Gene, 49 (1986) 199-205

Elsevier

GEN 01842

199

Linear DNA replication: inverted terminal repeats of five closely related ExherMia coli bacterio- phages

(Lipid-containing coliphages; PRDl; PR4; PR5; L17; PR722; terminal repeats; DNA replication initiation; conserved sequences)

Harri Savilahti and Dennis H. Bamford *

Department of Genetics, University of Helsinki, Arkadiankatu 7, SF-00100 Helsinki (Finland) Tel. (0)4027298

(Received August 26th, 1986)

(Accepted October 6th, 1986)

SUMMARY

The closely related lipid-containing bacteriophages PRDl, PR4, PR5, PR722 and L17 isolated from different parts of the world have double-stranded DNA genomes which replicate in a linear form. The nucleotide (nt) sequences of the genome termini of these viruses reveal 110-l 1 I-bp-long inverted terminal repeats (ITRs). Both ends of the viral DNA are identical. The first 18 bp and the last 35 bp of the ITRs are totally conserved in all viruses. Between these conserved nt sequences there is a variable sequence, which enables us to divide the phages into two groups. Comparison of the virus ITRs led also to the identification of a IO-bp-long A + T stretch, where the only changes observed were transversions between A and T. The tern&i of the PRD 1 virus family genomes exhibit sequence similarities to those of $29 and Cp-1 families.

INTRODUCTION

The lipid-containing bacteriophages PRD 1, PR4, PRS, L17 and PR722 are closely related; they have been isolated in the U.S.A., Australia, Canada, the U.K. and South Africa, respectively, and all have a similar genome size and share a quite uniform restric- tion map and protein pattern (Bamford et al., 198 1; Coetzee and Bekker, 1979). They infect a variety of Gram-negative bacteria, including Escherichia coli

* To whom correspondence and reprint requests should be

addressed

Abbreviations: bp, base pair(s); ds, double-stranded; ITR,

inverted terminal repeat; nt, nucleotide(s); PolIk, Klenow (large)

fragment of E. coli DNA polymerase I; RF, replicative form.

and Salmotiella typhimurium, harboring an appropri- ate plasmid (Olsen et al., 1974). Two of the phages (PRDl and PR4) have been studied in more detail in order to develop a membrane assembly model (for PRDl see Mindich et al., 1982a,b; Bamford and Mindich, 1982; for PR4 see Lundstrom et al., 1979; Davis et al., 1982; Muller and Cronan, 1983; Davis and Cronan, 1983). The phage particles have an outer protein coat. Inside this layer is a membrane vesicle which encloses the linear dsDNA genome of about 14700 bp (Barnford and Mindich 1982, McGraw et al., 1983).

Characteristic of viral DNA replicating in a linear form is a covalently bound protein associated with the 5’-termini of the DNA. This kind of terminal protein is found in adenoviruses (Desiderio and

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

200

Kelly, 1981); bacteriophages $29 (Hermoso and Salas, 1980); Cp-1 (Garcia et al., 1983) and PRDl (Bamford et al., 1983). Recently, linear plasmids with blocked 5’-termini have also been reported in prokaryotic and eukaryotic organisms (Kemble and Thompson, 1982; Sor et al., 1983, Hirochica et al., 1984). The function of the terminal protein is to serve as a primer in replication. During initiation, the terminal nt is covalently linked to the protein, thus creating a free 3’ hydroxyl group for the DNA polymerase (for adenovirus see Lichy et al., 198 1; for $29 see Watabe et al., 1982; Penalva and Salas, 1982; for PRDl see Bamford and Mindich, 1984; for Cp-1 see Garcia et al., 1986). All linear DNA replication systems so far studied have ITRs of different lengths (for adenoviruses see Alestrlim et al., 1982; $29 family, Yoshikawa et al., 1985; Cp-1 family, Escarmis et al., 1985; linear plasmids, Levings and Sederoff, 1983; Sor et al., 1983; Hirochica et al., 1984; Hishinuma et al., 1984). The ITRs (or part of them) seem to function as replication origin.

