Analysis of a Gene That Suppresses the Morphological Defect of

JOURNAL OF BACTERIOLOGY, May 1996, p. 2867–2875 Vol. 178, No. 100021-9193/96/$04.0010Copyright q 1996, American Society for Microbiology

Analysis of a Gene That Suppresses the Morphological Defectof Bald Mutants of Streptomyces griseus

LEE ANN MCCUE,† JANGYUL KWAK, JOHN WANG, AND KATHLEEN E. KENDRICK*

Department of Microbiology, Ohio State University, Columbus, Ohio 43210

Received 20 November 1995/Accepted 2 March 1996

When present in multiple copies, orf1590 restored sporulation to class IIIA bald mutants of Streptomycesgriseus, which form sporulation septa and thick spore walls prematurely. The orf1590 alleles from class IIIAbald mutants restored sporulation upon introduction at a high copy number into those same mutants, and thenucleotide sequence of one of these alleles was identical to that of the wild-type strain. We conclude thatoverexpression of orf1590 suppresses the defect in class IIIA bald mutants. Previous nucleotide sequence andtranscript analyses suggested that orf1590 could encode two related proteins, P56 and P49.5, from nestedcoding sequences. A mutation that prevented the synthesis of P56 without altering the coding sequence forP49.5 eliminated the function of orf1590, as did amino acid substitutions in the putative helix-turn-helixdomain located at the N terminus of P56 and absent from P49.5. To determine the coding capacity of orf1590,we analyzed translational fusions between orf1590 and the neo gene from Tn5. Measurement of the expressionof fusions to the wild-type and mutant alleles of orf1590 indicated that P56 was the sole product of orf1590during vegetative growth. Attempts to generate a nonfunctional frameshift mutation in orf1590 were unsuc-cessful in the absence of a second-site bald mutation, suggesting that orf1590may be required during vegetativegrowth by preventing early sporulation. Our results are consistent with the hypothesis that P56 at a high leveldelays the premature synthesis of sporulation septa and spore walls in class IIIA mutants.

Members of the bacterial genus Streptomyces are soil organ-isms that undergo both biochemical and morphological differ-entiation in response to environmental signals. The life cycle ofstreptomycetes is marked by the formation of chains of uni-nucleoidal spores from multinucleoidal, branched filamentsknown as hyphae. A key event in the formation of the spores isthe relatively synchronous deposition of sporulation septa atregular intervals throughout a sporulating hypha, delimitingthe developing spore compartments. In concert with septumformation, unit copies of the DNA partition into the compart-ments to ensure that each developing spore is viable. Afterseptum formation, the walls of the spore compartment thickenduring spore maturation.When differentiation occurs on a solid surface such as a

medium solidified with agar, the onset of spore formation ismarked by the projection of aerial hyphae upward from thesubstrate hyphae that were formed during vegetative growth.The spores are generally believed to develop only from theaerial hyphae. Mutants that cannot make aerial hyphae aredesignated as bald because their colonies lack the hairy ap-pearance characteristic of differentiating colonies. Four classesof bald mutants of Streptomyces griseus were identified by Bab-cock and Kendrick (3). Class I mutants appear to be blocked atthe earliest stage of sporulation. Class II mutants, which arealso blocked at an early stage, are conditionally bald, i.e.,sporulation occurs when a class II mutant is plated adjacent tothe wild-type strain or a class III or IV mutant. Class IIImutants have been divided into three genetic complementationgroups: class IIIA mutants sporulate when transformed withorf1590 (3), class IIIB mutants sporulate in the presence of

bldA (26, 33), and class IIIC mutants do not sporulate in thepresence of either orf1590 or bldA.The gene orf1590 was identified on the basis of its ability to

restore sporulation to some class III bald mutants of S. griseus(3). Subsequently, the highly similar orf1590 counterpart wasisolated from Streptomyces coelicolor (33). The nucleotide se-quences of both genes and transcription studies of the S. griseusorf1590 suggested that two proteins could be encoded byorf1590: the longer one, P56, contains a potential helix-turn-helix motif at its amino terminus, and the shorter one, P49.5,lacks the helix-turn-helix motif but otherwise is identical to P56(Fig. 1) (4, 33). The presence of an in-frame TTA within theregion of orf1590 common to both P56 and P49.5 suggestedthat translation of the orf1590 transcript would require thebldA gene product, a tRNAUUA

Leu that is a regulator of sporu-lation (28, 29, 34).Analysis of the expression of orf1590 from S. griseus revealed

two apparent transcripts, one of which could encode both P56and P49.5 and the other of which could encode only P49.5 froma putative translation initiation site located within the P56coding sequence (Fig. 1) (4). It remained to be determined,however, whether both proteins were produced from orf1590 invivo. Moreover, the previous results had not clearly indicatedwhether orf1590 was complementing or suppressing the mor-phological defect in the class IIIA bald mutants of S. griseus,which form their sporulation septa and spore walls prema-turely and subsequently do not complete the sporulation pro-cess (25, 33). Because our earlier results indicated that orf1590was transcribed at very low levels during both vegetativegrowth and sporulation (4), we examined the coding capacityof orf1590 by in vivo techniques that would permit detectionand functional analysis of small amounts of gene product. Inthis study, we present evidence that overexpression of orf1590suppresses rather than complements the morphological defectof the class IIIA bald mutants and that production of P56 isnecessary for this effect.

* Corresponding author. Mailing address: Department of Microbi-ology, Ohio State University, 484 West 12th Ave., Columbus, OH43210. Phone: (614) 292-1440. Fax: (614) 292-9195 or (614) 292-8120.Electronic mail address: [email protected].† Present address: Wadsworth Center for Laboratories and Re-

search, New York State Department of Health, Albany, NY 12201.

