AUD4, a new amplifiable element from Streptomyces lividans

Downloaded from www.microbiologyresearch.org by

IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19

Microbiology (1999), 145, 3331–3341 Printed in Great Britain

AUD4, a new amplifiable element fromStreptomyces lividans

Esther Schmid,† Christa Bu$ chler‡ and Josef Altenbuchner

Author for correspondence: Josef Altenbuchner. Tel : 49 711 685 7591. Fax: 49 711 685 6973.e-mail : joe!gensun.biologie.uni-stuttgart.de

Institute of IndustrialGenetics, University ofStuttgart, Allmandring 31,70569 Stuttgart, Germany

After transformation of the Streptomyces lividans chloramphenicol-sensitive,arginine-auxotrophic mutant strain AJ100 with a derivative of plasmid SCP2,some of the regenerated protoplasts contained an 8<2 kb DNA sequenceamplified to several hundred copies per chromosome. The corresponding non-amplified sequence, called AUD4, was isolated from a λ phage genomic libraryof S. lividans 1326. Two cytosine residues were the only directly repeatednucleotides at the ends of the element, indicating that AUD4 is a class Iamplifiable sequence. The element mapped in the AseI-D fragment of the S.lividans chromosome, where other class I amplifications have been described.The complete element was sequenced and 10 ORFs were identified. Some ofthe deduced proteins are highly conserved in other organisms but a putativefunction could be attributed to only a few of them. Duplication of AUD4 byintegration of an Escherichia coli plasmid carrying various parts of AUD4 and athiostrepton-resistance gene in S. lividans AJ100, ZX7 or TK64 inducedamplification of the integrated plasmid, AUD4 or both at high frequency.

Keywords : DNA amplification, deletion, genetic instability, class I amplifiable elements

INTRODUCTION

Genetic instability is a widespread phenomenon in thegenus Streptomyces : mutants affected in various pheno-typic properties like morphological differentiation, sec-ondary metabolism or antibiotic resistance arise atfrequencies between 0±1% and 1% of colony-formingspores (for reviews see Leblond & Decaris, 1994;Dharmalingam & Cullum, 1996; Volff & Altenbuchner,1998). Most of the spontaneous mutations are causedby large chromosomal deletions removing the endsof the linear chromosome. The chromosome be-comes circular and is then even more unstable thanthe linear chromosome (Fischer et al., 1997;Lin & Chen, 1997; Volff et al., 1997). This explainsthe hypervariability and progressive degenerationof such strains, in particular in S. lividans and S.ambofacienswhere circularized chromosomes have beenstudied..................................................................................................................................................

†Present address: Cardiovascular Biology, Pfizer Central Research, PfizerLimited, Sandwich, Kent CT13 9NJ, UK.‡Present address: Klinische Chemie, Franz Josef Strauß-Allee 11, 93053Regensburg, Germany.

Abbreviations: ADS, amplified DNA sequence; AUD, amplifiable unit ofDNA; HT, Hickey–Tresner.

The GenBank accession number for the sequence reported in this paper isAF072709.

Frequently, the large deletions are accompanied byDNA amplifications. Amplifications occur from specificchromosomal sequences called AUDs (amplifiable unitsof DNA). The AUDs amplify to several hundred copiesof head-to-tail arranged units, called ADSs (amplifiedDNA sequences) (Fishman & Hershberger, 1983). Twoclasses of AUD elements were identified. Class II AUDsconsist of two directly repeated sequences flanking aninternal single-copy sequence. Independent mutantsshow the same amplified unit consisting of one of thetwo direct repeats and the internal sequence each. ClassI amplifications arise in certain chromosomal regions ofa strain. The size of these regions is about 100 kb. Eachindividual mutant amplifies a different segment of thisregion; some of the amplified DNAs overlap, some haveno sequence in common with other amplified segments(Hu$ tter & Eckhardt, 1988).

Several amplifiable elements were found in S. lividans.The strain segregates chloramphenicol-sensitive (CmlS)mutants with a frequency of 0±5% of spores and thesemutants again segregate arginine-auxotrophic deriva-tives (Arg–) at about 25% of spores. The double mutantsusually show amplification of AUD1, a class II element,which consists of three 1 kb and two 4±7 kb directrepeats in the order 1 kb–4±7 kb–1 kb–4±7 kb–1 kb. Theamplified unit consists of a 4±7 kb and a 1 kb repeat(Altenbuchner & Cullum, 1985). Another class II

0002-3480 # 1999 SGM 3331


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19

E. SCHMID, C. BU$ CHLER and J. ALTENBUCHNER

element encoding mercury-resistance genes (AUD2) hasa size of 90 kb and is flanked by two directly repeatedcopies of the IS element IS1372. It was amplified togetherwith AUD1 in a CmlS Arg– mutant (Eichenseer &Altenbuchner, 1994). AUD1 is located about 800 kbaway from the end of the linear chromosome(Redenbach et al., 1993). A class I AUD region of about70 kb was described by Rauland et al. (1995) in 2-deoxygalactose-resistant mutants about 300 kb awayfrom the other end of the chromosome. The 24 kbAUD3 sequence, which was found amplified togetherwith AUD1 in a CmlS Arg– mutant (Altenbuchner et al.,1988), and a 4±3 kb sequence found by Betzler et al.(1997) might be part of this AUD class I region.

In this paper we describe the characterization andmapping of a further 8±2 kb amplifiable element from S.lividans, which was found after transformation of aCmlS Arg− mutant with an SCP2 plasmid derivative.

