Vol. 174, No. 3 JOURNAL OF BACTERIOLOGY, Feb. 1992, p. 889-898 0021-9193/92/030889-10$02.00/0 Copyright C 1992, American Society for Microbiology Molecular Cloning and Physical Mapping of the otsBA Genes, Which Encode the Osmoregulatory Trehalose Pathway of Escherichia coli: Evidence that Transcription Is Activated by KatF (AppR) INGA KAASEN, PAL FALKENBERG, OLAF B. STYRVOLD, AND ARNE R. STR0M* The Norwegian College of Fishery Science, University of Troms0, N-9000 Tromso, Norway Received 16 July 1991/Accepted 26 November 1991 It has been shown previously that the otsA and otsB mutations block osmoregulatory trehalose synthesis in Escherichia coli. We report that the transcription of these osmoregulated ots genes is dependent on KatF (AppR), a putative sigma factor for certain stationary phase- and starvation-induced genes. The transcription of the osmoregulated bet and proU genes was not katF dependent. Our genetic analysis showed that katF carries an amber mutation in E. coli K-12 and many of its derivatives but that katF has reverted to an active form in the much-used strain MC4100. This amber mutation in katF leads to strain variations in trehalose synthesis and other katF-dependent functions of E. coli. We have performed a molecular cloning of the otsBA genes, and we present evidence that they constitute an operon encoding trehalose-6-phosphate phosphatase and trehalose-6- phosphate synthase. A cloning and restriction site analysis, performed by comparing the cloned fragments with the known physical map of the E. coli chromosome, revealed that the otsBA genes are situated on a 2.9-kb HindIII fragment located 8 to 11 kb clockwise of tar (41.6 min). Trehalose, a nonreducing disaccharide of glucose, is a stress metabolite in various organisms (29). Saccharomyces cerevisiae accumulates trehalose when exposed to an ele- vated temperature of growth (3, 24) or to hazardous chemical agents such as ethanol, copper sulfate, or hydrogen peroxide (3). Rhizobia accumulate trehalose when stressed with low- oxygen (e.g., 1%) tension, regardless of the composition of the growth medium (23). Many phototrophic and heterotro- phic bacteria, including Escherichia coli, accumulate trehalose in response to osmotic stress (18, 31, 45, 50). Trehalose is shown to preserve the function and integrity of biological membranes exposed to conditions of low water activity (14) and to confer desiccation tolerance to yeasts (24), to spores of Streptomyces sp. (36), and to nematodes (14); frost tolerance to insects (2) and yeasts (22); and osmotic tolerance to E. coli (19). In yeasts, trehalose accu- mulation during growth in liquid culture coincides with an increased plating efficiency on agar plates of low water activity (34). In E. coli, the osmoregulatory trehalose pathway consists of a trehalose-6-phosphate synthase which converts UDP- glucose and glucose-6-phosphate to trehalose-6-phosphate and a phosphatase which dephosphorylates this metabolic intermediate (reference 19 and this study). Two insertion mutations, named otsA and otsB, which block the synthesis of the synthase, have previously been mapped to 42 min, but the trehalose-6-phosphate phosphatase activity of these mu- tants was not reported (19). However, a point mutation named otsP which causes accumulation of trehalose-6-phos- phate in stressed cells, presumably because of a defective phosphatase, was mapped near otsA (27). This mutation appears to be allelic with otsB (reference 27 and this study). Trehalose accumulation in stressed cells of E. coli is regulated at several levels. Experiments with lac fusions have shown that the otsA and otsB genes are transcription- * Corresponding author. ally activated by osmotic stress, and biochemical data have shown that the synthase is activated by potassium glutamate (19). At the cellular level, trehalose accumulation is regu- lated by a futile cycle. Trehalose is overproduced in the cytoplasm, and the excess is excreted and split to glucose by a periplasmic trehalase (TreA) and then reutilized (52). Similarly to the ots genes, the treA gene is transcriptionally activated by osmotic stress but not by external trehalose (8). treA maps at 26 min (8). Rod et al. (47) reported that E. coli K-12 and most of its derivatives carry an amber mutation that leads to decreased accumulation of trehalose in osmot- ically stressed cells. Styrvold and Str0m (52) showed that this amber mutation was not in the otsA and otsB genes themselves but that it caused decreased transcription of these genes. Trehalose can serve as the sole source of energy and carbon in E. coli. At high osmolarity, trehalose utilization depends on the TreA activity (20). But at low osmolarity, trehalose induces a transport protein, ElITre (TreB), and a pathway consisting of at least a catabolic trehalose-6-phos- phate phosphatase (TreE) and an amylotrehalase (TreC). treB, treC, and presumably treE map at 96.5 min (9). Their internal inducer is trehalose-6-phosphate; therefore, these genes are not expressed under osmotic stress when the osmoregulatory trehalose-6-phosphate phosphatase is oper- ative (27). A mutation which seems to influence expression of otsA, otsP (otsB), and treC, but not treB, has been mapped to 84 min, but the nature of this gene remains unclear (27). In this investigation, we have cloned the otsBA genes. We report that they constitute an operon which encodes the phosphatase and the synthase, respectively, of the osmoreg- ulatory trehalose pathway. The amber mutation which influ- ences otsBA expression was mapped at 59 min and shown to be in katF, which encodes a putative sigma factor (39) for starvation- and stationary phase-induced genes (30, 37). 889 on May 22, 2021 by guest http://jb.asm.org/ Downloaded from
Transcript
Vol. 174, No. 3JOURNAL OF BACTERIOLOGY, Feb. 1992, p. 889-8980021-9193/92/030889-10$02.00/0Copyright C 1992, American Society for Microbiology
Molecular Cloning and Physical Mapping of the otsBA Genes,Which Encode the Osmoregulatory Trehalose Pathway of
Escherichia coli: Evidence that Transcription IsActivated by KatF (AppR)
INGA KAASEN, PAL FALKENBERG, OLAF B. STYRVOLD, AND ARNE R. STR0M*The Norwegian College of Fishery Science, University of Troms0, N-9000 Tromso, Norway
Received 16 July 1991/Accepted 26 November 1991
It has been shown previously that the otsA and otsB mutations block osmoregulatory trehalose synthesis inEscherichia coli. We report that the transcription of these osmoregulated ots genes is dependent on KatF(AppR), a putative sigma factor for certain stationary phase- and starvation-induced genes. The transcriptionof the osmoregulated bet and proU genes was not katF dependent. Our genetic analysis showed that katF carriesan amber mutation in E. coli K-12 and many of its derivatives but that katF has reverted to an active form inthe much-used strain MC4100. This amber mutation in katF leads to strain variations in trehalose synthesis andother katF-dependent functions of E. coli. We have performed a molecular cloning of the otsBA genes, and wepresent evidence that they constitute an operon encoding trehalose-6-phosphate phosphatase and trehalose-6-phosphate synthase. A cloning and restriction site analysis, performed by comparing the cloned fragments withthe known physical map of the E. coli chromosome, revealed that the otsBA genes are situated on a 2.9-kbHindIII fragment located 8 to 11 kb clockwise of tar (41.6 min).
