Recombination between adenovirus type 12 DNA and a hamster preinsertion sequence in a cell-free system

.J. Mol. Biol. (1992) 226, 117-126

Recombination Between Adenovirus Type 12 DNA and a Hamster Preinsertion Sequence in a Cell-free System Patch Homologies and Fractionation

of Nuclear Extracts

J&g Tatzelt, Beate Scholz, Katja Fechteler, Rolf Jessbergert and Walter Doerflerl Institute

for Genetics, r,Tniversity

(Received 13 November

of Cologne, Cologne, Germany

1991; accepted 12 February

1992)

We have previously described a cell-free recombination system derived from hamster cell nuclear extracts in which the in vitro recombinat’ion between a hamster preinsertion sequence, the cloned 1768 base-pair p7 fragment, and adenovirus type 12 (Ad12) DNA has been demonstrated. The nuclear extracts have now been subfractionated by gel filtration on a Sephacryl S-300 column. The activity promoting cell-free recombination elutes from the Sephacryl S-300 matrix with the shoulder and not the peak fractions of the absorbancy profile. By using these protein subfractions, in vitro recombinants have been generated between the p7 preinsertion sequence and the 60 to 70 map unit fragment of Ad12 DNA, which has previously shown high recombination frequency. In all of the analyzed re combinants thus produced in vitro, striking patchy homologies have been observed between the p7 and Ad12 junction sequences, and between Ad12 DNA or p7 DNA and pBR322 DNA. The patchy homologies are similar to those found earlier during the analyses of some of t’he junction sequences in integrated Ad12 genomes in Adl2-induced hamster tumor cell lines. Proteins in the shoulder fractions of the gel-filtration experiment can form specific complexes with double-stranded synthetic oligodeoxyribonucleotides corresponding to several p7 and Ad12 DNA sequences. These sequences participate in the recombination reactions catalyzed by the same column fractions in the shoulder of the absorbancy profile. Such proteins have not been found in the peak fractions. Further work will be required to ascertain that the cell-free recombination system mimics certain elements of t’he mechanisms of integrative recombination and to purify the cellular components essential for recombination. Keywords: Cell-free recombination; insertional recombination; integration t’ype 12 D?jL4: fractionation of cell-free recombination system: patch integration

1. Introduction

mouse, and human cells (for a review. see Doerfler et al., 1983). There is evidknce that adenovirus DNA integrates preferentially into transcriptionally active sequences of the mammalian genome (Gahlmann et al., 1984; Schulz et al., 1987). A cell-free system from nuclear extracts of hamster cells has been developed to study the mechanism of integrative recombination of cellular DNA with foreign adenovirus type 12 (Ad129) DNA (Doerfler et al.. 1987; Jessberger et al., 1989). In previous work; we have found that the 60 to 7O’map

The mechanism of integrative recombination of foreign DNA into mammalian DNA is only partly understood. Foreign DNA can recombine with the host DNA in mammalian cells and can thus become integrated into the cellular genome by what has been termed heterologous recombination. As a model, we have studied the integration of human adenovirus DNA into the genomes of hamster,

t Present address: Department

of Biochemistry.

Stanford University. Medical Center, Stanford. California 94305. 1T.S.A. 1 Author to whom all correspondence should be addressed at: Tnstitut fiir Genetik, Weyertal 121. I)-5000 Kiiln 41, Germany. 0022--2X36/92/130115-10

$03.00/O

of adenovirus homologies and

Q Abbreviations used: Ad12, adenovirus type 12; Hepes, N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid; DTT, dithiothreitol; PMSF, phenylmethylsulfonylfluoride; ampR. ampicillin resistant. 117

0 1992 Academic l’ress Limited

118

J. Tatzelt

unit fragment of the Ad12 genome recombines at an increased frequency with the hamster preinsertion sequence p7 (Stabel & Doerfler, 1982) or ~16 (Lichtenberg et al., 1987), both identified in Ad12induced tumor cells. However, recombinants are not formed at a comparable frequency with randomly selected hamster DNA sequences (Jessberger et al.. 1989). In the following discussion, p7 will designate the 1768 base-pair preinsert,ion hamster DNA sequence that has been employed in recombination experiments as a pBR322 clone (Jessberger et al., the 1768 base-pair preinsertion 1989). Within sequence ~7; two recombinants (p7-R5, p7-R6) have been found previously close to the nucleotidr sequence 5’-CCTCTCCG-3’ (Jessberger et aZ., 1989), and this sequence also occurs at the Ad12 integration site in the p7 sequence in the Adl2-induced hamster tumor line CLACl (Stabel & Doerfler. 1982). Extensive control experiments have made it unlikely that the described recombinants have been coli transgenerat,ed or modified in the Escherichia feetion system that has been used for their identification (Jessberger et al., 1989). Here, the subfractionation of the recombination system from hamster cell nuclear extracts will be described. Moreover. we have determined the nucleotide sequences at, the sites of junction of eight new in vitro-generated recombinants bet’ween the 60 to 70 map unit fragment of Ad12 DNA and the 117 hamster preinsertion sequence or pBR322 DNA. Tt is likely that patchy nucleot,ide sequence homologies play a role in the generation of these recombinants. In one of the new recombinants, as in three previously analyzed sites of recombination (CLAP1 1 p7-R5 and p7-R6), the 5’.CCTCTCCG-3’ nucleotide sequence motif has been found at, or very close to, the site of recombination. Lastly, subfractions of the hamster cell nuclear extracts. which are active can form in recombination. promoting D%A-protein complexes with double-stranded synthetic oligodeoxyribonucleotides that include and flank several of the recombinat,ion sites.