METHODS

Bacterial strains, media and culture conditions. The bacterialstrains used in this work are listed in Table 1. Escherichia coliJM109 (Yanisch-Perron et al., 1985) was used as the host forconstruction of plasmids and DNA sequencing. It was grownin dYT liquid medium or on dYT agar plates (Sambrook et al.,1989) at 37 °C. For transformation, the protocol of Chung etal. (1989) was used. Transformants were selected by adding100 µg ampicillin ml−" to the growth medium. For infectionwith λ phages, the E. coli strain Q358 (Karn et al., 1980) wasgrown in L broth supplemented with 10 mM MgCl

#and 0±4%

maltose (Sambrook et al., 1989). Plaques were obtained in thefollowing way: 0±1 ml of an overnight culture of Q358 wasincubated with 0±1 ml phage dilutions for 10 min, transferredto 3 ml prewarmed (45 °C) liquid L broth soft agar (L brothcontaining 0±7% agar, maltose and MgCl

#), poured onto L

agar plates and incubated at 37 °C overnight. Streptomyceteswere grown at 30 °C in YEME liquid medium supplementedwith 27% w}v sucrose, 0±5% glycine and 5 mM MgCl

#

Table 1. Bacterial strains used in this study

Strain Marker Plasmid Reference

E. coli strains

JM109 recA1 endA1 supE44 relA1 hsdR17 thi

gyrA96 ∆(lac–proAB)

F« [traD36 proAB+ lacIq lacZ∆M15] Yanisch-Perron et al.

(1985)

Q358 supE hsdR φ80 R Karn et al. (1980)

HB101}F«lac : :Tn1739 tnpR

proA2 lacY1 galK2 rpsL20 xyl-5 mtl-1

recA13 ara-14 supE44 hsdS20 (r−B, m−

B)

F« [lac+ proAB+ λcI+ tnpR+, CmR] Altenbuchner (1993)

S. lividans strains

1326 Wild-type SLP2, SLP3 Hopwood et al. (1983)

TK19 Wild-type SLP3 Hopwood et al. (1983)

TK20 Wild-type SLP2 Hopwood et al. (1983)

TK21 Wild-type Hopwood et al. (1983)

TK24 str-6 Hopwood et al. (1983)

TK64 pro str-6 Hopwood et al. (1983)

ZX7 pro str-6 rec-46 Zhou et al. (1988)

AJ100 pro str-6, CmlS Arg− Altenbuchner &

Eichenseer (1991)

(Hopwood et al., 1985). To prepare spore suspensions, thestrains were grown on Hickey–Tresner (HT) agar plates(Pridham et al., 1957) and the spores filtered through cottonwool. S. lividans protoplasts were transformed according toHopwood et al. (1985) and regenerated on R2YE plates.Transformants were selected by overlaying the agar after 12 hwith 3 ml 0±7% soft agar containing 1±25 mg thiostrepton(kindly provided by Hoechst AG). Otherwise, thiostreptonwas used at 50 µg ml−" in YEME medium or in HT and R2YEplates.

DNA manipulation. Restriction enzymes and DNA modifyingenzymes were purchased from Boehringer Mannheim. Forrestriction-enzyme analysis and cloning experiments, standardprocedures were used as described by Sambrook et al. (1989).Plasmid DNA was isolated using the QIAwell 8 plasmidpurification kit (Qiagen). Total DNA of Streptomyces strainswas extracted from cultures grown for 2 d in 15 ml YEMEliquid medium by CsCl}EtBr density-gradient centrifugationas described by Sedlmeier & Altenbuchner, 1992).

Plasmid constructions. The plasmid pJOE907 was constructedfrom the SCP2- and pJOE810-derived shuttle vector pJOE850(Altenbuchner et al., 1988) by insertion of a 15 kb EcoRIfragment from λMT686 (Altenbuchner & Cullum, 1985),containing the complete AUD1 region of S. lividans TK64,into the EcoRI site of the vector. The plasmid pJOE803 is aderivative of pIC20H (Marsh et al., 1984) containing the1±05 kb BclI fragment with the thiostrepton-resistance genefrom pIJ702 (Katz et al., 1983) integrated in the NruI site(Altenbuchner & Eichenseer, 1991). The plasmids pEI53 andpJOE920-1 were constructed by inserting ADS4 DNA,obtained from total DNA of a S. lividans AJ100}pJOE907transformant, as an 8±2 kb BglII fragment, into the BglII site ofpJOE803 and pIC19H (Marsh et al., 1984), respectively. Theplasmids pEI52, pSC30 and pSC22 contained an incompletecopy of ADS4, which was a 5±9 kb BamHI fragment (pEI52), a3±8 kb XhoI–SacI fragment (pSC30) or a 2±63 kb SacI–MluIfragment (pSC22) inserted into pJOE803. The plasmidspSC23, pSC28 and pSC27 were derived from pEI53 by deletinga 5±9 kb KpnI fragment (pSC23), a 4±5 kb PstI–XhoI fragment(pSC28) or a 1 kb SacI fragment (pSC27). The plasmidspEI574 and pEI584-2 contain a 6 kb BamHI fragment from

3332


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19

AUD4 from Streptomyces lividans

λEI33 or a 2±5 kb MluI fragment from λEI32 inserted into theBamHI or EcoRI site of pJOE867, respectively (Altenbuchner,1993). The plasmids pEI584-3 and pEI573 contain a 2±7 kbMluI fragment from λEI32 and a 2±7 kb BamHI fragment fromλEI33, respectively, inserted into pIC19H. The plasmid pEI586is a SmaI-deletion derivative of pEI573, and pEI585 contains a290 bp SmaI–NcoI fragment from pEI584-3 inserted intopIC19H.

Genomic library of S. lividans 1326 in λRESI. Total DNA of S.lividans 1326 was partially digested with Sau3AI, separatedthrough a low-melting agarose gel, fragments in the range of8–12 kb were purified by phenol extraction and ligated to thearms of the λRESI vector (Altenbuchner, 1993) which had beendigested with BamHI and purified in the same way. Thepackaging was done as described by Sambrook et al. (1989).After plaque hybridization, plasmids were generated from theinserts in the phages by infection of E. coli HB101 F«lac : :Tn1739 tnpR (Altenbuchner, 1993).

Southern and plaque hybridization. The blotting of DNAfrom agarose gels onto nitrocellulose filters was done ac-cording to Smith & Summers (1980). λ DNA was transferredfrom plaques to nitrocellulose filters as described by Sambrooket al. (1989). DNA fragments (about 100 ng) were labelledwith [α-$#P]dCTP using the random-primed DNA labelling kitfrom Boehringer Mannheim. Filters were hybridized in buffercontaining 50% formamide and washed under the conditionsdescribed by Hopwood et al. (1985).

