Vol. 173, No. 3

Effects of Structural Chahges in the dsdA-dsdC Intergenic Regionon D-Serine Deaminase Synthesis

ELIZABETH McFALL,l* SAVITA S. NIKAM,2 AND SUNIL PALCHAUDHURI2

Department of Microbiology, New York University School of Medicine, New York, New York 10016,1 andDepartment of Immunology and Microbiology, Wayne State University School of Medicine,

Detroit, Michigan 482012

Received 4 June 1990/Accepted 26 November 1990

Single-base-pair changes well upstream of its transcription initiation site resulted in partially to fullyconstitutive expression of the D-serine deaminase structural gene, dsdA, independently of the cyclic AMP-cyclicAMP-binding protein complex and of the specific D-serine deaminase activator protein. These promotermutations appear to define a consensus sequence that is repeated several times. Basal expression of dsdA+ was

also strongly enhanced by subcloning on multicopy plasmids, by the DNA gyrase inhibitor novobiocin, and indsdC(Con) mutants by increasing growth temperature. These results suggest that activation of dsdA+expression by the dsdC-encoded protein involves distortion of promoter DNA. A dsdA translation start at bp-731 was verified by subcloning of dsdC+. Plasmid-specified activator at a high concentration interfered withchromosomal dsdC(Con) expression, and the interference was enhanced by deletion of most of the intergenicregion from the plasmid. Even at a high concentration, however, plasmid-specified activator did not activateexpression of chromosomal dsdA+, and in one case it was actually repressive. These results confirm the strongcis tropism of plasmid-specified dsdC-encoded protein and suggest that it is mediated by multiple sites in thedsd4-dsdC intergenic region.

The two genes specific to the inducible D-serine deaminasesystem of Escherichia coli K-12, the structural gene dsdAand the activator gene dsdC, are adjacent on the chromo-some and are transcribed with opposite polarity from theintergenic region (3). The DNA sequence of the entire locushas been determined; it occupies 3.3 kbp of a SalI-EcoRIfragment (13, 19, 24). The dsdA transcription start (bp +1)lies in an A+T-rich region 81 bp upstream of the translationstart in the wild type and in the three dsdAp mutantsdiscussed below (3). The size of the dsdC protein (33 kDa)indicates that the most likely translation start would be at bp-731 to give a gene product with a molecular weight of32,900 (24). The untranslated region between the two genes

is thus very large, on the order of 800 bp. This is similar tothe distance between the malT-malPQ operons, where thereis also a divergent activator gene-structural gene arrange-

ment (5).Although there is a good match (TACTAT) to the consen-

sus Pribnow box at the dsdA transcription start site and a

TTGCGG sequence at -45, there is no consensus -35sequence. As might be expected, the basal level of D-serinedeaminase is very low. The dsdC activator protein, with theinducer D-serine, enhances it about 5,000-fold in the pres-ence of cyclic AMP and cyclic AMP-binding protein (cAMP-CAP) and 1,000-fold without cAMP-CAP (7, 16). The dsdCprotein is apparently strictly an activator of dsdA expres-sion, in vivo and in vitro (19), and strictly a repressor of itsown synthesis (20).

Several findings suggest that the activation mechanism iscomplex and delicately balanced. Mutations in the dsdCgene that result in constitutive synthesis of D-serine deami-nase [dsdC(Con)] are readily isolated, but all require cAMP-CAP for expression of constitutivity. In two cases, this

* Corresponding author.

constitutivity was found to be thermosensitive (1, 14). dsdAp(promoter-constitutive) mutants were also readily isolatedand found to be independent of dsdC activator and cAMP-CAP to the extent of the constitutivity in vivo and in vitro(19). Low-constitutivity dsdAp dsdC+ strains are hyperin-ducible, requiring cAMP-CAP for maximal expression, likedsdAp+ strains. Basal D-serine deaminase activity is identi-cal in dsdC+ and lAdsdC strains (7). Curiously, inducedexpression of dsdAp+ DNA in vitro in response to dsdCactivator and D-serine was temperature sensitive, whereasconstitutive expression of dsdAp DNAs was not (19) andinduced synthesis in vivo also was not (14). The dsdApmutations did not create new promoters (3), which indicatedthat they affected a site(s) concerned with the activationmechanism. Plasmid constructs carrying dsdC+ and theentire intergenic region failed to complement chromosomaldsdC mutations for induction of dsdA+ or repression ofdsdC(Con)-mediated constitutivity (4, 18), although dsdC+(F') constructs did both (2, 14, 15). These results suggestedthat activator might be trapped by sites on the plasmids, as

has been shown to occur with cis-acting transposases andthe A gene product of single-stranded DNA bacteriophages(18, 22, 29).To analyze effects of DNA structural changes and inter-

actions of dsdC activator with DNA on dsdA expressionfurther, we located two dsdAp mutations by sequencing,examined the effect of temperature on constitutivity inseveral more dsdC(Con) mutants, subcloned dsdA + anddsdC+ together with various amounts of the intergenicregion, and examined the effect of a DNA gyrase inhibitor.The results indicate that multiple sites are involved in dsdAactivation, that small changes in DNA structure may havemajor effects on dsdA expression, and that the degree ofDNA supercoiling has a significant effect on dsdA expres-sion.