Trehalose, a nonreducing disaccharide of glucose, is astress metabolite in various organisms (29). Saccharomycescerevisiae accumulates trehalose when exposed to an ele-vated temperature ofgrowth (3, 24) or to hazardous chemicalagents such as ethanol, copper sulfate, or hydrogen peroxide(3). Rhizobia accumulate trehalose when stressed with low-oxygen (e.g., 1%) tension, regardless of the composition ofthe growth medium (23). Many phototrophic and heterotro-phic bacteria, including Escherichia coli, accumulatetrehalose in response to osmotic stress (18, 31, 45, 50).Trehalose is shown to preserve the function and integrity ofbiological membranes exposed to conditions of low wateractivity (14) and to confer desiccation tolerance to yeasts(24), to spores of Streptomyces sp. (36), and to nematodes(14); frost tolerance to insects (2) and yeasts (22); andosmotic tolerance to E. coli (19). In yeasts, trehalose accu-mulation during growth in liquid culture coincides with anincreased plating efficiency on agar plates of low wateractivity (34).
In E. coli, the osmoregulatory trehalose pathway consistsof a trehalose-6-phosphate synthase which converts UDP-glucose and glucose-6-phosphate to trehalose-6-phosphateand a phosphatase which dephosphorylates this metabolicintermediate (reference 19 and this study). Two insertionmutations, named otsA and otsB, which block the synthesisof the synthase, have previously been mapped to 42 min, butthe trehalose-6-phosphate phosphatase activity of these mu-tants was not reported (19). However, a point mutationnamed otsP which causes accumulation of trehalose-6-phos-phate in stressed cells, presumably because of a defectivephosphatase, was mapped near otsA (27). This mutationappears to be allelic with otsB (reference 27 and this study).
Trehalose accumulation in stressed cells of E. coli isregulated at several levels. Experiments with lac fusionshave shown that the otsA and otsB genes are transcription-
* Corresponding author.
ally activated by osmotic stress, and biochemical data haveshown that the synthase is activated by potassium glutamate(19). At the cellular level, trehalose accumulation is regu-lated by a futile cycle. Trehalose is overproduced in thecytoplasm, and the excess is excreted and split to glucose bya periplasmic trehalase (TreA) and then reutilized (52).Similarly to the ots genes, the treA gene is transcriptionallyactivated by osmotic stress but not by external trehalose (8).treA maps at 26 min (8). Rod et al. (47) reported that E. coliK-12 and most of its derivatives carry an amber mutationthat leads to decreased accumulation of trehalose in osmot-ically stressed cells. Styrvold and Str0m (52) showed thatthis amber mutation was not in the otsA and otsB genesthemselves but that it caused decreased transcription ofthese genes.
Trehalose can serve as the sole source of energy andcarbon in E. coli. At high osmolarity, trehalose utilizationdepends on the TreA activity (20). But at low osmolarity,trehalose induces a transport protein, ElITre (TreB), and apathway consisting of at least a catabolic trehalose-6-phos-phate phosphatase (TreE) and an amylotrehalase (TreC).treB, treC, and presumably treE map at 96.5 min (9). Theirinternal inducer is trehalose-6-phosphate; therefore, thesegenes are not expressed under osmotic stress when theosmoregulatory trehalose-6-phosphate phosphatase is oper-ative (27). A mutation which seems to influence expressionof otsA, otsP (otsB), and treC, but not treB, has beenmapped to 84 min, but the nature of this gene remainsunclear (27).
In this investigation, we have cloned the otsBA genes. Wereport that they constitute an operon which encodes thephosphatase and the synthase, respectively, of the osmoreg-ulatory trehalose pathway. The amber mutation which influ-ences otsBA expression was mapped at 59 min and shown tobe in katF, which encodes a putative sigma factor (39) forstarvation- and stationary phase-induced genes (30, 37).
Bacterial strains and growth conditions. The bacterialstrains used are listed in Table 1. The growth media (medium63 [38], low-osmolarity minimal medium [LOM] [11], andLB medium [38]) and the growth conditions used wereexactly as described previously by us (52).
Genetic procedures and strain constructions. Conjugationby mating on agar plates and P1 transduction were per-formed as described by Miller (38). When transferring TnJOinsertions and lacZ fusions generated with A placMu, weselected for resistance against tetracycline and kanamycin,respectively. TnJO was deleted by using the positive selec-tion method described by Bochner et al. (7). The presence ofa recA mutation was tested by determining UV sensitivity asdescribed previously (35). For strains carrying otsA-lacZ orotsB-iacZ operon fusions, the otsX (katF) genotype wasroutinely checked by scoring their Lac phenotype by growthon medium 63 with lactose as an energy source (see below).The pedigrees of the strains constructed in this study aregiven in Table 1. Further description of strain construction isgiven in this section and in Results.A culture of MC4100 carrying a random selection of TnlO
insertions was prepared by infecting MC4100 with the phageX NK561 essentially as described previously (49). After thetransposition, the cells were spread on 50 agar plates con-taining LB medium with 25 ,ug of tetracycline per ml and 4mM sodium citrate. Approximately 50,000 Tcr coloniesrepresenting individual transposition events were pooled andused for production of P1 lysate.We verified that our strain of MC4100 did not carry an
amber suppressor in the following way. A TnJO insertionwas introduced near an amber mutation by transducingBL50 [metF(Am)] with a lysate prepared from GD1 (zih::TnJO) by selecting for Tcr and screening for Met-. A P1lysate of a resulting strain, PF11 [metF(Am) zih::TnJO], wasthen used to transduce MC4100 to Tcr. Of 50 Tcr transduc-tants tested, 19 were Met-, showing that the metF(Am)mutation was not suppressed in MC4100.FF1005 was constructed by conjugational crossing of
JC10240 with MC4100 by selecting for srl::TnJO (Tcr) andscreening for recA (UVS). By also checking that the trans-ductant gave an osmotically tolerant phenotype on agarplates with medium 63 with 0.5 M NaCl added, we ensuredthat FF1005 carried the otsX+MC4100 allele of the recipientand not the otsX allele of the donor.
In order to verify that LCB107 carries a supE mutationand to construct a suppressor-free LCB107 derivative, a P1lysate prepared from BL50 [metB+ metF(Am)] was first usedto transduce the LCB107-derived IK5 (metBl metF+ supE?)to Met+. As metB and metF are neighbor genes on thechromosome (4), Met+ transductants that carry an ambersuppressor could have either a metB+ metF(Am) or a metB+metF+ genotype. To investigate their genotype, four trans-ductants from this cross were transduced to Tcr with a P1lysate prepared from JW1071 (supo zbf::TnJO), which carriesa TnlO marker within cotransductional distance of glnV, i.e.,the native supE gene (4). The progeny from each of the fourcrosses displayed both Met+ (30%) and Met- (70%) geno-types. This showed that their parental strains were supEmetF(Am). One of the Met+ transductants of strain IK5 wasnamed IK58. Strain IK59 was a supo metF(Am) derivative ofIK58.To introduce the OtSX+MC4l00 allele into an LCB107
background, we first transduced the LCB107 derivativeIK58 with a P1 lysate prepared from IK68 (cysC mutS::
TnJO), selecting for Tcr and screening for Cys-. The result-ing strain, IK60, was transduced with a P1 lysate preparedfrom FF1112 (cysC+ mutS+ otsX+Mc4l00), selecting forCys+ and screening for Tcs. As otsX is located between thecysC and the mutS genes (see below), a Cys+ Tcs transduc-tant such as IK61 should carry the otsX MC4100 allele.
General recombinant DNA procedures and cloning of otsAand otsB. Transformations were done by the method ofChung et al. (13) or a standard CaCl2 method. Plasmidisolation, restriction cleavage of DNA, and ligation weredone by standard methods.
