viroLocy 66, 53-69 (1975) Evolutionary Variants of Simian Virus 40: Cloned Substituted Variants Containing Multiple Initiation Sites for DNA Replication! THERESA N. H. LEE, WILLIAM W. BROCKMAN anp DANIEL NATHANS Department of Microbiology, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205 Accepted February 11, 1975 Two cloned evolutionary variants of simian virus 40 (SV40) containing substitutions of cellular DNA have been characterized by restriction endonuclease analysis, electron microscopic heteroduplex mapping, and DNA-DNA hybridization. Each variant genome is made up of a small, tandemly repeated segment of DNA consisting of cellular DNA and that portion of the SV40 genome containing the initiation signal for viral DNA replication. Cellular DNA sequences are different in the two variants, indicating that recombination between cell DNA and SV40-DNA can occur at more than one site. However, one end of the SV40 segment (0.68 SV40 map-units) is the same in each variant. The data suggest that substituted variants arise by integration of SV40-DNA into cellular DNA followed by excision of a small substituted genome which is subsequently amplified to a size com- patible with encapsidation; the presence of multiple initiation signals in each molecule re- sults in selective replication. INTRODUCTION Serial passage of SV40? at high multi- plicity of infection leads to the evolution of new species of viral DNA containing dele- tions, duplications, and substitutions with cellular DNA (Uchida et al., 1968; Yosh- iike, 1968; Lavi and Winocour, 1972; Tai et al., 1972; Brockman et al., 1973; Rozen- blatt et al., 1973; Martin et al., 1973). After many serial passages, dominant species evolve with genomes that contain DNA se- quences derived largely from host-cell DNA but retain a small amount of SV40- DNA (Brockman et al., 1973). Since these variant genomes replicate in cells coin- 'This is publication No. 15 in a series on the genome of simian virus 40. Publication No. 14 is Brockman et al. (1975). 2The following abbreviations are used: SV40, simian virus 40; DNA form I, covalently closed-circu- lar duplex DNA; DNA form Il, circular duplex DNA with one or more single-strand scissions; SV40-DNA- Lar, linear SV40-DNA prepared with Endo R. EcoRI. Abbreviations for restriction endonucleases follow the conventions described in Smith and Nathans (1973). Copyright © 1975 by Academic Press, Inc. All rights of reproduction in any form reserved. fected with SV40, we postulated that the segment of SV40-DNA retained includes nucleotide sequences specifying all cis functions needed for viral DNA replication, including the specific initiation site. Progress in understanding the evolution of SV40 variants has depended to a large extent on the isolation and characteriza- tion of genetically homogeneous variants, i.e., cloned populations of single variants, from both early passage and late passage stocks (Brockman and Nathans, 1974; Mertz and Berg, 1974). In the preceding paper, Brockman et al. (1975) described the characterization of cloned variants from early passage virus which can express some SV40 genes. In this communication we describe the properties of cloned sub- stituted variants present in late passage stocks. In contrast to early passage virus (Brock- man and Nathans, 1974; Brockman et al., 1974; Mertz et al., 1974; Brockman et al., 1975), the genomes of late variants are made up of short, tandemly repeating seg- 54 LEE, BROCKMAN AND NATHANS ments of DNA, consisting of cellular DNA and that portion of the SV40 genome which contains the initiation signal for viral DNA replication. Therefore, the initiation site is present in multiple copies, probably ac- counting for the predominance of these variants in late passage stocks. Since ge- nomes with very small segments of SV40- DNA replicate in cells coinfected with helper virus, it is likely that no cis func- tions other than the initiation signal are required for SV40-DNA replication. More- over, the two substituted variants exam- ined in detail contained nonhomologous cellular DNA sequences, indicating that recombination between SV40-DNA and cell DNA can occur at more than one site. A preliminary account of part of this work has been published elsewhere (Brockman and Nathans, 1974; Brockman e¢ al., (1974). MATERIALS AND METHODS Cells and viruses. These have been de- scribed in prior publications (Danna et al., 1973; Brockman and Nathans, 1974). Briefly, small-plaque SV40 (from strain 776) was grown in BSC-1 cells infected at an input multiplicity of 0.001 PFU/cell, starting with a stock made from a single plaque. Serial undiluted passage of this plaque-purified stock has been described (Brockman et al., 1973); P13 and P20 are the 13th and 20th serially passed stock virus, respectively. The procedure for clon- ing variants from P13 and P20 has been described (Brockman and Nathans, 1974). Viral DNA. DNA labeled with *?P or with ['*C]thymidine was prepared from infected cells by the procedure of Hirt (1967), as detailed earlier (Danna et al., 1973). Following the isolation of DNA form I by equilibrium centrifugation, in CsCl- ethidium bromide solution, the dialyzed sample was subjected to electrophoresis in 1.4% agarose gel as described previously (Brockman and Nathans, 1974). In the case of [??P DNA the variant and helper SV40- DNA were detected: by autoradiography and in the case of [1*C DNA by ethidium bromide staining (Sharp et al., 1973). The short, variant DNA and the SV40-DNA were recovered by electrophoresis into dial- ysis sacs. In some cases, equilibrium cen- trifugation was omitted in the purification of variant DNA. 'C-labeled cell DNA was prepared from BSC-1 cells as described (Brockman et al., 1973). Digestion of viral DNA. Digestion with various restriction endonucleases (Endo R-Hindll + II, R-HindIH, R-EcoRI and R- EcoRI) and isolation of DNA fragments by electrophoresis in acrylamide-gel slabs have been described (Danna et al., 1973). Endo R.HindII + II and R- HindIII were prepared as described earlier (Lai and Nathans, 1974); Endo R-EcoRI was pre- pared by P. Geshelin, following the proce- dure of Yoshimori (1971); Endo R-EcoRII was prepared through the ammonium sul- fate step by the procedure of Yoshimori (1971) and then chromatographed on dia- minobutane-substituted Sepharose (Shal- tiel and Er-El, 1973) as described for R-Hpa I (Sack, 1974). To quantitate [?