2867

on April 9, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

MATERIALS AND METHODS

Bacterial strains and plasmids. The wild-type strain of S. griseus, NRRLB-2682, and the class IIIA sporulation-deficient strains of S. griseus, SKK1014,-1015, and -1016, were described previously (3, 22). Streptomyces lividans TK24 (agift of D. A. Hopwood) was used as the host for the construction of plasmids onstreptomycete replicons and for translational fusion analyses, and S. coelicolorM145 (provided by M. J. Bibb) was used as the host for gene disruption exper-iments and as the source of RNA for the transcription studies. Escherichia coliTB1 was used for routine plasmid constructions and analysis, E. coli JM109 wasused for production of single-stranded DNA for site-directed mutagenesis, andE. coli BMH 71-18 mutS was used to suppress mismatch repair during themutagenesis.The plasmids constructed for this study are listed in Table 1. To mutagenize

orf1590, the 2.76-kb BamHI fragment containing the wild-type orf1590 allele wasligated to BamHI-digested pSelect-1 (Promega Corp., Madison, Wis.) and intro-duced into E. coli JM109. The mutated alleles were also used as templates for thePCR amplifications prerequisite to the translational fusion studies. For func-tional analysis of orf1590 from S. coelicolor and of the mutated alleles, fragmentswere moved from E. coli replicons (Table 1) to BglII-digested pIJ702, introducedinto S. lividans TK24, and then transferred to S. griseus. In all cases, orf1590 wasinserted in the orientation opposing the mel promoter.For construction of the translational fusions, the appropriate regions of the

orf1590 locus were amplified by PCR with an upstream oligonucleotide (synthe-sized by the Biochemical Instrument Center [The Ohio State University, Colum-bus] or Ransom Hill Bioscience, Inc. [Ramona, Calif.]) that contained a BglII site

and a downstream oligonucleotide that contained a BamHI site (Table 2). Theamplified fragments were digested with these two restriction endonucleases andligated to similarly digested pIJ2920 (20); each mutation was confirmed byanalysis of the nucleotide sequence. After digestion with BamHI and BglII andligation to BamHI-digested pIJ686, the DNA was introduced into S. lividansTK24. The orientations of the fragments in pIJ686 were determined by digestionwith BamHI and EcoRV.The 2.76-kb BamHI fragment from pKK330 was ligated to BamHI-digested

pXE4 (19) to place orf1590 from S. griseus on a low-copy-number vector. The2.76-kb BamHI fragments from the mutant strains SKK1014, -1015, and -1016were isolated by ligation to pUC18, and the nucleotide sequence of the fragmentfrom SKK1015 was determined. The BamHI fragments from the mutant strainswere ligated to BglII-digested pIJ702 and introduced into SKK1014, -1015, and-1016 to determine their ability to suppress the bald phenotype.Growth and transformation of bacteria. Strains of S. griseus were maintained

and cultivated essentially as described previously (3). Spore suspensions of S.lividans and S. coelicolor were prepared from R2YE agar (17) cultures supple-mented with 40 mg of thiostrepton per ml as necessary. The spores were har-vested by rolling sterile 3-mm-diameter glass beads over the surface of the agar(16). Liquid cultures of S. lividans and S. coelicolor were inoculated with sporesthat had been heat activated and germinated (17), or spores were inoculateddirectly into twofold-concentrated liquid medium and an equal volume of 30%(wt/vol) polyethylene glycol 8000 was added after 7 h of incubation at 308C. E.coli cells were made competent to take up plasmid DNA with CaCl2 and trans-formed by standard methods (38).Protoplasts of S. griseus were prepared by a modification of the procedure of

Babcock and Kendrick (3). Spores or mycelia from 1 to 2 colonies were inocu-lated into 3 ml of Trypticase soy broth (TSB) containing 1.0% glycine and 10 mMMgCl2 in a culture tube (18 by 150 mm) containing a 6-mm-diameter glass bead.Cells were harvested after 24 h of incubation with shaking at 308C by centrifu-gation in a microcentrifuge tube for 10 s. The cells were washed once with 1 mlof P buffer (17). After suspension of the cells in 1 ml of P buffer, the protoplaststhat formed during 15 min of incubation in the presence of 1 mg of lysozyme perml were diluted with 4 ml of P buffer and filtered through sterile glass wool.Protoplasts were collected by centrifugation for 10 min at 3,000 rpm at roomtemperature, suspended in P buffer to 4 3 109 protoplasts per ml, and usedimmediately for transformation. Protoplasts of S. lividans were prepared asdescribed earlier (3).Transformation of Streptomyces protoplasts with plasmid DNA was essentially

as described before (3, 17). Efficient transformation of the wild-type strain of S.griseus was precluded by restriction of DNA originating from other Streptomycesspecies or E. coli. Therefore, DNA ligation reactions were screened for thecorrect construction in S. lividans, and clones were transferred to S. griseus bymethylating the plasmid DNA sequentially with SssI methylase and AluI meth-ylase (New England Biolabs, Inc., Beverly, Mass.). Regeneration of transformedS. griseus protoplasts occurred on SpMR medium (3) buffered with TES[N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid]-KOH (pH 7.5) andcontaining 0.3% casein hydrolysate (HyCase; ICN Pharmaceuticals, Inc., CostaMesa, Calif.) for the bald mutants, whereas S. lividans and S. coelicolor proto-plasts were regenerated on R2YE medium. The ability of orf1590 to restoresporulation when at a high but not a low copy number was evident when thetransformants were incubated at 338C after application of the antibiotic selec-tion, at which temperature the bald phenotype of class IIIA mutants was mostclearly visualized.Isolation and manipulation of DNA. Minipreparations of plasmid DNA from

E. coli were prepared by standard methods (3). Streptomyces cultures were grownin TSB containing 5 mg of thiostrepton per ml in a culture tube (18 by 150 mm)containing a 6-mm-diameter glass bead. Plasmid DNA was isolated essentially asdescribed previously (3). DNA fragments used for cloning and as probes wereisolated after excision of the appropriate ethidium bromide-stained restrictionfragment by the phenol-freeze-fracture method (18). The orf1590 alleles fromthe class IIIA mutants were isolated by ligation of 2.5- to 3.0-kb BamHI frag-ments of genomic DNA to pUC18 and screening by colony hybridization (38)with pKK112 (3) that had been labeled by the random primer method (Prime-a-Gene; Promega) as a probe. Single-stranded phagemid DNA was isolated bythe procedure of Lin et al. (30).PCR. DNA fragment amplification by PCR was performed essentially by the

method of Guilfoile and Hutchinson (13), with 20 pmol of each oligonucleotideprimer (Table 2), 2 ng of linearized plasmid template, and 5 U of Taq polymerase(Perkin-Elmer Cetus, Norwalk, Conn.). The template was initially denatured byheating at 968C for 5 min followed immediately by 30 cycles of amplification, i.e.,denaturation at 968C for 1.5 min, annealing at 608C for 1.0 min, and extension at728C for 1 to 4 min (dependent on the fragment length) with a 5-s autoextensionperiod.Cloning of the PCR-generated fragments for construction of translational

fusions was facilitated by proteinase K treatment (8). PCR fragments used asprobes for S1 nuclease protection experiments were labeled with T4 polynucle-otide kinase (Gibco-BRL, Gaithersburg, Md.) (2). The end-labeled fragment wasisolated with a QIAquick-spin PCR purification column (QIAGEN Inc., Chats-worth, Calif.).DNA sequence determination. Sequences were determined by the dideoxynu-