Restriction maps of the chromosomally derived parts ofthe constructed plasmids are .shown in Fig. 2, and theplasmids are listed in Table 1. A gene library of strain CSH7was constructed in the pBR322-derived cosmid vector cos4(46) as described by Andresen et al. (1). Plasmids wereintroduced into the osmotically sensitive strain FF4169(otsA: :TnlO), and osmotically tolerant, trehalose-producingclones were selected by plating on medium 63 with 0.45 MNaCl added. The plasmid pFF1, selected in this way,contained a 43-kb chromosomal fragment (see Fig. 2). Theleft end of the fragment is proximal to the SalI site of thevector. The plasmid pFF1 was maximally shortened by Sallto give pFF101 and by BamHI to give pFF102. A 2.9-kbHindIII fragment from pFF101 was subcloned into theHindIII site ofpACYC184 (12) to give the subclone pFF106.The left HindIII site of the chromosomal insert of pFF106(see Fig. 2) is proximal to the ClaI site of the vector. HindIIIshortening of pFF102 yielded pFF109. In the construction ofpFF114, plasmid pFF106 was shortened by EcoRV so that0.8 kb of the insert and 0.2 kb of the vector were deleted.
Trehalose-6-phosphate synthase and trehalose-6-phosphatephosphatase activities. The trehalose-6-phosphate synthaseand phosphatase enzyme activities were measured in vitroby using cells permeabilized with toluene (19). In the syn-thase assay, trehalose-6-phosphate formed from UDP-glu-cose and glucose-6-phosphate was dephosphorylated by atreatment with alkaline phosphatase (19, 52) and thetrehalose formed was trimethylsilylated and determined bygas chromatography exactly as described previously (19).The standard reaction mixture for determination of thephosphatase activity contained (in a 150-,ul volume) thefollowing: 1.5 ,umol of trehalose-6-phosphate, 5 ,umol ofTris-hydrochloride (pH 7.4), 0.4 ,umol of MgCl2, and 125 to500 ,ug of cell protein. The reaction mixture was incubatedup to 12 min at 37°C, and the reaction was terminated byheating for 5 min in a boiling-water bath. Sucrose (0.25 ,molin a 25-,ul volume) was then added as an internal standard,and denatured protein was removed by centrifugation. Fordesalting, a sample of 150 ,ul was applied to a column (0.5 by2 cm) packed with equal amounts of Dowex 50X4-200 in H+form and Dowex 1X8-400 in formate form. Free sugars werewashed through the column with 1 ml of water, and theeluate was freeze-dried. Gas chromatographic determinationof trimethylsilylated trehalose was then performed exactlyas described previously (19). One unit of synthase or phos-phatase activity equals 1 nmol of trehalose produced per minat 370C.Other methods and special chemicals. Cell protein was
determined by the biuret method as modified for wholebacterial cells (21). The p-galactosidase activity of lacZfusion mutants was determined quantitatively by using cellstreated with chloroform and sodium dodecyl sulfate asdescribed previously (19, 38). One unit of 3-galactosidaseactivity equals 1 nmol of o-nitrophenol produced per min at28°C. Catalase activity and glycogen accumulation were
38D. P. Clark (47)D. P. Clark (47)D. P. Clark (47)CGSC 7049MC4100 x JC10240191919P1(FF1608) x FF41691919P1(FF2032) x FF401919Tcs Kms of FF4012P1(UE5) x FF4031Tc' of FF4037P1(FF1005) x FF4049P1(IK26) x FF4035Tcs of FF4055P1(FF4171) x FF40561952B. LowskyP1(DC906) x FF1112Tcs of OS18P1(FF2032) x IK2P1(FF1112) x IK2Tc' of OS36P1(IK10) x FF1112P1(FF1005) x PF2P1(N3002) x FF1112P1(IK6) x IK4P1(ES1481) x IK4P1(IK12) x IK4P1(IK12) x IK4P1(LCB107) x IK25P1(DC943) x PF2P1(IK26) x FF2032P1(FF1112) x IK43P1(FF4169) x FF2032P1(FF4169) x IK44MLE914 x IK45MLE914 x IK46MLE413 x IK45MLE413 x IK46P1(BL50) x IK5P1(JW1071) x IK58P1(IK68) x IK58P1(FF1112) x IK60P1(JW1071) x IK61P1(RH90) x FF1112P1(ES1481) x KM78P1(RH90) x IK4P1(RH90) x IK41CGSC 6074CGSC 6391
Plasmidscos4 Apr 46pACYC184 Tcr Cmr 12pFF1 Apr; vector cos4 This studypFF101 Apr, vector cos4 This studypFF102 Apr; vector cos4 This studypFF106 Cmr; vector pACYC184 This studypFF109 Apr; vector cos4 This studypFF114 Cmr; vector pACYC184 This study
a otsX is allelic with katF, appR, csi-2, and rpoS. otsX carries an amber mutation in many K-12 strains (see text). The otsX genotype is shown where it isrelevant, and when known, the origin of the otsX allele is indicated by a subscript. All strains described as Alac carry a deletion generated by selecting a Tcsderivative from a strain carrying A(argF-Iac)U169 zah-735::TnIO. The A(argF-lac)UI69 deletion is 100%o cotransducible with the transposon (48). The symbol 4)indicates that the strain carries a lacZ operon fusion generated by a A placMu53 or A placMu55 insertion.
b CGSC, E. coli Genetic Stock Center, Yale University, New Haven, Conn. All CGSC strains were obtained from B. J. Bachmann.c These strains carry a A placMuS5 insertion in which the kanamycin resistance marker of the prophage is deleted together with otsA::TnlO.d We do not know whether the A(cys::TnIO) deletion encompasses the otsXwl 5^ allele.The katF allele rpoS359::TnIO is identical with csi-2::TnIO. This mutation was generated in MC4100 (30).
assayed on agar plates (30). Trehalose-6-phosphate was a giftfrom Marine DNA, Troms0, Norway.
RESULTS
Identification of otsX. Most suppressor-free K-12 strains(e.g., sup' derivatives of W1485) display a lower expressionof otsA-lacZ and otsB-lacZ operon fusions than their coun-terparts with an amber suppressor (i.e., supD, supE, orsupF). However, the much-used sup0-containing strainMC4100 is an exception; its derivatives display the samehigh expression of ots-lacZ fusions as strains containingsup' (52). To make certain that our strain of MC4100 did notdisplay a high expression of the otsA and otsB genes becauseof an unknown amber suppressor, we verified that a met-F(Am) mutation was not suppressed in this genetic back-ground (see Materials and Methods).
In order to insert TnJO near the gene which activated otsAand otsB in MC4100, we prepared a culture of MC4100 withrandom TnlO insertions. A P1 lysate grown on this culturewas then used to transduce PF2 [W1485 sup' otsB-lacZcysI(Am)], which displayed a low expression of its lacZfusion, to Tcr. The Tcr transductants were scored for growthon plates with medium 63-lactose. On this agar medium,strains with high or low expression of lacZ fusions in otsA orotsB are known to display a Lac' or Lac- phenotype,respectively (52). Out of 800 Tcr transductants tested, 6 wereLac'.
