Analysis of Herpesvirus DNA Substructure by means of Restriction Endonucleases

Transcription

1 J. gen. ViroL (t976), 30, Printed in Great Britain 243 Analysis of Herpesvirus DNA Substructure by means of Restriction Endonucleases By J. B. CLEMENTS, RITA CORTINI AND N. M. WILKIE Institute of Virology, University of Glasgow, Glasgow GII 5JR. Scotland (Accepted Io October I975) SUMMARY The rnol. wt. and molar ratios of the Hind III and Hpa I fragments of HSV-I DNA and the Eco RI fragments of HSV-2 DNA have been determined. Results obtained suggest that DNA isolated from both HSV-I and HSV-2 consists of molecules with four different sequence arrangements which are present in similar amounts. Our explanation of the cleavage patterns of these four genome arrangements with the different restriction enzymes is presented. Some of the possible implications of these four genome arrangements for genetic recombination are discussed. INTRODUCTION Sequence specific endonucleases (restriction enzymes) which cleave DNA only at specific sets of sequences have been of great value in the construction of physical maps for a variety of animal virus DNAs. For example, with SV4o (Danna, Sack & Nathans, I973), polyoma (Fried et ak I974) and adenovirus (Mulder et al ~974)- If a population of linear DNA molecules are homogeneous with respect to size and sequence arrangement, the locations of cleavage sites on each molecule will be identical and the same fragments will be produced from each molecule. If these DNA fragments are separated (this is normally done on a size basis) the molar ratio of each different fragment should be unity. DNA isolated from herpes simplex type I (HSV-I) and herpes simplex virus type 2 (HSV-2) consists of linear duplex molecules with mol. wt. of 95 to Ioo IO s (Becker, Dym & Sarov, I968; Kieff, Bachenheimer & Roizman, I97Ia; Wilkie, 1973). HSV-I and HSV-2 DNAs show approx. 5o % sequence homology (Kieff et al I971 b). In addition, molecules of HSV-I DNA have been shown to be terminally redundant (Grafstrom et al. I974; Sheldrick & Berthelot, 1974) and also contain internal inverted repetitions of sequences present at both ends of the molecules (Sheldrick & Berthelot, 1974). In the analysis of HSV DNA with restriction endonucleases, the Eco RI, HindIII and Hpa I enzymes have proved to be the most useful as a limited number of fragments with varying sizes are generated (Wilkie et al. I974; Hayward, Frenkel & Roizman, I975; Skare, Summers & Summers, 1975). However, a consistent feature of the Eco RI and Hind III fragment patterns produced with DNA isolated from different strains of HSV-I and HSV-2 has been the detection of bands present in submolar amounts (Wilkie et al. 1974; Hayward et al. 1975; Skare et al. 1975)- The present study deals with the size and molarities of the Eco RI, Hind III and Hpa I restriction enzyme fragments produced with both HSV-I and HSV-2 DNAs. Results are interpreted in terms of a model proposed by Sheldrick & Berthelot (I974) for the structure of HSV-I DNA.

