. Roth , and B. N. Ames .
Procedure for identifying nonsense mutations .
Bobik , T. Unpublished results .
Casadaban , M. J. , and S. N. Cohen .
Lactose genes fused to exogenous promoters in one step using a Mu-lac bacteriophage : an in-vivo probe for transcriptional control sequences .
Castilho , B. A. , P. Olfson , and M. J. Casadaban .
Plasmid insertion mutagenesis and lac gene fusion with mini-Mu bacteriophage transposons .
Chan , R. K. , D. Botstein , T. Watanabe , and Y. Ogata .
Specialized transduction of tetracycline resistance by phage P22 in Salmonella typhimurium .
Properties of a high-frequency-transducing lysate .
Chen , P. , and J. R. Roth .
Chumley , F. G. , R. Menzel , and J. R. Roth .
Hfr formation directed by TnlO .
Chumley , F. G. , and J. R. Roth .
Rearrangement of the bacterial chromosome using TnlO as a region of homology .
Davis , R. W. , D. Botstein , and J. R. Roth .
A manual for genetic engineering : advanced bacterial genetics .
Cold Spring Harbor Laboratory , Cold Spring Harbor , N.Y. 14 .
Doak , T. , and J. R. Roth .
Elliott , T. , and J. R. Roth .
Characterization of TnlOdCAM : a transposition-defective TnlO specifying chloramphenicol resistance .
Escalante-Semerena , J. C. , M. G. Johnson , and J. R. Roth .
The Cobll and CobIlI regions of the cobalamin ( vitamin-B12 ) biosynthetic operon of Salmonella typhimurium .
Escalante-Semerena , J. C. , and J. R. Roth .
Regulation of cobalamin biosynthetic operons in Salmonella typhimurium .
Escalante-Semerena , J. C. , S.-J .
Suh , and J. R. Roth .
cobA function is required for both de novo cobalamin biosynthesis and assimilation of exogenous corrinoids in Salmonella typhi-murium .
Forage , R. G. , and M. A. Foster .
Glycerol fermentation in Klebsiella pneumoniae : functions of the coenzyme B12-depen-dent glycerol and diol dehydratases .
Hughes , K. , and J. R. Roth .
Directed formation of deletions and duplications using Mud ( Ap , lac ) .
Hughes , K. T. , and J. R. Roth .
Conditionally transposi-tion-defective derivative of Mu dl ( Amp Lac ) .
Hughes , K. T. , and J. R. Roth .
Transitory cis complementation : a method for providing transposition functions to defective transposons .
Cobalamin-dependent 1,2-propanediol utili-zation by Salmonella typhimurium .
Jeter , R. M. , B. M. Olivera , and J. R. Roth .
Salmonella typhimurium synthesizes cobalamin ( vitamin-B12 ) de novo under anaerobic-growth-conditions .
Jeter , R. M. , and J. R. Roth .
Cobalamin ( vitamin-B12 ) biosynthetic genes of Salmonella typhimurium .
Kleckner , N. , D. Botstein , and J. R. Roth .
Genetic engineering in-vivo using translocatable drug-resistance ele ¬ 27 .
Experiments in molecular genetics .
Cold Spring Harbor Laboratory , Cold Spring Harbor , N.Y. 28 .
Richter-Dahlfors , A. A. , and D. I. Andersson .
Analysis of an anaerobically induced promoter for the cobalamin biosynthetic genes in Salmonella typhimunum .
Rondon , M. R. , and J. Escalante-Semerena .
Thepoc locus is required for 1,2-propanediol-dependent transcription of the cobalamin biosynthetic ( cob ) and propanediol utilization ( pdu ) genes of Salmonella typhimunium .
Roth , J. R. , S. Kieffer-Higgins , M. Rubenfield , and G. Church .
A method for detection of phage mutants with altered transducing ability .
Toraya , T. , S. Honda , and S. Fukui .
Fermentation of 1,2-propanediol and 1,2-ethanediol by some genera of Entero-bacteriaceae , involving coenzyme B12-dependent diol dehydratase .
Vogel , H. J. , and D. M. Bonner .
Acetylornithase of Escherichia coli : partial purification and some properties .
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 .