Our rationale for using PF2 [cysl(Am)] as the recipient inthis cross was to detect whether a Lac' phenotype of thetransductants was caused by a spontaneous amber suppres-sor mutation. But fortuitously, all six Tcr Lac' transduc-tants selected appeared to be Cys+. When P1 lysates pre-pared from these Tcr Cys+ transductants were used totransduce DC931 [cysl(Am) lacZ(Am)] to Tcr, we obtainedTcr Cys+ transductants at a high frequency, but none wasTcr Lac'. Furthermore, when the same lysates were used totransduce AT2427, which carries a cysJ point mutation, wealso obtained Tcr Cys+ transductants. Thus, these crossesestablished that the gene in MC4100 that had the ability toprovoke a high expression of the lacZ fusions in a sup'background maps near cysIJ at 59 min. We tentativelynamed this gene otsX, and for clarity we name the MC4100allele otsX+MC4100 and the W1485 allele otsXw1485. In otherwords, alleles that caused elevated transcription of the otsAand otsB genes in a supo background are designated otsX+.
otsX genotype and strain construction. In order to know theorigin of the otsX allele of the constructed strains, we firstchecked the otsX genotype of the basic strains. This wasdone by transducing their otsX allele together with a neigh-bor selectable marker into two recipient strains, one carryingthe otsX+MC4l,0 allele and the other the otsXW1485 allele.Both recipients were sup0 Alac and carried either otsA-lacZor otsB-lacZ. When the donor and recipient had the sameotsX genotype, all transductants had the same Lac pheno-type as the recipient. When the donor and the recipient had
GENES GOVERNING TREHALOSE SYNTHESIS IN E. COLI 893
58
sri mutS otsX cys
+ t i59
fuc
60 min
0.44
0.01
0.58
0.134
FIG. 1. Mapping of otsX by cotransductional analysis. The val-ues given on the arrows show the P1 cotransduction frequenciesbetween otsX and various markers in the 59-min region. The arrowspoint at the selected marker, which in all cases was a TnIO insertion.All recipients were Alac and carried an otsB-iacZ operon fusion, andthe otsX genotype of the Tcr transductants was scored by determin-ing their Lac phenotype on agar plates with medium 63-lactose.From each cross, at least 100 Tcr transductants were screened.Cross A, donor IK12 (cys::TnJO otsX+Mc4l00) and recipient IK4(OtSXw1485); cross B, donor IK1 (fuc::TnlO otsX MC4100) and recip-ient PF2 (OtSXw148s); cross C, donor IK21 (mutS::TnlO OtSXw1485)and recipient FF1112 (otsXMC4100); cross D, donor IK6 (srl::TnIOotsX+Mc4,00) and recipient IK4 (OtSXw14s5)-
differing types of otsX alleles, the transductants would be ofboth Lac phenotypes. By setting up crosses in which theotsX genotype of the donor and the recipient differed, wecould always pick a progeny with the desired otsX allele.This strategy was therefore used in the strain constructionslisted in Table 1. Beside MC4100, we found that ES1481carries an otsX+ allele, whereas JC10240, N3002, and DC906are otsX.
otsX+-containing strains that also were otsA+ otsB+ couldbe distinguished from isogenic otsX mutants by their osmot-ically somewhat more tolerant phenotype, but this pheno-type was not strong enough to be used for positive selectionin a transductional cross. We also avoided a direct selectionfor a Lac' phenotype of otsA-lacZ or otsB-lacZ mutants,because strains carrying a fusion together with an otsX allelegave rise to spontaneous Lac+ mutants when streaked onagar medium with lactose as an energy source. Ten suchmutations, examined by transductional analysis, appeared tobe linked to the fusion; thus, they seemed to be promotermutations (data not shown).
Transductional mapping of otsX. We constructed strainswhich carried fuc::TnJO, cysC::TnJO, mutS::TnlO, orsrl: :TnJO together with either the otSX+Mc410 or theOtSXwi485 allele. These strains were used as donors intransductional crosses in which the recipient carried an otsXallele of the opposite category and an otsB-lacZ fusion. Thetransductants were selected for the TnWO marker (i.e., Tc9)and the cotransduction frequency of otsX and TnJO wasdetermined by scoring the Lac phenotype. The four crossesoutlined in Fig. 1 placed otsX almost exactly at 59 min.
In a three-factor cross, a P1 lysate prepared from strainIK68 (otsX+ mutS: :TnlO cysC) was used to transduce strainIK4 (otsXW1485 otsB-lacZ) to Tcr. Of 175 transductantstested, 54 had received only mutS::TnJO, 89 had receivedmutS::TnJO and otsX+, and 32 had received mutS::TnJO,otsX+, and cysC. The lack of transductants which hadreceived mutS::TnJO and cysC, but not otsX+, showed thatthe gene order was mutS-otsX-cysC. This is in accordancewith data presented in Fig. 1.
In addition to otsX, katF (32), appR (53), and csi-2 (rpoS)(30) have been mapped to 59 min (4). It has been shown that
TABLE 2. Influence of the otsX (katF) allele and geneticbackground on P-galactosidase expression from
a All strains are AIac.b One unit of enzyme activity is 1 nmol of o-nitrophenol formed per min at
28°C. Each value is an average of at least three independent measurements.Standard deviations were within ±15%.
katF is allelic with appR (54), as well as with csi-2 (rpoS)(30), and our results show that otsX is another allele of thislocus (see below).
Influence of the otsX allele and the genetic background onotsA and otsB expression. To measure quantitatively theinfluence of the otsX alleles on the expression of otsA andotsB, we assayed the ,-galactosidase activity of lacZ fusion-containing strains with known genetic backgrounds. Theactivity was assayed after growth in a medium with orwithout 0.3 M NaCl added. otsA-lacZ or otsB-lacZ fusions inthe same genetic background displayed similar values for,-galactosidase (19, 52). Therefore, in most cases only onetype of fusion is included for each combination of geneticbackground and otsX allele (Table 2).
In accordance with data presented before (52), strain IK4,which is a suppressor-free strain with a W1485 backgroundand the native otsXW1485 allele, displayed low P-galactosi-dase activity. Furthermore, the activity was much higher inW1485 derivatives that carried either a supE mutation or theOtSX+MC4100 allele, i.e., IK41 and IK19, respectively (Table2).The difference between W1485 and MC4100 derivatives
with respect to their expression of otsA-lacZ and otsB-lacZfusions was solely due to the otsX alleles, since the geneticbackground, MC4100 or W1485, was of no importance whenthe otsX allele remained the same. Thus, IK44 (MC4100OtSXw148s) displayed a P-galactosidase activity similar tothat of IK3 (W1485 otsXW1485), and FF1112 (MC4100otsX+Mc4l00) displayed an activity similar to that of IK19(W1485 otSX+MC4100) (Table 2).We have reported previously that LCB107 derivatives
display a higher expression of otsA and otsB than MC4100derivatives, and we have concluded that this property is notdue to the otsA and otsB genes themselves (52). The otsXallele of LCB107 was not the cause of the elevated expres-sion of the otsA and otsB genes, since the isogenic otsB-lacZfusion-containing strains IK4 (W1485 otsXW1485) and IK35
(W1485 otsXLcB107) displayed similar low ,B-galactosidaselevels.From the construction of LCB107, it was uncertain
whether it carried supE44 (5). Our study showed thatLCB107 carries a supE mutation (see Materials and Meth-ods). The supo-containing strain IK59, derived fromLCB107, displayed strongly reduced expression of the otsA-lacZ fusion compared with that of the parental supE-con-taining strain IK58. However, supE was not responsiblefor the superexpression of the otsA and otsB genes inLCB107 (i.e., expression above the MC4100 level), sinceIK63 (LCB107 sup' otsX+Mc4l00) expressed the otsA-lacZfusion at the same very high level as IK58 (LCB107 supEotsXLcBl07). Apparently, LCB107 carries a mutation in an asyet unidentified gene which influences the expression of theotsA and otsB genes.