2 244 J. B. CLEMENTS~ R. CORTINI AND N. M. WILKIE METHODS Baby hamster kidney cells (BHK cells). BHK 21 (CI3) cells were grown as monolayers in Eagle's medium supplemented with 1o ~ tryptose phosphate and IO ~ calf serum as described previously (Macpherson & Stoker, 1962). For preparation of unlabelled viral DNA, cells were grown in slowly rotating 80 oz bottles. For preparation of labelled DNA, infected monolayers in 50 mm diam. plastic Petri dishes were incubated in Eagle's medium containing one tenth the normal phosphate concentration in the presence of IOO #Ci/m132PO 4. Infection with herpes virus. Cell monolayers were infected with herpes simplex virus type I (HSV-I, Glasgow strain 17) or herpes simplex virus type 2 (HSV-2, strain HG 52) at low input multiplicities (o.oi to o'o3). Monolayers were incubated at 32 C for 3 days. Preparation of viral DNAs. HSV DNA was isolated from virus released from pelleted infected cells by phenol/sds extraction or by phenol/sds extraction of infected cell cytoplasm as described by Wilkie (1973). Preparations were also made from HSV-I virus particles and nucleocapsids after purification on gradients of dextran T-Io (Spear & Roizman, 1972). In some cases, DNAs were further purified by isopycnic banding in CsC1 (Wilkie, 1973). All preparations used when analysed on CsC1 gradients yielded a single virus DNA component with a density characteristic of either HSV-I or HSV-2 DNA (Halliburton, Hill & Russell, 1975). Phage lambda (R) DNA was obtained by phenol/sds extraction of CsCl-banded virus from a heat-inducible lysogenic host strain, Escherichia coli 803 (CI 857 $7). Restriction endonucieases. Hind III was isolated from Haemophilus influenzae strain d supplied by Dr S. Glover. Cells were sonicated in o.oi i-tris, 0"0I i-mercaptoethanol (ph 7"9) then centrifuged at rev/min for 2 h. Solid NaC1 was added to the supernate to a final concentration of I M and the extract applied to a Biogel A- o'5 m column equilibrated with the same buffer. Fractions containing endo- and exonuclease activity were combined and dialysed against o-oi M-potassium phosphate, Io-%t-EDTA, o.oi M-mercaptoethanol, ph 7"4, containing IO ~ glycerol. This extract was applied to a DEAE cellulose (DE 52) column and the column washed with the same buffer. A gradient of o to 0"3 M-KC1 was then applied and the Hind III appeared in the column flow through, free of contaminating exo- and endonucleases. Enzyme activity was concentrated either by ammonium sulphate precipitation or by dialysis against polyethylene glycol and stored in 50 ~ glycerol at C. Eco RI was prepared essentially as described by Yoshimori (1971). Hpa I obtained from Haemophilus parainfluenzae cells was kindly supplied by Dr R. Kamen. Enzyme digestions with Hind III and Eco RI were performed in o.oi M-tris, 0"I M-NaC1, o.oo6 M-MgCI~, ph 7"4 for I h at 37 C. Digestions with Hpa I were performed in o-oi M-tris, o'o6 M-MgC12, ph 7"4, for I h at 37 C. Mol. wt. estimates of DNAfragments. Size estimates for HSV DNA fragments were made using slab gels of 0.2 ~ or 0"3 ~ agarose which gave a reasonable separation over the tool. wt. range 2 to 30 x IO G (Wilkie et al. I974). Intact A DNA, the A DNA+Eco RI fragments (Allet et al. 1973) and the A DNA+Hind III fragments (Allet & Bukhari, 1975) were used as reference tool. wt. markers. The gel apparatus was essentially as described by Studier (1973). Slab gels were I6 cm long, 8 mm deep with a total volume of 23o ml and had 6 sample tracks each of 16 mm width. Gel sandwiches were prepared with a sheet of Visking tubing across the bottom, then bottom and sides were sealed with a thin layer of 0"9 ~ agarose before filling with 0.2

3 Herpesvirus DNA substructure 245 agarose. Setting was for I h at 20 C then the top surface was carefully overlayered with electrophoresis buffer (36 mm-tris, 3 mm-sodium dihydrogen phosphate, I mm-edta, ph 7"8, containing o. 5 #g/ml ethidium bromide). Samples (5 to Ioo #1) in electrophoresis buffer containing IO ~ (w/v) sucrose were applied directly to the gel surface, and electrophoresis was carried out for ~6 h at 4 o ma[slab at room temperature, DNA bands visualized by u.v. fluorescence (Sharp, Sugden & Sambrook, I973) were photographed with a Polaroid CU5 camera using positive/negative film type ~o5. Estimation of molar ratios of DNA fragments. Molar ratios were based on a tool. wt. of 97 I o6 for both HSV-I and HSV-2 DNAs. Two methods were used for visualization of fragments. Autoradiography. Slab gels of o'3 ~ agarose containing [a2p]-labelted fragments were dried overnight at 37 C and autoradiographs were prepared using Kodak KD54T film. Autoradiographs were scanned using a Joyce-Loebl double-beam scanning microdensitometer. Peaks from traces were excised and weighed to estimate molar ratios. DNA/ethidium bromide u.v. fluorescence. For this procedure, o'3 ~ agarose disc gels were used (2o cm long, I'5 cm internal diam. with o'5 to 1"5/zg of sample DNA) which gave slightly superior resolution compared to slab gels. Negatives of stained gels were scanned and molar ratios were calculated as described for autoradiographs. RESULTS MoI. wt. and molarities of HSV-r DNA + Hind III fragments and HSV- 2 DNA + Eco RI fragments Densitometer traces of electrophoretic separations of these fragments on o'3 ~ agarose disc gels with bands visualized by ethidium bromide staining are shown in Fig. I. On the traces, fragments are labelled alphabetically beginning with the fragment of highest mol. wt. In cases where single bands are indicated as containing more than one fragment the evidence to support this comes from data presented below on the molar ratios of bands. Mol. wt. of bands were estimated as described in Methods, and estimates are shown in Tables I and 2 together with molarity determinations. The molarity determinations shown were obtained both by autoradiography of [32P]-labelled fragments and by scanning negatives of ethidium bromide stained gels. Molarity estimates for the more intense bands tended to be lower than values predicted from the cleavage models outlined below. This effect was probably due to a non-linear response of the films. However, since the estimates made by these two methods, in general, showed good agreement and moreover were similar to estimates obtained by slicing and counting individual gel slices (data not shown) then it was assumed that there was a linear response on the film for most of the bands measured. Considering the HSV-I DNA+Hind III fragments, mol. wt. ranged from ~'9 to 23 IO n and the sum of the individual values yielded a total of , a value considerably in excess of estimates for the mol. wt. of this DNA. Molar ratios of bands ranged from approx, o'25 M (HSV-I DNA bands c, f) through values around o. 5 M (HSV-I DNA band n; HSV-2 DNA bands f, m) to ratios greater than T M from regions which were clearly composed of more than one component but which were not resolved enough to allow molarities of individual components to be estimated. The size, number or molarities of bands were not affected when a Ioo-fold excess of enzyme was used, even upon overnight digestion. Heating to 65 C in enzyme buffer containing o-oi M-EDTA followed by rapid cooling to o C prior to electrophoresis did not affect the pattern. Pre-digestion of 2 to 4 #g of DNA with 5o/zg/ml of pre-incubated pronase or

