Gene, 48 (1986) 133-144

Elsevier

133

GEN 01799

Structure of P&ZYUM actin gene locus ardA: a nonpalindromic sequence causes inviability of phage lambda and recA-independent deletions

(Recombinant DNA; exon; intron; chloroacetaldehyde; exonuclease I and V; recBC -sbcB - mutants; S l- hypersensitive sites; slime mold)

Werner F. Nadera*, Gerhard Isenberg” and Helmut W. Sauerb

a Max-Planck-Institutefor Psychiatry, Department of Neurochemistry, 8033 Planegg-Martinsried (F.R. G.) Tel. (089)85781, and b Department of Biology, Texas A &M University, College Station, TX 77843 (U.S.A.) Tel. (409)845-7760

(Received July 29th, 1986)

(Revision received September 19th, 1986)

(Accepted September 22nd, 1986)

SUMMARY

Previously we reported that approx. 80% of the genome from the plasmodial slime mold Physantm poly-

cephafum, including all the actin genes, can be cloned only in recBC- sbcB - Escherichiu coli hosts [Nader et al., Proc. Natl. Acad. Sci. USA 82 (1985) 2698-27021. We have now sequenced the actin gene locus arti. The nucleotide sequence of its coding region is flanked by the typical putative regulatory sequences for transcription initiation and polyadenylation. The coding region is interrupted by five introns, all located at novel positions with regard to those of previously analysed actin genes. Within the ardA gene we have located a 360-bp fragment which comprises exon V and parts of its flanking introns. This region suppresses plaque formation of recombinant I phages and causes recA-independent deletions in phages and plasmids. In contrast to our previous hypothesis, this sequence is not a DNA palindrome, but consists of five (dA) * (dT)- and (dG) - (dC)- homopolymers. Both termination of replication and partial unwinding of duplex DNA under torsional stress were detected within the unstable 360-bp region in vitro.

INTRODUCTION

Most eukaryotic DNA sequences can be stably propagated in E. coli. Recently, however, it was

* To whom correspondence and reprint requests should be sent

at his present address: Orpegen GmbH, Im Neuenheimer Feld

517, 6900 Heidelberg (F.R.G.) Tel. (06221)27082-5.

reported that certain eukaryotic DNAs are refractive to common cloning procedures. Significant fractions of the Drosophila (5 y0 ; Petri and Wyman, 1985), the human genome (8.9%; Wyman et al., 1985) and most DNA from the plasmodial slime mold Physarum polycephalum (80% ; Nader et al., 1985) can only be detected and maintained in E. coli, lacking both exonucleases V and I (recBC_sbcB -

Abbreviations: aa, amino acid(s); bp, base pair(s); kb, kilobases

or 1000 bp; nt, nucleotides; PolIk, Klenow (large) fragment of

E. coli DNA polymerase I; Pu, purine; Py, pyrimidine; ss, single-

stranded.

mutant strains). It was suggested (Nader et al., 1985; Wyman et al., 1985) that this instability was due to palindromic DNA sequences because they were the only structures known to suppress plaque formation

0378-I 119/86/$03.50 0 1986 Elsevier Science Publishers B.V. (Biomedical Division)

134

of phage I (Leach and Stahl, 1983) or to cause recA independent deletions (Collins et al., 1982). Further- more, secondary structures consistent with foldback sequences were observed in ss and self-annealed Physarum and human DNA from unstable clones in electron-micrographs (Nader et al., 1985; Wyman et al., 1985).

During analysis of a recombinant phage carrying the Physarum actin gene ardA we have detected a nonpalindromic sequence, which caused the same deletion phenomena as those observed for the artifi- cially constructed palindromes. This 360-bp element prevented the detection of the ardA gene in an EMBL3 library of genomic Physarum DNA (Nader et al., 1985). This element contains a series of five (dA) * (dT)- or (dG) . (dC)-homopolymers and is part of the actin gene itself. It comprises exon V and parts of the two flanking introns.

In our first cloning attempts with Physarum we

concentrated on the actin gene family because the four unlinked loci have been well defined by Mende- lian analysis of polymorphic restriction fragments (Schedl and Dove, 1982) and the chronological order of their replication has been determined during the naturally synchronous mitotic cycle of a Physarum plasmodium. Thus it is known that ardB,

ardC and ardD replicate early whereas ardA repli- cates late (Pierron et al., 1984). The nucleotide sequence of a&A contains putative transcription initiation and polyadenylation signals, and the coding region conforms to the known aa sequence of Physarum actin (Vanderkerckhove and Weber, 1978). However, the coding region is interupted by five introns, which is an unusually high number for a lower eukaryote. Moreover, the locations of all the introns are different from the positions determined for any other actin gene (Breathnach and Chambon, 198 1; Buckingham, 1985).

MATERIALS AND METHODS

(a) Strains and DNA clones

Recombinant phages ;1PpAlO and IPpAlO’ and subclones in pBR322 and their propagation in the E. coli host strain CES200 were described previously (Nader et al., 1985). For sequencing with the chain-

termination method, subclones from the DNA in those vectors were constructed in the bacteriophages M13mp18 and M13mp19 (Norrander et al., 1983), transformed into E. coli strain JM109 and pro- pagated either in that host or in strain JMlOl (Yanisch-Perron et al., 1985).

(b) DNA preparation and subcloning

Phage 1 and plasmid DNAs were isolated as described (Nader et al., 1985). Ml3 ss DNA was prepared according to Messing (1983). DNA frag- ments were isolated from agarose gels by electro- elution onto NA45 DEAE membranes (Schleicher & Schuell, Dassel, F.R.G.) according to Young and Davis (1985) and from polyacrylamide gels with a procedure by Maxam and Gilbert (1980). DNA frag- ments, generated by Hi& or HhaI digestion, were blunt-end ligated into Ml3 after degrading their 5’ ss ends with T4 DNA polymerase (Maniatis et al., 1982).