1161

JOURNAL OF BACTERIOLOGY, Feb. 1991, p. 1161-11670021-9193/91/031161-07$02.00/0Copyright ©D 1991, American Society for Microbiology

on Septem

ber 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

1162 McFALL ET AL.

TABLE 1. Bacterial strains

Strain Relevant Derivation or

genotype reference

EMlOOOa dsdA+ dsdC+ W3828; 3JM109 A(lac-proAB) gyrA96 recA(F' lacIq 31

lacZ AM15)EM61601 X dsdApl 3EM614005 A dsdAp4005 3EM61606 X dsdAp6 3YMC12 glnG:: TnS B. MagasanikYMC18 glnF:: TnlO B. MagasanikEM1900 dsdA+ dsdC(Con)9 W3828; 1FB5O11 dsdA+ dsdC(Con)1J 1FB5015 dsdA+ dsdC(Con)15 1FB5019 dsdA+ dsdC(Con)19 1EM1309b dsdA+ dsdC(Con)3 recA EM1300; 14EM145-1 dsdA+ dsdC::Mu dl(Apr lac) 20FB6001 dsdA+ dsdC::A 2EM1100 dsdApl dsdA+ dsdC+ W3818; 3EM1600 dsdAp6 dsdA+ dsdC+ 3AC6082 dsdA dsdC+ 4

a EM1000 is a lac+ transductant of strain W3828.b EM1309 is a recA derivative of strain EM1300 isolated as described by

Bloom and McFall (1).

MATERIALS AND METHODS

Bacterial strains and plasmids. The strains and plasmidsused are described in Tables 1 and 2. The procedures used incloning and agarose and acrylamide gel electrophoresis weredescribed by Maniatis et al. (12).Media, methods of cultivation, and D-serine deaminase

assay. Cells were, in most cases, grown in LB broth (17)supplemnented with appropriate antibiotics when necessaryto maintain plasmid selection. For RNA isolation, cells weregrown in modified M9 medium (30). Cultures were supple-mented with novobiocin (Sigma) at 200 ,ug/ml when indi-cated. Cells were cultivated to a density of 4 x 108/ml forassay of D-serine deaminase (16).

Sequencing of dsdAp mutations. The products of partialHaeIII digests of an SimaI-Hindlll fragment (carries theentire dsd intergenic region) of X bacteriophages carryingdsdApl, dsdAp4005, and dsdAp6 were cloned into the SmaIsite of M13mpl8 (31). Ligation products were transformedinto strain JM103, and progeny bacteriophages were exam-ined for the presence of dsd intergenic region sequences by

Southern blotting (28) to appropriate probes. DNA se-quences of the cloned fragments were determined by thedideoxy method of Sanger and Coulson (27). Both strandswere sequenced, including overlaps across restriction sites.Measurements of dsdC mRNA and protein. Total RNA was

isolated from log-phase cells by following the protocol ofVenkatesan et al. (30). For induction, bacteria were grownwith 1 mM isopropyl-p-D-thiogalactopyranoside (IPTG) and5 mM D-serine for 1 h at 37°C before harvesting. RNA wasglycosylated and slot blotted onto Bio-Rad Zetaprobe mem-branes. It was then hybridized to a 32P-labeled probe, thedsdC SphI-EcoRI fragment (Table 2) from pSB192. Thehybrid amounts were compared and quantitated by densitom-etry scans.

Nucleotide sequence accession number. The sequence inFig. 1 has been assigned GenBank accession no. JB M19035.

RESULTS

Promoter-constitutive mutations and repeat sequences. Welocated the mutational changes for two of the three dsdApmutants, representatives of each of the three phenotypicclasses, whose transcription start sites were determinedpreviously (Fig. 1 and 2) (3). One, dsdAp6, which resulted infully constitutive D-serine deaminase synthesis-4,000-foldbasal activity in the uninduced wild type-and completeindependence of cAMP-CAP, was a GC-*AT transition atbp -95. The mutation created a theoretically excellenttranscription initiation sequence: TTGCGA-16 bp-TACAAT. No transcription from this region was detected in theprevious study, however (3). The second, dsdAp4005, whichresulted in low constitutivity-1,000-fold basal activity-independently of dsdC protein and cAMP-CAP to the extentof the constitutivity, was a TA--AT transversion at bp-120. We were unable to locate the site of the thirdmutation, dsdApl, as mpl8 clones containing it are highlyunstable.When we examined the context of mutations dsdAp6 and

dsdAp4005, we saw that they occurred on opposite sides ofthe single sequence in the region with significant homologyto the consensus CAP recognition sequence (6). They alsooccurred in sequences with considerable homology to eachother. In fact, they appear to define a repeated sequencecentered at bp -95, -115, -149, -172, -256, and -735 (thelatter two on the opposite strand) (Fig. 3).Both dsdAp mutations, especially dsdAp6, decreased the