4 246 J. B. CLEMENTS~ R. CORTINI AND N. M. WILKIE (a) b, a k,j, i, h e d A I 2,9 5.0 Mol. wt. x 10-6 I I I "0 (b) a i,h,g cbj! n m J!!! I Mol. wt. x 10-6 Fig. i. Examples of traces obtained from electrophoretic separations of fragments on o"3 % agarose disc gels. (a) HSV-I DNA+HindIII fragments. (/5) HSV-2 DNA+Eco RI fragments. Bands were visualized by DNA/ethidium bromide u.v. fluorescence. of proteinase K also had no effect on the pattern. Virus was re-purified genetically by four successive cycles of picking single plaques of HSV-I strain 17 with either syncytial or nonsyncytial plaque morphology, and grown into separate stocks. The DNAs from four unrelated clones of non-syncytial virus and from two unrelated clones of syncytial virus were found to have the same Hind III pattern as shown in Fig. t. Thus, six genetically re-purified virus stocks reproduced the same pattern of major and minor bands as that obtained with strain 17 stock virus. Explanation of HSV-I DNA + Hind III and HSV-2 DNA + Eco Rl fragment patterns in terms of a model proposed for the structure of HSV-I DNA The presence of bands in less than molar amounts generated from a DNA of uniform size implies that the distribution of enzyme sites on the DNA is non-uniform with respect to the population of molecules restricted. However, an explanation of these molarity values can be provided in terms of a model for the structure of HSV-I DNA. Sheldrick & Berthelot (i 974)

5 Herpesvirus DNA substructure 247 Table I. Molecular weights and molar ratio values for the HSV-I DNA + Hind lli fragments* Fragment Mol. wt.(xlo-0 ~oftotaldna Molarity a,b 23'o±o'4 26'9±0"3 1.12±o'o2 C 20"5±0" 5 5"0±0"3 0"28~0'02 d e I7"5±o'2~ I7-o±o'2) 13"z±o'4 0"73±0-02 f I2"5±o'2 3"5±o'z 0"27±o'02 g 8-3±o-I) h 7-5+o'I~ 3o-7±o-8 3"88±o-t4 i-k 7-O~0"2) 1 5'5+O'I~ m 5"o~o'I! Io'9±o'4 1'99±o"o8 n 4"5±o-1 3"o±o'3 o'64±o'o7 o 2.9±0"1 3.7±0.3 1"22+0'II p I'9±0"I 1'8±0"2 0"9I±O'II No. ofsamples * The values shown are the mean estimates for the number of samples indicated together with standard error estimates. Molar ratio estimates were calculated from estimates obtained by autoradiography of la2pl-labelled DNA and from values obtained by DNA/ethidium bromide u.v. fluorescence. Table 2. Molecular weights and molar ratio values for the HSV-2 DNA + Eco Rl fragments* Fragment Mol. wt.( Io-0 ~oftotaldna Molarity a b,c I8"5~o'7 I7-6±o ±o'8 lo.9±1"o o.8±o'o5 o.6±o.o6 d,e f 14"5±o'7 IO'9±o'4 IO'3±o'5 6-3±o-6 o.6±o.~ o'5±o'05 g,h,i 9"9±o'3 24-9±o'8 2.4±o.Iz j 8"6±0'3 IO'9±o"4 1'2±o"o7 k, 1 m 7"5±0"3 4'I±O"4 I3"8±o"7 2"7±0"2 1'7±o"12 0-6±0'08 n 2"9±0"4 z'6±o'4 o'8±o-i9 o NDt 2-3±o-I No. ofsamples * Mean estimates are shown together with standard error estimates. Molar ratio values were calculated from estimates obtained by autoradiography of [a2p]-labelled DNA and from values obtained by DNA/ ethidium bromide u,v. fluorescence. 1" Not done. proposed that the DNA consisted of two unique regions each flanked by two repeated regions which were inverted relative to each other. In this structure, the two terminal repeats are present adjacent to each other internally within the molecule in an inverted form. Inherent in this configuration is the possibility that either terminal repeat can pair with its internal complement. Assuming this pairing then, as indicated by Sheldrick & Berthelot (1974), a reciprocal recombination event will result in the inversion of the entire unique region between the participating repeats. A high frequency of occurrence for this event, together with no preference for recombination between two of the four repeats wii1 generate the four genome arrangements shown in Fig. 2. In this figure both the long and short unique regions in genome arrangement (1) have been designated as L, S, and genome arrangement (4), with both unique regions inverted relative to arrangement (1), has been designated as l, s. The terminal repeats and inverted repeats are numbered appropriately. Sheldrick &