*PJDNA fragments in electrophoresis gels, autoradiograms were traced with a Joyce-Loebl microdensitometer and the areas of expanded peaks measured. DNA-DNA _ hybridization. **P-labeled variant DNA was hybridized to 'C-la- beled SV40-DNA, to ‘C-labeled SV40- fragments, or to '*C-labeled BSC-1 (cellu- lar)-DNA, essentially as described previ- ously (Brockman et al., 1973). In some in-, stances the DNA was nicked and dena- tured by heating at 100° in Yoo SSC for 15 min.; in other instances the DNA was boiled in 0.2 N NaOH for 30 min. To follow rates of reannealing of denatured °?P-la- beled variant or SV40-DNA, alkali-boiled DNA was incubated under oil in 0.3 M NaCl, 20 mM Tris-Cl, pH 7.6, 1 mM EDTA at 68°, and samples were removed periodically and chilled. The percentage of DNA remaining single stranded at various times was determined by adsorption of double-stranded DNA onto hydroxyapatite at 65° in 0.14 M sodium phosphate, pH 6.8-0.15 M NaCl and counting an aliquot of the supernatant fraction (Brockman et al., 1973). In some experiments S1 nu- clease (Sutton, 1971) was used to assess single-stranded DNA (Brockman et al., 1973). Electron microscopy of DNA. For length SUBSTITUTED VARIANTS OF SV40 55 measurements of variant DNA, variant form II DNA isolated by electrophoresis in agarose was mixed with SV40-DNA-La: and mounted on grids as described by Davis et al. (1971). For heteroduplex map- ping, an Endo R fragment of variant DNA was mixed with SV40-DNA-L,, (1 ug/ml) at a molar ratio of 10-20 (fragment to Lx), and the mixture denatured, renatured, and mounted according to the procedure of Davis et al. (1971). Heteroduplexes be- tween DNA fragments were prepared in a similar way, as described in the text, except that the molar ratio was about 1:1. To measure DNA lengths, projections of electron micrographs were traced with a Graphics calculator (Numonics Corp.). Replicative intermediates of ev-1101. Replicative intermediates, pulse-labeled with [8H |Jthymidine as described elsewhere (Brockman et al., 1975), were separated from cellular DNA by the method of Hirt (1967) followed by treatment with heated RNase, phenol extraction and alcohol pre- cipitation. The viral DNA was then di- gested to completion with Endo R- HpaI to cleave wild type helper DNA (Sack and Nathans, 1973) and centrifuged through a 5-30% sucrose gradient at 40,000 rpm and 10° for 7 hr in a Spinco SW41 rotor. Fractions containing radioactive DNA (about 24 S) were pooled and further proc- essed as described in the text. RESULTS Isolation of Cloned Variants from Late Passage Stocks To isolate cloned variants of SV40, which have retained only asmall portion of the SV40 genome, we used virus stocks from the 13th and 20th serial passages (P13 and P20). Cells infected with either of these stocks yield viral DNA consisting of a few dominant and abnormal species with mostly cellular DNA sequences (Brockman et al., 1973). In the cloning procedure, P13 or P20 lysate (containing defective and helper viruses) was plated on BSC-1 cells at various dilutions, and _ individual plaques were screened for the presence of variant genomes, as described by Brock- man and Nathans (1974). Viral DNA pre- pared from cells infected with plaqued virus stocks were surveyed in two ways: 1) By agarose-gel electrophoresis to detect short genomes, and 2) by digestion with Endo R-Aindil + III to detect new DNA fragments expected from a variant genome. Of 34 plaqued stocks from P13, four had evidence for short genomes, and two of five plaqued stocks from P20 had variants, which in this case were identical to each other. Of the five variants detected, we chose three for further characterization: ev-1101 and -1103 isolated from P13, and ev-1102 isolated from P20 (Fig. 1). Length of Variant DNA’s Variant genomes isolated from infected cells and separated from helper genomes by agarose-gel electrophoresis were mea- sured in the electron microscope. The re- sults, presented in Fig. 2, indicate that ev-1101 was 85 + 5% (SD) of the length of SV40-DNA present on the same grid, ev-1102 was 79 + 5%, and ev-1103 was 78 + 4%, As seen in Fig. 1, however, ev-1103 genomes were not uniform in length, the above measurements being done on the dominant species. As will be indicated below, the different size-classes of DNA derived from this variant are related, since they yielded the same Endo R- Hind frag- ments. Endo R Digests of Variant DNA’s As already noted, SV40 variants from Pi3 and P20 virus were detected by the Endo R. Hind digest pattern of DNA con- taining variant plus helper virus genomes. For a clearer analysis, short genomes iso- lated by agarose-gel electrophoresis were digested and the resulting fragments sub- jected to electrophoresis. Typical results are shown in Fig. 3. As seen in the figure, ev-1101 DNA yields a single Hin fragment which is 17% of the length of SV40-DNA, as determined by mobility (Danna et al., 1973); ev-1102 DNA yields two Hin frag- ments which are 20 and 7%, respectively, of SV40-DNA; and ev-1103 yields two frag- ments, which are 4.6 and 4.2%, respec- tively, of SV40-DNA. These results indi- cate that each of these DNA’s consists of tandemly repeated segments. For further 56 LEE, BROCKMAN AND NATHANS Wild Type Ol N02 103 —— Origin — SV40 — Variant FormIL. FormIL — SV40 FormT Variant FormI Fic. 1. Separation of short, variant DNA molecules from SV40-DNA by electrophoresis in agarose: ethidium bromide stain of viral DNA purified as described in Materials and Methods. characterization, we concentrated on ev-1101 and ev-1103, the two simpler vari- ants. Since ev-1103 yielded two Hin frag- ments, we could not be certain that these were derived from the same variant DNA, even though the two fragments were pro- duced in equimolar amounts (Table 1). Therefore, we looked for a restriction endo- nuclease which yielded only one fragment from ev-1103 DNA. Endo R-£EcoRII yielded a single fragment which was 8.8% of the length of SV40-DNA as judged by electrophoretic mobility (Fig. 4). This frag- ment as well as partial RII digest products (17.5 and 26% of unit length; see Fig. 4) were redigested with R- HindIII to yield the products also shown in Fig. 4. The amount of each fragment was then estimated from densitometer tracings and molar