otsX is allelic with katF and the amber-mutated gene of E.coli K-12. We found that all sup'-containing strains listed inTable 2 carrying the otsXW1485 or the otSXLCB107 allele, andalso E. coli K-12 itself, displayed the glycogen-negative andcatalase-negative phenotype that is found in katF mutants(30), while all strains containing OtSX+MC4100 or supE did notshow these defects. Also, when a katF::TnJO mutation wasintroduced into FF1112 (MC4100 supo otsX+Mc4l00), theresulting strain, IK65, did not display the high expression ofthe lacZ fusion seen in the parental strain. In fact, theP-galactosidase activities found were even lower than that ofIK4 (W1485 sup' otsXW1485), indicating that the katF::TnJOmutation reduced otsA and otsB gene transcription evenmore than otsXW1485 (Table 2). The simplest explanation ofthe present data is that otsX, katF, and the amber mutationdescribed previously (47, 52; see above) are allelic and thatthis gene has reverted to an active form in MC4100.An alternative explanation is that the amber mutation was
situated in another locus and was suppressed by the otsX+(katF+) allele of MC4100. If so, the otsX+MC4100 allele andtRNA amber suppressors would represent two differentroutes for intergenic suppression of the amber mutation.However, the latter explanation was ruled out by the findingthat IK70 (W1485 supE katF: :TnJO) expressed the otsA-lacZfusion at the same low level as IK69 (W1485 supokatF: :TnlO) (Table 2). Apparently, in these strains thekatF::TnlO allele had replaced the amber-mutated gene sothat suppression via supE was no longer possible.
It is well known that many K-12 strains lose their viabilityrather rapidly when stored as colonies on LB agar in thecold. We found that this phenotype was linked to the otsX(katF) genotype. This loss of viability was not caused bydecreased trehalose synthesis, since otsX+ (katF+) strainscarrying an otsA or otsB mutation remained viable for aprolonged period (data not shown).
otsX (katE) does not activate proU and bet. To investigatethe influence of otsX on the transcription of the osmoticallyregulated betB, betT (17), and proU (15) genes, we usedsuppressor-free strains which were blocked in trehalosesynthesis by an otsA::TnlO mutation (19). Thereby we couldeliminate any indirect effect of otsX mediated via trehalosesynthesis. As shown in Table 3, the ,3-galactosidase activi-ties of cells with a betB-lacZ or betT-lacZ operon fusionwere essentially the same whether the cells carried theotsX MC4100 or the otSXw1485 allele. The same lack of effectof otsX was obtained for strains carrying a proU-lacZ fusion(data not presented). Thus, otsX is not a general activator ofosmotically regulated genes.
Cloning of otsA and otsB. Strains which are defective intrehalose synthesis display an osmotically sensitive pheno-
TABLE 3. Influence of otsX (katF) allele on the osmoticinduction of ,-galactosidase activity of betT-lacZand betB-IacZ operon fusion mutants of E. coli
" All strains are Alac.b One unit of enzyme activity is 1 nmol of o-nitrophenol formed per min at
28°C. Each value is an average of at least three independent measurements.Standard deviations were within ±15%.
type (19). By introducing a cosmid library into FF4169 (otsA)and selecting for strains growing on agar with medium63-glucose with 0.45 M NaCl added, we obtained plasmidpFF1. This plasmid carried a chromosomal insert of 43 kb(Fig. 2). From pFF1 we constructed plasmid pFF106 carry-ing a 2.9-kb HindIII chromosomal fragment. This was thesmallest fragment obtained that restored trehalose synthesisin both otsA and otsB mutants, as well as in A(otsA otsB)mutants. We also constructed the subclones pFF109 andpFF114, which carried overlapping chromosomal fragmentswith lengths of 1.85 and 2.1 kb, respectively, covering the
tar(A) (41.6) ots
BE
EV Ii E,I, l,II ;
Pv
1,960 1,970 1,980
(B)
(C)
Coordinates
1,990 2,000 kb
SH BBSH S HB H
SI Ps EHB H EEE
__--- _-- ~B'BgHPvE EC-EV )I---I
HrHc pFF106 (otsA+ otsBB)pFF114 (otsA+ otsB8)
m@ . pFF109 (otsA otsB+)
pFF1
pFF1 01
pFF1 02
FIG. 2. Restriction site analysis of the otsBA region. (A) Restric-tion map of the 41- to 42-min region of the E. coli chromosome aspublished by Kohara et al. (28); the coordinates are in kilobasesfrom thr. The localization and the direction of transcription of theotsBA operon are indicated. (B) Restriction map of the chromo-somal part of plasmid pFF1 and its subclones pFF101 and pFF102,aligned with the map of Kohara et al. (C) Enlarged restriction mapof the 2.9-kb HindIII fragment containing the otsBA operon. Theextensions of the chromosomal part of plasmid pFF106, pFF109,and pFF114 are shown, and their genotypes are indicated. Abbre-viations used for restriction sites are as follows: B, BamHI; Bl, BgII;Bg, BglIl; C, ClaI; E, EcoRI; EV, EcoRV; H, HindIlI; Hc, HincII;K, KpnI; Ps, PstI; Pv, PvuII; and S, Sall.
a All strains were derived from MC4100. otsX is allelic with katF.b Enzyme activities were measured in toluene-treated cells. One unit of
enzyme activity is 1 nmol of trehalose formed per min at 37'C. Each value isan average of at least four independent measurements. Standard deviationswere within ±10%o for the synthase activities, except for FF4050(pFF114) (seetext), and within ±20% for the phosphatase activities. ND, not determined.
c The synthase activity displayed large variations (see text).
insert of pFF106 (Fig. 2). Properties of these three plasmidsare described below.
Expression of synthase and phosphatase activities. Wepreviously described an assay for trehalose-6-phosphatesynthase in E. coli (19, 52). In this work we have developedan assay for the phosphatase of the trehalose pathway, usingsynthetic trehalose-6-phosphate as a substrate and determin-ing, by using gas chromatography, the amount of trehaloseformed. In order to measure the osmotic induction of theenzymes, we grew the cells in medium 63 and then increasedthe osmolarity of the medium by addition of 0.4 M NaCl.Strains with intact trehalose synthesis which grew in thepresence of 0.4 M NaCl were incubated for one generation,whereas strains with a defective trehalose synthesis wereincubated in the high-osmolarity medium for 2 h. All strainsused were derivatives of MC4100 and carried a treA muta-tion to prevent in vitro degradation of trehalose (19). Exceptfor FF4057, they were all otsX+ (katF+).
Strain FF4171 is wild type with respect to trehalosesynthesis. As shown previously (19, 52), FF4171 has anosmotically inducible synthase, whereas related strains car-rying an otsA, otsB, or A(otsA otsB) mutation lack thissynthase activity completely (19) (Table 4). In addition,FF4171 displayed an osmotically induced phosphatase activ-ity. The phosphatase activity was also osmotically inducedin FF4026 (otsA), but not in FF4025 (otsB) and FF4050[A(otsA otsB)], in which it was the same whether the cellswere osmotically stressed or not (Table 4). Apparently, theotsB mutation blocked the synthesis of the osmoticallyinducible phosphatase, but it did not influence the back-ground phosphatase activity also found in nonstressed cells.It is well known that it is often difficult to measure accuratelyseparate phosphatase activities in wild-type E. coli. This isbecause of the multiplicity of phosphatases in E. coli and thelack of specificity of some of them (42).