TABLE 2. Physical properties of dsd hybrid plasmids

Parent FragmentPlasmid(s) vector Cloned dsd fragmenta size (kb) Reference

pAC51 pACYC177 SmaI-EcoRI (dsdC+) 2.2 4pSB8W1 pUC8 SmaI-EcoRI (dsdC+) 2.2 7pUC18, pUC19 31pUC8 31pSB183 pUC18 SphI-EcoRI (dsdC+) 1.6 2a, 18pSB192 pUC19 SphI-EcoRl (dsdC+) 1.6 2a, 18pEM12 pUC19 BglII-EcoRI (dsdC+) 1.2 This workpMM1 pUC8 SalI-EcoRI (dsdA+ dsdC+) 4.0 This workpMM2 pUC19 SalI-EcoRI (dsdA+ dsdC+) 4.0 This workpMM1O pUC18 SalI-SphI (dsdA+ AdsdC) 2.4 This workpMM12 pUC19 SalI-SphI (dsdA+ AdsdC) 2.4 This work

a Cloning sites: Sail, 1.2 kb beyond the C terminus of dsdA; Smal, bp +666; SphI, bp -211; BglII, bp -630; EcoRI, bp -1824 (348 bp beyond the C terminusof dsdC). The dsdC+ genes of pSB8W1 and pSB183 were cloned with the C terminus adjacent to lacP, and the dsdC+ genes of pSB192 and pEM12 were clonedwith the N terminus adjacent to lacP. The dsdA+ genes of pMM1 and pMM12 were cloned with the N terminus adjacent to lacP, and the dsdA+ genes of pMM2and pMM10 were cloned with the C terminus adjacent to lacP.

J. BACTERIOL.

on Septem

ber 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

dsdA-dsdC INTERGENIC REGION 1163

-750 -740 -730 -720 -710 -700 -690S' - GGGCTTAGATGGCTATGCAGCTGCATTACGTGTACGTGCCTAGCTTTTCCCGTGTCACTmCAGTGGTCA3 - CCCGAATC.ACCGATACGTCGACGTAATGCACATGCACGGATCGAAAAGGGCACAGTGAAAGTCACCAGT

-680 -670 -660 -650 -640 -630 -620CGACTGGCCGCCGTTGGTAATCTGGTGAACCCGGACGAGTTACAGAGTTTGCAGATCTCAAAGTCGCGAGCTGACCGGCGGCAACCATTAGACCACrTGGGCCTGCTCAATGTCTCAAACGTCTAGAGTCAAGCGCT

-610 -600 -590 -580 -570 -560 -550TGAAAAACCACTACCGTCATTAGGCGCAACGTTATTCTCCATACTGCTACCCATTGCGCTGAGTATTGG

-540 -530 -520 -510 -500 -490 -480TTAAACGATTGCCGAATTGAATATGGCGCGTGAGAGTGGTTTGTATATCTTGGTTGGAGTTTATTGGCA

-470 -460 -450 -440 -430 -420 -410ACCCTATCACTGCCATGTTTATCGCCGTGTTTGTCGCCTATTATGTGTTGGGTATACGGTCCAGCATATG

-400 -390 -380 -370 -360 -350 -340AGCATGGGGACGATGCTCACACATACGGAAAATGGCTTCGGTTCTATTGCTAATATTTTGCTGATTATCGTCGTACCCCTGCTACGAGTGTGTATGCCTTTACCGAGCCAAGATAACGATTATAAAACGACTATAGC

-330 -320 -310 -300 -290 -280 -270GGGCCGGAGGCGCATTCAACGCATTTTAAAAAGCAGCAGTCTCGCTGATACGCTGGCAGTTATTCTCTCC

-260 -250 -240 -230 -220 -210 -200JMTATGCATATGCACCCGATTCTTCTGGCCTGGTTAGTGGCTCTTATTCTGCATGCGGCAGTGGGCTCCGTTATACGTATACGTGGGCTA.AGAGACCGGACCAATCACCGAGAATAAGACGTACGCCGTCACCCGAGGC

-190 -180 -170 -160 -150 -140 -130CTACCGTGGCAATGATGGGGGCAACGGCAATTGTTGCACCCATGCTGCCGCTGTATCCCGACATCAGCCCGATGGCACCGTTACTACCCCCGTTGCCGTTAACMCGTGGGTACGACGGCGACATAGGGCTGTAGTCGGG

[20 -110 -100 -90 -80 -70 -60GGAA4TATTGCGATTGCTATCGGTTc1GGTGCAATTGGCTGCACTATCGTTACGGACTCGCTTTTCTGGcc CGCTACGATAGCCAA CCACGTTACCGACGTGATAGCATGCCTGAGCGGACC