yields calculated (Table 1). From these results we conclude that ev-1103 DNA has a repeating unit of about 8.8% of the length of SV40- DNA, each unit having two Hind sites and one EcoRII site, and that the RII site is within the Hin-I, segment of the molecule (Fig. 5). Also, the results presented in Table 1 confirm the presence of tandemly repeated segments in ev-1103 DNA. As pointed out earlier, ev-1103 DNA isolated from infected cells consists of more than one size-class, as indicated by agar- ose-gel electrophoresis of the open-circular form (Fig. 1). When each of the minor size-classes was isolated from the gel and analyzed by Endo R-Hind digestion, the electrophoretic pattern of fragments was the same in all cases and corresponded to that shown in Fig. 3. Since ev-1103 DNA has a repeating unit of about 8.8% of SUBSTITUTED VARIANTS OF SV40 57 SV40-DNA length, the shorter molecules probably lack one or more repeating units relative to the major size-class. Rates of Reassociation of Denatured Vari- ant DNA As a further test of the complexity of variant DNA’s, we compared the rates of reassociation of denatured variant DNA to that of helper SV40-DNA isolated from the same infected cells. If variant DNA is made up of n repetitious segments of identi- cal sequence per SV40-DNA equivalent, the rate constant for renaturation should be n times that of SV40-DNA. In order to compare variant and SV40-DNA as pre- cisely as possible, *?P-labeled variant and helper virus DNA’s were isolated from the same batch of infected cells by agarose-gel electrophoresis of form I DNA (see Mate- rials and Methods). (In each case the isolated helper virus DNA yielded a wild- type Hin digest pattern.) Each DNA sam- ple was then boiled in alkali to fragment the DNA, and the rates of renaturation were measured as described, using hydrox- yapatite to adsorb duplex DNA. The results are shown in Fig. 6 as a plot of the reciprocal of the fraction remaining single-stranded (1/fss) vs the product of initial DNA concentration (cpm/ml) and time at 68° (C,t) divided by the product of initial DNA concentration and time for 50% annealing of the reference SV40-DNA (Coty, (SV40)). From the slope of each line we have estimated the relative rate con- stant for ev-1101 DNA as five times that of SV40-DNA and the rate constant for ev-1103 DNA as 10-12 times that of SV40- DNA. These values agree well with the results of Endo R-Hin digestion, i.e., ev-1101 DNA has a repetitious segment of DNA which is slightly more than ¥ of the length of SV40-DNA, and ev-1103 DNA has a repetitious segment which is about \, of the length of SV40-DNA. Therefore, both sets of data indicate that each variant genome is made up of identical or nearly identical repetitious segments of DNA. The Presence of SV40 and Cell DNA Se- quences As pointed out earlier, P13 and P20 were known to contain predominantly non- repetitive cell DNA and relatively little SV40-DNA or reiterated cell DNA (Brock- ev-Hol ev-1o2 ev-1103 84.874.7 79.0150 78.1444 307 3 3 3 10 204 20 = Pe) 3 5s 10 1o+ li | ul At Lt 60 70 80 90 160 70 0 90 60 70 80 90 160 % SV4O DNA length Fic. 2. Lengths of variant DNA’s compared to SV40-DNA: Electron microscopic measurements. Form II variant DNA’s were isolated by electrophoresis in agarose, mixed with SV40-DNA-L,, and mounted for microscopy. 08 LEE, BROCKMAN AND NATHANS P13 , 13 13 P20 Q A i ik | it ' jf | fit ifiye i | | wy ! UM xy ji "A L | A 8 cD EF G Hivk \ Jl gv~ Ot aM ev—1t02 A Fi A o8 ev~ : Aisk Fic. 3. Endo R- Hin digests of variant DNA’s: Tracing of autoradiograms of slab gels after electrophoresis. On the left is the digest pattern of P13-DNA, of ev-1101 DNA isolated by electrophoresis in agarose, and of ev-1103 DNA isolated by electrophoresis in agarose. On the right is the digest pattern of P20-DNA and of ev-1102 DNA isolated by electrophoresis in agarose. In each case the origin is at the left, and A,B,C ... K represent the positions of SV40 Hin fragments in the same gel. TABLE 1 Propucts oF Hind anp EcoRII Dicestion oF ev-1103 DNA Digested DNA Enzyme Fragments Relative yield (% of SV40-DNA) {% cpm (moles) ] ev-1103 Hindll + II 4.6 (I1) 51 (1) or HindIII 4.2 (J1) 49 (0.96) ev-1103 EcoRIl 8.8 (monomer) _ 17.5 (dimer) _ 26.0 (trimer) _ EcoRII monomer Hindlll 4.2 (J1) 50 (1) 3.1 (Ila) 34 (0.89) 1.5 (Ib) 16 (0.90) EcoRI dimer Aindill 4.6 (11) 27) (1) 4.2 (J1) 47 (1.90) 3.1 (Ila) 17 (0.92) 1.5 (1b) 8.6 (0.97) EcoRII trimer Hindlll 4.6 (11) 34 (2.0) 4.2 (J1) 48 (3.1) 3.1 (Ha) 11.5 (0.98) 1.5 (I1b) 6.5 (1.2) man et al., 1973). We now wished to denatured *2P-labeled variant DNA with analyze the DNA of cloned variants for the excess denatured '*C-labeled SV40-DNA types of sequences present. For this pur- or ?*C-labeled BSC-1-DNA, as described in pose we incubated low concentrations of Materials and Methods, and measured the SUBSTITUTED VARIANTS OF SV40 59 percentage of [?7PJDNA remaining single stranded after 1 hr at 68°. (BSC-1-DNA was used under conditions where only re- petitive sequences renatured, i.e., at Cot of 0.3 mole sec/liter.) The percent of variant 123456 Origin <— . —___ 34.0 Trimer — 225 Dimer — ~~ 20.5 . ——-10.5 Monomer —— e — 85 Ir j ae 8 & eo t —— Tia —— * Iip — Fic. 4. HindHIl digests of EcoRII fragments of ev-1103 DNA: Autoradiograms of acrylamide gels after electrophoresis. Column 1, EcoRII trimer; col- umn 2, HindIII digest of trimer; column 3, EcoRII dimer; column 4, HindIII digest of dimer; column 5, EcoRI monomer; column 6, HindIII digest of mono- mer. I1, J1, 11a and I1b are digest products as noted in the text and Table 1. On the right are the positions of HindIII fragments of SV40-DNA and their respective lengths, as percent of SV40-DNA. ev —II0l DNA hybridized was assessed in two ways. In one procedure, heat-nicked and -dena- tured DNA was used, and hybridization was assessed by hydroxyapatite adsorption of duplex DNA, as detailed in Materials and Methods. This procedure detects even small amounts of SV40 or reiterated host DNA, since the heat-nicked DNA varied from % to full length (Brockman et al., 1973), and partial duplexes adsorb to hy- droxyapatite. In the second procedure DNA was denatured and broken by boiling in alkali, and duplexes were assessed by digestion with the single-strand-specific nuclease S1 (Ando, 1966). This method measures the percentage of sequences de- rived from SV40 or reiterated host DNA. Eco Ri Eco RIE Monomer ‘7’ In wi Iva Hin d or RI Eco Rit Dimer Ye Eso Ind tT 3 To Hindm Eco Rit EcoRm Trimer 9 $4~¥——>—+-4———_. —+-+—_.