Considerable interest has recently been focused on the understanding of the function of the ITRs in linear DNA replication. We report here the com- parison of the ITRs of the genomes of five closely related coliphages.

MATERIALS AND METHODS

(a) Nucleoside triphosphates and enzymes

The labeled nucleoside triphosphates, [ IX-~~P]- dCTP, [ a-32P]dGTP (approx. 3000 Ci/mmol) and [ a-35S]dATP (approx. 600 Ci/mmol), were obtained from Amersham or New England Nuclear. Restric- tion endonucleases were obtained from Boehringer Mannheim or Bethesda Research Laboratories. PolIk, T4 ligase, and pronase were from Boehringer Mannheim.

(b) Preparation of phage DNA

The phages PRDl, PR4, PR5, L17 and PR722 (Barnford et al., 1981) were grown on S. typhimurium LT2 harboring the plasmid pLM2 (Mindich et al., 1976), concentrated, and purified as described previ-

ously (Bamford et al., 1981). DNA was extracted with or without pronase treatment, as described by McGraw et al. (1983).

(c) Alkali treatment and cloning of terminal restric- tion fragments

Terminal restriction fragments were treated with 0.1 M NaOH, as described for $29 by Escarmis and Salas (198 1). Alkali-treated and reannealed restric- tion fragments were filled-in with dNTPs with PolIk, and blunt-end ligated to the M13mp18 sequencing vector (Yanisch-Perron et al., 1985) which was digested with Hind11 restriction endonuclease. The standard DNA manipulations were done according to Maniatis et al. (1982).

(d) Labeling of terminal fragments

In preliminary experiments it was shown that the 3 ‘-termini of the protease-treated PRD 1 DNA could be labeled with [a-32P]dCTP by PolIk or T4 poly- merase, using the exchange reaction, since it was already known that the terminal 5’ nt for these phages is G (Bamford and Mindich, 1984). This strategy was applied for the specific labeling of the isolated terminal fragments. Because the terminal 3’ nt of the phage genome was C, and the 3’ nt at the other end of the AluI fragment was G, it was possible to label the terminal restriction fragments specifically at each end with [ a-32P]dCTP or [ a-32P]dGTP. The same strategy was used to label the terminal @aI1 fragments at their HpaII sites with [a-32P]dGTP. This approach made additional DNA cutting or strand separation during Maxam-Gilbert sequencing unnecessary.

(e) Nucleotide sequence analysis

Nucleotide sequence determinations were per- formed either by the method of Maxam and Gilbert (1980) or by Ml3 dideoxy sequencing, as described by Messing et al. (1981). Gels containing 6% or 25 % polyacrylamide and 8 M urea were prepared as described by Sanger and Coulson (1978). Also, 30%

polyacrylamide gels without urea were used to deter- mine the very last nt.

201

RESULTS AND DISCUSSION

(a) Isolation of terminal DNA fragments

The inability of protein-linked DNA to enter DNA gels enabled us to identify the terminal DNA frag- ments of the phage genomes. Phage DNAs were isolated with and without protease treatment and subjected to A/u1 digestion. The digests were analyzed in agarose and polyacrylamide gels. The DNA from all phages showed one terminal fragment migrating at a position of approx. 260 bp (Fig. 1B). Phages PRDl, PR4, PR5 and PR722 had another approx. 2300-bp-long terminal fragment (Fig. lA), whereas the other terminal fragment for L17 was found to be about 300 bp (Fig. 1B). Since the

Fig. 1. Identification of the terminal Ah1 fragments of the

genomes of phages PRDl, PR4, PR5, L17 and PR722. Panel A

represents a 0.8% agarose gel and panel B represents an 8%

polyacrylamide gel. The same samples were electrophoresed in

both gels. The arrows indicate the locations of the terminal

fragments. In the following list protease-treated and untreated

samples are denoted by + and - , respectively. Lanes (1) and

(12), M, standard; (2), PRDl + ; (3), PRDl - ; (4), PR4 + ; (5),

PR4-; (6), PR5+; (7), PR5-; (8), L17+; (9), L17-; (lo),

PR722 + ; (1 I), PR722 - The M, standards (lanes 1 and 12) in

panel A are phage 1 DNA digested with EcoRI-HindHI; and in

panel B pBR322 digested with HpaII.