Materials and Methods (a) Preparation of NHK%l cell nuclear rxtracts (Shapiro et al., 1988) BHK21 cells adapted to suspension culture were propagated to a density of about 1 x lo5 to 5 x lo5 cells/ml. The cells were then pelleted at 170 g for 10 min at 4”C, washed once in phosphate-buffered saline (Dulbecco & Vogt. 1954), repelleted at 300 g for 10 min and resuspended in 5 packed cell vols of buffer A (lo-mM ,V-2-hydroxyethylpiperazine-IV’-2-ethanesulfonic acid (Hepes) (pH 7.9). 075 mi%-spermidine, 015 mM-spermine, @l miv-each of EDTA and EGTA. 1 mM-dithiothreitol (DTT). 10 mM-KC]). After a 10 min incubation on ice. the cells were pelleted and resuspended in 2 packed cell vols of buffer A. The cytoplasmic membranes were broken by 10 to 15 strokes in a Dounce homogenizer with a pestle 13. Subsequently. @I vol. of buffer B (50 mw-Hepes 0.75 mw-spermidine, @15 mM-spermine, (PH 7.9), 10 mm-Kcl, @2 mM-EDTA, 1 mm-DTT, 7.5% (w/v) sucrose) was added. The liberated nuclei were then

et al pelleted at 16,000 g for 30 s at CC’, resuspended in 3 ml (for an equivalent of JO9 ~11s) of buffer (’ (20 IIIM-Heprs (v/v) glycerol. 0.75 my spermidiIrr. (pH i.9). ZO’, Cl5 mM-spermine; 0.2 rnM-EDTA. 2 ~wEGTA. 2 m&lDTT, IO?;-saturated (NH,),SO,) and gently swirled for 30 min on ice. Bfter pelleting the nuclei at 150,OOOg for 90 min at, 2°C. t,he supernatant was dialyzed for !I0 min against 2 changes, each of greater than 200 ~01s. of buffel D (20 mM-Hepes (pH 7.9), 20% (x7/v) glycerol. lfiOm~-NaU. 0.2 rnM each of EDTA and EGTA, 2 mM-DTT, 1 rnM-phenylmethylsulfonylfluoride (PMSF)). The dialyzed fractions were frozen in liquid N2 and stored at -80°C for further use.

(b) Gel jiltration

on a Sephacryl S-300 column

.4 100 ml Sephacryl S-300 (Pharmacia) column of 50 cm length was equilibrat’ed at, 4°C with buffer D. A volume of 3 ml dialyzed nuclear extract in buffer D was loaded onto the column, which was eluted with buffer D in 1 ml fractions at a flow ratr of &Timl/min. The A,,, was measured in the eluate. and protein-containing fractions were assayed for recombination activity.

The 60 to 70 map unit fragment) of ,4dl2 lIS.4 or I); DNA was incubated in buffer lC (but%+ D plus IO mw-Tris’ HCI (pH 7.9). 15 mM-MgCl,. 1 m.n-%n(‘lz) with Jjrotein extract purified by gel filtration on a Sepha~ryl S300 rolumn (see Fig. I (a)). and DNA-protein complext~s were sedimented directly in a Heraeus Biofugr at 16.000 revsimin for 10 min. The pellet was resuspended in buff&r E and subsequently inrubat,ed wit)h the ot)her recom bin;ltion partner (p7 TINA or AdI2 DNA. resp~(‘tivc~l~) without adding further frac*tions containing protrirl.

Thr standard I,rc~o,nbination assay ( 100 1~1).c~onsistinp 20 mw-Heprs (pH 7.9). IO”,, (v:\-) ylyc~t~rol. 75 mw-Na(l1. 15 mM-MgCl,. 1 mM-Znc112. 0.1 tn>f t,wc,h of EDTA and E(:TA. 1 mnl-DTT, 0.5 mM-PMSF. 0.1 IIIXI each of the four dWl’s. 1 n1n1-ATP. 5 m&v-creat,ine phosphatr. I pg I’stI~~J) fragment (60 t,o 70 map units) of Ad 1-L IjB.4. 0.2 fig EcoRT-cut p’i DNA and 50 PI-protein f’raction (protein content about 0.3 ,ug/pl) to be trstcbd. ~vas inc*ubatrd for 30 min at 37 Y’. Subsrquclltly. tht, 1)iv.J was repurified from the rea&ion mixturcl using the SDSproteinase-K-phenol : chloroform (1 : I hy vol.) method (Sutter et al., 1978). precipitated using ethanol and resuspended in 10 mrcl-Tris’HCl (pH 7 5). 1 mM-EDTA (TE buffer). Recombinants were isolated uia transfection of the reext’racted DNA intIo t,he recAstrain HBlOl/LM1035 of E. coli and by hybridization (Grunstein & Hogness, 1975) to “‘P-labeled Ad12 DNA. as described earlier (Jessberger rt al.. 1989). of

Double-stranded plasmid (pBR322) l)XA c,arr,ving the recombinants was used with synthetics oligodeoxyribonucleotides as primers. which were produced using an Applied Biosvstems 381A DNA synthesizer. The dideoxy rhain termination method @anger et al., 1977) was applied.

V’iral-Cellular

(f) Determination

DNA Recombination in a Cell-free System

of DNA-protein complex formation the ye1 shift assay

by

Details of this method have been described (Hermann et a,Z., 1989). The 50 nucleotide fragment comprising nucleotide positions 225 to 274 in the p7 DNA sequence (underlined in Fig. 4) with the sequence 5’CACTCCACCGACGC GGCCTCTCCGCACGCTTGCACAAGCAGCAACCAGCT-3’ was prepared in a 381A DNA synthesizer. This sequence contains the p7 DNA site of recombination with Ad12 DNA in recombinants rec22 (this paper) and p7-R5, and p7-R’6 (Jessberger et al.. 1989). Nuclear extracts from RHK21 cells were fractionated by gel filtration on Sephacryl S-300 columns, and fractions 32 to 48 correspondmg to the trailing shoulder. which contained recombinational activities (shaded area in Fig. l(a)), were incubated with the synthetic double-stranded oligodeoxyribonucleotide (see Fig. l(b)). In a total volume of 20~1 of 10 mw-Hepes (pH 7.9), 80 mM-KC1, 1 mM-EDTA, 2 mM-DTT. 2 mM-MgCI,. and 1 pg unspecific doublestranded competitor poly(dI-dC) poly(dI-dC) per assay, about 10.000 cts/min, equivalent to about 140 pg of the

I+------I9090 0

119

32P-labeled specific oligodeoxyribonucleotide, and I pg protein from column fractions as indicated. were incubated at 20°C for 40 min. In some of the experiments, increasing amounts of the unlabeled oligodeoxyribonucleotides were added as specific competitors. After incubation. DNA-protein complex formation was monitored by electrophoresis on 4% (w/v) polyacrylamide gels in TEB buffer (89 mM-Tris.HCl, 89 mM-boric acid. 2 mu-EDTA) at 200 V for about 2 h. After electrophoresis. the gels were dried and autoradiographed. (g) Assay for he&case activity The displacement of a 17 base-pair oligodeoxyribonucleotide annealed to bacteriophage M13mp18 singlestranded DNA in a partially double-stranded structure was measured electrophoretically as a test of helicase activit,“v (Seo et al., 1991). The 17.mer universal primer 5’.GTAAAACGACGGCCAGT-3’ of Ml3 DNA was labeled at the 5’ end using polynucleotide kinase (Richardson, 1965) and [Y-~‘P]ATP, and was then annealed, by incubat’ion at 65°C for 20 min, to single-stranded M13mp18