(c) Nucleotide sequence analysis

Sequencing was performed by the chain-termi- nation method of Sanger et al. (1977) using 35S- labeled dATP and the modifications proposed by Biggin et al. (1983). Specifically NaCl was omitted from the reaction buffer, reaction temperatures were routinely increased to 55 “C, and dithiothreitol was replaced by 7 mM /?-mercaptoethanol. Radionucleo- tides were obtained from New England Nuclear (Dreieich, F.R.G.) and biochemicals and enzymes from Boehringer Mannheim (F.R.G.).

(d) Detection of single-stranded sites in supercoiled

DNA

The DNA was alkylated at ss sites and cleaved with S I-nuclease as described by Kohwi-Shigematsu et al. (1983) with the variation that chloroacetalde- hyde (Fluka, Buchs, Switzerland) was used (Kohwi- Shigematsu et al., 1986) instead of bromoacetalde- hyde (Kohwi-Shigematsu et al., 1978). After electro- phoresis in a 2% agarose gel, the DNA was electroblotted onto a cationized nylon-membrane (Zeta-ProbeTM, BioRad, Richmond, CA) and hybridized to a radioactively labeled probe, following the procedures recommended by the manufacturer.

The M13-derived probe was radioactively labeled by elongation from a hybridization-probe-primer, an- nealed to Ml3 (+) strand DNA containing the cloned exon V (Hu and Messing, 1982).

RESULTS

(a) Sequence analysis of the actin ad4 gene

In a previous report we described the isolation and some characteristics of ten recombinant phages har- boring actin-related sequences from a genomic library of Physarum in the vector/host system A

0 1

135

EMBL3IE. coli CES200 (recB21 recC22 sbcB 15) (Nader et al., 1985). All of those phages required the recBC - sbcB - host for growth and even then deletions of 300 to 400 bp occurred at one or more sites within the inserted DNAs. One of recombinant phages, APpAlO, contained a complete EcoRI fragment of the actin gene an&l, previously defined by Mendelian analysis of polymorphic restriction fragments (Schedl and Dove, 1982). A deletion of 360 bp in close proximity to the actin-related sequence of lPpAl0 was observed after plating on a ret + host. This new derivative had become stabilized in that host and was designated APpAlO’ (Nader et al., 1985).

The structure of the actin arat.4 locus was

2

I kb

6 4 35 5 7 25 83 1725 7 33 * 3 225 8

L -_ I II II ill I Illm I --

I b I I

-

A c TC CT

Fig. 1. Restriction map, sequencing strategy and structure of the arti gene locus. The restriction map was constructed by double

digestion (Lawn et al., 1978) of deleted and undeleted HindIII-EcoRI and the adjacent KpnI-Hind111 fragments, isolated from phages

lPpAl0 and IPpAlO’ or the pBR322 subclones. Restriction sites are labelled with numbers: 1 = AvaI; 2 = HueIII; 3 = HhaI;

4 = HindIII; 5 = Hid; 6 = KpnI; 7 = SacI; 8 = Sau3A. Various Ml3 subclones were constructed from smaller restriction fragments

and the arrows denote their position in the map and orientation on the (-) strand of phage M13. All Ml3 subclones were stable except

those where cloning of the complete 360-bp region, which had already caused problems with recombinant I phages and plasmids (Nader

et al., 1985), was attempted. In this case, only a few colourless plaques were detected but they were derived from phages with deleted

cloning sites and did not contain any ar& sequences. However, after this region was further subdivided into smaller fragments, mainly

by utilizing restriction sites for Sau3A in exon V, the resulting subclones could be stably maintained in both directions as recombinant

Ml3 phages (designated as a-f), even on the ret + host JMlOl. The Ml3 subclones harbouring deleted DNA from phage IPpAIO’ are

labelled as g and II and the two vertical lines, which interrupt those arrows, mark the approximate location of the deletion breakpoints.

The structure of the ardA gene is given on the bottom line. Open boxes indicate the six exons I to VI and the hatched boxes the location

ofthe homopolymers in the introns. The characters A, C, and T indicate poly(dA), poly(dC) and poly(dT) sequences on the coding strand.

136

5’

-97

-37

23

83

143

203

263

323

TCACGTAGCAATTCAGAATGACAAAGGATAAAAGTTGCCGTTTGTCTAGCCTTTTTAGTT

MetGluGlyGluAspValGlnA TCCAAAAGGACAAGCAACTAGCAAACATCACATCACATAAGCAATGGAAGGAGAAGACGTTC~G

EXON I laLeuVaLIleAspAsnGlySerGlyMetCysLysAlaGlyPheAlaGlyAspA CTTTGGTTATTGACAATGGCTCTGGCATGTGCAAGGCTGGCTGGATTT~AGGCGATGGTACTT -

INTRON I spAlaProArg

TTCTTTTGCTCTCAACATAATCTAATGTTTAATATTAATAATATATTAGATGCGCCCCGT -

AlaValPheProSerIleValGlyArgProArgHisThrGlyValMetValGlyMetGly GCAGTGTTCCCCTCTATCGTAGGACGTCCCCGCCACACTGGTGTGATGGTTGGTATGGGA

EXON II GlnLysAspSerTyrValGlyAspGluAlaGlnSerLysArgGlyIleLeu~rLeuLys CAGAAGGACTCCTACGTTGGTGATGAGGCTCAGTCCAAGCGTGGTATCCTCACCCTC~G

TyrProIleGluHisGlyIleValThrAsnTrpAspAspMetGluLysIleTrpHisHis TACCCCATTGAGCACGGCATTGTTACCAACTGGGATGATATGGAGAA~TCTG~ACCAC