Strain FF4050 [A(otsA otsB)] carrying plasmid pFF106displayed strongly elevated activities of both the synthaseand the phosphatase. Both activities, after subtraction of thebackground phosphatase activity, increased about fivefoldwhen FF4050(pFF106) was exposed to osmotic stress. Thesynthase and phosphatase activities of strains carrying plas-
mid pFF106 (otsA+ otsB+) were much lower in the otsX(katF) mutant strain FF4057 than in its otsX+ counterpartFF4050 (Table 4).Subclone pFF114, which restored trehalose synthesis in
synthase-defective otsA mutants, conferred only synthaseactivity to the A(otsA otsB) mutant FF4050. SubclonepFF109 conferred only phosphatase activity to FF4050 (Ta-ble 4). As would be expected, pFF109 did not restoretrehalose synthesis in either otsA or otsB insertion mutants,since these mutants lacked the synthase activity. The ele-vated enzymic activities displayed by the present plasmid-carrying strains and the finding that the enzymes can beexpressed separately indicated that the plasmids carriedstructural rather than regulatory genes. Presumably, otsAand otsB were the structural genes of the synthase and thephosphatase, respectively, and the lack of synthase activityin otsB mutants was because of a polar effect.
Difference in copy number may explain why pFF109(pBR322 derived) appeared to confer a much higher phos-phatase activity to osmotically stressed cells than didpFF106 (pACYC derived) (Table 4). The lower ratio ofosmotic induction of the phosphatase seen with pFF106compared with pFF109 may be because of a constitutivecontribution to otsBA transcription from a vector promoterin pFF106. Stuber and Bujard (51) have shown thatpACYC184 contains a strong promoter situated just up-stream of the otsBA insert in pFF106.
Addition of spectinomycin (300 jig ml-') reduced stronglythe osmotic induction of the enzyme activities in all plasmid-carrying strains (data not shown). Thus, the observed os-motic effect depended on de novo protein synthesis. Itshould, however, be noted that the synthase activity ofstressed cells of FF4050(pFF114) was very variable (Table4). This was probably due to a toxic effect of intracellulartrehalose-6-phosphate, which accumulated in the absence ofthe specific phosphatase activity. This notion is supportedby the finding that when FF4050(pFF114) was subjected toan osmotic upshock, the survival of the cells, as determinedby counting CFU on LB plates, was only 50% after 2 h and10% after 4 h. It has been reported previously that intracel-lular trehalose-6-phosphate is inhibitory to both S. cerevisiae(41) and E. coli (27).
Physical mapping of otsBA. The 43-kb chromosomal frag-ment of pFF1 was mapped with BamHI, HindIII, and SalI.The subclone pFF101 was in addition mapped with PstI andEcoRI, and the shorter subclone pFF102 also with BglII,PvuII, and ClaI. In Fig. 2, the restriction maps of thechromosomal fragments of pFF1 and its subclones havebeen aligned with the 41- to 42-min region of the physicalmap of the E. coli chromosome prepared by Kohara et al.(28). A very good agreement was found for all enzymes usedby both Kohara et al. and us. The only exception was theHindIII site furthest to the right in pFF1, which was notreported by Kohara et al. This is the HindIII site limiting theright-hand side of the chromosomal insert of pFF106 (Fig. 2).
In conclusion, the 2.9-kb HindIII fragment which comple-ments A(otsA otsB) mutants is localized around the coordi-nate 1,992 kb in the physical map of the E. coli chromosome(28). That is 8 to 11 kb clockwise of tar (41.6 min). On thebasis of the localization of the chromosomal fragment ofsubclones pFF109, conferring only OtsB (phosphatase) ac-tivity, and pFF114, conferring only OtsA (synthase) activity,it was evident that otsA is located proximal to tar withrespect to otsB. This localization and gene order is inagreement with our previous mapping data (19). However,on the latest chromosomal linkage map of E. coli these ots
genes are wrongly placed between fliA and uvrC (4) and notin the correct position between uvrC and flhD (flbB) (19).
It has been shown previously that a TnJO insertion in otsAdoes not influence the expression of a lacZ fusion in otsB(19). However, the present biochemical data showed that aninsertion in otsB prevented the expression of otsA (Table 4).This polar effect suggests that otsBA constitutes an operon.The operon model is also supported by the finding that whenpreparing deletion mutants from FF4012 (MC4100 otsA::TnJO otsB-lacZ otsX+Mc4l00) by selecting for Tcs and scor-ing for the Km Lac phenotype, we obtained a few strains,e.g., strain FF4035, which were Tcs Kms but remainedLac'. The otsB-1acZ fusion in FF4035 remained under thecontrol of the native otsX (katF)-dependent promoter since itbecame Lac- when an otSXwi485 allele was introduced togenerate strain FF4055. Apparently, the deletion of otsA::TnlO from FF4012 had been accompanied by a deletion ofthe kanamycin resistance marker, but not the lacZ gene ofthe X placMu insertion. The kanamycin marker is locateddistal to the promoter controlling the expression of lacZ inthe intact A placMu prophage. Thus, these data showed thatthe otsB gene is transcribed in the direction towards otsA,i.e., in the counterclockwise direction on the chromosome.We assume that the transcription of otsA in pFF114 (Table 4)starts from a promoter which is not functional on thechromosome.
DISCUSSION
The otsA and otsB genes have been mapped previously(19). In this study we have cloned these genes and deter-mined their localization on the physical map of the E. colichromosome prepared by Kohara et al. (28). We presentevidence that otsB and otsA constitute an operon in whichotsB, encoding trehalose-6-phosphate phosphatase, is prox-imal to the promoter, and otsA, encoding trehalose-6-phos-phate synthase, is distal. Thus, a recently described pointmutation named otsP, which maps near otsA and seems toinfluence only the expression of the phosphatase, is probablyallelic with otsB (27).For S. cerevisiae, the synthase and the phosphatase of the
trehalose pathway have been copurified as a complex andthe individual enzymic activities have not yet been resolved(33). This enzyme complex may not exist in E. coli, since theOtsA and OtsB activities could be expressed separately in anotsBA deletion mutant. Also supporting this is the findingthat we were unable to regain the synthase activity from cellsdisrupted in a French pressure cell, while the phosphataseactivity remained stable (data not shown).
It has been shown previously that the variation amongK-12 strains with respect to their capacity to synthesizetrehalose and to express otsBA is because of an amber-mutated gene in E. coli K-12 and many of its derivatives (47,52). In this study we have identified this mutated gene askatF, which has recently been reported to be allelic withappR (54) and csi-2 (rpoS) (30). The DNA sequence indicatesthat it encodes a sigma factor (39). Our demonstration thatkatF carries an amber mutation in E. coli K-12 and W1485(wild type), but not in MC4100, is in agreement with earlierreports that appR is inactive in many K-12 strains (16, 53)and that MC4100 is among the strains in which appR isactive (16).Genes which display increased expression when the cell is
exposed to a particular external stimulus or stress are oftengrouped into stimulons. A stimulon can comprise severalregulons, and many genes belong to more than one stimulon.