6 248 J.B. CLEMENTS, R. CORTINI AND N. M. WILKIE Restriction enzyme fragments: no cuts in the repeated regions 1234 a e a e b b e e b+d. f c 4'3'2'1'r 1'7'6' 5' b+c f d L, S I, a+d- f c L,s 3 a + e ~ f d i, S I 4 1, s 1.0 molar fragments: e, f 05 molar fragments: a, b, c, d 0.25 molar fragments: b+d. b+c, a+d,a+c Fig. 2. A model which illustrates the four possible genome arrangements which can result from inversions of the long and short unique regions of HSV DNA. Terminal repeats, inverted internal repeats and the terminal redundancy of the DNA are shown. The molar ratios of the fragments produced, and the size relationships of the submolar fragments are indicated for a restriction enzyme which cleaves the four genome arrangements only in both of the unique regions. Berthelot (~974) proposed that all four repeats were identical. However, electron microscopic studies (Wadsworth, Jacob & Roizman, I975; J. B. Clements & H. Delius, unpublished data) indicate that only the two repeats flanking each unique region are identical. This also is shown in genome arrangement (1) together with the terminal redundancy of the DNA. Assuming that the four genome arrangements in Fig. 2 are present in equal amounts then, for a restriction enzyme that cuts only in both of the unique regions and not in the repeated regions, the following deductions can be made regarding the molar ratios of the fragments produced and the size relationships of the submolar fragments. (i) There will be I M fragments (e, f) produced from the unique regions of the DNA which are unaffected by any inversions of these regions. The number of these fragments will depend upon the number of enzyme sites. (ii) There will be four o'5 M fragments (a, b, c, d) which are generated from the ends of the molecules. The mol. wt. of each o. 5 M fragment must be equal to or greater than the mol. wt. of the smallest repeat. (iii) There will be four o'25 M fragments (b + d, b + c, a + c, a + d) each of which spans the two internal repeats. The tool. wt. of each o'25 M fragment must therefore be greater than the size of the two internal repeats. Each o.25 M fragment will be composed of two 0"5 M fragments hence, the sum of the tool. wt. of the o'25 M fragments will equal twice the sum of the mol. wt. of the 0'5 M fragments. (iv) The mol. wt. of a genome can be calculated from this model as: E I M fragments +?~ ½(o" 5 M fragments)+ E ¼(o-25 M fragments). As both the HSV-I DNA+Hind III and the HSV-2 DNA+Eco RI patterns contained bands with molar ratios of approx, o'25 and o. 5, we have interpreted our molarity values in the light of the model shown in Fig. 2. Our postulated molarity values are shown in Table 3 together with the obtained values. When the six possible combinations of the sum of two 0. 5 M mol. wt. are calculated, according to the model four of these should equal the obtained mol. wt. of the o.25 M bands. These calculated values are shown in Table 4 together

7 Herpesvirus DNA substructure 249 Table 3. Obtained and postulated molar ratio values of the HSV-I DNA +Hind and HSV-2 DNA + Eco RI fragments Sample HSV-I DNA+Hind III HSV-2 DNA+E o RI Fragment Molarity obtained Molarity postulated* a, b I.i2 I +0"2 5 c o'28 o'25 d, e o'73 o'5 +o"25 f 0"27 0"25 g, h, i, j, k 3'88 I + I + I +o'5+o'5 1, m 1"99 I+i n o'64 o'5 o I "22 I p o'91 I a 0"8 I b, c 0.6 0'25 + 0'25 d, e 0"6 0'25+0"25 f o'5 o'5 g, h, i 2"5 1+I+O'5 j 1"2 I k, 1 1"7 1+o'5 m o'6 o'5 n 0.8 i * Postulated molar ratio values were interpreted in terms of the cleavage model shown in Fig. 2. III Sample HSV-I DNA+Hind III Table 4. Molecular weights of o'5 M and 0"2 5 M fragments Obtained tool. wt. of CaIculated mol. wt. of Obtained mol. wt. of o-5 M fragments o'25 M fragments o'25 M fragments ( I0 6) (X10 6)* ( i0-n) d = I7"3 d+g = 25"6 a = 23"0 g = 8"3 d+h = 25'ot c = 2o'5 h -- 7"7 d+n = 2I-8? e = 17"o n --= 4"5 g+h = I6"Ot f = I2"5 g+n = I2'8t h+n = 12.2 Mol. wt. genome (x lo-6)$ = 6o-8(Z I M)+ I8-9[Z ½(O'5 ~f)]+ I8"25[Z ¼(0"25 M)] = 97"9 lo 6. HSV-2 DNA+Eco RI f--- IO'9 f+g = 20'8 b = I7"6 g --= 9"9 f+k = I8"4t c = I7"6 k = 7"5 f+m= I5"Ot d = I4"5 m -- 4"1 g+k = I7"4t e = 14'5 g+m= I4'O~" k+m= II.6 Mol. wt. genome ( IO 6)+ = 57.3(y. I M)+ I6.2[X ½(0"5 M)]+ 16"O5[X ¼(0"25 M)] = 89'6 lo 6. * The tool. wt. of the o'25 M fragments are calculated from size values of the o'5 M fragments. t Marks the four values in the two sets cf calculated o'25 M sizes which satisfy the requirements of the model shown in Fig. 2. $ Genome size estimates calculated from the formula shown in Results. with the obtained values. The two sets of four indicated values have sizes very similar to the obtained values. Furthermore, from the model, the composition of each of the o.2 5 M bands in a series should be of the form b + d, b + c, a + c, a + d. Both sets of indicated values are of this composition. Also shown in Table 4 are the calculated mol. wt. for the HSV-I and HSV-2 genomes obtained using the formula shown above. :7 VrR 3