ThrPheTyrAsnG ACCTTCTATAACGGTACATTTTTTGCTCTCTTGTTGTGAGTCCATTTTTTATACACATCA

383 TCTACATATTATTAAGCGAACGTCAGCCCTAAAAGCCACTCTTATTTTTTAAT INTRON II

443 TTTCTTGGTATAAAATTACGTTACAAAAATTTTGTCTACTTTAGCTACAACTTCACATCTC

503 TAAACAAAAACAATAATTAAAAAAAAAA AAAAAAAAAAATTGATGTTATGCTCGCAACTA

1uLeuArgVa 563 TATTGCTAGAACTTTGTGTTAATGGAATTTTGCTAATAATATTGTTCTTGCAGAGCTCCGTGT -

lAlaProGluGluHisProValLeuLeuThrGluAlaProLeuAsnProLysAlaAsnAr 623 TGCCCCCGAAGAGCACCCTGTTCTCTTGACCGAGGCTCCACTCAACCCCAAG~TAACAG

gGluLysMetThrGlnIleMetPheGlu~rPheAsn~rProAlaMetTyrValAlaIl 683 AGAGAAGATGACCCAAATCATGTTCGAGACCTTCAACACTCAT

eGlnAlaValLeuSerLeu~rAlaSerGlyArgThrThrGlyIleValMetAspSerGl 743 CCAGGCCGTGTTGTCCCTCTATGCCTCCGGACGCACCGGACGCACCACCGGTATTGTGATGGACTCTGG

E X 0 N III yAspGlyValSerHisThrValProIleTyrGluGlyTyrAlaLeuProHisAlaIleLe

803 TGATGGTGTCTCCCACACCGTGCCCATCTACGAAGGATAT~CCTCCCCCACGCCATCCT

uArgLeuAspLeuAlaGlyArgAspLeuThrAspTyrLerGluAr 863 CCGTCTCGACTTGGCTGGACGTGATCTAACTGACTACCTG~TGAAGATCCTCACT~GCG

gGlyTyrSerPheThrThrThrAlaGluArgGluArgGluIleValArgAspIleLys 923 CGGATACTCCTTCACCACCACCGCTGAGCGTGAAATCGTTCGCGACATCAAGGTGTGTCT -

INTRON III 983 CAGTTCTTTTCCCCCCCCCCCCCCCCCCCCTTTGCTTCCATTTGTTTCGCGTTTA~TT

GluLysLeuAlaTyrValAlaLeuAspPheGluGlnG 1043 GCAAGTCCTAATCATAATTACAGGAGAAGCTCGCCTACGTTGCCCTCGACTTCGAGCAGG -

Fig.2. Nucleotide sequence ofthe a&4 gene. The putative CAAT and Goldberg-Hogness boxes and the eight times repeated signal sequence for cleavage and polyadenylation ofthetranscribed RNA are markedby lines above and below the sequence. Underlined are

the exon-intron splice junctions, which define the 5 putative introns (I-V) and 6 exons (I-VI). The arrows A-E indicate the start

positions of the Ml3 subclones a to e, which were defined in the restriction map in Fig. 1 and were used for the sequencing reactions

1103 luMetGln~rAlaAlaSerSerSerAlaLeuGluLysSerTyrGluLeuProAspGlyG AGATGCAAACCGCTGCCTCTTCCTCCGCTCTCGAGAAGAGCTACGAGCTCCCAGACGGAC

1163

1223

lnValIleThrIleGlyAsnGluArgPheArgCysProGluAlaLeuPheGlnProSerP AGGTCATCACCATCGG4AACGAACGTTTCCGTTGCCCTGAGGCCCTCTTCCAACCCTCCT

EXON IV heLeuGlyMetGluSerAlaGlyIleHisGlu~rThrTyrAsnSerIleMetLysCysA TCTTGGGTATGGAGTCCGCTGGAATCCACCTAG

1283

1343

1403

1463

1523

1583

heProGlyIleAlaAspArgMetGlnLysGluLeuThrAlaLeu TCCCTGGTATTGCTGACCGTATGCAAAAGGAGCTCACTGCTCACTGCCCTTGTATGTTATTACTTTT -

TTCTTTTTCTCCTTTTTTTCTTTATCCTTCTCCCCTGTTTCTACAGTCTTACGTTTTCCT

GCTTCTTTTTGCGTTTTCCCATTTTTTACCCTACCCTTAAGAGTTTTTTTTTTTTTTTTT INTRON IV

TTTTTTTTTTTTTTTTTTTTTTTTTTTGCGCGGAGTGAATATCACTCTCACCTTCTCCAA

1643

1703

1763

. . . . . . . . . . . . . . . CACTCACCTCTTGCGCCGCGTGCCCCCCCCCCCCCCCCCCCCCTTTTTTTTTAATTCACTT

C AlaProSerThrMetLysIleLysIleIleAlaProProGlu

TTTCTAACCTTTCTCCAGGCACCCTCTACCATGAAGATCAAGATCATCGCACCTCCTGAG - EXON V

ArgLysTyrSerValTrpIleGlyGlySerIleLeuAlaSerLeuSerThrPheGlnGln CGTAAGTACTCTGTCTGGATCGGTGGCTCTATTCTTGCCTCCCTTTCTACCTTCC~CAA

AL MetTrpIleSerLysGluGluTyrAspGlu ATGTGGATCTCGARAGAGGAGTATGATGAGGTATCCTTAATTTAGCGTCTCCATCCCCCC -

1823 CCCCCCCCACCCACCGCTCCCCCCCCCCCCCCCCTTCTCTTTCCCCTTTTCTTTTCCCTT

B-WE INTRON V 1883 TTTTGCGCTTCACTTCCTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT

