Cloning of Genes Governing DeoxysugarPortion ... · tant andthe properties ofclonedDNAadjacentto the eryD gene suggest the existence of regulatory regions governing the expression

Vol. 171, No. 11JOURNAL OF BACTERIOLOGY, Nov. 1989, p. 5872-58810021-9193/89/115872-10$02.00/0Copyright C 1989, American Society for Microbiology

Cloning of Genes Governing the Deoxysugar Portion of theErythromycin Biosynthesis Pathway in Saccharopolyspora erythraea

(Streptomyces erythreus)JESUS VARA,t MARLENA LEWANDOWSKA-SKARBEK, YI-GUANG WANG,4

STEFANO DONADIO,§ AND C. RICHARD HUTCHINSON*School ofPharmacy and Department ofBacteriology, University of Wisconsin, Madison, Wisconsin 53706

Received 1 May 1989/Accepted 28 July 1989

Genes that govern the formation of deoxysugars or their attachment to erythronolide B and 3a-mycarosylerythronolide B, intermediates of the biosynthesis of the 14-membered macrolide antibiotic erythromycia, werecloned from Saccharopolyspora erythraea (formerly Streptomyces erythreus). Segments ofDNA that co entthe eryB25, eryB26, eryB46, eryCI-60, and eryD24 mutations blking the formation of B or3a-mycarosyl erythronolide B, when cloned in Escherichia coli-Streptomyces shuttle cosmids or pla d vectorsthat can transform S. erythraea, were located in a ca. 18-kilobase-pair region upstream of the erythromycinresistance (ermE) gene. The eryCI gene lies just to the 5' side of ermE, and one (or possibly two) eryB gene isapproximately 12 kiobase pairs farther upstream. Another eryB gene may be in the same region, while anadditional eryB mutation appears to be located elsewhere. The eryD gene lies between the eryB and eryCI genesand may regulate their function on the basis of the phenotype of an EryD- mutant.

The use of macrolide antibiotics in human medicine stemsfrom the discovery of erythromycin A in 1952 (22) and thesubsequent clinical experience showing that it is a valuableanti-infective drug (23). Because of its medical and commer-cial importance, erythromycin is considered the archetype ofmacrolide antibiotics, as evident from the extensive studiesof its production and biological properties. We have beeninvestigating its biosynthesis by Saccharopolyspora eryth-raea (formerly Streptomyces erythreus) to determine thenature of the genes that govern its formation, with theintention that the cloned genes could then be used in studiesof the enzymology and regulation of erythromycin biosyn-thesis.Erythromycin is made by a three-stage pathway (see Fig.

5) (7): assembly of the 14-membered macrolactone 6-deoxy-erythronolide B (6DEB) from propionate and 2-methyl-malonate followed by its hydroxylation to erythronolide B(EB), formation of the deoxysugars mycarose and desos-amine from glucose and their addition to EB to makeerythromycin D, and then C-12 hydroxylation and C-3"0-methylation of the latter compound to produce erythro-mycin A.We isolated four phenotypically distinguishable, antibiotic

nonproducing S. erythraea mutants (EryA-, EryB-,EryC-, and EryD-) and used these to demonstrate cluster-ing of the erythromycin (ery) production genes in the chro-mosome of the NRRL 2338 strain by genetic mappingexperiments (36). Stanzak et al. (28) then showed that atleast three of the ery genes were linked to the erythromycinresistance (ermE) gene, which had previously been cloned

* Corresponding author.t Present address: Centro de Investigaciones Biomddicas-

Consejo Superior de Investigaciones Cientfficas, Facultad de Me-dicina, Universidad Aut6noma, Arzobispo Morcillo 4, 28029Madrid, Spain.

t Present address: Institute of Medicinal Biotechnology, ChineseAcademy of Medical Sciences, Tiantan, Beijing, China.

§ Present address: Department 93D, Corporate Molecular Biol-ogy, Abbott Laboratories, Abbott Park, IL 60064.

from S. erythraea (6, 29). Their conclusion was based on thefact that the pKC488 clone with a 35-kilobase-pair (kbp)DNA insert containing the ermE gene complemented theeryA34, eryB25, and eryCI-60 mutations in strains of S.erythraea we had isolated (36). This clone also conferredproduction of an erythromycin-like compound on Strepto-myces lividans transformants (28). Weber and Losick sub-sequently confirmed the close linkage of the eryA34 anderyB25 mutations to the ermE gene (34). More recently,Weber et al. (35) have shown that the eryG gene, whichgoverns the C-3' 0-methylation of erythromycin C, liesabout 7 kbp downstream of ermE, and researchers at AbbottLaboratories have shown that the eryAl gene is about 12 kbpdownstream of ermE (8).Our work has been focused on the biological processes

leading to the hydroxylation of 6DEB and on the formationand addition of deoxysugars to EB and 3a-mycarosyleryth-ronolide B (3MEB). We have purified 6DEB hydroxylase(26) and two electron transport proteins that are involved inthe conversion of 6DEB to EB (27), as well as a thymidinediphospho-D-glucose 4,6-dehydratase that may catalyze thefirst step in deoxysugar biosynthesis (31), from S. erythraea.We report here that by cloning segments of DNA from theregion surrounding the ermE gene into an Escherichia coli-Streptomyces shuttle cosmid or integrative plasmid vectorcapable of transforming S. erythraea, we were partially ableto overcome the problem of vector instability that previouslyhad prevented the cloning of ery genes by the shotgunmethod (37). Using this approach, we have found that someof the eryB genes, the eryCI gene, and the eryD gene arewithin an 18-kbp region immediately upstream of the ermEgene. The location of eryCI has been confirmed by Dhillon etal. (N. Dhillon, R. S. Hale, J. Cortes, and P. F. Leadlay,Mol. Microbiol., in press) in nucleotide sequencing and genedisruption experiments. The phenotype of the EryD- mu-tant and the properties of cloned DNA adjacent to the eryDgene suggest the existence of regulatory regions governingthe expression of some of the eryB and the eryCI genes.