8 250 J.B. CLEMENTS, R. CORTINI AND N. M. WILKIE 8-3 9"9 8"3 9"9 17,0 10' " l L, S I L, s "3 +l 4"5 7"0 1, S " V4-5 1, s Fig. 3. The four genome arrangements illustrated in Fig. 2 are shown with the obtained mol. wt. of the 0'5 M fragments inserted. Values above the lines refer to the HSV-I DNA + Hind III fragments, values below the lines are for the HSV-2 DNA+Eco RI fragments. The composition of the o'25 M fragments can be seen. d, c d, c s, r, q, p, o, n m kj i h Ill It r, :/ ) f I I I I I Mol. wt Fig. 4- A trace obtained from an electrophoretic separation of the HSV-I DNA+Hpa I fragments. Bands were visualized by DNA/ethidium bromide u.v. fluorescence. In Fig. 3 the sizes of the HSV-I and HSV-2 0"5 M fragments have been substituted into the four genome arrangements shown in Fig. 2 and the composition of the o'25 M fragments is shown. Mo[. wt. and molarities of the HSV-I DNA + Hpa I fragments A densitometer trace of an electrophoretic separation of fragments on a o'3 ~ disc gel is shown in Fig. 4. The insert on the trace shows a better separation of the high tool. wt. fragments obtained from another gel and the criteria for labelling bands is as described previously. Mol. wt. and molarities of bands are shown in Table 5. No o'25 M bands were detected; however, there was clear indication that bands were present in submolar amounts. For example, bands a and 1 were present in less than molar amounts and bands c, d and f, g

9 Herpesvirus DNA substructure 25I Table 5. Molecular weights and molar ratio values for the HSV-I DNA + Hpa I fragments* Fragment Mol. wt.( io 0 ~oftotaldna Molarity a b 137±o'2~ I2.O±O.i! 15'5±o'4 I'16±o'o3 c,d 99±o2 I4"5±o'2 I'4O±O'O3 e 8-o±o.1 8"1±o'3 o'97±o'o4 Lg 69±o'I lo'7±o'2 P48±o'o4 h 58±o'I 6"9±o'2 I'I4±o'o4 i 5z±o'I 5"8±o'3 1"o7±o"o6 j 4"4±o2 5"3±o'4 I'I5± o'io k 4-o±oq 5"4±o'4 I'29±o'Io 1 3-7±o'2 3"o±o'4 o'77±o'io m 3-1±o'I 4"5±o'3 1"39± o'io n-s 2"5~0"2 I4"2±0'4 5"68±0'27 t,u ~7±o'3 3"4±o'4 I'9z±o'4o v o-7±o-4 NDt w 0-6±0"4 ND x ND ND y ND ND No. ofsamples * Mean estimates together with their standard errors are shown. Molar ratio values were calculated with estimates obtained by autoradiography of [a2p]-labelled DNA and from values obtained by DNA/ethidium bromide u.v. fluorescence. "~ Not done. were approx. 1.5 M. We feel that the Hpa I data is best explained in terms of the model outlined below. Explanation of the HS V-I DNA + Hpa I fragment pattern in terms of a model proposed for the structure of HSV-I DNA Assuming the four possible genome arrangements described previously, the predictions for the molar±ties of fragments produced by a restriction enzyme that cuts only in those two repeats that flank a unique region and also within both unique regions are shown in Fig. 5. Again, this model assumes that the four gen ome arrangements are present in equal amounts. The predictions are as follows: (i) There will be I M fragments (i, e, g, h, f) generated from the unique regions and from the end repeat that is cut by the enzyme. (ii) There will be four o'5 M fragments two of which (c, d) are generated from the ends containing the uncut repeats and two of which (h + c, h + d) are generated from the cut internal repeat to the first cut in that unique region which has no cuts in its flanking repeats. From Fig. 5 it can be seen that genome arrangements (I) and (3) and (2) and also (4) are the same as inversions of the unique region containing the flanking cut repeats and have no effect on the molar ratios of fragments produced from that side of the molecule. The distance h (Fig. 5) from the end of a molecule to the cut (or first cut if there is more than one) in the repeat can be calculated. This distance is obtained from the six possible values calculated by subtracting the sizes of each o'5 M fragment from each other. The distance h is indicated by those two values out of the six possible values which are the same size and which are comprised of all four o'5 M fragments. (iii) There will be no o'25 M fragments. (iv) The mol. wt. of a genome can be calculated from this model as: Z I M fragments+ Z ½(0"5 M fragments). I7-2