137

. . . . . . . . . . . . . . . EXON VI SerGlyProSerIleValHisArgLysCysPhe

1943 TTTTTTTTTTTTTTTGACTACCCTAAGAGCGGACCTTCCATCGTCCATCGCAAGTGTTTC -

End WD 2003 TAAAAAAGGAAAGCCCCTTTCCTTTTTCGTACCGAGCCCGGCCACACTAGGATGCCCCTT

2063 TTCCTAATTTTCATTTTTGGCTTCTCCATTTTTGTACTTTCGCATCCCTCTAATTGGCCT

2123 GGATAAGATAGAGTATCTTTTTCCTTGTTTGGTTAGTTGGAGCTGT~TAGATTTAAAA

2183 TTTTTCAAAAGATATTTTTTAATTTACTTGCCGACTCACTAAGCCACCAT~C~CACAT

2243 TTTGAATAAGCAAAAAAATAAAAATAAAAATAAATAAAAATAAAAATAAAAATAAAATAT -

2303 TAAAATGGGATACCTTTGTAGCTCTGTTGCCTCTTAGTTTTTCA~GTTACGCGTT~TC

2363 TTTACAACGATC- 3’

shown in Fig. 3. The homopolymeric regions in and around the unstable region are underlined. As discussed in the text and also shown in Fig. 3, the accurate length of these homopolymers in introns IV and V, and the long poly(dA) . (dT) stretch of IPpAlO’ generated following deletion of exon V and parts ofits flanking regions, has not been determined. The approximate deletion breakpoints of LPpAlO’ are marked by dotted lines which also indicate the remaining uncertainty of the exact size of introns IV and V.

138

elucidated by restriction mapping and subcloning the EcoRI-Hind111 fragments of lPpAl0 and LPpAlO’, which carried the actin-related sequence, and the adjacent KpnI-Hind111 fragment, which contained the promoter region of ani4 (Fig. 1). The sequencing strategy which was necessary to obtain stable Ml3 subclones (Fig. 1, a-e), spanning the 360-bp region already known to be unstable in recombinant 1 phages and plasmids (marked by the two vertical lines in Fig. 1) is detailed in the legend to Fig. 1.

The nt sequences derived from Ml3 subclones (Fig. 2) confirmed the previously determined aa se- quence of Physarum actin (Vanderkerckhove and Weber, 1978). The coding sequence is interrupted by five intervening sequences (Figs. 1 and 2) which are located within codons 26 and 94 and between codons 2141215, 3211322 and 3651366. All introns start with a 5’-GT and end with a 3’-AG and the presumed 5’ and 3’ splice junctions of ani4 are similar to the respective eukaryotic consensus sequences (Breathnach and Chambon, 198 1).

The ard4 gene includes sequences in its 5’- and 3’-flanking regions, which can be interpreted as signals for transcription initiation and poly- adenylation of the transcribed RNA. Starting at nt position -72 from the AUG start codon we found

a putative Goldberg-Hogness box, 5’-GATAA- AA-3’, which resembles similar structures from vertebrate p-globin genes (Efstratiadis et al., 1980) and from an Acanthamoeba actin gene (Nellen and Gallwitz, 1982). 59 nt upstream from this position we further located a potential CAAT-box (5’-CCAATCT-3’), which matches the 9-bp eukaryotic consensus sequence (5’-GGCCAATCT-3’) at seven positions (Breathnach and Chambon, 1981). At the 3’ end of adA, a stretch of 55 (dA) 1 (dT)s starts from position 2254. This sequence can be interpreted as an eight times repeated 5’-AATAAA-3’ signal for cleavage and polyadenylation of mRNA (Proudfoot and Brownlee, 1976).

The nt sequence further reveals that the 360-bp element, responsible for the cloning problem and previously localized by restriction mapping in close proximity to the actin-related sequence (Nader et al., 1985), is indeed part of the ad4 gene (Figs. 1 and 2). The unstable region of ad4 includes exon V and is flanked by a series of five homopolymers. Intron IV contains a stretch of poly(dA). (dT) followed by

poly(dG) . (dC), whereas intron V contains two regions of poly(dG) * (dC) followed by a long stretch of poly(dA) - (dT). It is also apparent that a large segment of intron V (between nt positions 18 10 and 1957 of Fig. 2) represents a very long homoPu- homoPy stretch, which is irregular in only six po- sitions. Sequence analysis of that region from the deleted phage APpAlO’ (Ml3 subclones g and h in Fig. 1) indicates a continuous stretch of 80 to 90 (dA) * (dT)s, which presumably arose by deletion of the intervening region between nt position 1550 and 1900 and portions of the flanking (dA). (dT) homopolymers. The approximate location of the deletion breakpoints is indicated by the dotted lines in Fig. 2.

Introns II and III each contain one homopolymer (labelled A and C in Fig. 1) which, like the poly(dA) . (dT) of IPpAlO’, does not seem to cause instability of recombinant DNA during cloning.