For instance, carbon starvation (37) and osmotic stress (10)are reported to increase the synthesis of at least 54 and 41proteins, respectively, and 5 osmotic-stress proteins arereported to be among the starvation proteins (25). In exper-iments with strain MC4100, it has been shown that katFpositively controls 15 to 20 genes expressed during carbon-starvation-induced entry into stationary phase (30). How-ever, the characterization of stimulons in E. coli has beencarried out with different strains. Thus, the fact that E. coliK-12 and many of its derivatives carry an amber mutation inkatF may have made investigators overlook katF-dependentinduction. This has to be taken into account in furtherevaluations of E. coli stimulons.The katF gene is known to be transcriptionally regulated,
and it has been suggested that slow growth is a signal for itstranscription (30, 40). Since osmotic stress slows the growthof E. coli, it probably causes katF induction, and this maypartly explain the observed osmotic regulation of otsBAtranscription. However, it is unlikely that this is the onlymechanism, since the otsBA transcription, albeit being low,was osmotically regulated also in strains carrying katF: :TnlOor the amber-mutated katF gene.proU transcription is dependent on the housekeeping
sigma factor °70 (26), and the osmotic regulation ofproU canbe explained by a stimulatory effect of potassium glutamateon the transcriptional complex (43), independent of DNAsupercoiling (44). Apparently, an intrinsic property of theosmotic stress-dependent proU promoter elicits this potas-sium glutamate effect without the need of protein factors, atleast in vitro (43). A similar effect of potassium glutamate onthe otsBA promoter could explain the osmotic induction ofotsBA seen in strains mutated in katF.Trehalose protects organisms against a variety of stresses
(see the introduction); The finding that transcription ofotsBA, as opposed to proU and bet, is katF dependentnourishes a notion that trehalose, in addition to being anosmoprotectant, may serve as a more general stress pro-tectant in E. coli.
ACKNOWLEDGMENTSWe thank W. Boos and R. Hengge-Aronis for useful discussions.
We are grateful to B. J. Bachmann, E. Bremer, D. P. Clark, W.Epstein, M. W. Eshoo, R. Hengge-Aronis, and B. Lowsky forproviding bacterial strains and to P. I. Larsen and P. Kolsaker atMarine DNA, Troms0, Norway, for providing trehalose-6-phos-phate.
This work was supported by a grant from the Norwegian FisheriesResearch Council.
REFERENCES1. Andresen, P. A., I. Kaasen, 0. B. Styrvold, G. Boulnols, and
A. R. Str0m. 1988. Molecular cloning, physical mapping andexpression of bet genes governing the osmoregulatory choline-glycine betaine pathway of Escherichia coli. J. Gen. Microbiol.134:1737-1746.
2. Asahina, E., and K. Tanno. 1964. A large amount of trehalose ina frost-resistant insect. Nature (London) 204:1222.
3. Attfield, P. V. 1987. Trehalose accumulates in Saccharomycescerevisiae during exposure to agents that induce heat shockresponse. FEBS Lett. 225:259-263.
4. Bachmann, B. J. 1990. Linkage map of Escherichia coli K-12,edition 8. Microbiol. Rev. 54:130-197.
5. Bachmann, B. J. Personal communication.6. Becerra de Lars, L., J. Ratouchniak, and F. Cae. 1977.
Chromosomal location of gene governing the trehalose utiliza-tion in Escherichia coli K12. Mol. Gen. Genet. 152:105-108.
7. Bochner, B. R., H.-C. Huang, G. L. Schieven, and B. N. Ames.1980. Positive selection for loss of tetracycline resistance. J.
GENES GOVERNING TREHALOSE SYNTHESIS IN E. COLI 897
Bacteriol. 143:926-933.8. Boos, W., U. Ehmann, E. Bremer, A. Middendorf, and P.
Postma. 1987. Trehalase of Escherichia coli. Mapping andcloning of its structural gene and identification of the enzyme as
a periplasmic protein induced under high osmolarity growthconditions. J. Biol. Chem. 262:13212-13218.
9. Boos, W., U. Ehmann, H. Forki, W. Klein, M. Rimmele, and P.Postma. 1990. Trehalose transport and metabolism in Esche-richia coli. J. Bacteriol. 172:3450-3461.
10. Botsford, J. L. 1990. Analysis of protein expression in response
to osmotic stress in Escherichia coli. FEMS Microbiol. Lett.72:355-360.
11. Cairney, J., I. R. Booth, and C. F. Higgins. 1985. Salmonellatyphimurium: proP gene encodes a transport system for theosmoprotectant betaine. J. Bacteriol. 164:1218-1223.
12. Chang, A. C. Y., and S. N. Cohen. 1978. Construction andcharacterization of amplifiable multicopy DNA cloning vehiclesderived from the P1SA cryptic miniplasmid. J. Bacteriol. 134:1141-1156.
13. Chung, C. T., S. L. Niemela, and R. H. Miller. 1989. One-steppreparation of competent Escherichia coli: transformation andstorage of bacterial cells in the same solution. Proc. Natl. Acad.Sci. USA 86:2172-2175.
14. Crowe, J. H., L. M. Crowe, and D. Chapman. 1984. Preserva-tion of membranes in anhydrobiotic organisms: the role oftrehalose. Science 223:701-703.
15. Csonka, L. 1982. A third L-proline permease in Salmonellatyphimurium which functions in media of elevated osmoticstrength. J. Bacteriol. 151:1433-1443.
16. Diaz-Guerra, L., F. Moreno, and J. L. San Milian. 1989. appRgene product activates transcription of microcin C7 plasmidgenes. J. Bacteriol. 171:2906-2908.
17. Eshoo, M. W. 1988. lac fusion analysis of the bet genes ofEscherichia coli: regulation by osmolarity, temperature, oxy-
gen, choline, and glycine betaine. J. Bacteriol. 170:5208-5215.18. Galinski, E. A., and R. M. Herzog. 1990. The role of trehalose as
a substitute for nitrogen-containing compatible solutes (Ec-tothiorhodospira halochloris). Arch. Microbiol. 153:607-613.
19. Gisever, H. M., 0. B. Styrvold, I. Kaasen, and A. R. Str0m.1988. Biochemical and genetic characterization of osmoregula-tory trehalose synthesis in Escherichia coli. J. Bacteriol. 170:2841-2849.
20. Gutierrez, C., M. Ardourel, E. Bremer, A. Middendorf, W.Boos, and U. Ehmann. 1989. Analysis and DNA sequence of theosmoregulated treA gene encoding the periplasmic trehalase ofEscherichia coli K12. Mol. Gen. Genet. 217:347-354.
21. Herbert, D., P. J. Phipps, and R. E. Strange. 1971. Chemicalanalysis of microbial cells, p. 209-344. In J. R. Norris and D. W.Ribbons (ed.), Methods in microbiology, vol. 5B. AcademicPress, Inc. (London), Ltd., London.
22. Hino, A., K. Mihara, K. Nakashima, and H. Takano. 1990.Trehalose levels and survival ratio of freeze-tolerant versus
23. Hoelzle, I., and J. G. Streeter. 1990. Increased accumulation oftrehalose in rhizobia cultured under 1% oxygen. Appl. Environ.Microbiol. 56:3213-3215.
24. Hottiger, T., T. Boiler, and A. Wiemken. 1987. Rapid changes ofheat and desiccation tolerance correlated with changes oftrehalose content in Saccharomyces cerevisiae cells subjectedto temperature shifts. FEBS Lett. 220:113-115.