cleotide chain termination method (39) with Sequenase version 2.0 DNA poly-

FIG. 1. (A) Features of orf1590. Numbers indicate the nucleotides at whichthe designated features occur. ATG, putative translation initiation site for P56;ATGGTG, putative translation initiation site for P49.5; P1 and P2, transcriptsidentified by Babcock and Kendrick (4); hth, putative helix-turn-helix domain;TTA, location of the codon that would be recognized by tRNAUUALeu . (B) ELISAanalysis of the production of NPTII fusion protein from translational fusions tofragments of orf1590. Regions containing the putative translational start sites(open boxes) and the neo reporter gene (shaded boxes) are diagrammed. Thelarge arrows represent the direction of transcription and translation fromorf1590: filled arrows designate the upstream translation initiation site (for P56),and open arrows designate the downstream translation initiation site (for P49.5).Fragments containing site-directed mutations are indicated as follows: p, frame-shift mutation; GTC, wild-type GTG changed to GTC; D, deletion of ATGGTG.The nucleotides marking the end points of the fragments are numbered asdescribed in Babcock and Kendrick (4). The amount of hybrid protein detectedin extracts prepared from transformants of S. lividans TK24 as measured byELISA is indicated as nanograms of NPTII per milligram of total cellular pro-tein. Values are reported as average 6 standard deviation of duplicate extracts,each assayed in triplicate, after subtraction of the background (,0.05 ng ofNPTII per mg of protein).

2868 MCCUE ET AL. J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

merase (United States Biochemical Corp., Cleveland, Ohio). Sequence laddersfor the high-resolution S1 nuclease protection experiments were generated byend labeling the oligonucleotide primers (oligonucleotides 41, 58, and 120)(Table 2) and by use of Sequenase in reactions containing each deoxynucleosidetriphosphate (Pharmacia Biotech Inc., Piscataway, N.J.) at 1.5 mM. Reactionproducts were resolved in 6% denaturing polyacrylamide gels (2).Site-directed mutagenesis. Site-directed mutagenesis was performed with the

Altered Sites Mutagenesis System (Promega) with mutagenic oligonucleotides(Table 2). Each mutation was confirmed by determination of the nucleotidesequence in the region surrounding the mutated site.RNA isolation. S. griseus cultures were grown to the mid-exponential phase in

glucose-ammonia minimal medium supplemented with 1% (wt/vol) casein hy-drolysate (24); S. coelicolor cultures were grown similarly, except that the FeCl3was replaced with 4 ml of trace elements solution per liter (17). RNA wasisolated by the procedure of Covey and Smith (17). The concentration of theRNA was determined by measuring the A260.High-resolution S1 nuclease protection experiments (17) were performed by

combining 40 mg of freshly prepared RNA in diethylpyrocarbonate-treateddeionized water with 200,000 cpm of DNA probe prepared as described above.The solution was frozen at 2708C and lyophilized in a Savant SpeedVac vacuumcentrifuge. The RNA and probe were dissolved in 20 ml of S1 hybridization buffer(2) and denatured at 908C for 30 min. The temperature was slowly cooled to 48Cabove the melting temperature of the DNA probe, and the mixture was allowed

to hybridize overnight. The upstream labeled 59 end of the DNA probe was nothomologous to the RNA template and was subsequently hydrolyzed by the S1nuclease treatment. The dried gel was exposed to Kodak X-OMAT AR film for30 to 48 h at 2708C with intensifying screens.NPTII ELISA. Extracts containing total protein from S. lividans harboring

pIJ686 derivatives were prepared by growing cultures to the late exponentialphase (A500 5 6.0) in 50 ml of TSB containing 15% (wt/vol) polyethylene glycol8000. Thiostrepton was excluded from these cultures because the antibioticcaused an approximately threefold decrease in the amount of neomycin phos-photransferase (NPTII) fusion protein present in the cells and reduced thegrowth rate slightly. Cells were prepared for enzyme-linked immunosorbentassay (ELISA) detection of NPTII as described in the manufacturer’s instruc-tions (5 Prime33 Prime, Inc., Boulder, Colo.) except that the mycelia weredisrupted at 14,000 lb/in2 with a French pressure cell. Each extract was decantedand stored in aliquots at 2708C. Protein concentrations in the cell extracts weredetermined by the method of Ehresmann et al. (11).Disruption of orf1590 in S. coelicolor M145. Plasmids were constructed in

which the aph gene prepared by PCR amplification from pIJ61 (17) as templatewas inserted as a BglII fragment in both orientations into the unique BamHI siteof orf1590 from S. coelicolor in the vector pIJ2925 (pKK780 and -781). The 3.2-kbBglII fragment containing the disrupted orf1590 was ligated to BamHI-digestedpDH5, a suicide vector containing the colEI replicon and thiostrepton resistancedeterminant (tsr) (15), to generate pKK786 and -787. Each plasmid was passaged

TABLE 1. Plasmids used in this study

Plasmid Construction and/or referencea

Plasmids in vector pUC18pKK209 ...............................................................................3-kb BamHI fragment from S. griseus containing entire orf1590 (4)pKK302 ...............................................................................7-kb SalI fragment from S. coelicolor M145 containing orf1590 homologpKK826 ...............................................................................2.76-kb BamHI fragment containing orf1590 allele from SKK1016pKK827 ...............................................................................2.76-kb BamHI fragment containing orf1590 allele from SKK1014pKK828 ...............................................................................2.76-kb BamHI fragment containing orf1590 allele from SKK1015

Plasmid in vector pXE4pKK760 ...............................................................................2.76-kb BamHI fragment from pKK330

Plasmids in vector pSelect-1pKK310 ...............................................................................Introduction of T at nt 1045 of orf1590; generates EcoRI site and frameshift mutationpKK311 ...............................................................................Change of C to G at nt 1065 of orf1590, changing Arg-28 to Gly in P56pKK312 ...............................................................................Change of A to G at nt 1050 of orf1590, changing Asn-23 to Asp in P56pKK313 ...............................................................................Change of A to T at nt 1050 of orf1590, changing Asn-23 to Tyr in P56pKK314 ...............................................................................Change of C to A at nt 1054 of orf1590, changing Ala-24 to Asp in P56pKK315 ...............................................................................Change of GC to CG at nt 1053 and 1054 of orf1590, changing Ala-24 to Arg in P56pKK330 ...............................................................................2.76-kb BamHI fragment (wild-type allele) ligated to similarly digested vectorpKK701 ...............................................................................Deletion of ATGGTG at nt 1148 to 1153 of orf1590pKK702 ...............................................................................Change of G to C at nt 1153 of orf1590