5872

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

DEOXYSUGAR BIOSYNTHESIS GENES IN ERYTHROMYCIN PATHWAY

MATERIALS AND METHODS

General. Conditions used to transform and isolate plasmidand total DNA from the strains described in this work weresimilar to those described for E. coli by Maniatis et al. (21)and for Streptomyces spp. by Hopwood et al. (13). Standardmethods (13, 21) or those described by the commercialsupplier were used for in vitro DNA manipulations. DNAfragments were labeled with [32P]dCTP by using the nicktranslation kit from Bethesda Research Laboratories, Inc.,Gaithersburg, Md. Unless specified otherwise, DNA hybrid-ization was done by using nitrocellulose filters in 5x Den-hardt solution (21)-0.1% sodium dodecyl sulfate-50% form-amide-2x NET (0.3 M NaCl, 30 mM Tris hydrochloride, 2mM disodium EDTA [pH 8]) overnight at 42°C. The filterswere washed at room temperature three times for 15 mineach in 2x NET-0.1% sodium dodecyl sulfate and then onceat 650C with 0.2x NET-0.1% sodium dodecyl sulfate beforeautoradiography. S. erythraea strains were grown as de-scribed by Weber et al. (36), and their transformants wereselected by growth in the presence of thiostrepton (25 ,ug/ml)or apramycin (15 ,ug/ml) unless specified otherwise. PlasmidspIJ702 (17), pWOR109 and pWOR191 (2), and pIJ30 andpU486 (33) were obtained from Mervyn Bibb, Keith Chater,or David Hopwood, John Innes Institute, Norwich, En-gland. pcos2EMBL (24) was obtained from Hans Lehrach,European Molecular Biology Laboratory, Heidelberg, Fed-eral Republic of Germany. pKC505 (25) and the Micrococ-cus luteus apramycin-resistant (Am') strain were obtainedfrom Richard Baltz, Eli Lilly & Co., Indianapolis, Ind. TheStaphylococcus aureus thiostrepton-resistant (Thio) strainwas obtained from James Tuan, Abbott Laboratories, Ab-bott Park, Ill.

Construction of pWHM1, pWHM3, and pWHM4. pIJ30was digested with BamHI, and the 2-kbp BamHI fragmentcontaining the tsr gene, which confers Thior, was recoveredfrom an agarose gel by the method of Langridge et al. (18),treated with S1 nuclease, and ligated to the isolated 2.7-kbpfragment resulting from partial PvuII digestion of pUC19.pWHM1 was isolated from one of 50 blue colonies screenedand digested with NcoI, the digest was treated with Sinuclease and then with ClaI, and the 4.5-kbp band wasisolated from an agarose gel. pIJ486 was digested with PstI,the digest was treated with Si nuclease and then with ClaI,and the 3-kbp band was isolated from an agarose gel. Thetwo DNA samples were ligated, and pWHM3 (Fig. 1) wasobtained from E. coli transformants. pWHM4 (Fig. 1) wasconstructed in the same way as pWHM3; the fragmentobtained from pWOR191 after digestion with BglII, Sitreatment, and digestion with ClaI was used as the source ofthe Streptomyces replicon.

Isolation of ermE clones from S. erythraea DNA libraries.pcos2EMBL DNA (1 jig [24]) was digested with PvuII,treated with calf intestine alkaline phosphatase (molecularbiology grade; Boehringer Mannheim Biochemicals, India-napolis, Ind.), digested with BamHI, and ligated to totalDNA (9 ,ug) from S. erythraea WMH22 that had beenpartially digested with Bglll and treated with calf intestinealkaline phosphatase. The resulting mixture was packaged invitro by using the Packagene system (Promega Biotec,Madison, Wis.) and used to transduce E. coli DH1 (BethesdaResearch Laboratories) to kanamycin resistance (30 ,ug/ml).About 800 of these transductants (DNA insert average size,38 ± 2.1 kbp) were screened by colony hybridization essen-tially as described by Maas (19) for clones that hybridized toa 32P-labeled 1.7-kbp KpnI fragment containing the ermE

gene (6). One of the two identical clones isolated (pWHM5)was chosen for further work.EMBL4 DNA (2 ,ug; Promega Biotech) was digested with

BamHI and SalI and ligated in a small volume (10 ,ul) toWMH22 total DNA (3 ,ug) that had been partially digestedwith MboI and fractionated into 15- to 23-kbp fragments bysucrose density gradient centrifugation. The resulting mix-ture was packaged in vitro as described above and used totransfect E. coli K803 (permissive host) and Q359 (restrictiveSpi+ host). About 5 x 104 plaques were produced in K803;about 30% of these also grew in Q359. From these recombi-nant clones, we isolated by serial plaque purifications threeclones (pWHM6 through pWHM8) that hybridized to the32P-labeled, ermE-containing fragment. Analysis of theirEcoRI, HindIII, and PstI restriction digests showed thatthese clones represented about 23 kbp of overlapping DNAsurrounding the ermE gene.pKC505 DNA (50 ,ug [25]) was linearized with BamHI,

treated with calf intestine alkaline phosphatase, and ligatedwith WMH22 total DNA (100 ,ug) that had been partiallydigested with Sau3A and size fractionated by gel electropho-resis and sucrose density gradient centrifugation into 15- to23-kbp fragments. The ligation mixture was packaged invitro as described above and used to transduce E. coli DH1to apramycin resistance (100 ,ug/ml). A total of 2,000 trans-ductants were transferred to Hybond-N filters (AmershamCorp., Arlington Heights, Ill.) and screened for hybridiza-tion to a ca. 4.5-kbp 32P-labeled SphI fragment (see Fig. 3,sites 7 to 10), or to the ermE-containing fragment. Twelveclones were obtained that hybridized to either or both probes(average insert size, 19.6 kbp). The DNA obtained from nineof these clones was mapped by single and double digestionwith various restriction enzymes. Four overlapping clones(pWHM35, pWHM38, pWHM40, and pWHM43) whoseDNA inserts covered the region between sites 2 and 30 inFig. 3 were chosen for further study.

Construction of ery DNA subclones. The 1.7-kbp KpnIermE-containing fragment from pIJ461 (6) was subclonedinto KpnI-digested pIJ702 by selecting Thior erythromycin-resistant transformants of Streptomyces lividans TK24 (14).Of the two possible constructs, only the clone carryingpWHM2, in which the ermE gene is transcribed in the samedirection as the mel gene of pIJ702 (4), grew normally; theclone carrying a plasmid with ermE in the opposite orienta-tion grew poorly.DNA fragments from the appropriate restriction digests of

pWHM5 or pWHM9 were isolated from agarose gels by thecetyl triethyl ammonium bromide (18) or DEAE-cellulosemembrane (11) method and subcloned into pWHM3,pWHM4, or pKC505 by using E. coli DH5a (BethesdaResearch Laboratories) or into pIJ702 by using Streptomy-ces lividans TK24. The structure expected for each subclonewas confirmed by digestion of DNA from minilysates withthe appropriate restriction enzymes.

Analysis of antibiotic production by transformants. Primarytransformants were isolated after ca. 14 days from R2T20plates containing thiostrepton or apramycin and transferredto R2T agar in microdilution plug plates (36) containingthiostrepton or apramycin (20 ,ug/ml). The agar plugs wereassayed for antibiotic production after 6 to 7 days of growthat 30°C by the method of Weber et al. (36) with theStaphylococcus aureus Thior or M. luteus Amr test organ-isms. Confirmation that erythromycin was the antibioticproduced by recombinant strains was obtained by analysis ofethyl acetate extracts of solid cultures by thin-layer chroma-tography on silica gel in isopropyl ether-methanol-concen-

VOL. 171, 1989 5873

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.5874 VARA ET AL.