—_ a Inv ot 7 oT fF Toe Heo Map Units — 10 20 30 (%SV40 Genome) Fic. 5. Diagram of Hin and EcoRII sites in ev-1103 DNA, deduced from the data shown in Figs. 3 and 4 and Table 1 (see text). ev- 1103 0.25 0.5 0.75 1.0 0.50 1.0 i) 2.0 C, Cty (SV40) Fic. 6. Rates of reannealing of variants DNA’s compared to SV40-DNA. fss and Cot/Cot., (SV40) are defined in the text, and are plotted following Wetmur and Davidson (1968). 60 The results are presented in Table 2. As shown in the table (heat, HA) ev-1101 DNA contains SV40 sequences but no re- iterated BSC-1 sequences; ev-1103 con- tains both. In the case of ev-1101, 47% of the DNA hybridized to SV40-DNA and <5% to repetitive cellular DNA (Table 2, alkali, S1). Since only 79% of the 14C- labeled SV40-DNA in the same tube hy- bridized under these conditions, we esti- mate that about 59% (47/79) of ev-1101 DNA corresponds to SV40-DNA se- quences, and <5% corresponds to reiter- ated cell DNA; the remainder presumably was derived from nonreiterated or less reiterated cell DNA. In the case of ev-1103, about 20% of the DNA was derived from SV40-DNA, about 15% from reiterated cell DNA, and about 65% presumably from nonreiterated or less reiterated cell DNA. Since ev-1101 DNA yielded two Hin fragments, similar experiments were done with the individual fragments of this vari- ant DNA. As shown in Table 2, Hin-I, contained both SV40 and reiterated cell sequences and Hin-J, contained little or no SV40 or reiterated cell sequences. There- fore, J, consists largely or entirely of non- reiterated or less reiterated cell DNA. The Presence of Specific SV40-DNA Se- quences Having demonstrated the presence of SV40 sequences in ev-1101 and ev-1103 DNA’s, we next determined which regions of the SV40 genome were retained. For this purpose denatured **P-labeled variant DNA was incubated with an excess of each SV40 Hin fragment, one at a time and labeled with '*C (see Materials and Methods). The percentage of [*7P]DNA remaining single stranded was then deter- mined after 1 hr at 68° by the hydroxyapa- tite method. The results are presented in Table 3. In the case of ev-1101, contiguous Hin fragments A and C hybridized to variant DNA, whereas in the case of ev-1103 only Hin-C hybridized. In some experiments there was slight hybridization of Hin-D to ev-1101 DNA (Table 3); this was probably due to the demonstrable contamination of this Hin-D preparation with Hin-C, since LEE, BROCKMAN AND NATHANS other preparations of Hin-D showed <5% hybridization. We conclude that ev-1101 DNA contains a segment of SV40-DNA derived from the Hin-AC region of the DNA, whereas ev-1103 DNA contains a segment of SV40-DNA derived from the Hin-C region. From these experiments, however, we cannot exclude the presence of other SV40-DNA segments which might be too short to result in binding to hy- droxyapatite. Indeed, nucleotide sequence analyses of ev-1103 DNA indicate that a small part of Hin-A is present (Dhar and Weissman, personal communication). Heteroduplex Mapping of SV40 Sequences in Variant DNA’s To confirm the results of hybridization of variant DNA with Hin fragments and to map the SV40 segments more precisely, we have examined by electron microscopy het- eroduplex molecules consisting of a linear strand of SV40-DNA made with Endo R. EcoRI (SV40-DNA-Lg:) and an Endo R fragment from ev-1101 or ev-1103 DNA. In the case of ev-1101 the fragment was the single Hin tragment present in an Hind digest (Fig. 3) which is 17% of the length of SV40-DNA. In the case of ev-1103, we used the EcoRI] cleavage products which were 8.8% of the length of SV40-DNA or mul- timers of this segment (Table 1). The results are shown in Figs. 7 and 8. In the figures are micrographs of representative heteroduplex molecules and length mea- surements of individual molecules. As seen in the micrograph of an ev-1101/ SV40 heteroduplex (Fig. 7), the ev-1101 fragment has cyclized on the SV40-DNA strand. This is the result expected if the Hin cleavage site is within the SV40 se- quence of the variant. The ev-1101 frag- ment has paired with single-strands of SV40-DNA-Lai at a unique site beginning at 55.7 + 4.5% of the length of SV40-DNA from one end and 32.3 + 4.7% from the other end (Fig. 7). Since ev-1101 DNA has Hin-A and -C sequences, we can orient the SV40-DNA strand with the Hind cleavage map (Fig. 7, bottom right). From the measured lengths of different parts of het- eroduplex molecules of this type, we con- clude that the duplex region extends from SUBSTITUTED VARIANTS OF SV40 61 TABLE 2 HYBRIDIZATION OF VARIANT DNA to SV40-DNA or BSC-1-DNA? Variant DNA Method Percent hybridized Percent Percent SV40 reiterated Noadded +SV40-DNA + BSC-1-DNA sequences BSC DNA sequences ev-1101 Heat, HA 12 64 (91) 0 (29) ev-1101 Alkali, $1 0 47 (79) 0 (16) 59 <5 ev-1103 Heat, HA 18 62 (84) 68 (47) ev-1103 Alkali, $1 9 18 (90) 15 (21) 20 15 ev-1103 11 Alkali, $1 0 38 (86) 23 (13) 44 23 ev-1103 J1 Alkali, S1 0 3 (87) 3 (14) <5 <5 7A mixture of *?P-labeled variant DNA and !*C-labeled SV40-DNA (or }4C-labeled BSC-1-DNA) was denatured as indicated under Materials and Methods (100° in 1.5 mM NaCl or boiling in alkali) and incubated at 68° for 1 hr. Percent hybridized is percent of **P removed by hydroxyapatite (HA) or resistant to S1 nuclease. Figures in parenthesis are percent ['*C]DNA hybridized in the same tube. Percent SV40 sequences = (% variant DNA hybridized in S1 experiment/% SV40-DNA hybridized in same tube) x 100. Percent reiterated BSC sequences = percent variant DNA hybridized in presence of BSC-1-DNA, using S1 nuclease. 290 cpm of ev-1101 DNA was used (7 ng/ml); 300-400 cpm of ev-1103 DNA (4-8 ng/ml); and 100 cpm of I1- or JI-DNA (2-5 ng/ml). TABLE 3 HysrIDIZATION OF VARIANT DNA To SV40 Hin FRAGMENTS® Hinfragment Percent variant DNA hybridized ev-1101 ev-1103 A 76 0 B 9 1 Cc 59 32 D 18 (0) 5 E 7 0 F 0 0 G 3 0 H 0 0 I 1 0 J 6 0 K 0 0 ¢ A mixture of heat-denatured *?P-labeled variant DNA (240 cpm for 1101 and 230 cts/min for 1103) and excess '4C-labeled Hin fragment were incubated at 68° in 0.3 M NaCl; duplex DNA was then adsorbed by hydroxyapatite (Brockman et al., 1973). The percent variant DNA hybridized is the percent of *?P removed by hydroxyapatite, taking as 100% the **P unadsorbed in the absence of added fragment (82% of input 1101 DNA and 83% of input 1103 DNA). In each experi- ment >90% of each '*C-labeled Hin fragment rean- nealed. A second experiment with ev-1101 DNA anda different D fragment preparation gave the value indicated in the parentheses. 