PFGE. DNA from the various Streptomyces strains wasprepared from cultures grown for 2 d in YEME liquid mediumat 30 °C, as described by Leblond et al. (1993), and the agaroseblocks digested with AseI for 12 h at 37 °C. The DNA wasseparated on a 0±8% agarose gel in 0±5¬TBE buffer with1±6 mM thiourea using a Chef Mapper from Bio-Rad. Therunning conditions were: initial switch time 1±65 s, finalswitch time 2±3 min, 6 V cm−" for 20 h. Size standards wereSaccharomyces cerevisiae chromosomes and λ concatemericDNA from Bio-Rad.

DNA sequence analysis. DNA sequencing of the AUD4fragment was carried out by the chain-termination methodwith double-stranded plasmid DNA. Various restrictionfragments of AUD4 and ADS4 were inserted into pIC20H andsequenced using Cy5-labelled M13 universal and reverseprimers with the ALFexpress AutoRead sequencing kit(Amersham Pharmacia Biotech). In addition, primer walkingwas performed using oligonucleotides from MWG Biotechand the Cy5-dATP labelling mix in combination with theALFexpress AutoRead sequencing kit. The DNA wasseparated on a 5±5% Hydrolink Long Ranger gel matrix in anALFexpress DNA sequencer for 12 h at 55 °C, 800 V and0±5¬TBE buffer. The nucleotide sequence was analysed withthe GCG program package (Devereux et al., 1984). Codonusage was analysed with a codon-usage table based on theanalysis of 8 Streptomyces genes as described by Sedlmeier &Altenbuchner (1992). Database searches were performed withthe and programs (Altschul et al., 1990) usingthe electronic mail server of the National Center forBiotechnology Information, Bethesda, MD, USA.

RESULTS

Identification and cloning of the AUD4 element

Strain AJ100 has been extensively used to study ampli-fication of AUD1 (Altenbuchner & Eichenseer, 1991). Itis a spontaneous CmlS Arg– mutant of S. lividans TK64

in which the amplifiable element AUD1 was deleted.Plasmid pJOE907 consists of a 15 kb BclI fragmentcontaining the complete AUD1 region, which wasisolated from λMT686 (Altenbuchner & Cullum, 1985)as an EcoRI fragment and inserted into the SCP2-derived vector pJOE850. Transformants of AJ100 con-taining this plasmid showed no amplification of AUD1.Instead, an 8±2 kb BglII fragment was amplified toseveral hundred copies per chromosome in five out ofnine transformants tested. The amplified element wascalled ADS4, and the corresponding non-amplifiedsequence, AUD4. A repetition of the transformationof independently prepared AJ100 protoplasts withpJOE907 failed to evoke amplification of AUD4. Fur-thermore, transformation of S. lividans TK64 withpJOE907 and analysis of twelve transformants byrestriction-enzyme digestion and agarose gel electro-phoresis failed to reveal any amplification of AUD4.Over several years, at least 50 transformations of AJ100,TK64 and ZX7 with SCP2 derivatives containingvarious parts of AUD1 or just AUD1 sequences on E.coli vectors have been carried out, as described byAltenbuchner & Eichenseer (1991) or Volff et al. (1996),and several hundred transformants analysed for DNAamplifications. In all these transformations, the ampli-fication of AUD4 was only seen once more. It occurredin AJ100, the transformants had AUD4 amplified with afrequency similar to that described above and againAJ100 was transformed with an SCP2 derivative con-taining parts of AUD1 as inserts.

The DNA of one of the earlier AJ100 transformants thathad AUD4 highly amplified was digested with BglII, theDNA separated through an agarose gel and the 8±2 kbfragment isolated and inserted into the E. coli vectorpIC19H. The resulting plasmid, pJOE920-1, was used toscreen a genomic library of S. lividans 1326 for thecorresponding AUD4 sequence. The library was con-structed by insertion of S. lividans 1326 DNA, partiallydigested with Sau3AI, between the BamHI sites of the λ

replacement vector λRESI. Three different phageshybridizing with pJOE920-1 were isolated. The phageswere converted into plasmids as described byAltenbuchner (1993) and the insertions mapped byrestriction analysis. The location of AUD4 in the phageinserts was determined by comparing the restrictionmaps from pJOE920-1 and the corresponding wild-typesequences in the phages λEI32, λEI33 and λEI35 (Fig. 1)as well as by Southern hybridization (not shown).

Deletions in strains with ADS4 and mapping of AUD4by PFGE

Spontaneous CmlS Arg– S. lividans mutants with ampli-fications of AUD1 usually have one or both chromo-somal ends deleted (Redenbach et al., 1993; Rauland etal., 1995; Volff et al., 1996). One side of the deletionends near or within the amplified DNA. To see if theamplification of AUD4 is accompanied by a deletion, a2±5 kb MluI fragment from λEI32 flanking AUD4 on theleft side and a 6 kb BamHI fragment from λEI33 flankingAUD4 on the right side (Fig. 1) were inserted into the

3333


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


.....................................................................................................

Fig. 1. Restriction map of the chromosomalregion containing AUD4. The location ofAUD4 is indicated by a solid arrow. Theregion was isolated in three overlappingfragments in the λ phages λEI32, λEI33 andλEI35. The position of restriction fragmentsisolated from the phages and inserted intothe pIC plasmids or pJOE867 are indicated.The deletion seen in AJ100 strains withAUD4 amplified is marked by a dottedarrow.

(a)

(b)

0 1 2 3 4 5 1 2 3 4 5

0 1 2 3 4 5 1 2 3 4 5

.................................................................................................................................................

Fig. 2. Agarose gel electrophoresis of total DNA from S.lividans TK64 (lane 1), three transformants of AJ100 withpJOE907 (lanes 2–4) and AJ100 (lane 5), digested with BamHI(a) or MluI (b), and Southern blot hybridization with [α-32P]dCTP-labelled pEI574 (a) and pEI584-2 (b). The size standardis λ DNA digested with HindIII (lane 0).

vector pJOE867 (to give pEI584-2 and pEI574, respect-ively). The plasmid DNA was labelled with [α-$#P]dCTPand hybridized to total DNA of TK64, three strains ofAJ100 transformed with pJOE907 containing amplifiedAUD4, and AJ100. The same 6 kb BamHI fragment

hybridized in all strains, regardless of whether theycontained an amplification. The 2±5 kb MluI fragmentwas present only in strains without an amplification(Fig. 2). This indicates that the amplification of AUD4 isaccompanied by deletions on one side of the element, asobserved with other amplifications.