BamHI

Ncol;

SI;

Oal

Pstl

Distance(kbp)0

200500800

205022252485250028873016

Positon Distance(kbp)11 330612 361913 367014 375815 493516 618417 668418 669419 680420 7194

(go X cn)8 I 's R1 3

13

I

pWHM4 fragment:

Posktion12345678910

12

AS 3is11-2,1icaco x CA

I I 11 -1 I

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

DEOXYSUGAR BIOSYNTHESIS GENES IN ERYTHROMYCIN PATHWAY

FIG. 1. Construction of the pWHM1, pWHM3, and pWHM4 vectors. These plasmids were constructed from pIJ30, pUC19, pIJ486, andpWOR191 as described in Materials and Methods. amp, tet, and tsr signify the genes for ampicillin, tetracycline, and thiostrepton resistance,respectively; a-gal indicates the gene for the a-peptide of the lacZ product. pWHM4 (6.6 kbp) is the same as pWHM3, except for replacementof the 4.6-kbp NcoI/PstI-*ClaI fragment with the 4.0-kbp BglII-ClaI fragment shown below the polylinker region of pWHM3. The uniquecloning sites in the polylinker region of pWHM3 and pWHM4 are shown in bold type. The distances between the numbered restriction sitesin pWHM3 are indicated; the distances between the sites in the BglII-ClaI fragment used in the construction of pWHM4 are given by Baileyet al. (2).

trated ammonium hydroxide (150:70:4, vol/vol). The eryth-romycins A, B, C, and D plus 6DEB and 3MEB are clearlyresolved under these conditions and can be detected by theirrelative mobility and color with a chromogenic reagent (36).

RESULTS

Construction of vectors for transformation of S. erythraea.We had found previously that most of the available Strepto-myces plasmid cloning vectors did not stably transform thewild-type WMH22 or WMH278 (met-ll eryD24) strains (37);this presumably was due to poor replication and inadequatesegregation of most of the vectors tested. pWOR109 (2) gavethe highest percentage of stable transformants, whereaspIJ702 (17) yielded less than 10-8 stable transformants if theprimary Thior colonies were allowed to sporulate. Sincevector integration into the host genome should circumventthis instability, we tested three approaches to integrativetransformation of S. erythraea.The 1.1-kbp HindIII-KpnI fragment from the 3' end of the

ermE gene (see Fig. 3, sites 23 to 24) was cloned into the

a A B1 2 3 1 2 3 b

multiple cloning site of pWHM1 (Fig. 1) to give pWHM1-1.This plasmid contains a selectable marker (Thior [13]) and asegment of homologous DNA, but lacks a Streptomycesorigin of replication. Homologous, reciprocal recombinationinvolving the HindIII-KpnI ermE fragment, which extendsbeyond the 3' end of ermE (6), should not interrupt ermEtranscription. Nevertheless, we were unsuccessful in threeseparate attempts to obtain Thior WMH22 transformantswith pWHM1-1 by using transformation with pIJ702 as acontrol.The 1.7-kbp KpnI fragment (see Fig. 3, sites 21 to 24)

containing the complete ermE gene (6) was cloned into theKpnI site of pIJ702, and one of the two possible constructs,pWHM2 (Fig. 2), was used to transform WMH22. Theprimary Thior colonies were screened by serial transfer onselective and nonselective media to identify stable transfor-mants. Plasmid DNA could not be detected in minilysates oftwo of them by gel electrophoresis or by transfer of the sameDNA sample to nitrocellulose and hybridization against32P-labeled pIJ702.

_0- CE EDcD Me enI) __ I I

-U.-

§A1- 23

9.43_ ,A.i5.

_RM--6.5I~~~~~~k..

- 4.3

nHI HindillermE KpnlpWHM2 Bcll

mel/

11 Bcl I-2.3-2.0 I=

: 7E Een m- cuSCc) Ye coI in mel -m tsr m

I

I1=E 'EE-co se cot._m vm

OCLUcSa..-CI(

0.56- ermE 0.66 ermE6.7 kbp 2.8kbp

kbp-0.56 probe A probe B probe A

FIG. 2. Hybridization of pIJ702 and ermE to total DNA from WMH22(pWHM2) transformants. (a) Hybridizations were performed for 12to 18 h at 42°C in 50% formamide-5 x SSC (1 x SSC is 0.15 M NaCl plus 0.015 M sodium citrate) (13)-S5 x Denhardt (13)-100 ,ug of denaturedsalmon sperm DNA per ml; the filters were then washed three times at 420C for 30 min in 0.1 x SSC containing 0.1% sodium dodecyl sulfate.The data for probe A, the 1.1-kbp HindIII-KpnI 32P-labeled ermE fragment, are shown in panel A; the data for probe B, the 2.3-kbp SstII[32P]pIJ702 fragment, are shown in panel B. BamHI-digested total DNA from WMH22 is shown in lanes 1, and BamHI-digested total DNAfrom two different WMH22(pWHM2) transformants is shown in lanes 2 and 3 (the arrowhead indicates the position of the faint 0.66-kbpfragment). (b) The structure of the region where pWHM2 has integrated into the WMH22 genome is shown above pWHM2. The positionsof the hybridizing bands and the regions within them that hybridized the to the two probes are shown below the partial restriction map of thegenome in the WMH22(pWHM2) transformant.

23-

9.4 -

6.5-4.3-

I I I IO.-

ermE-

Il I II I I I

VOL. 171, 1989 5875

I

I

z-3 -2.0 -

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

5876 VARA ET AL.

Hybridization of the BamHI-digested total DNA fromthese two transformants and from WMH22 to the 32P-labeled1.1-kbp HindIII-KpnI fragment containing ermE revealedthe presence of an extra fragment of 6.7 kbp in the DNAfrom the two transformants, in addition to the 2.8-kbpwild-type fragment (Fig. 2, probe A). This indicates thepresence of an additional copy of the ermE gene. The32P-labeled 2.3-kbp SstII fragment of pIJ702 hybridized onlywith the DNA from the two transformants, giving signals at6.7 and 0.66 kbp (Fig. 2, probe B). The strong signal from thegenomic Southern blot transfer and hybridization and thelack of detectable free plasmid in the transformants areconsistent with the interpretation (shown in the lower part ofFig. 2) that pWHM2 had integrated into the S. erythraeagenome as expected for homologous, reciprocal recombina-tion between the two regions of ery DNA. Weber and Losick(34), Weber et al. (35), and Dhillon et al. (in press) haveindependently reported similar results in transformations ofS. erythraea with pIJ702 clones carrying different fragmentsof ery DNA. Although it is therefore possible to obtain stableS. erythraea transformants by specifically integrating pIJ702carrying homologous DNA, the negative results withpWHM1-1 suggest that some amount of plasmid replicationis necessary for integration to take place in this host.Streptomyces ambofaciens, on the other hand, has beensuccessfully transformed with nonreplicating E. coli vectorscontaining inserts of homologous DNA, albeit quite largerthan in our case (R. Nagaraja Rao, personal communica-tion).To determine whether the comparative stability of S.