-I 16650 20

No. ampR colonies Ad12 ONA-positive

tested colon~ts

Figure 1. Gel filtration of BHK21 cell nuclear extracts on a Sephacryl S-300 column. Technical details are described in the text. (a) The graph presents the A,,, nm elution profile. V, marks the void volume of the column. The fractions comprising the shaded area were active in generating recombinants between p7 DNA and the 60 to 70 map unit fragment of Ad12 DNA, whereas the main peak fraction did not mediate the formation of recombinants. Fractions 26 to 60 (every other fraction tested) were assessed for the presence of proteins that could form complexes with a double-stranded 50base-pair oligodeoxyribonucleotide comprising nucleotide positions 225 to 274 of the p7 nucleotide sequence (see Fig. 4) and for helicase activity (H), measured as O/osingle-stranded 17-mer per pg protein per 30 min. (b) Autoradiographs of band shift experiments. (c) Results of experiments in which supercoiled circular p7 DNA in the vector pBR322 (a to d) or EcoRI-linearized construct (a’ to d’) was incubated at 37°C for 30 min in buffer E alone (a, a’) or in buffer E with fraction 26 (b, b’) or in buffer E with fraction 38 (c, c’). As control, untreated DNAs were also coelectrophoresed (d, d’). In a to c and a’ to c’. the DNAs were reextracted using standard procedures. All DNA preparations were then electrophoresed on a 0.8% (w/v) agarose gel. A photograph of the ethidium bromide-stained gel is presented.

J. Tatzelt et al.

120

DNA in annealing buffer (50 miv-Tris. HCl (pH 7%), 800 mM-NaCl, 1 mM-EDTA). This mixture was then allowed to cool to room temperature _over a 1 h period. Subsequently, about 2000 cts/min (Cerenkov) of this partly double-stranded DKA preparation was incubated for 30 min at 37°C with protein fractions as indicated (see Fig. 1(a)) in a total volume of 20 ~1 of 50 m,n-Tris. HCl (pH 8.0), 7 mM-MgCl,, 5 mivr-DTT, 1 mM-ATP and 100 pg bovine serum albumin/ml. The partly double-stranded construct was purified from single-stranded 17-mer DNA and from [y-32P]ATP by gel filtration on Sepharose CTAB. Finally, the reaction was stopped by adjusting the solntion t,o (final concentrations) 10 mM-El)TA. Wl”& (w/v) SDS, 4% (v/v) glycerol. The reaction products were separated by electrophoresis at 120 V for 1 h on an X0,, (w/v) polyacrylamide gel in TEB buffer. Finally. the gel was dried and autoradiographed directly. Autoradiographically located bands were excised from the dried gel. The radioactivity in each band was yuantified by scintillation counting (Cerenkov) in a Beckman scintillation counter.

3. Results (a)

Fractionation

of

BHK%l

nmclear

extracts

The proteins in the BHKdl nucalear ext’ract*s were size-fractionated by gel filtrat’ion on a Sephacryl S-300 column as described in Materials and Met)hods. The protein fractions with activity in an in vitro recombination assay rluted in a shoulder after the main peak of absorbancy (A) at 280 nm (Fig. 1(a)). Out of a total of 16,650 ampicillinresistant (ampR) colonies tested, generated after incubating the EcoRI-linearized pBR322-cloned p7 DNA and the 60 to 70 map unit fragment of =Zd12 DNA with the pooled fractions 32 to 48 from the shoulder region, and after subsequent transfection of the reext’racted DNA into the rec*A- strain HBlOl/I,M1035 of E. coli, 20 colonies contained Ad12 DNA recombinants. Since we had demonstrated previously that the 60 t)o 70 map unit fragment of Ad12 DNA recombined in an in htro system derived from BHK21 hamster cell nuclear ext’ract,s at a, higher frequency with the pBR322-cloned hamster preinsertion sequence p7 than with other. randomly cloned, segments of this viral genome (Jessberger et al.. 1989). most experiments reported here were performed with these two reaction partners. The 60 to 70 map unit fragment of Ad12 DNA4 was used as the gel-purified PstI-D fragment that had been excised from the vector. The pBR322-cloned p7 sequence was linearized by cleavage with EcoRT in the plasmid DNA. In contrast, among 9090 colonies produced after the incubation of the peak fractions from the Kephacryl S-300 column with the recombination targets and after transfection into E. coli. no recombinants were observed. Similar results on in vitro recombination were obt’ained in four independent gel-filtration experiments. In three of these instances, the columns were developed with buffer E (see Materials and Methods) containing 150 rnM-NaCl; in one experiment this buffer was made up in 250 mM-NaCl. The trailing fractions 49 to 56 from the gel-filtration experiment described in Figure l(a)