10 252 J.B. CLEMENTS, R. CORTINI AND N. M. WILKIE Restriction enzyme fragments: cuts in the repeated regions h i e g h+d f I h i e W' g h+c f-~, I h g e i ~, h+d I f h g e i v h+c t f 1"0 molar fragments: i, e, g, h, f 1 = 3 0,5 molar fragments: c, d, h+d, h+c 2= molar fragments: none c C d d 1 L,S 2 L, s 3 1, S 4 I, s Fig. 5. The molar ratios of the fragments produced and the composition of the 0"5 M fragments are illustrated for a restriction enzyme which cleaves the four genome arrangements within both unique regions and also within those two repeats that flank a unique region. The model shows cuts in the two repeats which flank the long unique region, in terms of the model these cuts could equally well be in those repeats which flank the short unique region. Table 6. The obtained molar ratios of the HSV-I DNA +Hpa I fragments together with postulated molar ratio values interpreted in terms of the cleavage model shown in Fig. 5 Sample HSV-~ DNA+Hpa I Fragment Molarity obtained Molarity postulated a, b ~-i6 i +0"5 c, d 1"4o I+o'5 e 0"97 i f, g I'48 I +0"5 h I'I4 i i i "07 I j I'I 5 I k i "29 I 1 o'77 0"5 m I "39 t n-s 5'68 t+i+i+i+i+i t, u I " i Our postulated molarity values for the HSV-I DNA + Hpa I fragments together with the obtained values are shown in Table 6. Fig. 6 shows the calculation of the distance h. In the case of these data there are two possible distances for h and the four calculated values containing the two possible distances are indicated appropriately. Also shown in Fig. 6 are the genome models obtained with the o'5 M fragments and the two mean values for h. As, from Fig. 4, genome models 0) and (3) and also (2) and (4) are similar, only two genome arrangements are shown for each value of h. Fig. 6 also shows the calculated genome mol. wt. based on the formula shown above. In Fig. 6 the two repeats with cleavage sites indicated are those flanking the short unique region. However, as outlined in the Discussion, it seems more likely that the cuts are in the repeats spanning the long unique region.

11 Herpesvirus DNA substructure r 3.5 ~L, S ~, I 1, s ] 3-7 ~ L,S 3' I 6.5 l,s J Obtained mol. wt. of Estimation of distance h Sample 0"5 M fragments ( lo -6) from Fig. 5 ( IO-~) HSV-I DNA+Hpa I a "7 a-c = 3"8* c --- 9'9 a-f = 6"8** f = 6'9 a-e = lo-o "7 c-f = 3"0 C--I "2** f--i = 3'2* Mol. wt. genome = 74.3(Y~ I M)+ I7"1[E ½(0"5 M)] = 91'4 X I06. Fig. 6. The mol. wt. of the o'5 M HSV-I +Hpa I fragments are shown together with estimations of the distance h (see Fig. 5)- The two possible sets of values for h are indicated and the two mean values for h are inserted into the four possible genome arrangements. DISCUSSION Results obtained with HSV-t DNA preparations suggest that the four possible genome arrangements resulting from inversions of the long and short unique regions of the DNA are present in approx, equal amounts. Evidence for these four genome arrangements was obtained with DNA isolated from six genetically re-purified HSV-I virus stocks. It appears that HSV-2 DNA preparations also contain these four possible genome arrangements which indicates a similar type of genome structure for HSV-t and HSV-2 DNAs (cf. Sheldrick & Berthelot, I974, and our unpublished observations). The fragment sizes were such that molar ratio estimates could not be made for all those species which were apparent from a gel scan. However three 0"5 M fragments and three o"25 u fragments can be seen directly from the traces of the HSV-I+Hind III fragments (Fig. I a). Similarly, three 0. 5 M fragments and four o'25 M fragments can be seen from traces of the HSV-2 DNA + Eco RI fragments (Fig. I b). It should be emphasized that the interpretation of the results in terms of the cleavage models presented in Figs. 2 and 5 depends not only on detection of fragments present in submolar amounts but also on the size and size relationships of the submolar fragments. Sheldrick & Berthelot (1974) estimated the size of the four HSV-I DNA repeats as 4"5 Io 6 daltons with sizes of 76 Io 6 and 11 lo 6 daltons for the long and short unique regions respectively. Somewhat different estimates were obtained by Wadsworth et al. (I975). They estimated that each of the two repeats flanking the long region comprised 6 ~ of the DNA contour length and each of the two repeats flanking the short region comprised 4'3 ~ of the DNA contour length. Their size estimates for the long and short unique regions were 7 and 9"4 ~ of the DNA contour length. In terms of the cleavage model proposed for HSV-I DNA+Hind III, these repeat size estimates imply a minimum size of some 4 lo6 daltons for the o'5 M fragments and a