erythraea(pWOR109) transformants noted earlier (37) mightalso have been due to the transient integration of pWOR109,we tested for hybridization between 32P-labeled pWOR109and total DNA from wild-type and S. erythraea(pWOR109)strains. The data (not shown) indicated that there is signifi-cant homology between a 1-kbp SstI-SstII fragment ofpWOR109 and the S. erythraea genome and that S.erythraea(pWOR109) transformants consist of a populationwith pWOR109 integrated into a 6-kbp SstI fragment and onein which this vector is autonomous. Although these resultspoint to the possibility that the favorable transformation ofS. erythraea by pWOR109 is in part due to its integration,this may not be a sine qua non requirement.The knowledge gained from the above experiments guided

the construction of two E. coli-Streptomyces shuttle vectorsfor use in S. erythraea. pWHM3 was constructed by insert-ing the 3-kbp PstI-ClaI fragment containing the replicationorigin of pIJ486 into pWHM1 (Fig. 1). (pIJ486 was madefrom the same Streptomyces replicon as pIJ702 [33].)pWHM4 was constructed in the same way by using afragment containing the pWOR191 origin of replication(pWOR191 is the parent of pWOR109 [2]) to give a plasmidwith the same structure as pWHM3, except for replacementof the NcoIIPstI-ClaI segment (asterisks in Fig. 1) with the4-kbp BgllI-ClaI fragment shown below pWHM3. The re-gion of pWOR109 that is homologous to S. erythraea DNAwas purposely excluded so that cloned rather than vectorDNA could drive the integration of pWHM4 constructs in S.erythraea. Both vectors allow rapid cloning in E. coli bychromogenic selection (32), followed by transfer into theappropriate Streptomyces host.

Cloning of DNA that complements the eryB, eryCI, anderyD mutations. Following the report of Stanzak et al. (28),we isolated clones from S. erythraea WMH22 DNA librariesthat hybridized to the ermE gene (6) as described in Mate-rials and Methods. Figure 3 shows the combined restriction

map of the region of the S. erythraea chromosome asdeduced from that of a 43-kbp insert in pcos2EMBL (sites 1to 28) and from that of a pKC505 clone, pWHM40 (sites 9 to30). This map is consistent with a restriction map of the eryDNA cloned in pKC488 (28), kindly provided by R. H.Baltz. The ermE gene is between sites 21 and 24 and istranscribed in the direction shown by the arrow. Threeoverlapping EMBL4 (10) clones containing S. erythraeachromosomal DNA covering about 23 kbp of DNA aroundermE were also isolated, and a 6.5-kbp BglII-HindIII frag-ment (Fig. 3, sites 11 to 23) was subcloned from one of thesebacteriophage clones into pUC19 (32) as pWHM9 (notshown). Nine overlapping pKC505 (25) clones coveringabout 36 kbp of DNA from the same region (sites 2 to 30)were also isolated.DNA was subcloned from pWHM5 or pWHM9 into

pWHM3, pWHM4, pKC505, or pIJ702, and the properconstruct was isolated in E. coli DH5a or Streptomyceslividans TK24 because plasmid DNA could not be easilyrecovered from S. erythraea. The S. erythraea blockedmutants WMH25 (eryB25), WMH26 (eryB26), WMH40(eryB40), WMH46 (eryB46), WMH57 (rif-57 ers-57 eryB46),WMH254 (eryC1-60), and WMH278 (met-ll eryD24) werethen transformed by the procedures of Yamamoto et al. (37).Approximately the same number of protoplasts and amountof plasmid DNA were used in each transformation, but onlya few Thior transformants were obtained in some experi-ments. The Thior or Amr transformants were not rigorouslydistinguished as stable or unstable by serial transfer andsporulation before they were tested for antibiotic produc-tion. The production of erythromycin by these transformantswas determined as described in Materials and Methods; thefollowing transformants (Table 1) exhibited the Ery+ pheno-type, which was assumed to be due to complementation ofthe ery mutation: WMH254 transformed with pWHM17,pWHM18, pWHM19, pWHM33, and pWHM34; WMH278transformed with pWHM21; WMH25 transformed withpWHM35 and pWHM44; WMH26 transformed withpWHM11, pWHM35, and pWHM44; WMH57 transformedwith pWHM35, pWHM38, and pWHM40. Only a few of thecolonies produced antibiotic activity in WMH26 transformedwith pWHM11 and WMH57 transformed with pWHM35,pWHM38, and pWHM40. (The term complementation isused here to indicate restoration of erythromycin productioneither by true complementation or by gene repair.) Repre-sentative Ery+ transformants were grown under nonselec-tive conditions (following protoplast formation and regener-ation in some cases), and the resulting Thios colonies weretested for antibiotic production by bioassay. These colonieshad lost the ability to produce antibiotic, indicating that theEry+ phenotype was due to the presence of the transformingDNA and not due to reversion of the ery mutation. Inexperiments giving no Ery+ transformants, we have listedonly the ones in which 20 or more Thior or Amr colonieswere isolated, and we believe these to be reliable negativeresults.The state of the transforming pWHM3 constructs was

established by hybridization of digests of total DNA from anEry+ and an Ery- WMH254(pWHM19) transformant with32P-labeled pWHM3 DNA or the 32P-labeled insert ofpWHM19. The SphI digests probed with the pWHM19 insertgave the results expected for the presence of two forms ofpWHM19 in both transformants (Fig. 4A); the 11-kbp bandrepresents free pWHM19, since this plasmid has only oneSphI site, and the bands at about 4.3 and 23 kbp representthe structure expected for the integration of pWHM19 by a

J. BACTERIOL.

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

DEOXYSUGAR BIOSYNTHESIS GENES IN ERYTHROMYCIN PATHWAY

26 24 2 2 20 18 16 14 12 10 8 6 4 -2 0 +2 4 6 8 10 12A I _I I I I I I I Ill I I I i

2 345 6 78 9 10 1112 19 21 2225 26 272829 30B EEB G SP B S GP S KBP B PGE B

A I 11I I 11 I 1I I III I III If I I_

eryB

pWHM44

pWHM1 1

pWHM1615

1314 16 17T ETP BI II I

ermEl|

kbp

eryD

eryC 1

1918 20CS T11 I

pWHM1522

21 23 24KBH K

1 2 kb

pWHM34

pWHM1 7

pWHM14 pWHM18

pWHM19

pWHM33

pWHM20

pWHM21

pWHM22 pWHM23 pWHM24

FIG. 3. Location of the eryB, eryC1, and eryD genes in the ery gene cluster cloned from S. erythraea. A partial restriction map of thecloned region surrounding the erythromycin resistance (ermE) gene in the S. erythraea chromosome is shown at the top; the region betweenthe PstI sites numbered 12 and 25 is expanded twofold in the middle portion of the figure. Differently shaded bars immediately below themapped region indicate the positions of the three deoxysugar biosynthesis genes in relation to the ermE gene, and the arrow indicates thedirection of transcription of ermE. The locations of the restriction fragments tested for complementation of the four eryB, the eryC1-60, andthe eryD24 mutations by using the subclone designations given in Table 1 are shown below the map. Complementation of one of these erymutations by a fragment is indicated by the same shading as shown above for the particular ery gene. The fragment that increased antibioticproduction in wild-type transformants is shown in bold type. Restriction sites are abbreviated as follows: B, BamHI; G, BglII; C, ClaI; E,EcoRI; H, Hindlll; K, KpnI; P, PstI; S, SphI; T, SstI. The restriction map is complete for BglII, EcoRI, HindIll, PstI and ClaI; for the othersites, only those relevant for subcloning are shown.