0.56-0.68 map-units in the SV40 cleavage map. Therefore, the ev-1101 Hin fragment has Hin-C sequences at one end and Hin-A sequences at the other end; cleavage by Endo R-Hind occurred at the Hin-A.C junction (Fig. 7). (As reported earlier (Danna et al., 1973), this site is an Endo R.- HindIII site, and, as expected, ev-1101 DNA is cleaved to the 17% fragment by purified Endo R- HindIII.) From measure- ments of the duplex region we estimate that about 65% of the sequences present in ev-1101 DNA are derived from SV40-DNA; this is close to the estimate of 59% SV40- DNA sequences based on extent of hybridi- zation with SV40-DNA (Table 2). The ev-1103 EcoRI fragments also pair with single strands of SV40-DNA-La: at a unique site (Fig. 8). This site is at about 65-68% from one end of the SV40-DNA, which is consistent with sequences around the Hin-A.C junction (Fig. 8), as predicted from the results already presented. From the measurements of heteroduplex mole- cules, we estimate that about 34% of the sequences present in ev-1103 DNA are derived from SV40-DNA; this compares with the estimate of 20% based on extent of hybridization to SV40-DNA (Table 2). Comparison of Cell DNA Sequences in ev-1101 and ev-1103 It was of interest to determine whether the cell DNA sequences present in ev-1101 and ev-1103 are related, since they might have arisen from a common ancestral mol- 62 LEE, BROCKMAN AND NATHANS 657t45 Ly L, 323447 Number of Molecules i 20 30 40 50 40 50 60 70 L, 12.4%2.0 L, 68213 10 18 | 5 10 ll ul 1S 20 5 % $V40 DNA Length tt 56 .68 Fic. 7. Analysis of SV40-DNA-La:/ev-1101 Hin fragment heteroduplexes. Shown are a representative heteroduplex molecule (upper left); a tracing of the molecule (upper right); length measurements on a series of such molecules (middle); and a model (bottom) indicating Hin cleavage sites, lengths measured, and map units of the duplex region (taken from the graphs of L1 and L2). ecule or they could have been generated by independent recombinations between SV40-DNA and cell DNA at different or identical sites. From the results already presented it is clear that the cell DNA sequences in the two variants are not identical, since ev-1103 has both reiterated and nonreiterated cell sequences whereas ev-1101 has no reiterated sequences (Table 2). For further comparison the Hin-I, and -J, fragments of ev-1103 DNA were hybrid- ized to ev-1101 DNA. As seen in Table 4, about the same percentage of Hin-I, se- quences hybridized to ev-1101 DNA as to SV40-DNA. Since ev-1101 DNA appears to contain all detectable SV40 sequences present in ev-1103 DNA (see also Figs. 7 and 8), within the limits of error of the hybridization measurements, Hin-I, DNA and ev-1101 DNA share only SV40 se- quences. In the case of Hin-J,, neither ev-1101 nor SV40-DNA hybridized with this fragment. Therefore, HinJ, and ev-1101 DNA do not share sequences de- tectable by this method. We conclude that most or all of the host sequences present in one variant differ from those in the other. A similar conclusion arises from examin- ing heteroduplex molecules consisting of one strand of ev-1101 Hin fragment and one strand of ev-1103 EcoRII fragment (Fig. 9). Although length measurements are imprecise with such small fragments and heteroduplex molecules were rare, the heteroduplex structures seen are consistent with duplex formation only at the shared SV40 sequences around the Hin-A.C junc- tion (Fig. 9). Origin of DNA Replication in Variant DNA Although every variant we have exam- ined to date retains the region of SV40- SUBSTITUTED VARIANTS OF SV40 63 Oa nate 4 SO ea Ss, eee bE er gE Li 64.7449 le 32.2445 Number of Molecules 3 oO ul Ls A %e SV 40 DNA Length 65 68 Fic. 8. Analysis of SV40-DNA-Lg:/ev-1103 EcoRII fragment heteroduplexes. Shown are a representative heteroduplex molecule (upper left); a tracing of the molecule (upper right); length measurements (middle); and a model (bottom) indicating Hin cleavage sites and map units of the duplex region (taken from the graphs). TABLE 4 HYBRIDIZATION OF ev-1103 Hin FRAGMENTS TO SV40 or ev-1101 DNA® ev-1103 DNA in Concentration Percent fragment Percent SV40 or Fragment excess (ug/ml) hybridized ev-1101 sequences ay SV40 2.9 31 (80) 39 1 $V40 5.8 35 (84) 42 I ev-1101 1.6 35 (82) 43 I ev-1101 3.2 30 (87) 34 Jl SV40 2.9 <5 (80) <5 Jl $V40 5.8 <5 (84) <5 Jl ev-1101 1.6 <5 (83) <5 Ji ev-1101 3.2 <5 (87) <8 * Approximately 115 cpm of *?P fragment (1 ng) was mixed with *H-labeled SV40 or *H-labeled ev-1101 DNA and the mixture incubated as detailed in Materials and Methods. The mixture was then treated with S1 nuclease and TCA-precipitable and -soluble radioactivity determined. In the absence of added DNA, 22 and 18% of I1- and Ji-DNA, respectively, was S1-resistant. Percent fragment hybridized is the percent of input *?P in the TCA precipitate, correcting for the control just indicated. Nearly identical results were obtained using either TCA-soluble or precipitable counts. The values in parentheses are the percent [7H]DNA which was S1 resistant in each tube; these values were used to normalize percent fragment hybridized, as indicated in the last column. DNA which contains the origin of DNA _ variant does begin replication at the SV40 replication (Danna and Nathans, 1972; signal and not within its cell DNA seg- Fareed et al., 1972), we wished to demon- ment. To determine the position of replica- strate more directly that a substituted tion initiation in ev-1101, this site was LEE, BROCKMAN AND NATHANS 4.142 1198 s 3 arc Ac AC AC 3. 5 g 4 tL uy 3 3 5s 2 lac 2 UAL 2 4 6 8 AC Ac AC Lo (% of SV40DNA Length) FLt Fic. 9. Analysis of ev-1101 Hin fragment/ev-1103 EcoRII fragment heteroduplexes. Shown are two representative heteroduplex molecules (upper left); a tracing of the molecules (upper right); measurements of several molecules of this type (lower left); and diagrams indicating Hin cleavage sites. The loop is derived from the ev-1101 fragment. located relative to the nearest HindIII cleavage site. For this purpose replicating molecules of ev-1101 were partially cleaved with Endo R.