Plasmids in vector pIJ2920pKK711 ...............................................................................PCR-amplified fragment from pKK209, with oligonucleotides 63 and 67ApKK712 ...............................................................................PCR-amplified fragment from pKK310, with oligonucleotides 63 and 67ApKK735 ...............................................................................PCR-amplified fragment from pKK209, with oligonucleotides 63 and 64pKK736 ...............................................................................PCR-amplified fragment from pKK209, with oligonucleotides 63 and 65ApKK737 ...............................................................................PCR-amplified fragment from pKK209, with oligonucleotides 66 and 67ApKK738 ...............................................................................PCR-amplified fragment from pKK701, with oligonucleotides 66 and 67ApKK739 ...............................................................................PCR-amplified fragment from pKK702, with oligonucleotides 66 and 67ApKK740 ...............................................................................PCR-amplified fragment from pKK310, with oligonucleotides 63 and 65A

Plasmids in vector pIJ2925pKK349 ...............................................................................3.3-kb SmaI fragment from pKK302, containing entire orf1590 from S. coelicolor as

well as downstream ORFpKK703 ...............................................................................1.9-kb fragment from pKK349, containing orf1590 from S. coelicolor but lacking

downstream ORF

Plasmid in vector pIJ702pKK200 ...............................................................................3-kb BamHI fragment from S. griseus, containing entire orf1590 (3)

Plasmids in vector pDH5pKK786 ...............................................................................3.2-kb BglII fragment from pKK780, containing disrupted orf1590 from S. coelicolor;

aph in same orientation as orf1590pKK787 ...............................................................................3.2-kb BglII fragment from pKK781, containing disrupted orf1590 from S. coelicolor;

aph opposing orf1590a ORF, open reading frame.

VOL. 178, 1996 SUPPRESSION OF BALD MUTANTS 2869


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

through ET12567 (31) to produce unmethylated DNA and subsequently intro-duced into S. coelicolor M145 by protoplast transformation with selection forneomycin and thiostrepton resistance. For both orientations of aph, cointegratesresulting from a single recombination event were identified by Southern analysis(2) by use of the Genius II kit (Boehringer-Mannheim, Indianapolis, Ind.). Onecointegrate of each type was subcultured on R2YE or SpM agar (with or withoutneomycin) for a total of five passages. Isolated colonies were examined toidentify strains that were resistant only to neomycin.

RESULTS AND DISCUSSION

Suppression of the morphological defect in class IIIA mu-tants of S. griseus. The introduction of pKK200, which containsorf1590 from S. griseus as a 2.76-kb BamHI fragment on amulticopy plasmid (pIJ702) (3), into the class IIIA bald mutantSKK1015 restored sporulation but not to wild-type levels. Li-gation of the BamHI fragment to the low-copy-number vectorpXE4 (19) did not restore sporulation upon introduction intoSKK1015. The orf1590 alleles from the class IIIA mutants,SKK1014, SKK1015, and SKK1016, were isolated and ligatedas 2.76-kb BamHI fragments to pIJ702 that had been digestedwith BglII. Each of these alleles at a high copy number was ableto restore sporulation to the homologous mutant strain. Thenucleotide sequence of orf1590 as well as the adjacent 323nucleotides (nt) of upstream DNA and 183 nt of downstreamDNA from SKK1015 showed no difference from that of thewild-type allele. These results indicate that overexpression oforf1590 suppressed the defect in class IIIA bald mutants.Mutational analysis of orf1590 function. The orf1590 ho-

molog from S. coelicolor, which also restored sporulation to theclass IIIA bald mutants when present at a high copy number,could encode proteins of 473 and 417 amino acids as homologs

of P56 and P49.5, respectively. This gene contains the internaltranslation initiation site as well as the region encoding theputative DNA-binding domain but lacks the coding sequencefor 56 amino acids that are present in the central region of bothP56 and P49.5 in S. griseus (33). Thus, this central region isunnecessary for the function of orf1590 in the suppressionassay. To determine whether P56 or P49.5 was required forfunction, we assessed the ability of mutated forms of orf1590 torestore sporulation to SKK1015. On the basis of the nucleotidesequence surrounding the proposed P49.5 translation initiationsite (. . .AAGGGGAUGGUGCCGCAG. . .), it was not clearwhether the AUG or the GUG might serve as the initiationcodon. We therefore altered both codons. The mutant allelecontaining GTC in place of GTG, which would not alter theprimary structure of P56 but could prevent synthesis of P49.5by disrupting its translation initiation site, was unimpaired in itsability to restore sporulation to SKK1015 (Fig. 2). In contrast,sporulation was prevented in the case in which the ATGGTGwasdeleted. It is most likely that the elimination of the Met or Valfrom the primary structure of P56 prevents function of thisprotein because the distance between the ATG and the puta-tive ribosome binding site (indicated above by the underline)upstream of the P49.5 coding sequence is only 2 nt. The min-imal distance between the translation initiation codon and theribosome binding site in E. coli is 3 nt (37). Our examination ofall sequences of Streptomyces genes from GenBank (version83) indicated that the minimal distance was 4 nt.To detect the possible function of P49.5 in the absence of

P56, we introduced a single nucleotide at position 1045, in theregion unique to P56. This frameshift mutation resulted in the

TABLE 2. Oligonucleotides used in this study

Oligonucleotide Sequencea Positionb Site

S. griseus orf1590site-directedmutagenesis:

14 59-CTCGCGGGAATTCAGTAACGCC-39 S. griseus nt 1034–1054 Frameshift16 59-CGCCAAGGGG(D)CCGCAGGGCG-39 S. griseus nt 1138–1163 Delete ATGGTG35 59-GGGCTCGCCGGGCGGGTCAA-39 S. griseus nt 1055–1074 Arg-28 to Gly36 59-GGGAATCAGTGACGCCGGGC-39 S. griseus nt 1039–1058 Asn-23 to Asp37 59-GGGAATCAGTTACGCCGGGC-39 S. griseus nt 1039–1058 Asn-23 to Tyr38 59-GGAATCAGTAACGACGGGCTCG-39 S. griseus nt 1040–1061 Ala-24 to Asp39 59-GGAATCAGTAACCGCGGGCTCGCC-39 S. griseus nt 1040–1063 Ala-24 to Arg40 59-AGGGGATGGTCCCGCAGGGCG-39 S. griseus nt 1143–1163 Change GTG to GTC

S1 nucleaseprotection:

41 59-CGGGACCGGCCGGCCGAGCT-39c S. griseus nt 1200–1219 Downstream of P258 59-CGGCACCGGGCGGCCGAGCT-39c S. coelicolor nt 480–499 Downstream of P2120 59-CGGCCCGCGTTGCTCCCTGG-39c S. griseus nt 981–1006; S.