single crossover event into a site within a 12.5-kbp SphIfragment of genomic DNA (Fig. 4B). Wild-type (WMH22)and EryC- (WMH254) DNA show only the 12.5-kbp SphIband. The signal from the 11-kbp band in lane 4 of the SphIdigest is slightly less intense than the other signals, suggest-ing that there is less than one copy of the nonintegratedplasmid per genome. The data from the PstI digests (Fig. 4A)and hybridization of the same blot with pWHM3 DNA (notshown) corroborated these results.To prove that no major changes had occurred in the

structure of the transforming plasmids in pWHM4 con-structs, strain WMH254 was transformed with pWHM33 andpWHM34 (Table 1). Plasmid DNA was rescued by trans-forming E. coli with total DNA recovered from two repre-sentative Ery+ transformants. The agarose gel profiles of thetransforming and reisolated plasmids, which showed fivebands ranging from 0.5 to 4 kbp after SstII digestion, wereindistinguishable in both cases. This proved that, at least insome of the Ery+ transformants, the plasmids existed as theunaltered free replicons.The locations of the eryB, eryCi, and eryD genes shown in

Fig. 3 were deduced from the data in Table 1 as follows.Complementation of the eryB26 mutation by pWHM11 at thelow frequency observed suggests that pWHM11 may carryonly part of the eryB gene and have restored the blockedfunction by homologous recombination with the genome(this plasmid was constructed in pIJ702 and should not haveexisted autonomously in the transformants [34]). This as-sumption is supported by the fact that in pWHM44, whoseinsert extends ca. 2 kbp farther to the left than the fragmentcloned in pWHM11 (Fig. 3), 100% of the transformantscomplemented the eryB26 and eryB25 mutations. Therefore,these two mutations map between sites 5 and 9 in Fig. 3. TheeryB46 mutation may lie to the right of these two on the basisof the low-frequency complementation of this mutation instrain WMH57 by pWHM40, whose insert is contiguous tothat of pWHM44, and by pWHM35 and pWHM38. Thisresult suggests that the eryB46 mutation lies in a differenteryB gene than the eryB25 and eryB26 mutations do. TheeryB40 mutation could not be complemented by any of thecosmid clones, suggesting that other eryB genes may be inthe region outside sites 2 and 30 in Fig. 3. (Large deletions

36

G

4/

VOL. 171, 1989 5877

I

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

5878 VARA ET AL.

TABLE 1. Results of complementation experimentswith S. erythraea Ery- mutants

No. of Ery+Plasmid Insert'

os strain b transformants/no.Plasmid (kbp) Host of Thior or Amr

transformantsc

pWHM11d 6-11 (10.6) WMH26 (eryB26) 1/7pWHM14e 13-15 (1) WMH254 (eryCl-60) 0/41pWHM1S 25-27 (4.8) WMH278 (eryD24) 0/47pWHM16 8-12 (7.9) WMH26 0/27pWHM17 16-25 (6.5) WMH254 102/112f

WMH278 0/33pWHM18 19-25 (4.3) WMH254 54/54

WMH278 0/45pWHM19 20-25 (3.8) WMH254 30/78

WMH278 0/34pWHM20 16-20 (2.6) WMH25 (eryB25) 0/40

WMH26 0/40WMH254 0/23WMH278 0/57

pWHM21 16-23 (4.9) WMH254 0/96WMH278 39/39

pWHM22 17-18 (1.3) WMH278 0/26pWHM23 20-21 (1.6) WMH254 0/26pWHM24 22-25 (2.1) WMH254 0/34pWHM339 20-22 (1.8) WMH254 30/30pWHM349 14-22 (5.2) WMH254 30/30

WMH278 0/30pWHM35h 5-26 (23.1) WMH25 90/90

WMH26 96/96WMH40 (eryB40) 0/45WMH46 (eryB46) 0/90WMHS7 (rif-57 1/30

ers-57 eryB46)pWHM38h 2-26 (27.1) WMH40 0/48

WMH46 0/48WMHS7 2/60

pWHM40h 9-30 (24.1) WMH40 0/48WMH46 0/48WMH57 6/60

pWHM43h 5-17 (14.4) WMH26 80/80pWHM44h 5-9 (7) WMH25 48/48'

WMH26 96/96

a"Insert" corresponds to the numbered sites in pWHM5 (Fig. 3). Thesubclones were constructed in pWHM3 unless indicated otherwise.

b The strain designation UW used in reference 36 has been replaced withWMH.

C The primary transformants were assayed for antibiotic production asdescribed in Materials and Methods.

d Cloned in pIJ702.e Caused overproduction of erythromycin in WMH22 transformants.f One insert orientation gave 85% complementation, and the other gave

100%o complementation.g Cloned in pWHM4.h Cloned in pKC505.i Both insert orientations gave 100% complementation.

removing many of the ery genes could be another explana-tion for the lack of complementation. However, none of theeryB strains appeared to have large deletions in the ermEregion, as indicated by the results [not shown] of suitablehybridization experiments.)The 91 to 100% complementation of the eryCl-60 mutation

by pWHM17, pWHM18, and pWHM33 indicates that thisgene lies between sites 20 and 22 in Fig. 3. Since the onlycomplete open reading frame between sites 21 and 24 is theone encoding the resistance gene product (6, 30), it is likelythat the eryCI gene actually lies between sites 20 and 21immediately upstream of ermE. The eryD gene must there-fore lie just to the left of eryCI, probably between sites 16

410~~~1I__ X

S~S

FIG. 4. Analysis ofthe state ofpWHM19 inpWMH254(pWHM19)transformants by DNA hybridization. (A) The results of hybridiza-tion of the 3.8-kbp 12P-labeled insert of pWHM19 to total DNA froma WMH254(pWHM19) Ery- transformant (lanes 1 and 5), WMH254(lanes 2 and 6), WMH22 (lanes 3 and 7), and a pWMH254(pWHM19)Ery' transformant (lanes 4 and 8) after digestion with Sphl (lanes 1to 4) or Pstl (lanes 5 to 8). Hybridization conditions were as statedin the legend to Fig. 2, except that the filter was washed once atroom temperature for 15 min with 2 x SSC and then twice at 65"C for30 min each with 0.1x SSC. Symbol: 0, bands due to partialdigestion. X indicates the sizes of fragments from a lambda Hindillldigest. (B) Structure of the region of DNA where pWHM19 hasintegrated into the WMH254 genome. The two Sphl restrictionfragments from this region that hybridized to the probe are indicatedbelow the restriction map.