To localize the AUD4 sequence on the S. lividanschromosome, cells were embedded in agarose plugs, theDNA purified for PFGE and digested with the rare-cutting enzyme AseI. Since there are considerabledifferences between S. lividans strains, DNA was pre-pared from S. lividans strains 1326, TK19, TK20, TK21,TK24, TK64 (Hopwood et al., 1983) and ZX7 (Tsai &Chen, 1987; Zhou et al., 1988). S. lividans 1326 wascured of the linear 50 kb plasmid SLP2 so this band ismissing in TK19 (Fig. 3). In a second step, TK19 wascured of SLP3, a plasmid identified only by the lethalzygosis phenotype; after the curing, a 380 kb AseI band(AseI-F*) was missing (TK21). Strains TK24 and TK64are derived from TK21 and showed no differences totheir progenitors in the AseI pattern. ZX7 is derivedfrom TK64 by NTG mutagenesis and has a 900 kb anda 80 kb fragment (AseI-B* and AseI-J*, respectively)deleted. The AseI and DraI map of S. lividans was madefrom ZX7 and so the positions of bands B*, F* and J* arenot known. TK20 is a derivative of 1326, independentlycured of the plasmid SLP3. In this strain, both the AseI-F* and AseI-D fragments are missing. A new band of670 kb is presumably the result of a deletion fusing theAseI-D and AseI-F* fragments.

In all strains, the AseI-D fragment hybridized with theADS4 sequence in plasmid pJOE920-1, except in TK20where no hybridization was found (Fig. 3). The AseI-Dfragment was mapped near the chromosomal end on theopposite side to AUD1. This region is also affected bygenetic instability. Class I amplifications have beendescribed in this region by Redenbach et al. (1993) andRauland et al. (1995). The class I AUD region wascloned by Rauland et al. (1995) in the cosmids A6,A7 and U31, altogether a region of about 90 kb.

3334


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


A

BB*

CD

E1–2–3F*F

GH1–2

I1–2J*J

SLP2

0 1 2 3 4 5 6 7 1 2 3 4 5 6 7

.................................................................................................................................................................................................................................................................................................................

Fig. 3. PFGE of AseI-digested DNA of S. lividans TK19 (lane 1), TK20 (lane 2), TK21 (lane 3), TK24 (lane 4), TK64 (lane 5),ZX7 (lane 6) and 1326 (lane 7), and Southern-blot hybridization with labelled pJOE920-1 containing ADS4. Thenumbering of the AseI fragments is according to Leblond et al. (1993). Fragments not present in the S. lividans ZX7 mapare marked by asterisks. Saccharomyces cerevisiae chromosomes (Bio-Rad) were used as size standards (lane 0).

(a)

(b)

.....................................................................................................

Fig. 4. (a) Comparison of the nucleotidesequences of the ends of AUD4 and thejunction fragment in ADS4. Inverted repeatsare marked by arrows and a rectangle isdrawn around the two cytosine residuescommon to both ends. (b) Comparison ofthe left and right sequence flanking AUD4.Nucleotides repeated in direct orientationare marked by rectangles. The two cytosineresidues at the borders of AUD4 are given inlower-case letters.

Hybridization of ADS4 DNA to these cosmids gave nosignal (data not shown).

Nucleotide sequence of AUD4

The nucleotide sequence of the 8±2 kb BglII fragment,comprising a complete copy of the ADS4 element, inpJOE907 was determined on both strands. To identifythe ends of AUD4 and the flanking chromosomalsequences, a 2±7 kb BamHI fragment from λEI33(pEI573) and a 2±7 kb MluI fragment from λEI32(pEI584-3) were inserted into pIC19H (Fig. 1), mapped

in detail and compared to the corresponding amplifiedADS4 junction fragment. Finally, a 290 bp SmaI–NcoIfragment from pEI584-3 and a 310 bp BamHI–SmaIfragment from pJOE573 were sequenced. Together withthe ADS4 sequence, the AUD4 sequence could bereconstructed. AUD4 was 8202 bp long and with theflanking chromosomal regions, a continuous sequenceof 8366 bp was obtained (GenBank accession no.AF072709). The nucleotide sequence of the ends ofAUD4 and the ADS4 junction is shown in Fig. 4. Thecrossover point leading to the amplification of AUD4must have been between two cytosine residues present

3335


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


MluI ClaI BglII BamHI EcoRV BamHI

ORF: 1 2 43

5 6 7 8 9 10

BamHI

.....................................................................................................

Fig. 5. Restriction map of the 8366 bpsequence containing AUD4 (black arrow)and the position of the ten ORFs (whitearrows) identified in the sequence.

Table 2. ORFs and deduced proteins in the AUD4 sequence

ORF Size

(aa)

Putative RBS* GenBank protein

identifier of

similar protein

Identity

(%)

Putative function

ORF1 145 CTGGAGGTTCAGCCatg gnlrPIDrd1011372 64±8 Unknown

ORF2 192 AAGGAAGAGGTCCCCatg gnlrPIDre1248764 36±9 Regulatory protein

ORF3 66 GTAGGAGCTCCCGGCatg sprP26910 60±3 Ferredoxin

ORF4 406 TCTGCGAGGTCTTCCatg sprP26911 68±5 Cytochrome P-450

oxidoreductase

ORF5 149 CAGGGACGGATGCACatg sprQ10772 70±1 Unknown

ORF6 338 CCGCAAGGCGGGGTgtg gir2654559 28±6 Unknown

ORF7 341 GGAAGGAAGGCAGGGGCatg gnlrPIDrd1019605 44±5 Unknown

ORF8 131 CCTTCGGAGGAGACatg gnlrPIDre317273 25±2 Unknown

ORF9 313 GTTGCGAGGTCTCGatg sprP43903 39±7 Quinone oxidoreductase

ORF10 128 GCTGGAGGCAAGGTTCgtg sprQ11035 25±0 Unknown

*The putative RBSs are underlined; start codons are indicated in lower case.

on both ends of AUD4. A 11 bp sequence was found16 nt from one end of AUD4 which is inversely repeatedon the other side, here 22 nt away from the twocytosines. A comparison of the right and left AUD4flanking region revealed long imperfect direct repeatswith 25 out of 39 nt identical (Fig. 4b). The ADS4sequence of three additional transformants was clonedin E. coli and the junction analysed by DNA sequencing.They all showed exactly the same sequence. All theamplified strains tested came from a single transform-ation of AJ100 with pJOE907. They should be in-dependent transformants, since immediately after theaddition of DNA, the protoplasts were embedded in softagar and spread on R2 agar plates.