25. Jenkins, D. E., S. A. Chaisson, and A. Matin. 1990. Starvation-induced cross protection against osmotic challenge in Esche-richia coli. J. Bacteriol. 172:2779-2781.
26. Jovanovich, S. B., M. T. Record, Jr., and R. R. Burgess. 1989. In
an Escherichia coli coupled transcription-translation system,expression of the osmoregulated gene proU is stimulated atelevated potassium concentration and by an extract from cellsgrown at high osmolarity. J. Biol. Chem. 264:7821-7825.
27. Klein, W., U. Ehmann, and W. Boos. 1991. The repression oftrehalose transport and metabolism in Escherichia coli by highosmolarity is mediated by trehalose-6-phosphate phosphatase.Res. Microbiol. 142:359-371.
28. Kohara, Y., K. Akiyama, and K. Isono. 1987. The physical mapof the whole E. coli chromosome: application of a new strategyfor rapid analysis and sorting of a large genomic library. Cell50:495-508.
29. Laere, A. V. 1989. Trehalose, reserve and/or stress metabolite?FEMS Microbiol. Rev. 63:201-210.
30. Lange, R., and R. Hengge-Aronis. 1991. Identification of acentral regulator of stationary-phase gene expression in Esche-richia coli. Mol. Microbiol. 5:49-59.
31. Larsen, P. I., L. K. Sydnes, B. Landfald, and A. R. Str0m. 1987.Osmoregulation in Escherichia coli by accumulation of organicosmolytes: betaines, glutamic acid, and trehalose. Arch. Micro-biol. 147:1-7.
32. Loewen, P. C., and B. L. Triggs. 1984. Genetic mapping of katF,a locus that with katE affects the synthesis of a second catalasespecies in Escherichia coli. J. Bacteriol. 160:668-675.
33. Londesborough, J., and 0. Vuorio. 1991. Trehalose-6-phosphatesynthase/phosphatase complex from bakers' yeast: purificationof a proteolytically activated form. J. Gen. Microbiol. 137:323-330.
34. Mackenzie, K. F., K. K. Singh, and A. D. Brown. 1988. Waterstress plating hypersensitivity of yeasts: protective role oftrehalose in Saccharomyces cerevisiae. J. Gen. Microbiol.134:1661-1666.
35. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.
36. Martin, M. C., L. A. Diaz, M. B. Manzanal, and C. Hardisson.1986. Role of trehalose in the spores of Streptomyces. FEMSMicrobiol. Lett. 35:49-54.
37. Matin, A. 1991. The molecular basis of carbon-starvation-induced general re- stance in Escherichia coli. Mol. Microbiol.5:3-10.
38. Miller, J. H. 1972. Experiments in molecular genetics. ColdSpring Harbor Laboratory, Cold Spring Harbor, N.Y.
39. Mulvey, M. R., and P. C. Loewen. 1989. Nucleotide sequence ofkatF of Escherichia coli suggests KatF protein is a novel cutranscription factor. Nucleic Acids Res. 17:9979-9991.
40. Mulvey, M. R., J. Switala, A. Borys, and P. C. Loewen. 1990.Regulation of transcription of katE and katF in Escherichia coli.J. Bacteriol. 172:6713-6720.
41. Piper, P. W., and A. Lockheart. 1988. A temperature-sensitivemutant of Saccharomyces cerevisiae defective in the specificphosphatase of trehalose biosynthesis. FEMS Microbiol. Lett.49:245-250.
42. Pradel, E., and P. L. Boquet. 1989. Mapping of the Escherichiacoli acid glucose-1-phosphatase gene agp and analysis of itsexpression in vivo by use of an agp-phoA protein fusion. J.Bacteriol. 171:3511-3517.
43. Prince, W. S., and M. R. Villarejo. 1990. Osmotic control ofproU transcription is mediated through direct action of potas-sium glutamate on the transcription complex. J. Biol. Chem.265:17673-17679.
44. Ramirez, R. M., and M. Villarejo. 1991. Osmotic signal trans-duction to proU is independent of DNA supercoiling in Esche-richia coli. J. Bacteriol. 173:879-885.
45. Reed, R. H., L. J. Borowitzka, M. A. Mackay, J. A. Chudek, R.Foster, S. R. C. Warr, D. J. Moore, and W. D. P. Stewart. 1986.Organic solute accumulation in osmotically stressed cyanobac-teria. FEMS Microbiol. Rev. 39:51-56.
46. Roberts, I., R. Mountford, N. High, D. Bitter-Suermann, K.Jann, K. Timmis, and G. Boulnois. 1986. Molecular cloning andanalysis of genes for production of K5, K7, K12, and K92capsular polysaccharides in Escherichia coli. J. Bacteriol. 168:1228-1233.
47. Rod, M. L., K. Y. Alam, P. R. Cunningham, and D. P. Clark.1988. Accumulation of trehalose by Escherichia coli K-12 athigh osmotic pressure depends on the presence of ambersuppressors. J. Bacteriol. 170:3601-3610.
48. Schweizer, H., and W. Boos. 1983. Transfer of the A(argF-lac)U169 mutation between Escherichia coli strains by selectionfor a closely linked TnlO insertion. Mol. Gen. Genet. 192:293-294.
49. Silhavy, T. J., M. L. Berman, and L. W. Enquist. 1984.Experiments with gene fusions. Cold Spring Harbor Labora-tory, Cold Spring Harbor, N.Y.
50. Str0m, A. R., P. Falkenberg, and B. Landfald. 1986. Genetics ofosmoregulation in Escherichia coli: uptake and biosynthesis oforganic osmolytes. FEMS Microbiol. Rev. 39:79-86.
51. Sttiber, D., and H. Bujard. 1981. Organization of transcriptionalsignals in plasmids pBR322 and pACYC184. Proc. Natl. Acad.Sci. USA 78:167-171.
52. Styrvold, 0. B., and A. R. Str0m. 1991. Synthesis, accumula-tion, and excretion of trehalose in osmotically stressed Esche-
richia coli K-12 strains: influence of amber suppressors andfunction of the periplasmic trehalase. J. Bacteriol. 173:1187-1192.
53. Touati, E., E. Dassa, and P. L. Boquet. 1986. Pleiotropicmutations in appR reduce pH 2.5 acid phosphatase expressionand restore succinate utilisation in CRP-deficient strains ofEscherichia coli. Mol. Gen. Genet. 202:257-264.
54. Touati, E., E. Dassa, J. Dassa, P. L. Boquet, and D. Touati. 1991.Are appR and katF the same Escherichia coli gene encoding anew sigma transcription initiation factor? Res. Microbiol. 142:29-36.
![A Tale of Two Sugars: Trehalose 6-Phosphate and Sucrose1[OPEN] · Topical Review on Sugars A Tale of Two Sugars: Trehalose 6-Phosphate and Sucrose1[OPEN] Carlos M. Figueroa and John](https://static.documents.pub/doc/80x56/6002290a227ae9108550bd85/a-tale-of-two-sugars-trehalose-6-phosphate-and-sucrose1open-topical-review-on.jpg)
![Trehalose 6-Phosphate Regulates Photosynthesis and · Trehalose 6-Phosphate Regulates Photosynthesis and Assimilate Partitioning in Reproductive Tissue1[OPEN] Maria Oszvald,a Lucia](https://static.documents.pub/doc/80x56/611f5cf3b1f8956e674d52f9/trehalose-6-phosphate-regulates-photosynthesis-trehalose-6-phosphate-regulates-photosynthesis.jpg)