were also assayed for recombination activity. In several tests evaluated so far, no recombinat,ion activity was found in these fractions. Control experiments, in which the p7 and Ad12 DNA recombination part,ners were jointSly transfected into the re& strain of E. roli without prior incubation with protein fractions, or after incubating t.he DNA partners separately with the prot’ein extracts, failed to yield recombinants. The results of a large number of similarly negative control experiment,s using unfraotionated nuclear extracts were reported previously (.Jessberger et nl.. 1989). These results provide the basis for the notion tha,t essential st.eps in the generation of recombinants proceeded during the in vitro incubation with nuclear ext.ra.cts. It is concluded that t,he recombina.tion-promoting fractions in the nuclear extract elute from a Sephacryl S-300 column after the main absorbanc) peak, possibly in a complex form that is stable to salt concrntrations of up to 250 mbf-Sac?]. All t hcb recaombinants described in this paper have bern rlic~it~t~tl hy using these caolurnn fractions (shadrd area in Fig. l(a)). It was c~onc~eivable that thr main peak in t.hr absorbancy profile (Fig. 1 (a)) caontained nuclfaasc ac+vitirs and that, for this reasoil. rec*ombinants were not tlet~ec~table after the inc~ubation of thth recombination partners with frac%ions from that peak. The, dat,a presented in Figurfl 1(P) dflmon st.ratc that there was no evidencta for rndonac~lco~ I\-tic activities on supercoiled or linearized I)SA in rlt)her the main peak or the shouldtar frac%ions, Tn preliminary tlxperimrnts. ii c~tuld br shown that incubation of p7 I)SA with t,hc. shoulder fra(ations from thca Sephacryl S-300 column (Fig. I (a)) led to the formation of a l)NvI1\- protein cbomplex that could bth srparatcad from pi l)?iX by \-c,locGty srdimentation in ii glycerol density gradifhnt (data not shown). The l)NA-protein c*ornplrs c~mltl also 1~ pfhlleted hy c~eritritugat,iorr at Iow “pt”“i. anti th incubat,ion of this pellet fraction with solely the 60 to 70 map unit fra,gmrnt of Ad12 l)N;\ generated the p7-Ad Id I)SA recombinant rec&! (SW below).

Thtb rectombinants were mapped t)J. rest,riction analyses. and the relative proportions of -4d12 DNA. i.(b. of subsegments of its 60 to 70 map unit fragment. and of KHKZl 1~7 I>?;A were derived from the nucleotide sequenctl determination. The charts in Figure 2 present the results of these mapping experiments for the rrcombinants recl0. recl8, rec22. rt:c27. rec9, rec20. recll and recl3. It was apparent that t,hese rrcombinants carried both Ad12 and p7 DNAs, while some of the pHR322 sequences were deleted. In rec*ombinants reclo. reel 8. recap:! and rec27. t’he p7 and Ad12 DNA fragments werta linked directly to racbh other. In recombinants rec9. rec20, reel 1 and recl3. short

Viral-Cellular

DNA Recombination in a Cell-free System

Figure 2. Maps of 8 different recombinant5 generated between the 60 to 70 map unit fragment of Ad12 DXA and the RHK21 preinsertion sequence p7 or pBR322 DKA. The recombination partners were incubat,ed in the Sepharryl S-300 subfractions of nuclear extracts as indicated by the shaded area in Fig. 1. In each chart the Ad 12 (H), p7 (8) and pBR322 (-) DKA sequences are designated. Details are explained in the text. At the sites of junctions (I), the numbers of the deleted nucleotides are indicated (a). The numbering of p7 nucleotides is that of the published sequence (Jessberger et al.. 1989), the numbering of pBR322 nucleotides is based on the published sequence (Watson. 1988). For the Ad12 sequence in the 60 to 70 map unit fragment, the entire sequence is not. yet, known. The open arrowhead in the black Ad12 sequence block designates the orientation (left to right) of this DNA segment.

pBR322 DNA sequences were interspersed between the two recombination partners, i.e. in the generation of these two recombinants the 60 to 70 map

121

unit fragment of Ad12 DNA was linked at either terminus to pBR322 DNA. In recombinants recl0, recl8, rec22, rec27, rec9, rec20 and recll, the Ad12 DNA segments had the same orientation (left to right arrowheads in the schemes in Fig. 2). In recl0, recl8, rec22 and rec27, the left ends of the Ad12 DNA segments had recombined wit’h p7 DNA, whereas the right termini were linked to pBR322 DNA. By definition, the left Ad12 terminus was the sequence closest to the 60-map-unit end, the right terminus the sequence closest to the 76-map-unit end of the PstI-D fragment. The p7 DNA sequences extending beyond the point of junction to Ad12 DNA were deleted in the process of recombination. The junction sequences at both the right and left termini of the inserted Ad12 DPL’A sequences were determined in the eight recombinants described. In Figure 3: the right and left junction sequences between p? and Ad12 DNA or between pBR322 and Ad12 DNA were reproduced for recombinants recl0, recl8, rec22. rec27, rec9 and rec20. In the schemes in Figure 3, the continuous sequences represent the actual junction sequences. The Ad12 DNA sequences are presented against, a black background, and the pBR322 (pBR) and p7 DNA sequences are designated as such. The data also include the original parts of the Ad12 DNA sequence that were deleted (black background) and parts of the deleted p7 DNA or of the pBR322 sequence. The designations of the nucleotide numbers correspond to the numbering in the published p7 (Jessberger et al., 1989) or pRR322 DSA sequence (Watson, 1988). As the Ad12 DNA sequence in the 60 to 70 map unit sequence of Ad12 D?U’A was only partly known (Kruijer et (zl., 1983), nucleotide numbers could not be assigned to the Ad12 sequences. However, the Ad112 DNA sequences occurring in the rerombinants had previously been determined in virion DNA (Kruijer et al.. 1983). The nucleotides marked by asterisks indicate sequence identities between the viral and cellular or plasmid DNA segments in the junction sequences and the deleted cellular, plasmid or viral segments, respectively. Among the plasmids analyzed, there was none containing only Ad12 DNA or p7 DNA sequences. Jn addition to the eight recombinants
(c) Evaluation

of the nucleotide sequences

junction

(i) Patchy

at the

sites in the recombinanta

homologies

One of the striking results in the sequence determinations on junction sequences of rrcombinants recl0, reel 8, rec22, rec27, rec9 and rec20 were patchy homologies in nucleotide sequences between p7 and Ad12 DNAs at the sites of recombination. These homologies were designated by asterisks in

J. Tatzelt et al.

122

right

left ,

recl0

trPBR

<

Ad12 Ad 12

recl8

PY

rec22

* l l ** CTTTCCTCTTCAAGAATT GGATGGCCTTCCCC~T;ATGA GCTGGCGTT$$CGACGCGAG * l l ** PBR

CAAGCAGCAACCAGCTCCT P7

PBR

rec27

ACCAAAAGGCAGCCCGAG D7

rec9

rec20

Ad12

PER

CTGCTCGCTTCG PBR

AAAAATAGG

PBR AAAAAIA

Ad12

BHK (~7)