- HindIII, which cleaves at the Hin-A-C junction. Since the length of the ev-1101 repeating unit is known, each A.C junction can be located along the cleaved molecule and the initiation site (taken as the midpoint of a replicating branch; (Schnés and Inman (1970); Fareed et al. (1972)) can be localized with respect to this junction. The stocks of ev-1101 used to prepare replicating intermediates, how- ever, contain wild-type virus in addition to the variant. Because the partial digest products of wild-type replicative interme- diates would make the interpretation of the digestion results difficult, wild-type repli- cative intermediates were removed prior to Hindlll digestion. This was accomplished by digesting to completion the pulse- labeled DNA preparation containing ev-1101 and wild-type replicative interme- diates with Endo R-Hpal. This enzyme cleaves wild-type SV40 into three frag- ments (Sack and Nathans, 1973; Sharp et al., 1973) but does not cleave ev-1101 DNA. When the Hpal-digested DNA was sub- jected to velocity centrifugation through a 5-20% sucrose gradient, the wild-type frag- ments were readily separated from repli- cating and mature molecules of ev-1101. The broad peak of ev-1101 replicative in- termediates sedimenting at about 24 S was isolated and partially digested with Hin- dIII so that the majority of cleaved mole- cules contained only one scission. The distance of the replication initiation site from the HindIII site in full-length ev-1101 replicating molecules was measured in the electron microscope (Fig. 10) and was ex- pressed in terms of the repeating unit, using the formulae given in the legend to Fig. 10. As shown in the figure, the major- ity of ev-1101 molecules analyzed started replication at a site close to a HindlIl SUBSTITUTED VARIANTS OF SV40 65 cleavage site. (Although the orientation of molecules visualized is not known, since the origin of replication in most molecules is very close to the A-C junction, the orientation has little effect on the length distributions.) The initiation point thus falls at or near the Hin-A-C junction, which corresponds to the origin of wild-type SV40-DNA replication (Danna and Na- thans, 1972). Since ev-1101 DNA contains five possi- ble initiation sites for DNA replication, we examined replicating molecules for multi- a ple replication loops. Of 30 replicating molecules measured, ten were <50% repli- cated. None of these molecules (nor those > 50% replicated) had more than one repli- cation loop. DISCUSSION In Fig. 11 we summarize our results on the structure of ev-1101 and ev-1103 DNA in relation to the Hin cleavage map of the SV40 genome. As shown in the figure, ev-1101 DNA is made up of a repeating segment which is 17% of the length of L3 Cc’ Cell DNA tig ptt Le L2 Li es L4 =O L3 8 a oO ® = 10 o o €5 Za Zz L? 3.4 51 68 85 Distance of rep-i from nearest Hin site (% of SV40 DNA) Fic. 10. Analysis of replicating molecules of ev-1101 DNA cleaved once by HindIII. (a), Representative molecule that has been cleaved in one of the replicated branches. In molecules of this type the unbroken circle was equal to the length of ev-1101 genomes present on the same grid. (b), Representative molecule that has been cleaved in an unreplicated branch. In molecules of this type the circle varied from a small fraction of the length of ev-1101 DNA to nearly twice the length of variant DNA. The arrows in the models indicate the midpoint of each replication loop, taken as the origin of replication. The distance of the initiation site (rep-i) from a HindIII site, expressed as a multiple of ev-1101 DNA Hin fragment, was computed as follows: For molecules of type (a), 5x [(L1 + L2)/2 + L1 + L4J(L1 + L4 + 12); for molecules of type (b), 5x [L1 + (L2 + L3)/4]/[L1 + (L2 + L3)/ 2 + L4]. For the histogram shown on the right, the distance of rep-i from the nearest HindIII site (Hin-A -C junc- tion) is expressed as percent of SV40-DNA length, taking the distance between A-C junctions as 17% of the length of SV40-DNA. At the top of the graph are shown the two possible orientations of ev-1101 DNA near the Hin cleavage site. 66 SV40 DNA ’ Initiation ev -1101 LEE, BROCKMAN AND NATHANS Reiterated cell DNA Non- reiterated cell DNA site for DNA Replication l Hin d_ site Fic. 11. The structure of the genomes of ev-1101 and ev-1103, compared to the Hin cleavage map of SV40 (Danna et al., 1973). SV40-DNA. This segment is repeated five times in the variant genome, which mea- sures 85% of the length of SV40-DNA. Each repeating unit consists of a segment of SV40-DNA derived from map positions 0.56-0.68, comprising about 65% of the repeating unit, and a segment derived from nonreiterated or less reiterated host cell DNA, making up about 35% of the repeat- ing unit. The DNA of ev-1103 is made up of a repeating unit which is 8.8% of the length of SV40-DNA. In the major size-class of ev-1103 genomes (about 78% of the length of SV40-DNA) there are nine tandemly repeated segments. Each unit has a seg- ment of SV40-DNA derived from the region around the Hin-A-C junction (about 0.65 map-units) and makes up about “4-4 of the repeating unit. The rest of the DNA is derived from cell DNA, including reiter- ated and nonreiterated sequences. The topologic relationship of the different cell DNA sequences in ev-1103 is not entirely clear and is drawn in Fig. 11 with continu- ous reiterated and nonreiterated cell se- quences. In comparing the structure of the two variant genomes it is interesting that the SV40 segment of ev-1103 is included in the SV40 segment of ev-1101 (see below). How- ever, the cell DNA present in ev-1101 does not correspond to that present in ev-1103. The latter findings suggest that the two variants probably arose by recombinations between SV40-DNA and cell DNA at dif- ferent sites in the cell DNA. In this regard, Frenkel et al. (1974) recently characterized DNA from different series of uncloned substituted variants of SV40 by DNA- DNA hybridization and presented evi- dence for nonrandom differences in host sequences present. Concerning the SV40- cell DNA junctions in ev-1101 and ev-1103 DNA, at least one of the SV40 sites next to cell DNA is different in the two variants (the Hin-A-.cell DNA junction). The