A AT T

-50 -40 -20 -10 +1 10CTAGTGAAG cTATTGCGGCGCTACGCTCAATGAAACATTTAAATACTATACGAC CGACATTTATCGGATCACTTCt TATACGCCGCGATGCGAGTTACTTTGTAAATTTATGATATGCTGT GCTGTAAATAGC

20 30 40 50 60 70 80CTTCAGTCGTCGCTCTGGCGGGCACATTCCTGCTGTCATTTATCATCTAAGCGCAAGAGACGTACTATGGAAGTCAGCAGCGAGACCGCCCGTGTAGGACGACAGTAAATAGTAGATTCGCGTTTCTCTGCATGATE.

FIG. 1. dsdA proximal portion of the dsd intergenic region.Vertical arrows at -120 and -95 indicate dsdAp4005 and dsdAp6mutations. Box centered at -23: dsdA transcription initiation se-quence. Box centered at -110: presumed CAP recognition se-quence. Arrow at +81: dsdA translation start. Repeat sequences areunderlined.

consensus. The repeats all contained sequences with homol-ogy to the consensus transcription initiation sequence forNtr promoters, CTGGPyAPyPuN4TTGCA. We found norecognition sequences for the Ntr activator, the glnG prod-uct (23, 26). To rule out the possibility of a role for the Ntrsystem in D-serine deaminase regulation, however, we ex-amined D-serine deaminase synthesis in rpoN:: TnJO(YMC18) and ginG:: TnS (YMC12) mutants. The basal levelsand induction ratios were identical to those in the wild type(data not shown); thus, the cr54 (Ntr) polymerase does notseem to be involved in induction control.

Thermosensitivity of constitutivity in dsdC(Con) mutants.We previously observed in two cases that the constitutivesynthesis of D-serine deaminase which resulted from thedsdC(Con) mutations {strains EM1300 [dsdC(Con)3] andEM1400 [dsdC(Con)4]} was thermosensitive and that consti-tutivity increased exponentially with temperature in thesemutants, as did constitutivity in the dsdApl promoter mutant(14, 19). This behavior could result from progressive meltingof a DNA-protein complex, as was previously suggested forthe dsdApl mutant (19). We therefore examined the phe-nomenon in more detail, with four more representativedsdC(Con) mutants. One of the mutants, the dsdC(Con)19mutant, is fully constitutive, and the other three have lowconstitutivity. Constitutivity was uniformly thermosensitivein all cases, increasing exponentially with temperature as in

FIG. 2. Temperature dependence (K) of constitutive D-serinedeaminase synthesis in dsdC(Con) and dsdApl mutant strains.Symbols: *, strain FB5019 [dsdC(Con)19]; *, strain FB5011 [dsd-C(Con)JJ]; A, strain EM1100 (dsdAp1); V, strain EM1900 [dsdC-(Con)9]; 0, strain FB5015 [dsdC(Con)15].

the mutants examined previously. Thermosensitive consti-tutivity thus seems to be a general property of these mutants.Arrhenius plots of the data over the range from 20 to 43°Cyielded parallel lines with a slope similar to that observedwith the dsdApl mutant (Fig. 2). The four mutants were allhyperinducible in the presence of D-serine at all tempera-tures; thus, the dsdC(Con)-encoded proteins were not inac-tivated at low temperature. Indeed, if the effect were strictlyon the structure of the protein, we would have expected asharper temperature demarcation. Rather, the data suggestthat the dsdC(Con)-encoded proteins affect dsdA expressionsimilarly to the dsdApl mutation.

Effects of subcloning and changes in superhelicity on dsdA

TTCAGGTGCAATTGGCTGC

AATTATTGCGATTGCTATC

CCATGCTGCCGCTGTATCC

AACAATTGCGGTTGCCCCC

ATCGGGTGCATATGCATAT

ACGTAATGCAGCTGCATAG

-95

-115

-149

-172

-256

-735

A-P-TGCAPPTGCAT-C consensusu UY

FIG. 3. Homology of sequences containing mutations dsdAp6and dsdAp4005.

VOL. 173, 1991

on Septem

ber 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

1164 McFALL ET AL.

dsdA

pSB183.pSBI92

pEN12

FIG. 4. Restriction map of the dsd region with fragments usedfor construction of dsdA and dsdC hybrid plasmids.

expression. We had previously observed low but significantexpression of dsdA+ in strains carrying two different dsdA+AdsdC plasmid constructs and tentatively assumed that itwas due to plasmid promoters (4). To investigate it further,we cloned the dsdA+ gene with and without the dsdC geneinto unique pUC plasmid cloning sites and transformed theconstructs into strain JM109 (Fig. 4). The lac promoter atone end of the pUC multiple cloning sites is repressed instrain JM109, which carries lacIq; there is no promoter at theother end (31). The results are presented in Table 3.There was considerable basal expression of dsdA with

dsdA+ AdsdC and dsdA+ dsdC+ constructs, irrespective ofthe orientation of dsdA+ with respect to lacP. The copynumber of the pUC-derived plasmids in each of the plasmid-bearing strains was similar in agarose gels of boiled TritonX-100-lysozyme lysates (12) from cells in the logarithmicgrowth phase, i.e., approximately 100 per cell. Thus, thebasal levels of dsdA+ expression in the dsdA+ dsdC+ anddsdA+ A dsdC strains were roughly 2 and 4 to 5% per dsdA +gene copy, respectively, of that in maximally induced cells.