Figure 3. Nucleotide sequences at the right and left junctions of p7 hamster DNA with ,4dl2 or with pBR322 DNA in the in vitro-produced recombinant5 recl0, recl8, rec22, rec27, rec9 and rec20. Ad12 DNA sequences are shaded. p7 DNA sequences and pBR322 DNA sequences are designated p7 and pBR, respectively. The nucleotide numbers are those of’ the published DNA sequences for p7 (Jessberger et al., 1989) and pBR322 (Watson, 1988). The Ad12 DNA sequences have not been numbered, since the nucleotide sequence of the 60 to 70 map unit fragment of Ad12 DNA is only partI) known (Kruijer et al.. 1983). The Ad12 DNA sequences in the recombinants are derived in part from unpublished da,ta. The left (60 map unit’) to right (70 map unit) orientation of the Ad12 sequence is designated by an arrowhead. The junction sequence between the left terminus of Ad12 virion DNA and hamster cell DNA in the Adl%-induced tumor c.cll line CLACl is also presented (Stabel & Doerfler, 1982). The nucleotides designated by asterisks represent regions of’ sequence identity between the react’ion partners. The Sephacryl S-300 protein fractions 32 to 46 (Fig. 1) were used f’or the recombination experiments.

the sequences in Figure 3. One of the junction sites, the right-terminal connection of the Ad12 DNA segment to p7 DNA in rec22, coincided with the site of insertion in the previously described in vitro

recombinants

p7-K5

et al., 1989), whereas rec22 was different isolated recombinants.

and p7-FM (Fig. 1 in ,Jesxberget the left junction sequence of from those of t,he previously

DNA Recombination in a Cell-free System

Viral-Cellular

123

EcoRI pBR322

BHKPl(p7)