and 19, in view of the complementation data for theWMH278(pWHM21) transformant.Complementation of an ery mutation was expected but not

observed in some cases when more than 20 transformantscould be tested (Fig. 3). pWHM17, pWHM34, pWHM35,and pWHM39 did not complement the eryD24 mutation,even though their inserts fully encompass those ofpWHM21. pWHM35, pWHM38, and pWHM40 did notcomplement the eryB46 mutation in strain WMH46; thus,these three pKC505 clones exhibited contrasting behavior indifferent genetic backgrounds: WMH46 (eryB46) versusWMH57 (rif-57 ers-57 eryB46). (The ers-57 mutation causesWMH57 to have about 1/10 the wild-type level of resistanceto erythromycin A [J. M. Weber, Ph.D. thesis, University ofWisconsin, 1984].) pWHM2O did not complement the eryD24mutation, and pWHM21 did not complement the eryCi-60mutation.

J. BACTERIOL.

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

DEOXYSUGAR BIOSYNTHESIS GENES IN ERYTHROMYCIN PATHWAY

Erythronolide BeryF

6-Deoxyerythronolide BeryA Propionate

+Methylmalonate

eryB 3a-Mycarosyl-erythronolide B

Erythromycin D Erythromycin C

\eryG

Erythromycin A

HO 0

HOWOH

OTDPCH3

eryB

0

H OTDP

OH OH

CH20H0

OHHO \L /OTDP

OHTDP-D-Glucose

epimerize

eryB

TDP-D-Desosamine

CH30

TDP-4-Keto-6-Deoxy- /OD-Glucose

OTDP

OH

CH20H

OT~~DPL- OH

]-

FIG. 5. Erythromycin biosynthetic pathway. Thymidine diphospho-L-mycarose (TDP-L-Mycarose) and thymidine diphospho-D-desos-amine (TDP-D-Desosamine) are made from thymidine diphospho-D-glucose and attached in sequo to EB, which is made from propionate andmethylmalonate, to give erythromycin D. Erythromycin D is hydroxylated at position 12 to give erythromycin C, and the latter is0-methylated at position 3" to give erythromycin A. Thymidine diphospho-D-glucose 4,6-dehydratase has been purified to homogeneity fromS. erythraea CA340 (31), and the epimerization of thymidine diphospho-4-keto-6-deoxy-D-glucose is assumed by analogy to how otherL-hexoses are formed from D-glucose (31). The nature of the other steps in the branches leading to thymidine diphospho-L-mycarose andthymidine diphospho-D-desosamine are unknown. Thin arrows between intermediates indicate single steps in the pathway, and thick arrowsindicate multiple steps. The steps blocked by the eryA, the hypothetical eryF mutation, and the eryG mutation are indicated above therespective arrows. Since the eryB and eryCI mutations could affect more than a single step, as discussed in the text, multiple locations areshown. The position of the eryD mutation is uncertain and is therefore not shown.

An overriding factor in some of these situations is thepossibility that complementation of the ery mutation re-quired plasmid integration to restore formation of a completetranscript, as has been seen in other cases (1, 20), or thatcomplementation was prevented by plasmid integration ifthis interrupted the formation of the normal transcript.Therefore, incomplete complementation or lack of comple-mentation in the case of the pWHM3-containing subclonescould be related to the existence of both free and integratedforms of the plasmid, as shown for pWHM19, with only oneform restoring erythromycin production in the mutant. Com-plementation would therefore depend on the fraction oferythromycin-producing cells in the mixed population. Itshould be noted that a significant difference was observed inthe complementation of the eryCI mutation when almostidentical fragments were cloned in pWHM3 as in pWHM21or in pWHM4 as in pWHM33 (Table 1). In the case of thepKC505-containing clones, their large size may have re-

sulted in instability of the plasmid and thus an apparent lackof complementation in some cases.

It appears that none of the Streptomyces vectors are wellsuited for complementation of blocked mutants of S. eryth-raea. For this reason, pWHM3, pWHM4, pIJ702, pKC505were all used in this work. The only Streptomyces vectorthat has proved useful in S. erythraea so far is pIJ702,although its application has been limited to gene disruptionexperiments (35; Dhillon et al., in press).

DISCUSSION

Mutations that block the conversion of 6DEB to erythro-mycin D in S. erythraea all have the Ery- phenotypebecause none of the accumulated pathway intermediates(Fig. 5) have antibiotic activity. Such mutants have beenclassified into only three groups (36): EryB- (which accu-mulate EB but biotransform 3MEB to erythromycin A),

VOL. 171, 1989 5879

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

5880 VARA ET AL.

EryC- (which accumulate 3MEB but biotransform erythro-mycin D to erythromycin A [M. Lewandowska-Skarbek,unpublished results]), and EryD- (which accumulate EB butdo not biotransform 3MEB or erythromycin D to erythro-mycin A [Lewandowska-Skarbek, unpublished]). TheeryB25, eryB26, eryB40, eryB46, eryCl-60, and eryD24 mu-tations could actually affect any one of the several probablesteps in the formation and attachment of L-mycarose andD-desosamine to EB and 3MEB, respectively (Fig. 5). Theycannot be distinguished further by antibiotic cosynthesistests because none of these steps involve diffusible interme-diates. Interruption of a function regulating the expression ofthe deoxysugar biosynthesis genes could be an alternativereason for the multiple effects of the eryD24 mutation or, infact, any of the eryB mutations. The limited number ofphenotypes and EryC- and EryD- mutants available (36)thus makes it difficult to estimate how much DNA encom-passes the number of possible eryB, eryCI, or eryD genes.The data reported here show only that several of these genesare clustered in an approximately 18-kbp region upstream ofthe ermE gene (Fig. 3).