Hin- C-cell DNA junctions, however, may be identical in the two variants (at 0.68 map- units); nucleotide sequence analyses will be required to settle this point. Obviously this result would have a bearing on the question of preferred sites on SV40-DNA for integration into cell DNA. In this con- nection it is interesting to point out that Khoury et al. (1975) have inferred from transcriptional analysis on SV40-trans- formed cells that Hin-C may contain a preferred site for integration. Probably the most striking feature of the genomes analyzed is retention and repeti- SUBSTITUTED VARIANTS OF SV40 tion of that part of the SV40 genome which contains the initiation site for DNA repli- cation (Fig. 11). This is reminiscent of the emergence of variants of coliphage Q£- RNA described by Spiegelman and his co-workers (Mills et al., 1973). In both instances, short polynucleotide segments evolve which retain the initiation site(s) for replication. In the case of ev-1103 only about 3% of the SV40 genome is retained, and this segment corresponds to the previ- ously mapped initiation site. In all likeli- hood, retention and reiteration of this site have selective advantage during serial pas- sage (see below). Taking into account the structure of specific early passage variants described in the preceding paper (Brockman et al., 1975) and those reported by Khoury et al. (1974) and by Mertz et al. (1974), we suggest that evolutionary variants arise by recombination, reiteration, and selection under conditions where helper viruses sup- ply trans functions. SV40-infected cells have an active recombination system re- sulting in extensive recombination be- tween and within SV40 genomes (Tai et al., 1972; Lai and Nathans, 1974) and between SV40 and cellular DNA (Lavi and Wino- cour, 1972). Hence new genomes are formed containing deletions, duplications, rearrangements, and covalently linked cell DNA. Also, as a result of a reiteration proc- ess not presently understood, a relatively small segment of DNA can be enlarged to a size compatible with encapsidation (Da- voli and Fareed, 1974). These various proc- esses are illustrated in Fig. 12, which shows a hypothetical scheme for the gen- eration of substituted variants. An alterna- tive mechanism for amplifying viral DNA near the origin of replication has been sug- gested by Chow et al. (1974), Robberson and Fried (1974), and Folk (1974). On the diverse pool of variant genomes generated by recombination, selective pressures must operate to increase the pool of those molecules that replicate faster and are large enough to be encapsidated and hence transmitted. The selective replica- tive advantage may be due simply to reiteration of the initiation site for viral DNA replication, since no other cis func- 67 tion for replication appears necessary (see below). As a consequence, early variants often contain duplicated initiation sites (Brockman and Nathans, 1974) and late variants contain many copies of the initia- tion site in each molecule and little SV40- DNA. However, most or perhaps ail mole- cules with multiple initiation sites actually begin replication at only one of the possible sites (Fareed et al., 1974; Brockman et al., 1975; ev-1101 results). Although this could be due to a conformational change in the DNA following initiation, it could also be due to a limiting factor required for initia- tion (e.g., gene A product; Tegtmeyer (1972)) or perhaps a topological restriction. In either case the rate of initiation of a given DNA species would likely be propor- tional to the number of initiation sites per molecule. The examination of replicating mole- cules of ev-1101 confirms the expectation that DNA replication begins in the SV40 Cell DNA J Integration co A Mt, 8B OG WF KE, DC ‘- 4 -68 56 Map units | Excision eo” ev-H1O0l_ Monomer | Amplification ev-l101 Fic. 12. A scheme for the generation of substi- tuted variants with tandemly repeated segments of DNA. In the figure the formation of ev-1101 is illustrated. 68 segment near the Hin-A-C junction and terminates about half-way around the cir- cular molecule. However, within the limits of sensitivity of the hybridization experi- ments reported in Table 2, neither ev-1101 nor ev-1103 DNA contain sequences de- rived from Hin-G where normal SV40- DNA replication terminates (Danna and Nathans, 1972). These data are thus com- patible with our previous conclusion that termination of DNA replication is not spec- ified by a nucleotide sequence (Brockman et al., 1975). Further, the paucity of SV40- DNA sequences in ev-1103 (about 150 nu- cleotide pairs) makes it unlikely that this segment codes for a functioning initiator protein. Therefore, the initiation signal is probably the only cis function involved in SV40-DNA replication. (We should point out, however, that a contribution of the host sequences found in substituted vari- ants has not been excluded.) If the initiation signal is the only re- quired cis function for replication, it fol- lows that in the presence of infection by helper SV40 to supply trans functions, any double-stranded circular DNA molecule containing the SV40 signal will replicate. For example, it should be possible to con- struct a variety of replicatable DNA mole- cules by linking them to an SV40 initiation segment and then cloning in the presence of helper virus as was done for substituted variants. If these are of appropriate size, they will be encapsidated in SV40 coats; in this way potential transducing particles can be generated. Further, a plausible mechanism for SV40-stimulated cell DNA synthesis is suggested, namely, that cell DNA contains nucleotide sequences corre- sponding to the SV40 signal, either by chance and/or as a result of recombina- tional exchanges with SV40-DNA, leading to initiation at those sites. One of the objectives of our studies of SV40 evolutionary variants has been to define the initiation signal for viral DNA replication. Clearly, molecules that retain this SV40 segment in functional form with- out any other viral DNA would be useful for nucleotide sequence analysis of the initiation site. We appear to be approach- LEE, BROCKMAN AND NATHANS ing that goal with ev-1103, which has about 150 nucleotide pairs from SV40; continued evolution appears to reduce this segment further (Gutai and Nathans, unpublished observations). ACKNOWLEDGMENTS