coelicolor nt 267–286Downstream of P1

121 59-CGATGCGATGATAGGGAGAACTACGCTGTG-39 S. griseus nt 805–825 Upstream of P1122 59-ACGTACGTACGTCGTGGCCGGAGAAGAGC-39 S. coelicolor nt 49–68 Upstream of P1123 59-GATCCGATCCTCGCCCTCGCGGGAATCAG-39 S. griseus nt 1028–1047; S.

coelicolor nt 308–327Upstream of P2

S. griseus orf1590translationalfusions:

63 59-AGCTAGATCTGGGAGCACGCCGCACCCCAGC-39 S. griseus nt 702–722 Upstream of P56 start64 59-TCGAGGATCCCGGCCCGCGTTGCTCCCTGG-39c S. griseus nt 987–1006 Downstream of P56 start65A 59-TCGAGGATCCGGCGAGCCCGGCGTTACTG-39c S. griseus nt 1045–1063 Downstream of P56 start66 59-AGCTAGATCTGGGAGCAACGCGGGCCGAAC-39 S. griseus nt 990–1009 Upstream of P49.5 start67A 59-TCGAGGATCCCGGCGCCGCGCCCTGCGG-39c S. griseus nt 1154–1171 Downstream of P49.5 start

aMutated nucleotides are underlined. (D), site of deletion.b Relative to the nucleotide sequences registered in GenBank (accession numbers M32687 and M80614).c Represents sequence of antisense strand.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

positioning of a stop codon (UAA) 5 nt further downstream.Translational fusion studies (see below) confirmed that thisframeshift mutation effectively prevented synthesis of P56. Theorf1590 allele containing the P56 frameshift mutation (pKK725)did not restore sporulation to SKK1015 in 95% of the transfor-mants (Fig. 2). Sporulation in the remaining 5% of the popu-lation was the result of the recombination of the plasmid-bornemutant allele (containing a unique EcoRI site as a conse-quence of the frameshift mutation) with the chromosomalwild-type allele such that one of the plasmid molecules ac-quired the wild-type allele by a process resembling markerexchange. Subsequent replication to generate multiple copiesof plasmid bearing the wild-type allele would be expected tosuppress the bald phenotype of SKK1015. (We speculate thatthe other mutant alleles described below and in Fig. 2, whichwould be expected to yield altered gene products, may not havegiven rise to sporulating recombinants because of dominanceor subunit mixing.)On the basis of these observations, we reasoned that we

could construct a frameshift mutation in the chromosomalcopy of orf1590 by introducing the mutant allele into the wild-type strain on a multicopy plasmid and screening transfor-mants for loss of the plasmid-borne thiostrepton resistance.Among several thiostrepton-sensitive isolates screened bySouthern hybridization analysis, only one isolate contained theframeshift mutation as noted by the acquisition of an EcoRIsite characteristic of the insertion mutation in the chromo-somal copy of orf1590 (data not shown). This transformantdisplayed a bald phenotype, was cross-fed to sporulate by the

wild-type strain and other class III bald mutants, and did notsporulate in the presence of orf1590 at either a high or a lowcopy number. Thus, the bald phenotype was a consequence ofa second-site mutation to generate a strain resembling thosecategorized as class II bald mutants (3). Attempts to constructa disrupted orf1590 in S. coelicolor by use of the nonreplicatingplasmid pDH5 (15) and a neomycin-resistance cassette (17)were unsuccessful: none of 800 cointegrate derivatives thatwere grown in the absence of selection for thiostrepton resis-tance (mediated by tsr, the plasmid-borne resistance determi-nant) was resistant only to neomycin. These results are consis-tent with orf1590 being a gene required for vegetative growth.Because the null mutation in S. griseus could apparently begenerated only in the presence of a second-site mutation thatblocked sporulation at an early stage, it seems likely that loss oforf1590 function is lethal in a cell capable of normal sporula-tion. This hypothesis predicts the facile disruption of orf1590 inclass I and II bald mutants of S. griseus that are blocked at anearly stage of sporulation. We presume that the sporulation ofthe orf1590 frameshift mutant (containing the second-site baldmutation) in response to the diffusible inducer occurred be-cause the inducer was not produced by the wild-type strainuntil after the mutant colony had formed. The most likelyinterpretation of our results is that orf1590 prevents prematuresporulation and has as its target one or more genes (or geneproducts) required early in sporulation.Analysis of the probability that P56 is a DNA-binding pro-

tein by the scoring system of Dodd and Egan (10) suggestedthat the helix-turn-helix nature of the N terminus of P56 was

FIG. 2. Restoration of sporulation to SKK1015 by mutated alleles of orf1590. (A) Diagram of the site and effect of the mutation on P56. Numbers written assubscripts indicate the amino acid residue of P56; a single underline indicates the putative N terminus of P49.5; a double underline marks the EcoRI site generatedby the frameshift mutation; an asterisk marks the C terminus of the truncated form of P56; large capital letters indicate the turn of the putative helix-turn-helix motif.(B) Functional assay of the mutated orf1590. The column marked mutation indicates the change imposed on orf1590, and the column marked sporulation indicates theability of the mutated gene to suppress the bald phenotype of SKK1015. Symbols: 1, sporulation comparable to that of SKK1015 containing the wild-type orf1590 onpIJ702; 2, no sporulation; 2/1, mixture of nonsporulating (95%) and sporulating (5%) transformants.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

only moderately probable, exceeding the average score for P56by 2.6 standard deviations. This score is similar to that of theTrp repressor, which requires interaction with the corepressorto bind to DNA (10). We thus determined whether this regionat the N terminus of P56 was required for function by intro-ducing missense mutations into the DNA that could encodethe second helical region. Mutations were constructed to gen-erate amino acid substitutions that would be highly likely todisrupt the contact between this helix and the DNA target.Earlier analyses have indicated that these contact points mostoften include the first and second amino acid residues of thesecond helix and frequently the fifth or sixth residue as well (5,6, 9, 21, 35, 43). Our results were consistent with these pat-terns. The change of Arg-28 to Gly did not affect the functionof P56; in contrast, the mutations that altered Asn-23 or Ala-24resulted in the loss of suppressor activity (Fig. 2). Since thisputative DNA-binding domain is absent from P49.5 but isrequired for suppressor function of P56, these results supportthe hypothesis that P56 is the critical product of orf1590.Translational fusion studies. To corroborate and extend our

mutant analysis, we determined whether both putative trans-lation initiation sites in orf1590 were functional by constructing