Clustering of antibiotic production and resistance genes inStreptomyces spp. is common (12, 15). Fishman et al. (9)have shown through gene cloning experiments that sevengenes governing the formation and addition of three deoxy-sugars to tylactone, the 16-membered macrolactone precur-sor of tylosin in Streptomyces fradiae, lie within a 38-kbpDNA segment along with two tylosin resistance genes.Weber and Losick (34), by comparing recombination fre-quencies in conjugational matings of S. erythraea Ery-mutants having the tsr marker chromosomally integrated 23kbp upstream of the ermE gene, found that the eryB25 locusmaps to the right of the tsr marker somewhere within a ca.20-kbp region around ermE. This is consistent with thelocation we show for the eryB25 mutation, but their sugges-tion that eryB25 is downstream of ermE is not consistentwith the data presented here. Donadio et al. have located theeryAl gene about 12 kbp downstream of the ermE genethrough complementation and gene disruption experiments(8). Weber et al. (35) and Donadio et al. (8) have used similarmethods to show that the eryG gene, which governs theconversion of erythromycin C to erythromycin A (Fig. 5), isabout 7 kbp downstream of ermE. Thus, it is clear thatseveral of the erythromycin production genes are clusteredabout ermE.The proximity of the ermE and eryC] genes hints at their

possible coordinate expression. Indeed, complementation ofthe eryCi-60 mutation was not observed if the transformingpWHM3-containing subclone did not also contain all ofermE; conversely, when pWHM4 was used, only the 5'portion of ermE was sufficient to confer complementation.Bibb and Janssen (5) reported that the promoter regions forthe ermE gene and the start of an open reading frame that istranscribed in the opposite direction overlap and that thelevels of expression of ermE and the divergently transcribedopen reading frame were inversely correlated in tests of theeffect of transition and deletion mutations in the ermEpromoter region. From our results, this incomplete openreading frame appears to be the start of the coding region forthe eryCI gene product. The results independently reportedby Dhillon et al. (in press) establish the validity of thisassumption. Since the eryCl-60 mutation blocks the forma-tion of erythromycin D and erythromycin D is the firstantibiotically active substance in the erythromycin pathway,the eryCI gene would be an ideal point for the organism toregulate erythromycin production. (The facts that the pro-

duction of EB can be increased much more than that of3MEB or erythromycin A by addition of n-propanol to thefermentation medium [K. Petzoldt and K. Kieslich, Germanpatent 1900647, January 1969] and that more EB accumu-lates in the wild-type strain than any other pathway inter-mediate [Petzoldt and Kieslich, patent] suggest that a step inthe conversion of EB to erythromycin D is the bottleneck inthe erythromycin pathway.) Any speculation about such amechanism, however, has to accommodate the knowledgethat the ermE gene is apparently constitutively expressed(29, 30), unlike some other Streptomyces macrolide resis-tance genes, whose expression is induced by the correspond-ing macrolide antibiotic (16).We believe that the eryD24 mutation affects a regulatory

function, rather than a step common to the formation ofthymidine diphospho-D-desosamine and thymidine diphos-pho-L-mycarose (Fig. 5), since the WHM278 strain containsnormal levels of thymidine diphospho-D-glucose 4,6-dehy-dratase activity, the enzyme responsible for the possiblecommon step in both pathways (31). Some of our results(J. A. Vara, E. Wendt-Pienkowski, and C. R. Hutchinson,unpublished results) support the existence of a possiblesecond regulatory region. In fact, a reproducible 10- to15-fold increase in erythromycin production was obtainedwhen the wild-type S. erythraea was transformed withpWHM14, which contains a 1-kbp DNA segment (sites 13 to15 in Fig. 3) that lies just to the left of the eryD regionbetween the eryB and eryD genes.The results described here and by Dhillon et al. (in press)

provide some insight into the location and function of genesthat govern deoxysugar biosynthesis in the erythromycinpathway. Genes with functions analogous to these have beenidentified in Streptomycesfradiae for tylosin biosynthesis (3,9), but the correspondence between such genes and theenzymes of deoxysugar biosynthesis implicit in the pathwayshown in Fig. 5 is still obscure.

ACKNOWLEDGMENTS

We thank Dick Baltz, Mervyn Bibb, Keith Chater, David Hop-wood, Leonard Katz, Hans Lehrach, and James Tuan for antibioticstandards, bacterial strains, and plasmids; Dick Baltz for thepKC488 restriction map; and Evelyn-Wendt Pienkowski for techni-cal assistance.

This research was supported in part by fellowships to J.A.V. fromthe Ministerio de Educacion EspaniollFullbright Foundation and toM.L.-S. from the Council for International Exchange of Scholars/Fullbright Foundation, by a grant from Abbott Laboratories, AbbottPark, Ill., and by Public Health Service grant GM31925 from theNational Institutes of Health.

LITERATURE CITED1. Anzai, H., R. Murakami, S. Imai, A. Satoh, K. Nagaoka, and

C. J. Thompson. 1987. Transcriptional regulation of bialaphosbiosynthesis in Streptomyces hygroscopicus. J. Bacteriol. 169:3482-3488.

2. Bailey, C. R., C. J. Bruton, M. J. Butler, K. F. Chater, J. E.Harris, and D. A. Hopwood. 1986. Properties of in vitro recom-binant derivatives of pJV1, a multi-copy plasmid from Strepto-myces phaeochromogenes. J. Gen. Microbiol. 132:2071-2078.

3. Baltz, R. H., and E. T. Seno. 1988. Genetics of Streptomycesfradiae and tylosin biosynthesis. Annu. Rev. Microbiol. 42:547-574.

4. Bernan, V., D. Filpula, W. Herber, M. Bibb, and E. Katz. 1985.The nucleotide sequence of the tyrosinase gene from Strepto-myces antibioticus and characterization of the gene product.Gene 37:101-110.

5. Bibb, M. J., and G. R. Janssen. 1987. Unusual features oftranscription and translation of antibiotic resistance genes in

J. BACTERIOL.

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

DEOXYSUGAR BIOSYNTHESIS GENES IN ERYTHROMYCIN PATHWAY

two antibiotic-producing Streptomyces species, p. 309-318. InM. Alacevic, D. Hranueli, and Z. Toman (ed.), Genetics ofindustrial microorganisms. Proceedings of the Fifth Interna-tional Symposium, part B. Pliva, Zagreb, Yugoslavia.

6. Bibb, M. J., G. R. Janssen, and J. M. Ward. 1985. Cloning andanalysis of the promoter region of the erythromycin-resistancegene (ermE) of Streptomyces erythraeus. Gene 38:215-226;E357-E368.

7. Corcoran, J. W. 1981. Biochemical mechanisms in the biosyn-thesis of the erythromycins, p. 132-174. In J. W. Corcoran (ed.),Antibiotics. IV. Biosynthesis. Springer-Verlag, New York.

8. Donadio, S., J. S. Tuan, M. J. Staver, J. M. Weber, T. J. Paulus,G. T. Maine, J. 0. Leung, J. P. Dewitt, J. A. Vara, Y.-G. Wang,L. Katz, and C. R. Hutchinson. 1989. Genetic studies onerythromycin biosynthesis in Saccharopolyspora erythraea, p.53-59. In C. L. Hershberger, S. W. Queener, and G. Hegeman(ed.), Genetics and molecular biology of industrial microorgan-isms. American Society for Microbiology, Washington, D.C.