The complete AUD4 sequence was analysed for codonusage with the CodonPreference program using aStreptomyces codon usage table as described bySedlmeier& Altenbuchner (1992). The upstream regionsof each ORF were analysed for sequences matchingStreptomyces ribosome-binding sites (Strohl, 1992). Tenputative geneswere identified (Fig. 5). The gene productswere compared with entries in various protein databasesusing the programs. The results are listed in Table2. A possible function can be attributed to ORFs 3 and4, which are highly similar in sequence to cytochrome P-450 oxidoreductase and ferredoxin, the correspondingelectron carrier. ORF9 shares a significant identity withquinone oxidoreductases andORF2might be a regulatorprotein. The genes encoding ORF1 and ORF5 seem tobe highly conserved in some bacteria but the function isunknown. Between 25% and 44% amino acid sequence

identities with hypothetical proteins, mainly identifiedin genome sequencing projects, were found for theremaining ORFs.

Induction of AUD4 amplification

When an E. coli plasmid with a DNA fragmenthomologous to the incomplete AUD1 region in strainAJ100 was integrated into the AJ100 chromosome byhomologous recombination, the integrated plasmid wasamplified to high copy numbers in most transformants.The only requirement for frequent, high-level am-plification was a minimum length of 2 kb for thedirect repeats generated by the integrating plasmid(Altenbuchner & Eichenseer, 1991). To see if ampli-fication of AUD4 in AJ100 could be evoked by similarexperiments, various fragments of ADS4 were insertedinto the polylinker sequence of pJOE803, a pIC20Hderivative containing a thiostrepton-resistance gene.Representative examples of the plasmids constructedare given in Fig. 6. Since the DNA for the plasmidconstruction was taken from the ADS4 sequence, therewere, in principle, two different sorts of plasmids. Someof them contained the junction from the left and rightends of AUD4, i.e. pEI53, pEI52, pSC27 and pSC30 andothers, i.e. pSC22, pSC23 and pSC28, that containedfragments from the internal region of AUD4. Afterintegration of the former group of plasmids into thechromosomal AUD4 sequence there will be two copiesof AUD4 in the chromosome, one AUD4 wild-type copyand a second copy with the AUD4 ends but modified

3336


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


.....................................................................................................

Fig. 6. Plasmid constructions containingpIC20H (white bars), the thiostrepton-resistance gene (hatched bar) and variousparts of ADS4 (black bars and arrows) whichwere used to transform S. lividans AJ100and TK64. At the top of the figure, therestriction map of ADS4 (represented by twoADS4 copies) is shown. On the right side theresults achieved with these plasmids inAJ100, ZX7 and TK64 are given: () weakamplification of the integrated plasmid; strong amplification of the integratedplasmid, AUD4 or both.

.....................................................................................................

Fig. 7. Structure of AUD4 in S. lividansAJ100 or TK64 after integration of theplasmids pSC22 or pSC27 (representative forall plasmids shown in Fig. 6) by homologousrecombination. For pSC27, the two possiblesites of integration are shown. Symbols arethe same as in Fig. 6; white arrows indicatethe sequence duplicated by the integration.MluI* indicates the site destroyed duringconstruction of the plasmid.

internally by the integrated plasmid sequence. With thelatter group of plasmids there will be just one AUD4 unitwith an internal duplication flanking the E. coli plasmid(Fig. 7). Strain AJ100 as well as the wild-type strains

TK64 and ZX7 were transformed with these plasmids.Thiostrepton-resistant transformants were streaked outon HT agar plates containing thiostrepton and fromthere transferred to 10 ml YEME liquid medium supple-

3337


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


0 1 2 3 4 5 6 7 8

ADS4

.................................................................................................................................................

Fig. 8. Examples of DNA amplifications found by restrictionenzyme analysis in S. lividans AJ100 and TK64 transformed withplasmids as shown in Fig. 6. Lanes: 0, λ¬HindIII size standard;1, AJ100/pSC22; 2, pSC22; 3, AJ100/pSC27; 4, pSC27; 5, TK64/pEI52no. 1; 6, TK64/pEI52, no. 2; 7, TK64/pEI52 no. 3; 8, pEI52. TheDNA in lanes 1 and 2 was digested with EcoRV and with SacI inlanes 3–8.

mented with thiostrepton. After growth for about 3 d,total DNA was extracted and characterized by re-striction analysis for DNA amplifications. With pSC22 aweak amplification of the integrated plasmid was seen inone in five AJ100 transformants and one in six ZX7transformants tested (Fig. 8). Similar results wereobtained with pSC23: one in five AJ100 and one in sixZX7 transformants showed weak amplification of theintegrated plasmid. With pSC28, which had the longestADS4 insert in this group, there was no amplificationseen in AJ100 (five transformants tested), weak ampli-fication in one out of five TK64 transformants and astrong amplification of the plasmid in three out of sixZX7 transformants. None of the transformants had thecomplete AUD4 element amplified together with theintegrated plasmid. With pSC27, pSC30, pEI52 andpEI53, some transformed colonies showed amplificationof just the integrated plasmid. Others showed ampli-fication of the integrated plasmid and, separately, theamplification of AUD4 at the same or higher copynumber. For example, with pSC30 one out of threecolonies of AJ100 had the plasmid weakly amplified,whereas from seven colonies of ZX7 tested, two had theplasmid highly amplified and two other colonies had alow plasmid amplification but in addition a highamplification of AUD4. Similar frequencies wereobtained with pSC27, pEI52 and pEI53. Some examplesare shown in Fig. 8.