translational fusions between the two potential N-terminalcoding regions of orf1590 and the neo gene in the vector pIJ686(23). Because pIJ686 is unstable in S. griseus, measurements oftranscription and translation initiation were made in transfor-mants of S. lividans TK24 containing regions of orf1590 fusedin-frame to the reporter gene. The transformants were firstassayed for the acquisition of neo-encoded resistance to kana-mycin; however, the levels of resistance were too low (#1mg/ml) to be reliable indicators of gene expression. Moreover,this method of measurement of expression was semiquantita-tive and required the synthesis of active protein. We thereforeused an ELISA to determine the amount of hybrid proteinencoded by these gene fusions independent of NPTII activity.Fusions were constructed to generate chimeric proteins con-

taining (i) 8 amino acids derived from the N terminus of P56,(ii) 27 amino acids from the N terminus of P56, (iii) 63 aminoacids from the N terminus of P56 and/or 8 amino acids fromthat of P49.5, or (iv) 8 amino acids from the N terminus ofP49.5. In each case, the hybrid proteins also included six aminoacids encoded by the plasmid linker region followed by theNPTII coding sequence. In these constructions, the endoge-nous orf1590 promoter(s) was required for expression of the

FIG. 3. High-resolution S1 nuclease protection studies to determine the 59 end of orf1590 transcript. (A) S. griseus orf1590 transcript determination. Lanes: 1 and3, full-length probes 1 and 2, respectively; 2 and 4, RNA isolated from a vegetatively growing culture of S. griseus B-2682 hybridized with probes 1 and 2, respectively,and treated with S1 nuclease. (B) S. coelicolor orf1590 transcript determination. Lanes: 1 and 3, full-length probes 3 and 4, respectively; 2 and 4, RNA isolated froma vegetatively growing culture of S. coelicolor M145 hybridized with probes 3 and 4, respectively, and treated with S1 nuclease. The DNA probes were generated byPCR as described in the text, and their map positions are indicated below the autoradiograms. The transcript end points and the 59 end labels are marked by asterisks.The sequence ladders (A, G, C, and T) were generated from an end-labeled oligonucleotide identical to that used as the downstream primer for PCR and reveal thesequence of the antisense strand but are represented as the sense strand on the side of the autoradiogram. E. coli tRNA, used as the negative control in each experiment,provided no protection (data not shown).



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

neo gene. Because all of the fusion junctions were locatedupstream of the in-frame TTA in orf1590 and the neo tran-script lacks UUA codons, translation of the fusion proteintranscripts would not require the bldA gene product. A rela-tively high level of NPTII was synthesized from the fusion tothe shortest N-terminal portion of P56 (pKK743) (Fig. 1). Inagreement with the studies of Reiss et al. (36), which demon-strated that fusions containing longer N-terminal regionsyielded substantially lower NPTII activity, extension of thisN-terminal region of P56 by 19 amino acids reduced the levelof NPTII 13-fold (pKK745). Further extension of the N-ter-minal portion of the fusion so that the junction occurred down-stream of the P49.5 initiation site led to a slight increase inNPTII (pKK751). Fusion of neo to the N-terminal coding re-gion of P49.5 alone did not increase the level of NPTII signif-icantly above background (pKK747); neither the substitutionof GTC for GTG nor deletion of the ATGGTG altered thebasal level of NPTII when only the P49.5 start site was present(pKK755 and pKK757). These last results suggested that dur-ing vegetative growth there was no significant promoter activitywithin the 158 nt upstream of the P49.5 coding sequence. Thisregion includes the putative downstream promoter identifiedin a previous study (4). Therefore, P49.5, if produced fromorf1590, would most likely be encoded by the transcript thatinitiates upstream of the P56 coding sequence.To distinguish translation initiating at the P56 start site from

that at the P49.5 start site, we introduced a frameshift mutationinto the P56 coding sequence at a position 72 nt upstream ofthe putative downstream transcription start site and 103 nt

upstream of the P49.5 translation start site. This mutationeliminated the synthesis of NPTII from the P56 start site(pKK749) (Fig. 1). Because this mutation also effectively abol-ished NPTII production from the fusion containing both pu-tative start sites (pKK753), we conclude that virtually all of theexpression of orf1590 emanated from the region upstream ofthe P56 coding sequence. Each fusion construction showed alow level of NPTII (,0.01 ng/mg of protein) when in invertedorientation. The amounts of immunoreactive protein detect-able by immunoblot of extracts prepared from the S. lividanstransformants (data not shown) paralleled the levels of NPTIImeasured in the ELISA.The use of the heterologous host S. lividans in these studies

raised the possibility that this species differs from S. griseus inits recognition of promoters and translation initiation sites. Wethink that this is unlikely, however, because of the following: (i)S. lividans, which contains DNA that hybridizes at high strin-gency to orf1590 (3), is closely related to S. coelicolor, whichcontains an orf1590 that is functional in S. griseus and is tran-scribed from an identical site (see below); (ii) S. lividans hasbeen used in numerous studies to identify promoters and over-express genes from a wide variety of streptomycete species (seereferences 14, 27, and 42 as examples); and (iii) the apparentorf1590 promoters that were originally identified by promoter-probe studies were evident during vegetative growth of S. livi-dans (4).Because S. lividans does not undergo submerged sporula-

tion, we were unable to examine the plasmid constructions forthe production of fusion proteins during synchronous differ-entiation. These experiments also do not exclude the possibilitythat translation of the P49.5 transcript may be coupled to thatof the P56 transcript, provided that the fusion junction be-tween the P49.5 start region and NPTII is sufficiently upstreamto exclude sites in orf1590 required for translational coupling incis, and the P56 homolog encoded by the S. lividans chromo-some cannot function in trans to promote translation from theP49.5 start site. It is alternatively possible that the introductionof the frameshift mutation in P56 had a polar effect on syn-thesis of P49.5.Analysis of transcription of orf1590 in S. coelicolor. In light

of our results that only P56 appeared to be produced fromorf1590 during vegetative growth and that there was only onefunctional promoter, we examined the transcription of orf1590in both S. griseus and S. coelicolor. In a previous study (4),apparent 59 ends of two transcripts were identified in RNAprepared from vegetative mycelia of S. griseus, but the tran-script originating further downstream diminished in amountduring sporulation. In the present study, two protected frag-ments were again observed in total RNA prepared from veg-etative mycelia of S. griseus (Fig. 3); although the signal orig-inating at the upstream site was evident in all samples, thepresence of the signal corresponding to the downstream originwas variable among RNA preparations. Total RNA from veg-etative mycelia of S. coelicolor showed only the transcript cor-responding to the upstream site (Fig. 3). Taking into accountthe translational fusion studies, we believe that the down-stream transcription start site evident in some preparations ofRNA from S. griseus is an artifact, possibly resulting from RNAdegradation or secondary structure. The orf1590 transcript hada 59 end 29 nt further upstream than previously reported (4).We attribute this discrepancy to slight degradation of RNA inthe earlier study, in which a short treatment of the mycelia withlysozyme to facilitate disruption of the cells may have beensufficient to permit limited degradation of the RNA.The 59 end of the orf1590 transcript from S. coelicolor

mapped to the same nucleotide as that from S. griseus (Fig. 4).