(b) Sequencing through homopolymers

Homopolymers cause problems for dideoxy- sequencing, as demonstrated in Fig. 3. Under stan- dard reaction conditions nonspecific termination was frequently observed in all of the homopolymers. However, when the sequencing reactions were incubated at 55 ‘C in a low ionic strength buffer and in the presence of /%mercaptoethanol, the stacking forces, particularly those of poly(dG) sequences, were at least partially overcome. Whereas primer- extension by PolIk proceeded quite well along a short poly(dC)-stretch on the template strand (Fig. 3, lane b; Figs. 1 and 2, arrows b/B), non- specific termination had already occurred after the fifth base in the (dG)-homopolymer of the same sequence subcloned in the opposite direction (Fig. 3,

lane 1; Figs. 1 and 2, arrows a/A). Terminations were less frequently observed in a longer poly(dC)- stretch (Fig. 3, lane c; Figs. 1 and 2, arrow c/C), or within the first 30 nt of a poly(dT) (Fig. 3, lane d; Figs. 1 and 2, arrows d/D) or a long poly(dA)-stretch (Fig. 3, lane e; Figs. 1 and 2, arrows e/E) on the template strand. Because of the remaining difficulties during dideoxy-sequencing of long homopolymers, the lengths of poly(dA). (dT) in nt positions 1506-1549 of intron IV and 1900-1957 of intron V and of poly(dG) * (dC) in nt position 1605-1624 of intron IV are estimates within 20 nt based on the

139

a e

ACGTACGTACGTACGTACGT

-

- -- S%_ _.* - -

111 “-- .” d

--

-

-

Fig. 3. Sequencing through homopolymers. Single-stranded template DNA was prepared from Ml3 subclones (Fig. 1, a-e and Fig. 2,

A-E) and contained poly(dG) (lane a), poly(dC) (lanes b and c), poly(dT) (lane d) and poly(dA) (lane e) stretches. Sequencing reactions

were performed at 55°C as described by Biggin et al. (1983), denatured with formamide and analyzed on a 6% polyacrylamide 8 M urea

sequencing gel.

140

analysis of sequencing gels and by size determination of the respective double-stranded restriction frag- ments after polyacrylamide gel electrophoresis.

The HaeIII-Sau3A fragments within introns IV and V (Fig. 1, a and c) contained (dA) . (dT)-homo- polymers followed closely by poly(dG). (dC). The intervening segment could only be sequenced after further reducing fragments a and c in size by HhaI digestion and subcloning into M13, thereby separat- ing the homopolymeric regions.

(c) Detection of single-stranded sites in the su-

percoiled a&4 DNA

The problems encountered in sequencing through homopolymeric regions suggested to us that stacking

+ + +

10 5 1 10 5 1

forces, particularly among purines, might alter the DNA conformation and lead to ss sites. It had been shown that the specific alkylation of unpaired bases with chloro- or bromoacetaldehyde and subsequent cleavage with Sl-nuclease allows the detection of ss sites in a supercoiled DNA molecule (Khowi- Shigematsu et al., 1983; 1986). Therefore, plasmid pBR322 carrying the cloned ardA gene was treated with chloroacetaldehyde and, after electrophoresis, blotted onto a cationized nylon membrane. The blotted DNA was hybridized with a Ml3 probe containing 96 bp from exon V (clone fin Fig. 1). One major band of 780 bp could be detected beside nicked-circular and linear plasmid DNA on the corresponding autoradiogram (Fig. 4). Cleavage of alkylated plasmid DNA with restriction endonu-

0.25 M CHLOROACETALOEHYDE

UNWyg DNA Sl NUCLEASE

2 600

1 600

780

360 300

Fig. 4. Single-stranded sites in supercoiled ardA DNA. Plasmid pBR322 carrying the cloned a44 gene was alkylated with 0.25 M

chloroacetaldehyde, cleaved with various amounts of Sl nuclease and, after electrophoresis in a 2% agarose gel and blotting to a

cationized nylon-membrane, hybridized to a 96bp fragment from exon V (Ml3 clone f from Fig. 1). As a control, plasmid DNA was

S I-nuclease-treated and blotted without prior alkylation. 200 ng of DNA were loaded per lane. The fragment sizes (in bp) are specified

on the right margin.

141

clease AvaI before Sl-nuclease treatment reduced the size of the 780-bp fragment in Fig. 4 by approx. 100 bp (not shown), thus indicating that the two ss sites, which generated that fragment, had to be located in intron III and either intron V or exon V. The minor 300-bp and 360-bp fragments can be explained by ss sites in the unstable region, com- prising exon V and its flanking introns. The minor bands at 1.6 and 2.6 kb suggest other Sl nuclease sensitive sites in a 600-bp region downstream from the ardA gene, which has not yet been sequenced, and in the three small inverted repeats in the replica- tion origin of pBR322, which were already found to be S l-hypersensitive by Lilley (1982). S 1-nuclease cleavage of the recombinant ara!A plasmid without prior alkylation only yielded linear and nicked- circular DNA (Fig. 4), indicating that all ss sites were induced by torsional stress.

DISCUSSION

(a) Functional implications of the ardA structure

A pseudogene for histone H4, which is interrupted by a putative transposon (Wilhelm and Wilhelm, 1984) and the a&i locus are the only genes from Physarum that have been completely sequenced. Because the Myxomycetes are rather isolated and primitive in the evolutionary context of eukaryotic organisms, a comparison of their actin genes with those of other organisms is of interest. None of the actin genes sequenced so far contain introns at the same positions as an&l (for review see Buckingham, 1985). However, after the recent isolation of the ardC locus of Physarum by Monteiro and Cox (1986), it is now known from preliminary sequence data (Gon- zales-y-Merchand et al., 1986) that at least three introns are located at the same positions as in a&,4,

although they differ considerably in size and sequence. In contrast to other primitive eukaryotes, like yeast (Gallwitz and Sures, 1980), Acanthamoeba (Nellen and Gallwitz, 1982) or Dictyostelium (Ro- mans and Firtel, 1985), the actin coding region in Physarum is interrupted by more than one intron. According to one hypothesis (Gilbert et al., 1986), numerous introns, and intron/exon junctions that break codons (as in introns I and II of arti), may be

indicative of the ancestral structure of a gene. It was also argued that only genes in slowly replicating cells of complex organisms still retain their numerous introns because of a lack of pressure toward rapid DNA synthesis. However, the members of the actin gene family of Dictyostelium have no introns (Romans and Firtel, 1985) and at least the actin genes a&A and ardC of Physarum contain multiple introns, although the cell cycle times of both organisms is quite similar, about 6-8 h. The struc- ture of ara!A may indicate that Physantm holds a critical intermediary position among eukaryotes. It was previously argued that Physarum is a multipotent developmental system that is no longer as simple as a single cell but not as complex as multicellular organisms (Sauer, 1982). Therefore, the analysis of the actin genes of Physarum may shed some light on the possible role of introns in relation to the timing of gene replication and their developmentally con- trolled expression in the distinct life cycle stages of Myxomycetes and different cell types of more com- plex organisms.