12 254 J.B. CLEMENTS, R. CORTINI AND N. M. WILKIE minimum size of IO x io 6 daltons for the 0"25 M fragments. The size values obtained for the submolar fragments were all within these limits. The size of the HSV-2 + Eco RI submolar fragments also fall within these limits set with data obtained with HSV-I DNA, and this may imply some similarity in repeat sizes between HSV-I and HSV-2 DNAs. The HSV-x DNA+Hpa I data imply that all four repeats are not identical. From Fig. 6 the smallest 0. 5 M fragment which must span the two repeats which are not cleaved is 3"7 x lo 6 daltons. Therefore, the two repeats which are not cleaved with Hpa I are more likely to be those which flank the short unique region. Furthermore, in view of a size of 6 x Io 6 daltons for the repeat flanking the long unique region the distance h (Fig. 5) from a molecule end to the cut within a repeat is more likely to be 3"5 x Io 6 daltons rather than 6"5 x Io 6 daltons. It is possible however that HSV-t strain I7 DNA does not have exactly the same repeat sizes as DNA from the HSV-~ strains F1 and Justin examined by Wadsworth et al. (I975). Based upon the contributions of the molar and submolar Hind III fragments, a genome mol. wt. of 98 x io 6 was obtained for HSV-I DNA. This value agrees well with a recent size estimate obtained with electron microscopic data (Wadsworth et al. ~ 975). Size estimates for HSV-I DNA obtained with the Hpa I fragments and for HSV-2 DNA obtained from the Eco RI fragments were somewhat lower. However, these size estimates did not include some of the smaller fragments for which accurate size estimates could not be made. In the case of the HSV-2 + Eco RI fragments these smaller fragments comprised more than 2 ~ of the DNA mass (Table 2). Inversions of the unique DNA regions can occur, as described previously, by intramolecular recombination events. Once molecules with both unique regions relatively inverted are generated, further relative inversions of unique regions can occur by intermolecular recombination events. The nature of these intermolecular recombination events and the possible structures of recombinant DNA molecules have been analysed (J. H. Subak-Sharpe, J. B. Clements & N. M. Wilkie, unpublished data). Here we present the consequences of the four genome arrangements for the measurement of recombination distances derived from intermolecular crosses. Highly efficient genetic recombination has been shown to occur in HSV-I systems (Brown, Ritchie & Subak-Sharpe, 1973; Schaffer, Tevethia & Benyesh-Melnick, ~974). There are no obvious consequences for recombination between markers within either the long or short unique regions; however, whenever viable recombinants arise by recombination events involving loci in both the long and short unique regions the apparent recombination distances can be determined as follows: In Fig. 2 genome arrangement (1) let the long unique region consist of a number of loci at distances/1... In from the terminal repeat flanking the long unique region (TRL). Similarly, let the short unique region consist of loci at distances sl...s~ from the terminal repeat flanking the short unique region (TRs). Let the internal inverted region have a length of IR~+IRs. Consider recombination between loci l~ and Sb. In terms of the four possible genome arrangements shown in Fig. 2, the distances between these loci can be written as" ln- l, + IRz + IRs + s~- sb, l,- l, + IRL + IRs + sb, l~ + IRL + IRs + sm- so, /~+IRL+IRs +sb. (0 (2) (3) (4)