FIG. 4. (A) Alignment of nucleotide sequences of the region upstream oforf1590 from S. griseus (SG) and S. coelicolor (SC). The 59 ends of the transcriptsare indicated (11), and the putative 210 and 235 regions upstream of these 59ends are marked with single underlines. Vertical lines signify identical nucleo-tides, and dots indicate gaps. The proposed translation initiation codon isshaded, and the ribosome-binding site is marked with double underlines. Thetranscription start site identified in the previous study (4) is marked with an X.(B) Proposed consensus sequence for the major sigma factor of Streptomyces spp.(7, 40).



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Examination of the DNA sequence upstream of the 59 endrevealed identical 210 (TACGCT) and 235 (TTAACG) sitesin both species, which suggested that synthesis of this transcriptmay be directed by RNA polymerase containing sHrdB, thesigma factor essential for vegetative growth of S. coelicolor (7)which is also present during vegetative growth of S. griseus (32).The presence of the in-frame TTA in orf1590 indicates that

translation of the orf1590 transcript would be dependent on afunctional bldA. Although the bldA product appears to be aregulator of morphogenesis in both S. coelicolor (34) and S.griseus (26), the mechanism by which this regulation occurs hasnot been elucidated. Recent evidence indicates that a signifi-cant amount of functional tRNAUUA

Leu is present during vege-tative growth of S. coelicolor (1, 12). If this is also true for bldAfrom S. griseus, then it is likely that the low level of orf1590transcript present during vegetative growth is translated toform P56.Ueda et al. (41) reported that overexpression of orf1590

suppressed the developmental defect of an A-factor nonpro-ducing mutant (afsA) of S. griseus. By using as hosts both theafsAmutant strain HH1 (41) and another A-factor nonproduc-ing strain (3), we have not observed suppression of the mor-phological defect by orf1590 at either a high or a low copynumber. This discrepancy may be the consequence of differentgrowth conditions used to examine suppression in the twolaboratories. It is clear that the class IIIA mutants of S. griseusare not afsA mutants because the class IIIA mutants are notphenotypically suppressed when plated adjacent to the wild-type strain (3). The class IIIA mutants prematurely deposittheir sporulation septa and spore walls (25, 33) and thus aredefective in the temporal regulation of sporulation. Recentevidence also suggests a spatial defect in these mutants (25).The defect in class IIIA mutants is partially overcome by theoverproduction of P56. Our inability to obtain a null mutationin orf1590 in the absence of a second-site bald mutation sug-gests that orf1590 encodes a negative regulator of sporulation.On this basis, we propose to rename orf1590 as nrsA.

ACKNOWLEDGMENTS

We thank T. Henkin and J. Reeve for critically reading the manu-script and helpful comments, M. Bibb for suggestions for the isolationof RNA, G. Janssen and T. Kieser for facilitating the use of pIJ686, S.Horinouchi for the gift of S. griseus HH1, and G. Kleman and W.Strohl for the use of the computer program for Dodd and Egananalysis.This research was supported by grants from the Ohio State Univer-

sity Office of Research and Graduate Studies and the National ScienceFoundation (DMB-8813143 and MCB-9210743).

REFERENCES

1. Angell, S., C. G. Lewis, M. J. Buttner, and M. J. Bibb. 1994. Glucoserepression in Streptomyces coelicolor A3(2): a likely regulatory role for glu-cose kinase. Mol. Gen. Genet. 244:135–143.

2. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A.Smith, and K. Struhl. 1987. Current protocols in molecular biology. JohnWiley & Sons, Inc., New York.

3. Babcock, M. J., and K. E. Kendrick. 1988. Cloning of DNA involved insporulation of Streptomyces griseus. J. Bacteriol. 170:2802–2808.

4. Babcock, M. J., and K. E. Kendrick. 1990. Transcriptional and translationalfeatures of a sporulation gene of Streptomyces griseus. Gene 95:57–63.

5. Benson, N., and P. Youderian. 1989. Phage lambda cro protein and cIrepressor use two different patterns of specific protein-DNA interactions toachieve sequence specificity in vivo. Genetics 121:5–12.

6. Brennan, R. G., and B. W. Matthews. 1989. Structural basis of DNA-proteinrecognition. Trends Biochem. Sci. 14:286–290.

7. Brown, K. L., S. Wood, and M. J. Buttner. 1992. Isolation and characterizationof the major vegetative RNA polymerase of Streptomyces coelicolor A3(2);renaturation of a sigma subunit using GroEL. Mol. Microbiol. 6:1133–1139.

8. Crowe, J. S., H. J. Cooper, M. A. Smith, M. J. Sims, D. Parker, and D.Gewert. 1991. Improved cloning efficiency of polymerase chain reaction(PCR) products after proteinase K digestion. Nucleic Acids Res. 19:184.

9. Dallmann, G., P. Papp, and L. Orosz. 1987. Related repressor specificity ofunrelated phages. Nature (London) 330:398–401.

10. Dodd, I. B., and J. B. Egan. 1990. Improved detection of helix-turn-helixDNA-binding motifs in protein sequences. Nucleic Acids Res. 18:5019–5026.

11. Ehresmann, B., P. Imbault, and J. H. Weil. 1973. Spectrophotometric de-termination of protein concentration in cell extracts containing tRNA’s andrRNA’s. Anal. Biochem. 54:454–463.

12. Gramajo, H. C., E. Takano, and M. J. Bibb. 1993. Stationary-phase produc-tion of the antibiotic actinorhodin in Streptomyces coelicolor A3(2) is tran-scriptionally regulated. Mol. Microbiol. 7:837–845.

13. Guilfoile, P., and C. R. Hutchinson. 1992. Sequence and transcriptionalanalysis of the Streptomyces glaucescens tcmAR tetracenomycin C resistanceand repressor gene loci. J. Bacteriol. 174:3651–3658.

14. Hale, V., M. McGrew, B. Carlson, and J. L. Schottel. 1992. Heterologousexpression and secretion of a Streptomyces scabies esterase in Streptomyceslividans and Escherichia coli. J. Bacteriol. 174:2431–2439.