These values are 50- and 100-fold higher than the wild-typebasal level in uninduced wild-type cells. Readthrough fromlacP does not seem to be involved, as induction of lacP byIPTG had only a slight effect.We suspected that differences in supercoiling between

plasmid and chromosomal dsdA regions might affect expres-

sion of the gene. We therefore tested the effect of the DNAgyrase inhibitor novobiocin on basal and induced D-serinedeaminase synthesis in JM109 and another strain, EM1000,that is wild type for the dsd system, on basal synthesis instrain JM109 carrying pUC plasmid constructs with a dsdA+AdsdC fragment (SphI-SalI; Fig. 4) in the opposite orienta-tion, and on constitutive synthesis in the three dsdApmutants described above. The results are presented in Table3. There was no effect on basal synthesis in the wild type or

in dsdAp mutants in which constitutive synthesis is thermo-stable. Induced synthesis and constitutive synthesis in ther-mosensitive mutant EM61601 were significantly enhanced.Basal synthesis from pMM12, the dsdA+ AdsdC plasmidconstruct of strain EM512, was three to four times higherthan that from pMM10, the analogous construct of strainEM510. In pMM12, the direction of dsdA transcription is thesame as that of the Ampr-encoding locus; in pMM10, it is theopposite. These results indicate a significant effect of theinhibitor on dsdA expression when active transcription isalready occurring.

cis and trans actions of dsdC activator and dsdC translationstart. We consistently observed that plasmid-specified dsdCactivator did not act in trans to activate or interfere withchromosomal dsdA+ expression when the dsd intergenicregion was intact on the plasmid. Thus, A(dsd)lpdsdA'dsdC+ and dsdA+ dsdC+lpdsdC+ strains were highly induc-ible and dsdA+ dsdClpdsdC+ strains were not (7, 18).

TABLE 3. Effects of D-serine, novobiocin, and IPTG on D-serine deaminase activity

Strain Genotype Conditions" Sp act of D-serinedeaminase (U)

EM1000 dsdA+ dsdC+ D-Serine 10Novobiocin + D-serine 18

JM109 gyrA (Nalr) dsdA+ dsdC+ D-Serine 10Novobiocin + D-serine 27Novobiocin 0.05No addition 0.05

EM 501 JM109(pMM1) dsdsA+ AXdsdC No addition 42

EM 502 JM109(pMM2) dsdA+ dsdC+ No addition 45

EM510 JM109(pMM10) dsdA+ AdsdC No addition 71Novobiocin 41

EM512 JM109(pMM12) dsdA+ AdsdC No addition 95Novobiocin 141IPTG 133

EM61601 AdsdlX dsdA+ Api No addition 3.1Novobiocin 12.3

EM614005 Adsd/X dsdA+ Ap4005 No addition 8.3Novobiocin 9.9

EM61606 AdsdIX dsdA+Ap6 No addition 17.4Novobiocin 15.8

a Additions to the growth media were D-serine at 500 ,ug/ml, novobiocin at 200 p.g/ml, 0.5% glucose and 10-3 M IPTG.