With their different DNA conformation (Arnott et al., 1983) and their ability to form single-stranded sites under torsional stress, homopolymers are potential regulatory sites for gene expression and chromatin assembly. The correlation between ex- pression of the chicken B-globin gene and ss sites in a 20-bp (dG). (dC)-homopolymer in its promoter region has been studied extensively (Weintraub, 1983; Emerson et al., 1985). Furthermore it was reported that nucleosomes will not form over long segments of homopolymers in a recombinant DNA molecule (Kunkel and Martinson, 1981). The long homopurine-homopyrimidine stretch observed around exon V of am!,4 might therefore form a nucleosome-free zone in the Physarum chromatin in vivo and thereby serve to phase nucleosomes along the actin gene. With the complete nucleotide sequence of the ara!A gene at hand, we will be able to investigate ss sites in that gene in isolated nuclei from different stages of the life and cell cycle of the organism (Sauer, 1982), following chemical moditi- cation of DNA in situ (Kohwi-Shigematsu et al., 1983; Benezra et al., 1986). Most interesting in this context might be the late replication of an&4 half way through the 3 h S-phase, whereas ardB, ardC and ardD have already completed replication within the first half hour (Pierron et al., 1984).

142

(b) Effects of homopolymeric regions on the viability

of recombinant DNA molecules in bacterial hosts

We have described a eukaryotic DNA sequence, which, like artificially constructed palindromic DNA (Collins et al., 1982; Leach and Stahl, 1983), causes inviability of phages and recA-independent de- letions. Although similar instabilities were observed with other eukaryotic DNAs (Wyman et al., 1985; Petri and Wyman, 1985; Murray et al., 1984), this is the first sequence analysis of an element, which causes such problems. In the case of the Physurum

actin gene ara!A, cloning in standard hosts seems to be prevented by a series of directly repeated (dA) . (dT)- and (dG) * (dC)-homopolymers, because disruption of this arrangement by cutting exon V with restriction endonuclease Suu3A allowed stable subcloning into bacteriophage Ml3 in all common E. coli hosts.

Nonspecific termination in the sequencing reaction, caused by stacking forces between the bases in a homopolymer on the template strand, was partially overcome by a high temperature and a low ionic strength in the reaction mixture. Preliminary results from Southern blot analysis further indicate that (dA) * (dT)- and (dG) . (dC)-homopolymers are widely distributed within the Physarum genome, may cause problems for cloning of other Physarum genes (for review see Nader, 1986) and might be a part of introns from other genes as well (W.F.N., unpub- lished results). Whether similar homopolymeric arrangements are the cause of the instabilities observed in other organisms (Wyman et al., 1985; Petri and Wyman, 1985; Murray et al., 1984) remains to be tested.

In our previous report we presented distinct secondary structures in electron micrographs of ss and self-annealed ar&t DNA in the unstable region (Nader et al., 1985). We now know that these structures were not correlated with the instability phenomenon, but were due to intramolecular annealing between two directly repeated (dA) . (dT)- homopolymers in nt positions 1505 and 1559 of intron IV and 1899-1957 of intron V and their ‘inverted repeat’, a nearly perfect (dA) . (dT)-homo- polymer in nt position 2254-2308, containing the multiple polyadenylation sites, which yielded the observed stem-loop and hairpin structures respec- tively.

Beside the more technical aspect of gene cloning, the structure of ardA might lead to a further under- standing of E. coli’s recombination system. It was speculated that DNA palindromes might form cruci- form structures in supercoiled phage or plasmid DNA, which resemble intermediates of DNA recom- bination and are thus cleaved by the bacterial recom- bination exonucleases I and V (Leach and Stahl, 1983). The fact that directly repeated homopolymers cause the same effect, may suggest a different mecha- nism. It is also unlikely that crossing over between the two poly(dA) . (dT) stretches, flanking the un- stable 360-bp element from Physarum, can explain its deletion because it occurs in recA - hosts as well. Therefore, we propose that ss sites in both structures might explain their similar behavior in E. coli. Such sites can be induced in palindromes (Lilley, 1982) as well as in homopolymers (Weintraub, 1983) by torsional stress. They also can be created by termi- nation of DNA replication, shown in vitro in in- verted repeats (Weaver and DePamphilis, 1984) and on a poly(dG) template (Fig. 3, lane a).

Partial unwinding of duplex DNA under torsional stress was shown to occur mainly in the unstable 360-bp region in a supercoiled plasmid carrying the ardA gene (Fig. 4). A plausible explanation for the instability phenomenon is the altered DNA confor- mation in homopolymers (Arnott et al., 1983), as was discussed by Kohwi-Shigematsu and Kohwi (1985) for (dG) * (dC)-homopolymers. To what degree DNA replication is terminated within homo- polymers in vivo, where the ss binding protein will reduce stacking forces remains to be tested (for review see Kowalczykowski et al., 1981).