15. Hillemann, D., A. Puhler, and W. Wohlleben. 1991. Gene disruption andgene replacement in Streptomyces via single stranded DNA transformation ofintegration vectors. Nucleic Acids Res. 19:727–731.

16. Hirsch, C. F., and J. C. Ensign. 1976. Nutritionally defined conditions forgermination of Streptomyces viridochromogenes spores. J. Bacteriol. 126:13–23.

17. Hopwood, D. A., M. J. Bibb, K. F. Chater, T. Kieser, C. J. Bruton, H. M.Kieser, D. J. Lydiate, C. P. Smith, J. M. Ward, and H. Schrempf. 1985.Genetic manipulation of Streptomyces. A laboratory manual. The John InnesFoundation, Norwich, Great Britain.

18. Huff, J. P. 1991. Rapid isolation and purification of DNA from agarose gels:the phenol-freeze-fracture method. BioTechniques 10:724.

19. Ingram, C., M. Brawner, P. Youngman, and J. Westpheling. 1989. xylEfunctions as an efficient reporter gene in Streptomyces spp.: use for thestudy of galP1, a catabolite-controlled promoter. J. Bacteriol. 171:6617–6624.

20. Janssen, G. R., and M. J. Bibb. 1993. Derivatives of pUC18 that have BglIIsites flanking a modified multiple cloning site and that retain the ability toidentify recombinant clones by visual screening of Escherichia coli colonies.Gene 124:133–134.

21. Kelley, R. L., and C. Yanofsky. 1985. Mutational studies with the trp repres-sor of Escherichia coli support the helix-turn-helix model of repressor rec-ognition of operator DNA. Proc. Natl. Acad. Sci. USA 82:483–487.

22. Kendrick, K. E., and J. C. Ensign. 1983. Sporulation of Streptomyces griseusin submerged culture. J. Bacteriol. 155:357–366.

23. Kieser, T., M. T. Moss, J. W. Dale, and D. A. Hopwood. 1986. Cloning andexpression of Mycobacterium bovis BCG DNA in ‘‘Streptomyces lividans’’. J.Bacteriol. 168:72–80.

24. Kroening, T. A., and K. E. Kendrick. 1987. In vivo regulation of histidineammonia-lyase activity from Streptomyces griseus. J. Bacteriol. 169:823–829.

25. Kwak, J., and K. E. Kendrick. Morphological analysis of class III baldmutants of Streptomyces griseus. Submitted for publication.

26. Kwak, J., L. A. McCue, and K. E. Kendrick. Identification of bldA mutantsof Streptomyces griseus. Gene, in press.

27. Lampel, J. S., J. S. Aphale, K. A. Lampel, and W. R. Strohl. 1992. Cloningand sequencing of a gene encoding a novel extracellular neutral proteinasefrom Streptomyces sp. strain C5 and expression of the gene in Streptomyceslividans 1326. J. Bacteriol. 174:2797–2808.

28. Lawlor, E. J., H. A. Baylis, and K. F. Chater. 1987. Pleiotropic morphologicaland antibiotic deficiencies result from mutations in a gene encoding a tRNA-like product in Streptomyces coelicolor A3(2). Genes Dev. 1:1305–1310.

29. Leskiw, B. K., E. J. Lawlor, J. M. Fernandez-Abalos, and K. F. Chater. 1991.TTA codons in some genes prevent their expression in a class of develop-mental, antibiotic-negative, Streptomyces mutants. Proc. Natl. Acad. Sci.USA 88:2461–2465.

30. Lin, J. J., M. Smith, J. Jessee, and F. Bloom. 1991. DH11S: a competent cellhost for preparation of single-stranded DNA from phagemid vectors, p.96–98. In Focus, vol. 13. Life Science Technologies, Inc., Gaithersburg, Md.

31. MacNeil, D. J., K. M. Gewain, C. L. Ruby, G. Dezeny, P. H. Gibbons, and T.MacNeil. 1992. Analysis of Streptomyces avermitilis genes required for aver-mectin biosynthesis utilizing a novel integration vector. Gene 111:61–68.

32. Marcos, A. T., S. Gutierrez, B. Diez, F. J. Fernandez, J. A. Oguiza, and J. F.Martin. 1995. Three genes, hrdB, hrdD, and hrdT of Streptomyces griseusIMRU 3570, encoding sigma factor-like proteins, are differentially expressedunder specific nutritional conditions. Gene 153:41–48.

33. McCue, L. A., J. Kwak, M. J. Babcock, and K. E. Kendrick. 1992. Molecularanalysis of sporulation in Streptomyces griseus. Gene 115:173–179.

34. Merrick, M. J. 1976. A morphological and genetic mapping study of baldcolony mutants of Streptomyces coelicolor. J. Gen. Microbiol. 96:299–315.

35. Pabo, C. O., and R. T. Sauer. 1984. Protein-DNA recognition. Annu. Rev.Biochem. 53:293–321.

36. Reiss, B., R. Sprengel, and H. Schaller. 1984. Protein fusions with thekanamycin resistance gene from transposon Tn5. EMBO J. 3:3317–3322.

37. Ringquist, S., S. Shinedling, D. Barrick, L. Green, J. Binkley, G. D. Stormo,



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

and L. Gold. 1992. Translation initiation in Escherichia coli: sequences withinthe ribosome-binding site. Mol. Microbiol. 6:1219–1229.

38. Sambrook, J., T. Maniatis, and E. F. Fritsch. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold SpringHarbor, N.Y.

39. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing withchain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463–5467.

40. Strohl, W. R. 1992. Compilation and analysis of DNA sequences associatedwith apparent streptomycete promoters. Nucleic Acids Res. 20:961–974.

41. Ueda, K., K. Miyake, S. Horinouchi, and T. Beppu. 1993. A gene clusterinvolved in aerial mycelium formation in Streptomyces griseus encodes pro-

teins similar to the response regulators of two-component regulatory systemsand membrane translocators. J. Bacteriol. 175:2006–2016.

42. Vara, J., F. Malpartida, D. A. Hopwood, and A. Jimenez. 1985. Cloningand expression of a puromycin N-acetyl transferase gene from Strepto-myces alboniger in Streptomyces lividans and Escherichia coli. Gene 33:197–206.

43. Wissmann, A., R. Baumeister, G. Muller, B. Hecht, V. Helbl, K. Pflei-derer, and W. Hillen. 1991. Amino acids determining operator bindingspecificity in the helix-turn-helix motif of Tn10 Tet repressor. EMBO J.10:4145–4152.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Documents

Analysis of a Gene That Suppresses the Morphological Defect of