’ 436v1

315/I

I

kA

A 20r

10&$;#

II

500

1

I

!!r

I,, A

p7-R5,p7-R6

‘ni,rst

I

L

I

+

CLAC 1

1000 I1 .

*+-pBR322 I

l& *:r

I

I

1500 #

I

I

1768 I

l&

A llr,13r

4381

576

ACGTCCTCTT

TGGCCCTGCC

TGTOCGWW

CCCW-

626

CTTTGGCCTT

GClTGGcCTT

GGCCTC-

GGGCTCTCAT

ACAGCCACTC TCTCACCCCA

67.5

GGCCGTCTCC

AACmTAGG

AMCACCACG

CCCTCGCCZT

AGlTCGGACC

726

ACCTCCTCCC

TCCCTTCC~

CTTTCCTTCC

TTCC~CAT

CCTfCCTTcC

776

TTCCTTCCTT

CCTTCCTTCC

TcCClTCCTT

CCTTCCTTCC

TTccTTccGT

826

CTTWTTXT

CTCTTCTCGT

TTACGCAGC

MTAATAGAA

ATAGTAGTM GCAAAAGTAC

‘$1 I

51

876

TAGCAGCAGT

AGTCGTGGGA

GTGGAALAAC

ATTATTAATA

101

926

TCCTAGCAAT

AGCAATACTG

GTCCTAbG

Aw3aAAaTTlTcATTT.r

151

CT-TGGTTATG

CCGGTACPX

CGGGCCTCTT

GCGGGATATC

GTCCATTCCG

976

CGGGCCTACT

CCCAGGCAGG

TA-

CTTGCCTTGC

Clw2TTcAT

201

ACAGCATCGC

CAGTCACTAT

GGCG’lWTGc

TAGCGCTATA

TGCGTTGATG

1026

TGGACATAG

GGA CAGc$AC

CACGGGGAGT

CGGTAGAATG

GGA’lWXCTG

251

C$TTWTAT

GCGCACCCGT

TCTC GGAGCA

CTGTCCGACC

GCTTTGGCCG

1076

GCCAGAGGGA

GCACG&G

CWLC-

GCAGCCCGAG

-Gc

301

C;GCCCAGTC

CTGCTCGCTT

AGCCACTATC

GACTACGCGA

1126

TAACAGGGAT

TGAGATAGGT

TGCAAWCTC

CGAGGCTTCG

CATCCTTTGA

351

TCAWGAC

CACACCCGK

CGCTACTTGG 375,l CTG--

CGAGTTCGAG

GCCAGCCTCG

1176

ATACAGACTC

AACEAGACG

TCGTGTTCGG

TTGAATGACA

GACTGAA’L-GA

26

GGCTACACAT

TTTTTTCTCC

TCKCTGGGG

CCTGGCTGCA

CTl.ZlTCGTC

1226

TWX’TTG-

GT-TTGTTTAG

TGTCCTTCCG

-

CCTTGGCrn

76

TCTCCCGTCT

GCGCTCGGCT

CC TTTGCTTC

GGCCTCTTTT

GCNCGGCCT

1276

CGTTCCCTGA

TGGAATTXG

AGTTCAGC

126

CGTCITCAGC

CCCGCTGCCC

GCITGXACA

ACCCCCCGTT

GCCG’PTGCCA

1326

GTGCTGCCAG

GTGCTGAGAC

CCTCGG’IGGA

AAGAAGCCGG

TTCAGZTCT

176

TTGTCCCTGC

CTCTGCCCCG

TKAGCCCGC CCGCCCGCCC 22r,p7-.yp7-m

GCCCGCCCGC_

1376

GTGGCTCTGC

AC’XGTCGTG

CCCAGTGTGG~

ATTCCCTFX

TGG’,TGT’l’CC

226

ACTCCACCGA

CGCGGCCTCT

CCGCACGCTT

-

CAACCAGXTC

1426

TGGGCTCCAT

GGCCTTGAGG

GTTGCACd

GGGGAGGCG

TCGGGTCGGG

276

CTCCCGCTTC

CCl”XCXTc

CGCGTCCCGC

GCTCAACCGT

GGCCCGCCCT

1476

TGCGGAGCAA

TATCCAGCCT

CGACCCCAGC

AGGCTTWXG

GCACCCTCTG

326

CGGGAGACCC

CACCGTGCCC

TTCAGCCCTT

TCTCCGGCGC

CTGCCTCGCC

1526

AGCCGGGTCC

TTGCCCTCCT

ATTGCTCCCC

GAGTTTGTCC

TCCAGCTCCC

376

CTGCCTCGCC

TCCCCGGCTC

CACCAGCAAG

CGCTCCCCGCC

AGACCGATGC

1576

GGTGTA’PTTC

ATGGAGGAGC

GACWTGTT

TPGTAACCTT

CChWTG

426

GCCGCAAGCC

ACGCCTCCTC

GTCCTCCCAA

GCCGCACGCC

GAGG’lYXGGC

1626

AGTGCACTTT

CTTAGG’XGG

CTl’GTCGCGA

GACCCCTGCT

GCT’XTGCTG

CCGAGCCTCC

CGCGCTCCGG

CTCCGACTTG

-GC

1676

ClGClGcTGc

TGGAGATGGA

GGGGGTGCGA

G’lWXACCGG

GGWCXCCCC

GACGACGCCC

1726

GATGGTlGGT

TGGATGGATG

m GGT’XGTGGG

GATGGATGGT TGGA’FSGATG

%-% GA+

476

CTCGGTTTCC

526

CGCC

TCC

GG GCCAXAAGG

GTCCCTGCCA

* CLAC,

Figure 4. Map and nucleotide sequence of p7 DNA or of pBR322 DNA and of the sites of recombination with the 60 to 70 map unit fragment of Ad12 DNA in in vitro-generated recombinants. The Ad12 DNA “termini” are the sequences closest to the 60 (left) or the 70 map unit (right) point in the Ad12 DNA fragment. The partial map at the top is that of circular pBR322 plasmid DNA containing the 1768 base-pair p7 DNA fragment derived from BHK21 DNA. The original site of Ad12 DNA insertion in the Adl2-induced tumor cell line CLACl is indicated by double-headed arrows. The right (r) and left (1) sites of linkage of the p7 DNA fragment to Ad12 DNA or to pBR322 DNA are designated with these letters and the numbers of the recombinants, both in the map and in the nucleotide sequence. The nucleotide sequence of p7 DNA and the site of insertion of the previously described in vitro recombinants p7-R5 and p7-R6 have been published (Jessberger et al.. 1989). The underlined nucleotide sequence, p7 nucleotides 225 to 274, was synthesized as an oligodeoxyribonucleotide, which has been used for DNA-protein binding studies (Fig. l(b)). In the map, the EcoRI cleavage site in pBR322 DNA is also designated. The sequence reproduces the first 375 base-pairs of pBR322 DNA (375/l) followed by the 1768 base-pair sequence of p7 DNA.

The hamster preinsertion sequence p7 was originally derived from the Ad12 insertion site in the Adl2-induced hamster tumor cell line CLACl, and at this site extensive patchy homologies between the nucleotide sequence of the hamster preinsertion site and some of the (deleted) left terminal 174 basepairs of Ad12 DNA were observed (Fig. 2 in Stabel & TIoerfler, 1982). This patchy sequence arrangement was also reproduced in Figure 3 and designated CLACl. The present data confirmed the previously formulated working hypothesis that insertional recombination of foreign (adenoviral) DNA in mammalian cells was aided by, but not strictly dependent on, the availability of patchy homologies between the foreign genome and the site of insertion in hamster DNA (Stabel & Doerfler, 1982). Patchy homologies, as compared to long

stretches of perfect homologies, might offer recognitional advantages, as structural features in the former might be more distinctly recognized by the recombination machinery in the cell. (ii) Localization

of the sites of recombination in the 1768 base-pair p7 hamster preinsertion sequence

The identification of nucleotide sequences at the sites of linkage between Ad12 and p7 DNA sequences in the recombinants recl0, recl8, rec22, rec27, rec9 and rec20 facilitated the unequivocal localization of the sites of recombination in the p7 hamster preinsertion or the pBR322 plasmid sequence (Fig. 4). The map in Figure 4 presents the nucleotide sequences at which the aforementioned recombinants had been generated in the cell-free recombination system. The map and the nucleotide

124

J. Tatzelt

sequence in Figure 4 located the points of the rightterminal Ad12 DNA junct’ions to p7 DNA (lOr, 18r. 22r. 27r) or t’o pBR322 DNA (9r, 20r, llr: 13r). As described earlier, t*he p7 or pBR322 sequences to the left of the designated points were deleted, and the left termini of Ad12 DNA were linked to pRR322 DNA (101, 181, 221, 271. 91: 201, 111, 131 in the maps in Fig. 4; see also Figs 2 and 3). These latter sites of linkage and their pBR322 map locations were clustered very close to t’he site of EcoRI cleavage in the plasmid pBR322 DNA. This clustering suggest,ed that precleavage of the plasmid DNA led to an endto-end joining of the right end of Ad12 DNA (closest’ to map unit 70) and pBR322 DNA at this site. The recombination between the left end of ,4d12 DNA (closest to map unit 60) and p7 or pBR322 acceptor DNA generated structures that suggested a different mechanism for t’hese recombination events. Tt, cannot be excluded that recombination at the left end of Ad12 DNA also involved end-to-end joining after extensive exonucleolytic cleavage. Deletions had occurred at the sites of linkage, as was apparent from the schemes in Figure 2. The map and nucleotide sequence in Figure 4 also contained the locat’ions of the previously determined nucleotide sequences at sites of junction in the recombinantx p7-R5, and p7-R6, and the Ad12 insertion sit’e in the (sell line CLACl. Several regions of the p7 DNA sequence had served as insertion targets. There were three independent recombination events at or around the which nucleotide sequence 5’-CCTCTCCG-3’. occurred twice in the p7 DNA fragment. i.e. between nucleotides 241 and 248 (sites of recombinants p7-R5, p7-R6 and rec22) and between nucleotides 529 and 536 (original site of Ad1 2 DNA insertion in cell line CLACl). Other recombinants were located at nucleotide sequences apparently unrelated to the 5’-CCTCITCCG-3’ motif. (d) Binding

of proteins from th,e kiephacryl S-300 fractions with activity in in vitro recombination to ccjunction sequence in, p7 IjNA

It was reasoned that proteins from the Sephacrvl S-306 fractions 32 t’o 48. which were active in promoting in vitro recombination between the 60 to 70 map unit fragment of Ad12 DNA and p7 DNA (Fig. l(a)), might specifically bind to a p7 DNA sequence t,hat was identified in junction sequences in some of the recombinants (underlined in Fig. 4: see also Materials and Methods). We used a synthetic double-stranded 50 nucleotide fragment from Ad12 DNA as substrate for binding experiment’s with the mentioned protein fractions and monitored complex formation by gel shift assays. The results in Figure l(b) demonstrate that proteins in the shoulder fractions (numbers 32 to 48) formed complexes with the oligodeoxyribonucleotidc composed of nucleotides 225 to 274 in p7 DNA. Only the fractions eluting in the shoulder (numbers 32 to 48) from the Sephacryl S-300 column. but not in the peak of the elution profile (numbers 26 to 30).