J. BACTERIOL.

!Llncli S"l Rsal §2hI A&lll+ 41 4 + +

EcoRl5a11l

dsdC

on Septem

ber 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

dsdA-dsdC INTERGENIC REGION 1165

TABLE 4. Effects of dsdC+ fragments on D-serine deaminase activities

D-Serine deaminase activity (U) Relative dsdC mRNAStrain dsdC+ level with IPTG and(plasmid or genotype) fragmenta Without IPTG or With IPTG and Devel IndD-serine D-serlneEM1000 (dsdA+ dsdC+) 0.02 18.8 bEM1000(pAC51) SmaI-EcoRI 0.02 19.2 NTcEMlOOO(pSB8W1) SmaI-EcoRI 0.05 17.9 20EMlOOO(pUC19) 0.03 18.1 NTEM1000(pSB183) SphI-EcoRI 0.03 15.1 100EM1000(pSB192) SphI-EcoRI 0.02 2.0 200EM1000(pEM12) BgII-EcoRI 0.01 14.0 50FB6001 (dsdA+ dsdC:: X) 0.48 0.26 NTFB6001(pAC51) SmaI-EcoRI 0.71 0.23 NTFB6001(pUC8) 0.2 0.3 NTFB6001(pSB8Wl) SmaI-EcoRI 0.75 0.32 NTFB6001(pSB183) SphI-EcoRI 0.45 0.32 NTFB6001(pSB192) SphI-EcoRI 0.5 0.5 NTFB6001(pEM12) BglII-EcoRI 0.3 0.3 NTEM1309 [dsdA+ dsdC(Con)3 recA] 3.0 3.3 10EM1309(pSB8W1) SmaI-EcoRI 3.1 4.0 20dEM1309(pUC19) 3.0 3.3 10EM1309(pSB183) SphI-EcoRI 1.7 1.5 loodEM1309(pSB192) SphI-EcoRI 0.05 0.05 200dEM1309(pEM12) BgIII-EcoRI 0.75 0.36 50dEM1600 (dsdA+ Ap6) 20.0 NT NTEM1600(pSB192) j,hI-EcoRI 17.6 NT NTEM1100 (dsdA+ Apl) 5.0 NT NTEM1100(pSB192) SphI-EcoRI 5.5 NT NT

a See Fig. 4.b -, Less than 2.c NT, Not tested.dRelative DsdC protein synthesis determined in maxicells (20) was similar.

Moreover, the basal (constitutive) rate of D-serine deami-nase synthesis was the same in dsdA+ dsdC(Con) haploidstrains and in the corresponding dsdA+ dsdC(Con)lpdsdC+merodiploids: i.e., plasmid-specified dsdC+-encoded proteindid not repress expression of chromosomal dsdC(Con) (19a).In contrast, dsdA+ DNA was readily activated in vitro bypartially purified activator (8).We considered it likely that newly synthesized activator

was trapped by binding sites in the plasmid-borne intergenicregion. We therefore constructed several dsdC+ subclonesthat lack parts of the intergenic region and yield high levelsof activator, as shown by measurement of dsdC mRNAlevels (Table 4). The extents of the deletions are shown inFig. 4 and Table 2.To test the effects of the new constructs on chromosomal

dsd expression, we transformed them into an induciblewild-type dsdA+ dsdC+ strain (EM1000), a noninducibledsdA+ dsdC::X strain (FB6001), a constitutive dsdA+ dsdC(Con) strain (EM1309), and two dsdAp (promoter-constitu-tive) strains (EM1100 and EM1600). The dsdA promoter isfunctional in strain FB6001, as A ddsdA phage DNA derivedfrom this strain programmed D-serine deaminase synthesis invitro upon activation by partially purified dsdC-encodedprotein (8). FB6001 was therefore an appropriate strain withwhich to test trans activation of dsdA+ by plasmid-bornedsdC+. dsdC(Con)3 strains synthesize D-serine deaminase ata low constitutive rate, and the enzyme level is not enhancedby induction (14). Thus, EM1309 is an appropriate strainwith which to examine both the effect of plasmid-bornedsdC+ on constitutivity mediated by a dsdC(Con) allele andto test again for trans activation. The transformations intoEM1000, EM1100, and EM1600 tested, respectively,

whether plasmid-borne dsdC+ affects induction in a dsdA+dsdC+ strain or constitutivity in two promoter mutants. Theresults are presented in Table 4.At low levels of dsdC+ expression (pAC51- and pSB8W1-

bearing strains), we observed neither induction of D-serinedeaminase expression nor interference with dsdC(Con)-me-diated constitutivity. As the level of dsdC+ expressionincreased, however (pSB183, pSB192, and pEM12 con-structs), we observed a progressive decrease in constitutiv-ity in dsdC(Con) strains. Since the dsdC protein represses itsown synthesis (20), it is likely that plasmid-specified activa-tor repressed chromosomal dsdC(Con) expression in thesecases. It is also possible that there was subunit mixingbetween dsdC(Con) and dsdC+ monomers. The repressionwas greatest at the highest activator levels, but was alsodisproportionately high with pEM12, the construct with thesmallest amount of the intergenic region. These resultsindicate that plasmid-specified activator can act in trans tointerfere with expression of dsdC(Con). Thus, as plasmidsites were deleted and activator production increased, thelevel of constitutivity decreased. In no case, however, didwe observe induction of D-serine deaminase expression in adsdC background. When the plasmids were introduced intoa dsdC+ dsdA+ strain, the transformants were inducible. Atthe highest activator level, however (dsdA+ dsdC+IpSB192),induced expression was reduced. This suggests a repressioncomponent to the action of the dsdC activator on dsdAp, butwe rather suspect that a nonspecific arrangement of activatormolecules interferes with activation (21).The fact that dsdC construct pEM12 (BglII-EcoRI frag-

ment; Table 2 and Fig. 4) expressed dsdC+ indicates,moreover, that the dsdC translation start is present in this

VOL. 173, 1991

on Septem

ber 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

1166 McFALL ET AL.

plasmid. Thus, the translation start is most likely at bp -731(Fig. 1).