0 1 2 3 4 5 6 7 8 9 10(a)

1 2 3 4 5 6 7 8 9 10(b)

.................................................................................................................................................

Fig. 9. Southern-blot hybridization of a labelled 1±2 kb KpnIfragment from the terminal inverted repeats of the S. lividanschromosome to chromosomal DNA of TK64 and ZX7 derivativestransformed with pSC27, pSC28, pSC30 and pEI52. The DNA onthe agarose gel (a) was digested with KpnI. The autoradiogram(b) shows the hybridizing 1±2 KpnI fragments from the strains.Lanes: 0, λ¬HindIII size standard; 1 and 2, ZX7/pSC30; 3 and 4,TK64/pSC27; 5 and 7, ZX7/pSC28; 6, TK64; 8–10, TK64/pEI52. Foreach pair of lanes, different colonies from a transformationexperiment were used for DNA preparations.

In general, the frequency of amplification was in a rangebetween 10% and 40% of all transformants tested, andin TK64 and ZX7 it was slightly higher than in AJ100 inwhich ADS4 had been originally found. This is incontrast to integration of plasmids in the AUD1 region,where the frequency of amplification was high in AJ100(50–75% of the transformants) and low in TK64 or ZX7(in the latter strains, amplification was first seen afterloss of the chloramphenicol resistance and argG at theusual frequencies). All the TK64 and ZX7 transformantsshowing high amplification of AUD4 were still chlor-amphenicol resistant and arginine prototrophs exceptone mutant of TK64 transformed with pEI52, which hadthe AUD1 element amplified in addition to AUD4 (notshown).

Several transformants of TK64 or ZX7 were also testedto see if they had still intact chromosomal ends. The

3338


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


total DNA of six transformants with high amplificationsofAUD4 and}or the integrated plasmids (pSC27, pSC28,pSC30), of three with low amplifications or justrecognizable amplification (pEI52) was digested withKpnI and hybridized with a 1±2 kb KpnI fragment frompLUS449, a plasmid containing DNA from the terminalinverted repeats of the S. lividans chromosome, just afew base pairs away from the chromosomal ends (Lin etal., 1993). From the six mutants with high ampli-fications, only one showed no hybridization, the othersgave faint hybridizing bands in comparison to S. lividansTK64 wild-type. The bands of the other mutants withlow amplification showed reduced intensities or hadabout the same intensity as the wild-type (Fig. 9). Thisindicates i, a correlation of the amplification of AUD4and the deletions of the chromosomal ends, and ii, thateach single mutant contains chromosomes at differentstages of deletions.

Finally, five TK64 derivatives transformed with pSC22,pSC27 or pSC28 without visible amplification weretested to determine whether the plasmids were correctlyinserted via homologous recombination. The DNA wasdigested with PstI and hybridized to detect ADS4 DNAby Southern blotting. The plasmids chosen had no PstIsite and in TK64 DNA the enzyme cuts outside theAUD4 DNA and gives a high molecular mass fragmentof greater than 20 kb. No additional fragment was seenin the transformants (data not shown), which indicatesthat the plasmids were integrated by homologousrecombination.

DISCUSSION

A class I AUD region was mapped in S. lividans about300 kb away from one chromosomal end (Redenbach etal., 1993; Rauland et al., 1995). This corresponds to theAseI-D band in which AUD4 was mapped in this workby Southern hybridization.Hybridization ofADS4DNAto cosmids containing the class I AUD region gave nosignal. Therefore, AUD4 seems to be a different class Ielement. The AUD4 sequence was found in all S. lividansstrains tested except in TK20 where the AseI-D andAseI-F* bands are missing. AseI-F* is also deleted inZX7 and so its position on the chromosome is unknown.However, it seems very likely that AseI-D and AseI-F*are fused in the new band seen in TK20, which wouldmean that AseI-F* is located on one or the other side ofthe AseI-D fragment.

DNA amplifications have been observed in many bac-teria. They can be interpreted as an adaptive andreversible response to the demand of higher resistance totoxic metals or antibiotics, to conditions of nutrientlimitations or to processes of pathogenic and symbioticinteractions (reviewed by Romero & Palacios, 1997).The number of amplified copies is low and often thereare only duplications. Unequal crossover or circleexcision and reinsertions at large tandem repeats wereproposed as mechanisms of amplification. The lowamplifications of plasmids observed in some S. lividansstrains after transformation and integration of the ADS4

containing pJOE803 derivatives into the chromosomemight have been selected by insufficient thiostreptonresistance mediated by just one copy of the gene. To ourknowledge, a precise relationship between thiostreptonresistance and copy number of the resistance gene in acell has not formally been tested.

The spontaneous high DNA amplifications in strepto-mycetes differ from amplifications in other bacteria inmany respects. Large deletions removing one or bothends of the chromosome precede the amplificationevents. DNA sequence analysis of amplified DNA hasrevealed the presence of putative genes and in a fewcases gene expression was demonstrated (Aigle et al.,1996; Betzler et al., 1997; this work) but there is noindication that the amplification of these genes is of anyselective advantage to the cells. Even in the case ofAUD2, which encodes the mercury-resistance genes,amplification was spontaneous and not selected byadding high mercury concentrations to the medium(Sedlmeier & Altenbuchner, 1992). Furthermore, theclass I amplifiable regions are quite large but in each ofthe mutants only small and different parts of theseregions are amplified. Therefore, it seems that ampli-fication of DNA in a certain region itself is selected butthe amplification of specific genes is not.