Both types of potentially ss homopolymeric regions in vivo in recombinant DNA carrying a&t might serve as target sites for the ss-specific endo- nuclease activity of the recBC enzyme, the key en- zyme not only for homologous DNA recombination, but also for the E. coli recombination repair system (for review see Smith, 1983). This enzyme might cut the 360-bp sequence at its symmetrically arranged homopolymer stretches. Further degradation of the flanking poly(dA) * (dT) regions might occur by exo- nuclease I, a normally ss-specific nuclease, which was previously shown to hydrolyze poly(dA) - (dT) duplex DNA (Lehman and Nussbaum, 1964) as well. As a result supercoiled recombinant phage DNA, such as lPpAl0 and plasmids derived from

143

it, may become deleted and degraded in E. coli.

However, when a deletion of the unstable region was established in a recombinant phage molecule, such as APpAlO’, it could be maintained on recB + recC +

sbcB’ E. coli hosts. It seems that single homopolymers like po-

ly(dC). (dG) in intron III, the very large po- ly(dA) * (dT) sequence in phage APpAlO’, or (dG) . (dC)-stretches which are generated with the (dG) . (dC)-tailing method (Villa-Komaroff et al., 1978), do not cause apparent instabilities, although they may induce ss sites. Therefore, we suggest that a particular arrangement of several homopolymers, like the one observed in the 360-bp element in an&t, is necessary to cause the observed phenomena.

ACKNOWLEDGEMENTS

This work is dedicated to the late Prof. Charles E. (“Ned”) Holt. We thank Dr. G.L. Shipley for critical reading of the manuscript, Dr. A. HUttermann, Gottingen for support and T. Llanes for typing the manuscript. We are further indebted to Dr. T. Kohwi-Shigematsu for communicating an im- proved method for detecting single-stranded DNA prior to publication. This work was supported, in part, by grant HU 141/13 from the German Re- search Council to A. HUttermann and by National Science Foundation Grant PCM-84 11124 to H.W.S.

REFERENCES

Arnott, S., Chandrasekaran, R., Hall, I.H. and Puigjaner, L.C.:

Heteronomous DNA. Nucl. Acids Res. 11(1983) 4141-4155.

Benezra, R., Cantor, C.R. and Axel, R.: Nucleosomes are phased

along the mouse B-major globin gene in erythroid and

nonerythroid cells. Cell 44 (1986) 697-704.

Biggin, M.D., Gibson, T.J. and Hong, G.F.: Buffer gradient gels

and 35S label as an aid to rapid DNA sequence determi-

nation. Proc. Natl. Acad. Sci. USA 80 (1983) 3963-3967.

Breathnach, R. and Chambon, P.: Organization and expression

of eukaryotic split genes coding for proteins. Annu. Rev.

Biochem. 50 (1981) 349-383.

Buckingham, M.E.: Actin and myosin multigene families: Their

expression during the formation of skeletal muscle. Essays

Biochem. 20 (1985) 77-109.

Collins, J., Volckaert, G. and Nevers, P.: Precise and nearly

precise excision of the symmetrical inverted repeats of TnS:

common features of recA independent deletion events. Gene

19 (1982) 139-146.

Efstratiadis, A., Posakony, J.W., Maniatis, T., Lawn, R.M.,

O’Connel, C., Spritz, R.A., de Riel, J.K., Forget, B.G.,

Weissmann, S.M., Slightom, J.L., Blechel, A., Smithies, O.,

Baralle, F.E., Shoulders, C.C. and Proudfoot, N.J.: The struc-

ture and evolution of the &globin gene family. Cell 2 1 (1980)

653-668.

Emerson, B.M., Lewis, C.D. and Felsenfeld, G.: Interaction of

specific nuclear factors with nuclease-hypersensitive region

of the chicken adult b-globin gene: nature of the binding

domain. Cell 41 (1985) 21-30.

Gallwitz, D. and Sures, I.: Structure of a split yeast gene:

complete nucleotide sequence of the actin gene in Saccharo-

myces cerevirriae. Proc. Natl. Acad. Sci. USA 77 (1980)

2546-2550.

Gilbert, W., Marchionni, M. and McKnight, G.: On the antiquity

of introns. Cell 46 (1986) 151-153.

Gonzales-y-Merchand, J.A., Monteiro, M.J. and Cox, R.A.: 7th

European Physarum Workshop, Canterbury, U.K., 1986.

Kohwi-Shigematsu, T., Enomoto, T., Yomada, M.-A., Nakani-

shi, M. and Tsuboi, M.: Exposure of DNA bases induced by

the interaction of DNA and calf thymus DNA helix-

destabilizing protein. Proc. Natl. Acad. Sci. USA 75 (1978)

4689-4693.

Kohwi-Shigematsu, T., Gelinas, R. and Weintraub, H.: De-

tection of an altered DNA conformation at specific sites in

chromatin and supercoiled DNA. Proc. Natl. Acad. Sci. USA

80 (1983) 4389-4393.

Kohwi-Shigematsu, T. and Kohwi, Y.: Poly(dG)-poly(dC) se-

quences, under torsional stress, induce an altered DNA con-

formation upon neighboring sequences. Cell 43 (1985)

199-206.

Kohwi-Shigematsu, T., Manes, T. and Kohwi, Y.: Z-DNA in-

duces neighboring sequences to adopt an unpaired confor-

mation. Proc. Natl. Acad. Sci. USA (1986) in press.

Kowalczykowski, S.C., Beer, D.G. and von Hippel, P.H.: Single-

stranded DNA binding proteins. In Boyer, P.D. (Ed.), The

Enzymes, Vol. 14. Academic Press, New York, 1981,

pp. 374-444.

Kunkel, G.R. and Martinson, H.G.: Nucleosomes will not form

on double-stranded RNA or over poly(dA)-poly(dT) tracts in

recombinant DNA. Nucl. Acids Res. 9 (1981) 6869-6888.

Lawn, R.M., Fritsch, E.F., Parker, R.C., Blake, G. and Maniatis,

T.: The isolation and characterization oflinked 6 and /$globin

genes from a cloned library of human DNA. Cell 15 (1978)

1157-1174.