2300-bp fragments were too long to be sequenced entirely from both ends, they were isolated from the gel and digested with HpaII. The gel analysis of these digests revealed a 315-bp terminal fragment for PRD 1. This was identified as a terminal fragment by comparing the + / - protease-treated HpaII- digested total genome of PRDl. Phages PR4, PR5 and PR722 showed a 3 15-bp doublet in HpaII

digestion. Using HpaII + RsaI double digestion, it was possible to cut the non-terminal 3 15-bp fragment without affecting the terminal fragment. All terminal fragments were isolated.

(b) Sequencing the terminal restriction fragments

The terminal restriction fragments were sequenced using the Maxam and Gilbert (1980) method for all five phage genomes. To verify the ultimate terminal nt, we attempted to clone the terminal HpaII frag- ments of phage PRDl (315 bp and 420 bp). Treat- ment with NaOH was used to remove the residual amino acids that remain after protease digestion. After renaturation the fragments were filled with dNTPs using PolIk. Cloning into M 13mp18 yielded only the 3 15-bp insert in one orientation. This might relate to the fact that McGraw et al. (1983) were unable to clone the extreme right-hand HaeII frag- ments of PRDl into pBR322 (420-bp HpaII frag-

ment is in the right-hand end). The difficulty of cloning the terminal fragments might be due to early phage promoters which are located in the genome term%. The clone obtained was sequenced using the Ml3 dideoxy method (Messing et al, 1981). The other strand was sequenced from the RF molecules with reverse sequencing primer using the plasmid sequencing method (Zagursky et al., 1985) with PolIk.

The M 13mp 18 clone sequence unambigously revealed four Gs at the terminus. On the contrary, with Maxam-Gilbert sequencing, the determination of the number of Gs was less accurate, and usually revealed three Gs at the ends. For this reason, in the sequence comparison the number of Gs is con- sidered to be four.

(c) The terminal nucleotide sequences of phages

PRDl, PR4, PR5, L17 and PR722

The overall comparison of the phage genome termini (Fig. 2) reveals ITRs in all viruses. The

PR5

. .

. .

. .

. .

. .

. .

. .

. .

. .

. .

. . . . . ..AC.

_.........

. . . . . . . . . . . . . . . . . . . . . ..G......

. . . . . . . . . . . . . . . . . . . . __........

. . . . . . . . . . .;"A;;;;;;;

',

PR4

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

..A. CG.C ......

A....CT ..C .AAA .............................

..A

CCCAAAACA

L

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

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

.GAATAAAAT

R

PI77

22

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

..A. CG.C

A....CT ..C .AAA.A

CCCAAAACA

L

......

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

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

.GAATAAAAT

R

L17

.

. .

. .

. .

. .

. . . . . . . ..A. C..C...GC.

A....CT..C

.AAA.A....

. ..GT.....

. ..G....GC

.TCATC....

. . . . . . . . . . I.........

. . . . . . . . . .

I

0

CCCAAAACA

L

'GAATAAAAT

R

Fig.

2.

T

erm

inal

nu

cleo

tide

sequ

ence

s of

pha

ges

PRD

I,

PR5,

PR

4,

PR72

2 an

d L

17.

The

fi

rst

111

bp s

peci

fy

the

ITR

s.

Aft

er

that

th

e se

quen

ces

dive

rge

in t

he

left

-han

d (L

) an

d ri

ght-

hand

(R)

term

inus

. D

ots

(.)

indi

cate

id

entic

al

nucl

eotid

es.