According to a hypothesis discussed by Volff &Altenbuchner (1998), which is based on observationsmade with artificially circularized S. lividans chromo-somes, the amplification of DNA in streptomycetesoccurs in mutants with circular chromosomes orinverted fused chromosomes. Due to the lack of rep-lication terminators, the DNA is overreplicated in thechromosomal region where the replication forks meet.The overreplicated DNA is ordered into tandem repeatsby illegitimate recombination. These amplifications areunstable and replaced by more stable ones like theAUD1 amplification. This needs deletions in the vicinityof the amplifiable region which direct the meeting pointof replication forks to the new positions. The frequencyof amplification of AUD1 is very high due to the presenceof long tandem repeats. In contrast, class I sequences areamplified rarely. The rate-limiting step seems to be aduplication of a sufficiently large single-copy sequenceby illegitimate recombination events. This might alsobe the case for AUD4. Only the duplication of largeparts of AUD4 through integration of the plasmidspEI52, pEI53, pSC27, pSC28 and pSC30 leads to anefficient amplification of AUD4 sequences regarding thecopy number of the amplified DNA and number oftransformants showing amplifications.

It seems unlikely that the AUD1 sequence on pJOE907induced amplification of AUD4 in a direct way, since nosignificant identity was found between the twosequences using the GCG program for com-parison. The high frequency of amplification of AUD4in AJ100 in one transformation experiment withpJOE907 and the failure to detect it in a secondexperiment, or in TK64, as well as in many othertransformations of AJ100, ZX7 or TK64 using deriva-

3339


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


tives similar to pJOE907 may be explained in thefollowing way. Presumably, a large part of the AJ100cells used for protoplasting had already spontaneouslyduplicated the AUD4 sequence and the protoplastingand regeneration evoked the high amplification. Otherstocks of AJ100 (as well as of TK64 or ZX7) had no suchduplication and therefore showed no amplification.Unfortunately, the original stock of AJ100 used in thefirst experiment could not be tested by Southern blothybridization and another stock of AJ100 which gave noamplification ofAUD4when transformedwith pJOE907contained no duplication of AUD4, as expected [thiswas shown by digesting AJ100 DNA with BglII andhybridization with labelled ADS4 sequence (data notpresented)].

S. lividans strain AJ100 has a deletion in the amplifiableAUD1 element and a circular chromosome (data notshown) that according to the hypothesis of Volff &Altenbuchner (1998) should lead to amplification ofother DNA sequences like the class I element AUD4.Strains TK64 and ZX7 have linear chromosomes. Theintegration of plasmids like pSC27 generates longduplications which may favour amplification of AUD4instead of AUD1 in these strains. In the mutants tested,the high amplification of AUD4 correlated with a highfrequency of deletions of the chromosomal ends asshown by Southern hybridization using DNA from thechromosomal ends. Again, it seems that the highamplification of AUD4 needed the deletion of thechromosomal ends. One can conclude from this ex-periment that each mutant was heterogeneous and themycelium represented different stages of chromosomaldeletions. It is tempting to speculate that also the degreeof amplification in each of the chromosomes of a specificmutant is different with no or low amplification in intactchromosomes and high amplification in chromosomeswhere the ends are deleted and the chromosomescircularized. The DNA rearrangements in such cellscould be occurring in various orders. The integration ofthe ADS4 DNA might stimulate first low copy ampli-fications, followed by deletion formation and highamplification or stimulate first deletions which theninduce the high copy amplification. All of the plasmid-induced amplifications can be explained by homologousrecombination at long direct repeats generated by theintegration. For the amplification of AUD4 in AJ100,transformed by pJOE907, there were no such long directrepeats. The first duplication must have occurredbetween two cytosine residues, which are the only directrepeated base pairs at the crossover site. There are alsotwo imperfect directly repeated sequences flankingAUD4 but the crossover point is at the end of one andthe beginning of the other repeat and not at the ends orwithin these direct repeats, as one would expect. Thetwo inverted repeats found inside AUD4 have differentdistances to the crossover point. It is therefore unclearwhich DNA sequences directed the recombiningenzymes to these crossover points.

It is also unclear what defines a chromosomal region asan AUD region. Is it the border between regions of

nonessential and essential genes where deletions in anunstable strain come to a halt since any furtherdeletions would be lethal? Are there other propertiesthat favour the amplification of a specific DNA sequencein a AUD class I region? For the class II element AUD1it was found that, besides the long direct repeats, thebinding of the regulatory protein encoded by the 1 kbrepeats to the binding sites up- and downstream of theright 1 kb repeat is necessary for efficient amplification(Volff et al., 1996). It is possible that the gene product ofone of the ORFs in AUD4 also favours the amplificationof the element.

REFERENCES

Aigle, B., Schneider, D., Morilhat, C., Vandewiele, D., Dary, A.,Holl, A.-C., Simonet, J.-M. & Decaris, B. (1996). An amplifiable anddeletable locus of Streptomyces ambofaciens RP181110 containsa very large gene homologous to polyketide synthase genes.Microbiology 142, 2815–2824.

Altenbuchner, J. (1993). A new λ RES vector with a built-inTn1721-encoded excision system. Gene 123, 63–68.

Altenbuchner, J. & Cullum, J. (1985). Structure of an amplifiableDNA sequence in Streptomyces lividans 66. Mol Gen Genet 201,192–197.

Altenbuchner, J. & Eichenseer, Ch. (1991). A new system to studyDNA amplification in Streptomyces lividans. In Genetics andProduct Formation in Streptomyces, pp. 253–264. Edited by S.Baumberg, H. Kru$ gel & D. Noack. New York & London:Plenum.

Altenbuchner, J., Eichenseer, Ch. & Bru$ derlein, M. (1988). DNAamplification and deletion in Streptomyces lividans. In Biology ofActinomycetes, pp. 139–144. Edited by Y. Okami, T. Beppu & H.Ogawara. Tokyo: Japan Scientific Press.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J.(1990). Basic local alignment search tool. J Mol Biol 215, 403–410.

Betzler, M., Tlolka, I. & Schrempf, H. (1997). Amplification of aStreptomyces lividans 4±3 kbDNA element causes overproductionof a novel hypha- and vesicle-associated protein. Microbiology143, 1243–1252.

Chung, C. T., Niemela, S. L. & Miller, R. H. (1989). One-steppreparation of competent Escherichia coli : transformation andstorage of bacterial cells in the same solution. Proc Natl Acad SciUSA 86, 2172–2175.

Devereux, J., Haeberli, P. & Smithies, O. (1984). A comprehensiveset of sequence analysis programs for the VAX. Nucleic Acids Res12, 387–395.