et al formed complexes with the DNA substrate derived from a frequent site of in vitro recombination (auto radiograph in Fig. l(b)). The peak fractions were not active in eli&ing the recombination reaction. 111 t,he experiment, described in Figure 1(b), every other fraction from the column, which was elutrd with a buffer containing 150 rnbf-NaCl. was tested for the presence of proteins c*apable of complexing with thch oligotleoxyribonu~leotitle. Bovine serum albumin did not elicit any complex with this I)NA fragment i even in Wfold excbess(dat)a not. shown). The specificity of complex formation was assessed by using various oligodeoxyribonucleotides (50 ng). in 356fold excess. as specific c:ompet.itors in tht, binding reaction (data not shown). (‘omplex formation was inhibited or abolished by the identical oligodeoxyribonucleotjdr. whereas other synthetic oligodeox,vrit~onucleotides cornpet’& to a limitc~d extent or not all. Although an extensive enzymat’ic analysis of’ thtb eluted fractions would he premature at this stage. it was encsouraging t.o find t,hat helicase activity was present in t,he fractions tested that were recombinat)ionally active (Fig. f(a)). The conventional >issa>liberating a 17.mer oligodeoxyribonuolrotidc~ from double-st,randed configuration with the partly single-stranded >I1 3mplX DNA was employed in these experiments (see Materials and Sfethods). It is c~oncluded that> prot)eins in the shoulder fixc.tions from the gel filtration experiments (Fig. f(a)). which can catalyze the in vitro recombination reac*tion. can form DNAprotein complexes with ii 1’7 1)NA segment that has been involved directly in the recombination reaction and has become part, of’ junction sequences. itt

(e) Randomly selected hamster cell D,liil sequemrs did not recombine with Ad12 D,VA

We previously described control experiments that demonstrated that randomly select,ed hamster DNA sequences, as pBR322 clones. did not recombine with Adf 2 DNA (Jessbergrr of n/. 1989). These experiments \v(re extended to ;I tot’al of’ eight randomly selected fsHK21 DNA sequences. fn over 11,000 new bacterial ampR colonies tested, which had been generated after transfcct,ion of the rec,4 strain of E. coli wit’h the reaction products of th<, eight, randomly selrptrd hamster DNA-plasmid clones p-BHK-1. -2, -3. -4. -22. -26. -27 and -29 with Pstl-cut Ad12 DNA. none contained XdlB-positive recombinants. In these experiments. t,he 0.3 &I-Na!‘l extract’s of BHK21 nuclei were used wit,hout suhfractionation, as previously described (Jesxberger it al., 1989). Tt. is concluded that t,heae earlier findings have now been confirmed in a, more extensive series of experiments that demonstrate that Ad12 DNA does not recombine in a cell-free nuclear extraczt system in r+tro with randomly sele&ed hamster DNA sequences at a frequency that is comparable to that of Ad1 2 f)NA recombination with sequences in the 117preinsertion sequence.

Viral-Cellular

DNA

Recombination

4. Discussion We have set out to use an in vitro system for studies on Ad12 DNA-hamster cell (pi’) DNA recombination in cell-free extracts, in order t,o elucidate at least some of the steps in the mechanism of insertional recombination of Ad12 DNA. In all these experiments, the pBR322-cloned hamster DNA preinsertion sequence p7 has been used as acceptor for the 60 to 70 map unit fragment of Ad12 DNA, which recombined efficiently with p7 DNA. The results presented point the way in which the further purification of the extract components essential for the reaction might be accomplished (Fig. l(b)). We adduced evidence for the binding of proteins from the gel-filtration fractions, which catalyze recombination, to a p7 cellular DNA sequence, and in preliminary experiments (data not shown) to other p7 and Ad12 DNA sequences that have directly become part of junction sequences. Another striking feature capable of promoting the recombination reaction are the patchy homologies between the reaction partners p7 and Ad12 DNAs at several of the sites of linkage. Such patches have been observed for many of the junction sites and are too frequent in this system to be dismissed as random events. Moreover, the analysis of the original Ad12 DNA integration site in the Ad12induced hamster tumor CLACl has also shown patches that are similar in structure to the junction sites described here, though different in sequence (Stabel & Doerfler, 1982). Thus, the cell-free system seems to mimic features of the situation in the living cell. We do not want’ to overrate the significance of patch homologies in directing the recombination process, but patches can be utilized when available at sites of insertional recombination. Sites of insertion for foreign DNA were often found in mammalian DNA sequences that were transcriptionally active (Gahlmann et al., 1984; Schulz et al., 1987; Thomas & Rothstein, 1989; Mooslehner et al., 1990: Scherdin et al., 1990). The results of a large number of control experiments argue against contributions that the E. coli system could have made to the generation of the in vitro recombinants (see Table II in Jessberger et al., we are still investigating 1989). Nevertheless, different possibilities to exclude a contribution of the E. coli host to the generation of the in vitro recombinants. Additional lines of evidence reasoning against the role of the prokaryotic host have now been deduced. (1) The experiments with randomly selected hamster DNA sequences have been extended. Eight different DNA segments of the unique or the repetitive sequence type do not recombine in this in vitro system with Ad12 DNA, at least not at a frequency comparable to that of the p7 preinsertion sequence (see Jessberger et al., 1989). This apparent preference for the p7 preinsertion sequence is probably not absolute. (2) The localization of the recombinatorial activity to certain protein fractions in the gel-filtration experiments further strengthens the interpretation that