DISCUSSION

The occurrence of dsdAp (promoter-constitutive) muta-tions in repeated sequences in the promoter region suggeststhat the repeats are regulatory sites. The simplest explana-tion for these mutations, which render D-serine deaminasesynthesis independent of D-serine-DsdC protein and cAMP-CAP to the extent of the constitutivity, would be that theDsdC protein and cAMP-CAP repress dsdA expression inthe absence of D-serine and the mutations affect the repres-

sor-binding sites. However, both activator and cAMP-CAPare required for optimal D-serine deaminase synthesis invivo and in vitro; the basal level of D-serine deaminase isidentical in dsdC+, dsdC::Mu dl(Ap lac), and dsdC+ cya

strains in vivo; and there was no repressive effect of DsdCprotein in vitro at low levels of D-serine (9, 20). Moreover,dsdA promoter-constitutive mutants are readily isolated,which argues that constitutivity does not require alterationof both repression and activation sites.dsdA expression is quite sensitive to manipulations that

are likely to affect DNA tertiary structure, such as temper-ature variations in dsdC(Con) mutants and in the in vitrosystem, cloning into a foreign environment, and inhibition ofDNA gyrase. We enhanced expression of dsdC+ on multi-copy plasmids by 2-, 5-, 10-, and 20-fold, a series of smallincrements (Table 4), and were able to demonstrate a transeffect in terms of inhibition of constitutive expression indsdA+ dsdC(Con) hosts at the higher levels. However, we

were not able to obtain trans activation in dsdA+ dsdChosts. Indeed, at the highest level of dsdC+ expression, instrain EM1000(pSB192) (dsdA+ dsdC+IpdsdC+) plasmid-specified activator interfered with induction. Very likely theconstitutivity in dsdC(Con) mutants was decreased becauseof repression on the dsdC(Con) allele, with induction inhib-ited because of nonspecific DsdC protein-DNA interactions(21). The failure to effect induction of dsdA+ expression atany in vivo plasmid-specified activator concentration that wecould attain, however, is puzzling in view of in vitro activa-tion and low-level trans activation in vivo with dsdC+(F')(2). It suggests that there is a critical arrangement of DsdCprotein and DNA that is necessary for induction and that wedid not obtain that in vivo for the plasmid trans situation. Itmay be that the DsdC protein, like certain other regulatoryproteins (10, 11, 25), recognizes multiple sites in the dsdApromoter-i.e., the repeated sequences-and that interac-tions between bound protein molecules promote transcrip-tion initiation.Our findings essentially rule out a repression component

for dsdA+ regulation and indicate an activation mechanismin which DNA tertiary structure plays a crucial role intranscription initiation. If the two repeats defined by dsdApmutations are dsdC protein-binding sites, they and theadjacent cAMP-CAP consensus sequence are, at bp -95 to

-120, too far above the transcription start for side-by-sideinteractions of bound activators with RNA polymerase.However, formation of a DsdC protein network in thewild-type promoter, as well as binding of cAMP-CAP, may

distort the DNA in the transcription initiation region suchthat initiation can occur. The other agents that influencedsdA expression-dsdAp mutations, temperature in dsdC(Con) mutants and in the dsdApl mutant, superhelicity andcontext-would also distort this DNA to a characteristic

extent, resulting in more or less efficient expression, accord-ing to how well the induced configuration is approximated.

ACKNOWLEDGMENTS

This work was supported by American Cancer Society grantNP-557 and National Science Foundation grant DMB8411383 toE.M.We thank Marion McGlynn for skillful technical assistance and

Amit Baneree for helpful discussions.

REFERENCES1. Bloom, F. R., and E. McFall. 1975. Isolation and characteriza-

tion of D-serine deaminase constitutive mutants by utilization ofD-serine as sole carbon or nitrogen source. J. Bacteriol. 121:1078-1084.

2. Bloom, F. R., E. McFall, M. C. Young, and A. M. Carothers.1975. Positive control in the D-serine deaminase system ofEscherichia coli K-12. J. Bacteriol. 121:1093-1101.

2a.Bornstein, S. M. 1986. Ph.D. thesis. New York University, NewYork.

3. Bornstein-Forst, S., E. McFall, and S. Palchaudhuri. 1987. Invivo D-serine deaminase transcription start sites in wild-typeEscherichia coli and in dsdA promoter mutants. J. Bacteriol.169:1056-1060.

4. Carothers, A. M., M. C. Heincz, and E. McFall. 1980. Positioneffect on expression of dsd genes cloned onto multicopy plas-mids. J. Bacteriol. 142:185-190.

5. Debarbouille, M., P. Cossart, and 0. Raibaud. 1982. A DNAsequence containing the control sites for gene malT and for themalPQ operon. Mol. Gen. Genet. 185:88-92.