Leach, D.R.F. and Stahl, F.W.: Viability of I phages carrying a

perfect palindrome in absence of recombination nucleases.

Nature 305 (1983) 448-451.

Lehman, I.R. and Nussbaum, A.L.: The deoxyribonucleases of

Escherichia coli: V. On the specificity of exonucleases I (phos-

phodiesterase). J. Biol. Chem. 239 (1964) 2628-2636.

Lilley, D.M.: Dynamic, sequence-dependent DNA structure as

exemplified by cruciform extrusion from inverted repeats in

negatively supercoiled DNA. Cold Spring Harbor Symp.

Quant. Biol. 47 (1982) 101-l 12.

144

Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning.

A Laboratory Manual. Cold Spring Harbor Laboratory, Cold

Spring Harbor, NY, 1982, p. 395.

Maxam, A.M. and Gilbert, W.: Sequencing end-labelled DNA

with base-specific chemical cleavages. Methods Enzymol. 65

(1980) 499-560.

Messing, J.: New Ml3 vectors for cloning. Methods Enzymol.

101 (1983) 20-78.

Monteiro, M.M. and Cox, R.A.: Characterization of a&B and

ardC actin gene loci of Physarum polycephalum. Mol. Gen.

Genet. (1986) in press.

Murray, M.G., Kennard, W.C., Drong, R.F. and Slightom, J.L.:

Use of a recombination-deficient phage lambda system to

construct wheat genomic libraries. Gene 30 (1984) 237-240.

Nader, W.F.: Genomic cloning and construction of genomic

libraries in Physarum. In Dove, W.F. (Ed.), The Molecular

Biology ofPhysarumpolycephalum. Plenum, New York, 1986,

pp. 291-300.

Nader, W.F., Edlind, T.D., Htlttermann, A. and Sauer, H.W.:

Cloning of Physarum actin sequences in an exonuclease-

deficient bacterial host. Proc. Natl. Acad. Sci. USA 82 (1985)

2698-2702.

Nellen, W. and Gallwitz, D.: Actin genes and actin messenger

RNA in Acanthamoeba castellanii. J. Mol. Biol. 159 (1982)

l-18.

Norrander, J., Kempe, T. and Messing, J.: Construction of

improved Ml3 vectors using oligodeoxynucleotide-directed

mutagenesis. Gene 26 (1983) 101-106.

Petri, W.H. and Wyman, A.R.: A significant fraction of the

Drosophila genome cannot be cloned in ret+ E. coli hosts.

Genetics 110 (1985) ~59 (abstr.).

Pierron, G., Durica, D.S. and Sauer, H.W.: Invariant temporal

order of replication of the four actin gene loci during the

naturally synchronous mitotic cycles of Physarum poly-

cephalum. Proc. Natl. Acad. Sci. USA 81 (1984) 6393-6397.

Proudfoot, N.J. and Brownlee, G.G.: 3’-Noncoding region se-

quences in eukaryotic messenger RNA. Nature 263 (1976)

21 I-214.

Romans, P. and Firtel, R.A.: Organisation of the actin multigene

family of Dictyostelium discoideum and analysis of variability

in the protein coding region. J. Mol. Biol. 186 (1985) 321-335.

Sanger, F., Nicklen, S. and Coulsen, A.R.: DNA sequencing with

chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74

(1977) 5463-5467.

Sauer, H.W.: Developmental Biology of Physarum. Cambridge

University Press, Cambridge, 1982.

Schedl, T. and Dove, W.F.: Mendelian analysis of the

organisation of actin sequences in Physarum polycephalum. J.

Mol. Biol. 160 (1982) 41-57.

Smith, G.R.: General Recombination. In Hendrix, R.W.,

Roberts, J.W., Stahl, F.W. and Weisberg, R.A. (Eds.),

Lambda II. Cold Spring Harbor Laboratory, Cold Spring

Harbor, NY, 1983, pp. 175-209.

Vanderkerckhove, J. and Weber, K.: The amino acid sequence

of Physarum actin. Nature 276 (1978) 720-72 1.

Villa-Komaroff, L., Efstratiadis, A., Broome, S., Lomedico, P.,

Tizard, R., Naker, S.P., Chick, W.L. and Gilbert, W.: A

bacterial clone synthesizing proinsulin. Proc. Nat]. Acad. Sci.

USA 75 (1978) 3727-3731.

Weaver, D.T. and DePamphilis, M.L.: The role of palindromic

and nonpalindromic sequences in arresting DNA synthesis in

vitro and in vivo. J. Mol. Biol. 180 (1984) 961-986.

Weintraub, H.: A dominant role for DNA secondary structure in

forming hypersensitive structures in chromatin. Cell 32

(1983) 1191-1203.

Wilhelm, M.L. and Wilhelm, F.-X.: A transposon-like DNA

fragment interrupts Physarumpolycephalum histone H4 gene.

FEBS Lett. 168 (1984) 249-254.

Wyman, A.R., Wolfe, L.B. and Botstein, D.: Propagation ofsome

human DNA sequences in bacteriophage 1 vectors requires

mutant Escherichia coli hosts. Proc. Natl. Acad. Sci. USA 82

(1985) 2880-2884.

Yanish-Perron, C., Vieira, J. and Messing, J.: Improved Ml3

phage cloning vectors and host strains: nucleotide sequence

of the Ml3mpl8 and pUCl9 vectors. Gene 33 (1985)

103-119.

Young, R.A. and Davis, R.W.: Immunoscreening lgtl 1 recombi-

nant DNA expression libraries. In Setlow, J.K. and

Hollaender, A. (Eds.), Genetic Engineering, Plenum, New

York, 1985, pp. 29-41.

Communicated by S.R. Kushner.