Dharmalingam, K. & Cullum, J. (1996). Genetic instability inStreptomyces. J Biosci 21, 433–444.

Eichenseer, Ch. & Altenbuchner, J. (1994). The very largeamplifiable element AUD2 from Streptomyces lividans 66 hasinsertion-sequence-like repeats at its ends. J Bacteriol 176,7107–7112.

Fischer, G., Decaris, B. & Leblond, P. (1997). Occurrence ofdeletions, associated with genetic instability in Streptomycesambofaciens, is independent of the linearity of the chromosomalDNA. J Bacteriol 179, 4533–4558.

Fishman, S. E. & Hershberger, C. L. (1983). Amplified DNA inStreptomyces fradiae. J Bacteriol 155, 459–466.

Hopwood, D. A., Kieser, T., Wright, H. M. & Bibb, M. J. (1983).Plasmids, recombination and chromosome mapping in S. lividans66. J Gen Microbiol 129, 2257–2269.

3340


IP: 54.237.57.119

On: Sat, 14 May 2016 15:19:19


Hopwood, D. A., Bibb, M. J., Chater, K. F. & 7 other authors(1985). Genetic Manipulation of Streptomyces : a LaboratoryManual. Norwich, UK: John Innes Foundation.

Hu$ tter, R. & Eckhardt, T. (1988). Genetic manipulation. InActinomycetes in Biotechnology, pp. 89–184. Edited by M.Goodfellow, S. T. Williams & M. Modarski. London: AcademicPress.

Karn, J., Brenner, S., Barnett, L. & Cesareni, G. (1980). Novelbacteriophage cloning vector. Proc Natl Acad Sci USA 77, 5172.

Katz, E., Thompson, C. J. & Hopwood, D. A. (1983). Cloning andexpression of the tyrosinase gene from Streptomyces antibioticusin Streptomyces lividans. J Gen Microbiol 129, 2703–2714.

Leblond, P. & Decaris, B. (1994). New insights into the geneticinstability of Streptomyces. FEMS Microbiol Lett 123, 225–232.

Leblond, P., Redenbach, M. & Cullum, J. (1993). Physical map ofthe Streptomyces lividans 66 genome and comparison with that ofthe related strain Streptomyces coelicolor A3(2). J Bacteriol 175,3422–3429.

Lin, Y-S. & Chen, C. W. (1997). Instability of artificially circularizedchromosomes of Streptomyces lividans. Mol Microbiol 26,709–719.

Lin, Y.-S., Kieser, H. M., Hopwood, D. A. & Chen, C. (1993). Thechromosomal DNA of Streptomyces lividans 66 is linear. MolMicrobiol 10, 923–933.

Marsh, J. L., Erfle, M. & Weykes, E. J. (1984). The pIC plasmid andphage vectors with versatile cloning sites for recombinantselection by insertional inactivation. Gene 32, 481–485.

Pridham, T. G., Anderson, P., Foly, C., Linderfelser, C. W.,Hesseltime, C. W. & Benedict, R. C. (1957). A selection of mediafor maintenance and taxonomic studies of Streptomyces. AntibiotAnnu 1956–1957, 947–953.

Rauland, U., Glocker, I., Redenbach, M. & Cullum, J. (1995). DNAamplifications and deletions in Streptomyces lividans 66 and theloss of one end of the linear chromosome. Mol Gen Genet 246,37–44.

Redenbach, M., Flett, F., Piendl, W., Glocker, I., Rauland, U.,Wafzig, O., Kliem, R., Leblond, P. & Cullum, J. (1993). TheStreptomyces lividans 66 chromosome contains a 1 Mb deleto-

genic region flanked by two amplifiable regions. Mol Gen Genet241, 255–262.

Romero, D. & Palacios, R. (1997). Gene amplification and genomicplasticity in prokaryotes. Annu Rev Genet 31, 91–111.

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). MolecularCloning : a Laboratory Manual, 2nd edn. Cold Spring Harbor,NY: Cold Spring Harbor Laboratory.

Sedlmeier, R. & Altenbuchner, J. (1992). Cloning and DNAsequence analysis of the mercury-resistance genes of Streptomyceslividans. Mol Gen Genet 236, 76–85.

Smith, G. E. & Summers, M. D. (1980). The bidirectional transferof DNA and RNA to nitrocellulose or diazobenzyloxymethylpaper. Anal Biochem 109, 123–129.

Strohl, W. R. (1992). Compilation and analysis of DNA sequencesassociated with apparent streptomycete promoters. Nucleic AcidsRes 20, 961–974.

Tsai, J. F.-Y. & Chen, C. W. (1987). Isolation and characterizationof Streptomyces lividans mutants deficient in intra-plasmidrecombination. Mol Gen Genet 208, 211–218.

Volff, J.-N. & Altenbuchner, J. (1998). Genetic instability of theStreptomyces chromosome. Mol Microbiol 27, 239–246.

Volff, J.-N., Eichenseer, C., Viell, P., Piendl, W. & Altenbuchner, J.(1996). Nucleotide sequence and role in DNA amplification of thedirect repeats composing the amplifiable element AUD1 ofStreptomyces lividans 66. Mol Microbiol 21, 1037–1047.

Volff, J.-N., Viell, P. & Altenbuchner, J. (1997). Artificialcircularization of the chromosome with concomitant deletion ofits terminal inverted repeats enhances genetic instability andgenome rearrangement in Streptomyces lividans. Mol Gen Genet253, 753–760.

Yanisch-Perron, C., Vieira, J. & Messing, J. (1985). Improved M13phage cloning vectors and host strains : nucleotide sequences ofthe M13mp18 and pUC19 vectors. Gene 33, 103–119.

Zhou, X., Deng, Z., Firmin, J. L., Hopwood, D. A. & Kieser, T.(1988). Site-specific degradation of Streptomyces lividans DNAduring electrophoresis in buffers contaminated with ferrous iron.Nucleic Acids Res 16, 4341–4352.

.................................................................................................................................................

Received 5 May 1999; revised 2 August 1999; accepted 27 August 1999.

3341

AUD4, a new amplifiable element from Streptomyces lividans

Documents