6. Ebright, R., P. Cossart, B. Gicquel-Sanzey, and J. Beckwith.1984. Mutations that alter the DNA sequence specificity of thecatabolite gene activator of E. coli. Nature (London) 311:232-235.

7. Heincz, M. C., S. M. Bornstein, and E. McFall. 1984. Purifica-tion and characterization of D-serine deaminase activator pro-tein. J. Bacteriol. 160:42-49.

8. Heincz, M. C., N. E. Kelker, and E. McFall. 1978. Positivecontrol of D-serine deaminase synthesis in vitro. Proc. Natl.Acad. Sci. USA 75:1695-1699.

9. Heincz, M. C., and E. McFall. 1978. Role of the dsdC activatorin regulation of D-serine deaminase synthesis. J. Bacteriol.136:96-103.

10. Huo, L., K. J. Martin, and R. Schleif. 1988. Alternative DNAloops regulate the arabinose operon in Escherichia coli. Proc.Natl. Acad. Sci. USA 85:5444-5448.

11. Lee, N., C. Francklyn, and E. Hamilton. 1987. Arabinose-induced binding of AraC protein to aral2 activates the araBADoperon promoter. Proc. Natl. Acad. Sci. USA 84:8814-8818.

12. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

13. Marceau, M., E. McFall, S. D. Lewis, and J. A. Shafer. 1988.D-Serine dehydratase from Escherichia coli: DNA sequence andidentification of catalytically inactive glycine to aspartic acidvariants. J. Biol. Chem. 263:16926-16933.

14. McFall, E. 1964. Pleiotropic mutations in the D-serine deami-nase system of Escherichia coli. J. Mol. Biol. 9:754-762.

15. McFall, E. 1967. "Position effect" on dominance in the D-serinedeaminase system of Escherichia coli K-12. J. Bacteriol. 94:1989-1993.

16. McFall, E. 1973. The role of adenosine 3',5'-cyclic monophos-phate and its specific binding protein in the regulation ofD-serine deaminase synthesis. J. Bacteriol. 113:781-785.

17. McFall, E. 1975. Escherichia coli K-12 mutant forming a tem-perature-sensitive D-serine deaminase. J. Bacteriol. 121:1074-1077.

18. McFall, E. 1986. cis-Acting proteins. J. Bacteriol. 167:429-432.19. McFall, E. 1987. The D-serine deaminase operon, p. 1520-1526.

In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik,M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli andSalmonella typhimurium: cellular and molecular biology, vol. 2.

J. BACTERIOL.

on Septem

ber 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

dsdA-dsdC INTERGENIC REGION 1167

American Society for Microbiology, Washington, D.C.19a.McFall, E. Unpublished data.20. McFall, E., and M. C. Heincz. 1983. Identification and control of

synthesis of the dsdC activator protein. J. Bacteriol. 153:872-877.

21. Mitraki, A., and J. King. 1989. Protein folding intermediates andinclusion body formation. Biotechnology 1:690-697.

22. Morisato, D., J. C. Way, H.-J. Kim, and N. Kleckner. 1983.TnJO transposase acts preferentially on nearby transposon endsin vivo. Cell 32:799-807.

23. Ow, D. W., V. Sundaresan, D. M. Rothstein, S. E. Brown, andF. M. Ausubel. 1983. Promoters regulated by the glnG (n:trC)and nifA gene products share a heptameric consensus sequencein the -15 region. Proc. Natl. Acad. Sci. USA 80:2524-2528.

24. Palchaudhuri, S., V. Patel, and E. McFall. 1988. DNA sequenceof the D-serine deaminase activator gene dsdC. J. Bacteriol.170:330-334.

25. Ptashne, M. 1986. Gene regulation by proteins acting nearby andat a distance. Nature (London) 322:677-701.

26. Reitzer, L. J., and B. Magasanik. 1985. Expression of glnA isregulated at tandem promoters. Proc. Natl. Acad. Sci. USA82:1979-1983.

27. Sanger, F., and A. R. Coulson. 1975. A rapid method fordetermining sequences in DNA by primed synthesis with DNApolymerase. J. Mol. Biol. 94:441-448.

28. Southern, E. M. 1975. Detection of specific sequences amongDNA fragments separated by gel electrophoresis. J. Mol. Biol.98:503-517.

29. Tessman, E. S. 1986. Mutants of bacteriophage S13 blocked ininfectious DNA synthesis. J. Mol. Biol. 17:218-236.

30. Venkatesan, M. M., J. M. Buysse, and D. J. Kopecko. 1988.Characterizatioh of invasion plasmid antigen genes (ipaBCD)from Shigella flexneri. Proc. Natl. Acad. Sci. USA 85:9317-9321.

31. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. ImprovedM13 phage cloning vectors and host strains: nucleotide se-quences of the M13mp18 and pUC19 vectors. Gene 33:103-119.

VOL. 173, 1991

on Septem

ber 10, 2020 by guesthttp://jb.asm

.org/D

ownloaded from