We are grateful to Dr. Thomas J. Kelly, Jr., for his help with electron microscopy and to John DeBoy and Douglas Welsh for skillful assistance. This research was supported by grants from the National Cancer Institute (No. CA11895) and the Whitehall Founda- tion. REFERENCES Anpo, T. (1966). A nuclease specific for heat- denatured DNA isolated from a product of Aspergil- lus oryzae. Biochim. Biophys. Acta 114, 158-168. Brockman, W. W., Leg, T. N. H., and Natuans, D. (1973). The evolution of new species of viral DNA during serial passage of simian virus 40 at high multiplicity. Virology 54, 384-397. Brockman, W. W., and Natuans, D. (1974). The isolation of simian virus 40 variants with specifi- cally altered genomes. Proc. Nat. Acad. Sci. U.S.A. 71, 942-946. Brockman, W. W., Ler, T. N. H., and Nartuans, D. (1974). Characterization of cloned evolutionary var- iants of simian virus 40. Cold Spring Harbor Symp. Quant. Biol. 39, in press. Brockman, W. W., Gurat, M. W., and Naruans, D. (1975). Evolutionary variants of simian virus 40: Characterization of cloned complementing vari- ants. Virology. in press. Cuow, L. T., Davipson, N., and Bere, D. (1974). Electron microscope study of the structures of \dv DNAs. J. Mol. Biol. 86, 69-89. Danna, K. J., and Natuans, D. (1972). Bidirectional replication of simian virus 40 DNA. Proc. Nat. Acad. Sci. U.S.A. 69, 3097-3100. Danna, K. J., Sack, G. H., Jr., and Naruans, D. (1973). Studies of SV40 DNA. VI. A cleavage map of the SV40 genome. J. Mol. Biol. 78, 363-376. Davis, R. W., Stmon, M., and Davinson, N. (1971). Electron microscope heteroduplex methods for mapping regions of base sequence homology in nucleic acids. Methods Enzymol. 21, 413-428. Davout, D., and Fargep, G. C. (1974). Amplification of a circular segment of SV40 DNA. Nature (Lon- don) 251, 153-155. FarEED, G. C., Garon, C. F., and Satzman, N. P. (1972). Origin and direction of SV40 DNA replica- tion. J. Virol. 10, 484-491. FareED, G. C., Byrne, J. C., and Martin, M. A. (1974). Triplication of a unique genetic segment in a simian virus 40-like virus of human origin and SUBSTITUTED VARIANTS OF SV40 evolution of new viral genomes. J. Mol. Biol. 87, 275-288. Fotk, W. R., and Wana, H.-C. E. (1974). Closed circular DNAs with tandem repeats of a sequence from polyoma virus. Virology 61, 140-155. FRENKEL, N., Lavi, S., and Winocour, E. (1974). The host DNA sequences in different populations of serially passaged SV40. Virology 60, 9-20. Hirt, B. (1967). Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol. Biol. 26, 365-369. Kuoury, G., FAREED, G. C., Berry, K., Martin, M. A., LEE, T. N. H., and Natuans, D. (1974). Charac- terization of a rearrangement in viral DNA: Map- ping of the circular SV40-like DNA containing a triplication of a specific one-third of the viral genome. J. Mol. Biol. 87, 289-301. Kuoury, G., Martin, M. A., Lee, T. N. H., and Natuans, D. (1975). A transcriptional map of the SV40 genome in transformed cell lines. Virology, 63, 263-272. Lal, C.-J., and Naruans, D. (1974). Deletion mutants of simian virus 40 generated by enzymatic excision of DNA segments from the viral genome. J. Mol. Biol. 89, 179-193. Lavi, S., and Winocour, E. (1972). Acquisition of sequences homologous to host deoxyribonucleic acid by closed circular SV40 DNA. J. Virol. 9, 309-316. Martin, M. A., Gets, L. D., Fareep, G. C., and Mustein, J. B. (1973). Reassortment of SV40 DNA during serial undiluted passage. J. Virol. 12, 748-757. Mertz, J. E., and BerG, P. (1974). Defective simian virus 40 genomes: Isolation and growth of individual clones. Virology 62, 112-124. Mertz, J. E., Carson, J., Herzperc, M., Davis, R. W., and Berc, P. (1974). Isolation and character- ization of individual clones of simian virus 40 mutants containing deletions, duplications and in- sertions in their DNA. Cold Spring Harbor Symp. Quant. Biol. 39, in press. Mitts, D. R., Kramer, F. R., and SPIEGELMAN, S. (1973). Complete nucleotide sequence of a replicat- ing RNA molecule. Science 180, 916-927. Rosserson, D. L., and Friep, M. (1974). Sequence arrangements in clonal isolates of polyoma defec- tive DNA. Proc. Nat. Acad. Sci. U.S.A. 71, 3497-3501. 69 Rozensiatt, S., Lavi, S., SiIncer, M. F., and Wino- cour, E. (1973). Acquisition of sequences homolo- gous to host DNA by closed circular simian virus 40 III host sequences. J. Virol. 12, 501-510. Sack, G. H., Jn., and Natuans, D. (1973). Cleavage of SV40 DNA by restriction endonuclease from He- mophilus parainfluenzae. Virology 51, 517-520. Sack, G. H., Jr. (1974). “The Use of Restriction Endonuclease Hpa I in the Analysis of the Genome of Simian Virus 40.” Ph.D. thesis, The Johns Hopkins University, Baltimore. Scunos, M., and Inman, R. B. (1970). Position of the branch points in replicating ) DNA. J. Mol. Biol. 51, 61-73. SuartieL, S., and Er-Er, Z. (1973). Hydrophobic chromatography: Use for purification of glycogen synthetase. Proc. Nat. Acad. Sci. U.S.A. 70, 778-781. Suarp, P. A., SuGpEN, B., and SamBroox, J. (1973). Detection of two restriction endonuclease activities. in Hemophilus parainfluenzae using analytical agarose-ethidium bromide. Biochemistry 12, 3055-3063. Smitu, H. O., and Natuans. D. (1973). A suggested nomenclature for bacterial host modification and restriction systems and their enzymes. J. Mol. Biol. 81, 419-423. Sutton, W. D. (1971). A crude nuclease preparation suitable for use in DNA reassociation experiments. Biochim. Biophys. Acta 240, 522-531. Tat, H. T., Smiru, C. A., Suarp, P. A., and Vinocrap, J. (1972). Sequence heterogeneity in closed simian virus 40 deoxyribonucleic acid. J. Virol. 9, 317-325. TEGTMEYER, P. (1972). Simian virus 40 DNA synthe- sis: The viral replicon. J. Virol. 10, 591-598. Ucnina, S., Yosuke, K., WATANABE, S., and Furuno, A. (1968). Antigen-forming defective viruses of $V40. Virology 34, 1-8. Wermovr, J. G., and Davipson, N. (1968). Kinetics of renaturation of DNA. J. Mol. Biol. 31, 349-370. YosuHuKE, K. (1968). Studies on DNA from low density particles of SV40. I. Heterogeneous defective virions produced by successive undiluted passages. Virology 34, 391-401. Yosuimort, R. N. (1971). “A Genetic and Biochemical Analysis of the Restriction and Modification of DNA by Resistance Transfer Factors.” Ph.D. dis- sertation, University of California at San Francisco, Medical Center, San Francisco.