in a Cell-free

System

125

essential steps in the recombination event had taken place in the cell-free system, particularly because the peak protein fractions from the column have been found to be devoid of this activity (Fig. 1). The peak fraction does not harbor significant nuclease activity that might have destroyed the DNA substrates. (3) Proteins from those column fractions that promote recombination can form complexes with Ad12 and p7 DNA sequences that have served as sites of recombination. In four documented recombination events leading to the in vitro recombinants p7-R#5, p7-R6 (Jessberger et al., 1989), rec22 (this study, Figs3 and 4), or to the AdlS-induced hamster tumor line CLACl (Stabel & Doerfler, 1982; double headed arrow in the sequence in Fig. 4), the cellular sequence 5’-CCTCTCCG-3’ is located at. or very close to, the site of recombination. The observation that the right end junction sites of the 60 to 70 map unit fragment of Ad12 DNA to pBR322 DNA exhibit less patchy homologies and shorter deletions suggests that these junctions might have been generated by end-to-end ligation after exonucleolytic trimming at, or close to, the EcoRI site in the pBR322 part of the pBR322cloned p7 DNA where the construct has been linearized prior to use in the recombination reaction. This linearization of the pBR322-cloned p7 sequence may, in part, be responsible for the occurrence of the Ad12 DNA linkage to pBR322 DNA. Such linkage was observed previously in one of four recombinants analyzed by the determination of the nucleotide sequence (see Fig. 3a in Jessberger et al., 1989). On the other hand, the left-hand junction between the 60 to 70 map unit fragment of Ad12 DNA and p7 DNA bears features reminiscent of an invasive (single-strand) mechanism searching for a set of short patchy homologies. Obviously, other mechanisms can not be ruled out. The analyses at the sites of junction between p7 and Ad12 DNA or between pBR322 and Ad12 DNAs demonstrate that larger sequence blocks can be delet*ed by an unknown mechanism. The fact that four out of eight of the recombinants (Fig. 2) had plasmid sequences interspersed between the p7 and the Ad12 DNA parts could indicate that rearrangements of DNA sequences might have occurred at some point during the generation or isolation of the recombinants. We thank Irmgard HGlker for synthesis of oligodeoxyribonucleotides and Hanna Mansi-Wothke for preparation of media. Petra BGhm rendered expert editorial assistance. B.S. was on leave from the Bayer AC;, Wuppertal. This research was made possible by a grant from the Bundesministerium fiir Forschung und Technologie, Bonn-Bad Godesberg through Genzentrum Kiiln (TP 2.03. BCT 0930/2).

References Doerfler, W., Gahlmann, R., Stabel, S., Deuring, R., Lichtenberg, U., Schulz, M., Eick, D. & Leisten, R. (1983). On the mechanism of recombination between adenoviral and cellular DNAs: the structure of

126

J. Tatzelt et al

junction sites. Curr. Topics Microbial. Immunol. 109. 193-228. Doerfler. W.. Spies. A.. ,Jessberger, El., Lichtenberg. U.. Zock, C. & Rosahl, R. (1987). Recombination of foreign (viral) DEA with the host genome. Studies in airv and in a cell free system. In Molecular Ra.uis of J’iml and Microbial Pathogenesis (Rott, R. & Goebel, W’.. eds), 38th Colloquium Mosbach 1987, pp. 60 72. Springer Verlag, Berlin & Heidelberg. Dulbecco. R. & Vogt. M. (1954). Plaque formation and isolation of pure lines with poliomyelitis viruses. J. Expt. Med. 99. 167-182. Gahlmann? R.. Schulz, &I. 8: Doerfier, W. (lQ84). Low molecular weight RKAs with homologies to cellular DIVA at sites of adenovirus Dh’A insertion in hamster or mouse cells. EMBO J. 3. 3263-3269. Grunstein, M. Br Hogness. I). S. (1975). C!olony hybridization: a method for the isolation of cloned DKAs that contain a specific gene. Proc. Xat. Acad. Sri.. I’.S.A. 72, 3961-3965. Hermann. R.. Hoeveler. A. 9i Doer&r. W. (19%)). Sequence-specific methylation in a downstream region of the late E2,4 promoter of adenovirus type 2 DEL4 prevents protein binding. .I. &foZ. Riol. 210. 41 131.?. Jessberger, R.. Heuss. 1). d Doerfler, IIT. (1989). Recombination in hamster cell nuclear extracts between adenovirus type 12 DNA and two hamster preinsertion sequences. EMBO J. 8. 869-878. Kruijer. W.. van Rchaik, F. M. A.. Bpeijer. .J. G. & Sussenbach.
Mooslehner. K., tiarls. I... & Harbrrs. K. (IQQO). Retroviral integration sites in transprniv mov micae frequently map in the vicinity of transcribed DK.4 regions. J. I’irol. 64. 3056-3058. Richardson. (‘. (‘. (1965). I’hosphorylation of nucleic acid by an enzymr from T4 bacteriophage-infected E. coli. Proc. ,\‘a!t. Acad. Ski., I’.S.A. 54, 158165. Sanger, F.. Xicklen, S. & (‘ouison. A. R. (1977). I)SA sequencing with chain-terminating inhibitors. Pro?. ,VTat. Acad. Sri.,

1’S.A.

74. .5463-q5467.

Scherdin. I’.. Rhodes. K. B Hrrindl. >I. (IQQO). Transc~riptionally active genome regions are preferred targets for r&rovirus integration. ,J. I’irol. 64. 9077!J12. Schulz. ,\l.. Freisem-Rabien. I... .Jessbrrger. I<. & I)oerfler. M’. (1987). Transcriptional activit,ies ot mammalian genomes at sites of rec~ombination with foreign DNA. ,I. Vi&. 61. 444-353. Seo. Y.-S.. Lee. S.-H. & Hurwitz. .J. (1991). Isolation of a JINX hrlirasr from HrLa cells requiring the multisubunit human single-stranded DXA-binding protein for activity. J. Riol. Chum. 266. 13161-13170. Shapiro. D. .I .. Sharp, p. A4.-1.. Wahl. W. W. & Keller, M. .J, (1988). A high-efIicienc&y He La (ell niicleai. 7, 47-55. transcription ext,ract. I)XA Stabel, S. Kr Dorrfler. WT. (1982). Xucleotidr sequencae at the site of junction between adenovirus type 12 DXA and repetitive hamster cell DKA in t,ransformed (~311 line CLA(‘1, .v~cl. Acids Rrs. 10, 8007-8023. Sutt,er. I).. \Vest,phal. M. & Dorrflf?r, LT. (I Q78). I’at~tt~rns of integration of viral DXA sequencaes in the yenomrs of adenovirus type I ‘L-transformed hamster cells. Crll. 14. 56!1-58.5. Thomas. 1%. .J. 8r Kot~hstein. K. (l!)XQ). Elrva~tetl recombination rat,es in transcriptionally active, l)IYA. t’ell. 56. 61!J-630. Watson X;. (1088). zA nflu’ revision of thcl srquenc~e of plasmid pBR322. NUX. 70, 399-403.