9. Fishman, S. E., K. Cox, J. L. Larson, P. A. Reynolds, E. T.Seno, W.-K. Yeh, R. van Frank, and C. L. Hershberger. 1987.Cloning genes for the biosynthesis of a macrolide antibiotic.Proc. Natl. Acad. Sci. USA 84:8248-8252.

10. Frischauf, A.-M., H. Lehrach, A. Poustka, and N. Murray. 1983.Lambda replacement vectors carrying polylinker sequences. J.Mol. Biol. 170:827-842.

11. Girvitz, S. C., S. Bacchetti, A. J. Rainbow, and F. L. Graham.1980. A rapid and efficient procedure for the purification ofDNA from agarose gels. Anal. Biochem. 106:492-496.

12. Hopwood, D. A. 1986. Cloning and analysis of antibiotic biosyn-thetic genes in Streptomyces, p. 3-14. In G. Szabo, S. Biro, andM. Goodfellow (ed.), Biological, biochemical and biomedicalaspects of actinomycetes. Akademiai Kiado, Budapest, Hun-gary.

13. 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 Innes Foundation, Norwich,England.

14. Hopwood, D. A., T. Kieser, H. M. Wright, and M. J. Bibb. 1983.Plasmids, recombination and chromosome mapping in Strepto-myces lividans 66. J. Gen. Microbiol. 129:2257-2269.

15. Hutchinson, C. R. 1988. Impact of genetic engineering on thecommercial production of antibiotics. Appl. Biochem. Biotech-nol. 16:169-190.

16. Kamimiya, S., and B. Weisblum. 1986. Inducible macrolide-lincosamide-streptogramin resistance in Streptomyces: cloningand characterization of inducible erm from Streptomyces viri-dochromogenes and Streptomyces fradiae, p. 169-175. In M.Alacevic, D. Hranueli, and Z. Toman (ed.), Genetics of indus-trial microorganisms. Proceedings of the Fifth InternationalSymposium, part B. Pliva, Zagreb, Yugoslavia.

17. Katz, E., C. J. Thompson, and D. A. Hopwood. 1983. Cloningand expression of the tyrosinase gene from Streptomyces anti-bioticus in Streptomyces lividans. J. Gen. Microbiol. 129:2703-2714.

18. Langridge, J., P. Langridge, and P. L. Bergquist. 1980. Extrac-tion of nucleic acids from agarose gels. Anal. Biochem. 103:264-271.

19. Maas, R. 1983. An improved colony hybridization method withsignificantly increased sensitivity for detection of single genes.Plasmid 10:296-298.

20. Malpartida, F., and D. A. Hopwood. 1986. Physical and geneticcharacterisation of the gene cluster for the antibiotic actinorho-din in Streptomyces coelicolor A3(2). Mol. Gen. Genet. 205:66-73.

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

22. McGuire, J. M., R. L. Bunch, R. C. Anderson, H. E. Boaz, E. H.Flynn, H. M. Powell, and J. W. Smith. 1952. Ilotycin, a newantibiotic. Antibiot. Chemother. 2:281-283.

23. Nakayama, I. 1984. Macrolides in clinical practice, p. 261-301.In S. Omura (ed.), Macrolide antibiotics. Academic Press, Inc.,New York.

24. Poustka, A., H.-R. Rackwitz, A.-M. Frischauf, B. Hohn, and H.Lehrach. 1984. Selective isolation of cosmid clones by homolo-gous recombination in Escherichia coli. Proc. Natl. Acad. Sci.USA 81:4129-4133.

25. Richardson, M. A., S. Kuhstoss, P. Solenberg, N. A. Schaus, andR. Nagaraia Rao. 1987. A new shuttle cosmid vector, pKC505,for streptomycetes: its use in the cloning of three differentspiramycin-resistance genes from a Streptomyces ambofacienslibrary. Gene 61:231-241.

26. Shafiee, A., and C. R. Hutchinson. 1987. Macrolide antibioticbiosynthesis. Isolation and properties of two forms of 6-deoxy-erythronolide B hydroxylase from Saccharopolyspora eryth-raea (Streptomyces erythreus). Biochemistry 26:6204-6210.

27. Shafiee, A., and C. R. Hutchinson. 1988. Purification and recon-stitution of the electron transport components for 6-deoxyeryth-ronolide B hydroxylase, a cytochrome P-450 enzyme of macro-lide antibiotic (erythromycin) biosynthesis. J. Bacteriol. 170:1548-1553.

28. Stanzak, R., P. Matsushima, R. H. Baltz, and R. N. Rao. 1986.Cloning and expression in Streptomyces lividans of clusterederythromycin biosynthesis genes from Streptomyces erythreus.Bio/Technology 4:229-232.

29. Thompson, C. J., J. M. Ward, and D. A. Hopwood. 1982.Cloning of antibiotic resistance and nutritional genes in strepto-mycetes. J. Bacteriol. 151:668-677.

30. Uchiyama, H., and B. Weisblum. 1985. N-Methyl transferase ofStreptomyces erythreus that confers resistance to the macro-lide-lincosamide-streptogramin B antibiotics: amino acid se-quence and its homology to cognate R-factor enzymes frompathogenic bacilli and cocci. Gene 38:103-110.

31. Vara, J., and C. R. Hutchinson. 1988. Purification of thymidine-diphospho-D-glucose 4,6-dehydratase from an erythromycinproducing strain of Saccharopolyspora erythraea by high reso-lution liquid chromatagraphy. J. Biol. Chem. 263:14992-14995.

32. Vieira, J., and J. Messing. 1982. The pUC plasmids, anM13mp7-derived system for insertion mutagenesis and sequenc-ing with synthetic universal primers. Gene 19:259-268.

33. Ward, J. M., G. R. Janssen, T. Kieser, M. J. Bibb, M. J.Buttner, and M. J. Bibb. 1986. Construction and characterisa-tion of a series of multi-copy promoter-probe plasmid vectorsfor Streptomyces using the aminoglycoside phosphotransferasegene from TnS as indicator. Mol. Gen. Genet. 203:468-478.

34. Weber, J. M., and R. Losick. 1988. The use of a chromosomeintegration vector to map erythromycin resistance and produc-tion genes in Saccharopolyspora erythraea (Streptomyceserythreus). Gene 68:173-180.

35. Weber, J. M., B. Schoner, and R. Losick. 1989. Identification ofa gene required for the terminal step in erythromycin A biosyn-thesis in Saccharopolyspora erythraea (Streptomyces eryth-reus). Gene 75:235-241.

36. Weber, J. M., C. K. Wierman, and C. R. Hutchinson. 1985.Genetic analysis of erythromycin production in Streptomyceserythreus. J. Bacteriol. 164:425-433.

37. Yamamoto, H., K. H. Maurer, and C. R. Hutchinson. 1986.Transformation of Streptomyces erythraeus. J. Antibiot. 39:1304-1313.

VOL. 171, 1989 5881

on February 2, 2021 by guest

http://jb.asm.org/

Dow

nloaded from