13 Herpesvirus D NA substructure 255 Assuming that the four genome arrangements are present in equal amounts, the average distance between the loci I. and l~ will be: ¼[(0+(2)+(3)+(4)] = ½(/.+s~)+IR~+IR~. That is, every locus in the long unique region would have the same apparent map distance from every locus in the short unique region. This map distance would be equivalent to 5o of the total length of the genome and therefore less than the map distance between loci like /1 and l.. This consequence of the four genome arrangements for genetic mapping studies illustrates just one of what might be many complexities inherent in this unusually structured animal virus genome. We wish to thank Professor J. H. Subak-Sharpe for the study on the implications of the HSV genome structure in genetic recombination and for his continued interest in this work. We thank Dr D. A. Ritchie for maintaining the bacterial strains used for isolation of restriction enzymes. We acknowledge the expert technical assistance of Mr Alan Revill. R. C. acknowledges receipt of a European Molecular Biology Organisation Fellowship. REFERENCES ALLET, B. & BHKHARI, A. L 0975)" Analysis of bacteriophage Mu and 2-Mu hybrid DNAs by specific endonucleases. Journal of Molecular Biology 92, ALLET, B., JrPPESrN, P. G. S., KATACARI, K. J. & DELIHS, H. (1973). Mapping the DNA fragments produced by cleavage of DNA with endonuclease RI. Nature, London z4i, I2o-I23. BECKER, V., DYM, H. & SAROV, L 0968). Herpes simplex virus DNA. Virology 36, I BROWN, S. M., RITCHIE, D. A. & SUBAK-SHARPE, J. H. (I973). Genetic studies with herpes simplex virus type I. The isolation of temperature-sensitive mutants, their arrangement into complementation groups and recombination analysis leading to a linkage map. Journal of General Virology i8, DANNA, K. J., SACK, G. H. & NATHANS, D. (I973)- Studies of simian virus 4o DNA. VII. A cleavage map of the SV 4o genome. Journal of Molecular Biology 78, FRIED, M., GRIFFIN, B. E., LUND, E. & ROBBERSON, D. L. (1974). Polyoma virus - a study of wild-type, mutant and defective DNAs. Cold Spring Harbor Symposia on Quantitative Biology 39, GRAFSTROM, R. H., ALWINE, J. C., STEINHART, W. L. & HILL, C. W. 0974). Terminal repetitions in herpes simplex virus type I DNA. Cold Spring Harbor Symposia on Quantitative Biology 39, IqALLmtJR~:ON, I. W., HILL, E. A. & RCSSELL, G. J. 0975). Identification of strains of herpes simplex virus by comparison of the density 6f their DNA using the preparative ultracentrifuge. Archives of Virology 48, I57-t68. HAYWARD, G. S., ERENKEL, N. & ROIZMAN, B. (1975). Anatomy of herpes simplex virus DNA: strain differences and heterogeneity in the location of restriction endonuclease cleavage sites. Proceedings of the National Academy of Sciences of the United States of America 72, 1768-I772. meff, E. D., BACHENHEIMER, S. L. & RO~ZMAN, B. (I971a). Size, composition, and structure of the DNA of subtypes 1 and 2 herpes simplex virus. Journal of Virology 8, 125-I32. KIEFF, E., UOVER, B., BACHENHEIMER, S. r. & ROIZMAN, B. 0971b). Genetic relatedness of type I and type 2 herpes simplex viruses. Journal of Virology 9, MACPHERSON, I. & STOKER, M. (1962). Polyoma transformation of hamster clones - an investigation of genetic factors affecting cell competence. Virology x6, I47-I5L MULDER, C., ARRAND, J. R., DEHUS, ri., KELLER, W., PETTERSSON, U., ROBERTS, R. J. & SHARP, P. A. 0974). Cleavage maps of DNA from adenovirus types 2 and 5 by restriction endonucleases Eco RI and Hpa I. Cold Spring Harbor Symposia on Quantitative Biology 39, 397-4oo. SCHAFFER, P. A., TEVETHIA, M. J. & BENYESH-MELNICK, M. 0974). Recombination between temperaturesensitive mutants of herpes simplex virus type I. Virology 58, SHARP, t'. A., SUGDEN, B. & SAMBROO~:, J. 0973)- Detection of two restriction endonuclease activities in Haemophilusparainfluenzae using analytical agarose-ethidium bromide electrophoresis. Biochemistry x2, 3o55-3o63. SHELDRICK, 1". & BERTHELOT, N. 0974). Inverted repetitions in the chromosome of herpes simplex virus. Cold Spring Harbor Symposia on Quantitative Biology 39, SKARE, J., SUMMERS, W. I". & SUMMERS, W. C. (I975)- Structure and function of herpesvirus genomes. I. Comparison of five HSV-I and two HSV-2 strains by cleavage of their DNA with Eco RI restriction endonuclease. Journal of Virology x 5,

14 256 J.B. CLEMENTS, R. CORTINI AND N. M. WILKIE SPEAR, P. 6. & ROIZMAN, B. (I972). Proteins specified by herpes simplex virus. V. Purification and structural proteins of the herpes virion. Journal of Virology 9, I STtJDmR, F. W. (2973). Analysis of bacteriophage T7 early RNA's and proteins on slab gels. Journal of Molecular Biology 79, WAOSWORT, S., JACOB, R. J. e~ ROrZMAN, a. (I975). Anatomy of herpes simplex virus DNA. II. Size, composition, and arrangement of inverted terminal repetitions. Journal of Virology xs, t487-i497. W2LKIE, N. M. (I973). The synthesis and sub-structure of herpesvirus DNA: the distribution of alkali-labile single strand interruptions in HSV-2 DNA. Journal of General Virology 2x, WILKIE, N. M., CLEMENTS, J. B., MACNAB, J. & SUBAK-SHARPE, J. H. (2974). The structure and biological properties of herpes simplex virus DNA. Cold Spring Harbor Symposia on Quantitative Biology 39, YOSrI2MORL R. N. (297I). Ph.D. thesis, University of California, San Francisco Medical Center. (Received I7 September I975)