The

de

letio

n (a

t po

sitio

n 40

) is

ind

icat

ed

by s

ymbo

l A

. T

he

only

di

ffer

ence

be

twee

n th

e le

ft-

and

righ

t-ha

nd

term

ini

of t

he

sam

e vi

rus

is a

t po

sitio

n 91

in

L17

(s

ee

the

box)

. T

he

two

vert

ical

lin

es

sepa

rate

th

e co

nser

ved

term

ini

of t

he

ITR

s fr

om

the

vari

able

ce

ntra

l ar

ea.

TABLE 1

Nucleotide differences between the ITRs of phages PRD 1, PR5,

PR4, PR722 and L17 a

PRDl PR5 PR4 PR722

PR5 3 0

PR4 12 15 0

PR722 12 15 2 0

L17 23 21 14 13

a The numbers include the deletion at position 40. The mismatch

in L17 (at position 91) is not included in these numbers. Phages

PRD 1 and PR5 form one close group and phages PR4 and PR722

the other. Phage L17 belongs to the latter group but is further

evolved from that (See also Fig. 2).

lengths of the ITR’s in PRDl and PR5 are 110 bp, and in all other cases 111 bp. The difference is due to a single nt deletion at position 40. In the same genome, both termini are identical, except in L17 where there is a T at position 91 in the left-hand terminus and an A in the right-hand terminus. The viruses are closely related and their ITRs contain conserved areas. The first 18 bp and the last 35 bp of the ITRs are totally conserved, except for the single mismatch in L17. Between these blocks there is a variable area of 58 bp. Comparing the region spanning bp 19-46 enabled us to divide the phages into two groups: (1) Phages PRDl and PR5, from the U.S.A. and Canada, respectively, and (2) Phages PR4, PR722 and L17, from Australia, South Africa and the U.K., respectively. The overall nt differences in the variable area between the viruses are given in Table I.

20

203

The conserved IO-bp A + T block (41-50) sug-

gests the importance of low-melting DNA in this region, since the only changes observed are trans- versions between A and T. The first conserved 18 bp might be needed for the replication initiation complex recognition, analogously to the $29 and adenovirus systems (Penalva and Salas, 1982; Challberg and Kelly, 1979). The ITRs do not include the binding sequences for the dnaA protein (Fuller et al., 1984, Matsui et al., 1985). The function of the other con- served 35-bp sequence in the PRDl system is com- pletely unknown.

The nt sequence analysis described here places the PRDl phage family in the group which has long ITRs (adenoviruses, Cp-1 phages, linear plasmids). The $29-type replication system dilfers from the group above in having only short ITRs (Yoshikawa and Ito, 1981).

Because the ITRs in the same genome of PRDl group phages are similar, there must be a mechanism leading of co-evolution of the opposite long termini of the same molecule. A panhandle structure, where the terminal pair in a single-stranded replication state, has been suggested for adenovirus (Lechner and Kelly, 1977) and plasmid systems (Hirochica et al., 1984). In the $29 phage family, the terminal conservation might be needed for protein recognition alone, and terminal pairing might not take place.

The comparison of PRDl family genome termini with those of the $29 and Cp-1 families (Fig. 3) reveals that the conserved terminal areas of PRDl phage group and the left-hand end of $29 family are

18 and 17 bp long, respectively. Close to the end of

40

PRDI GGGGATACGTGCCCCTCCCC ACCTACCCGCGCCCCTAACA - -- -- - -------- ----

029L AAAGTAAGCCCCCACCCTCA CATGATACCATTCTCCTAAT -- - ----_---

029R AAAGTAGGGTACAGCGACAA CATACACCATTTCCCCATTG -- - - -_-----

cp-1 AAAGC,ATGTACTCCCCCACC CCTTTTTCAAAAATTCCCCC ------- _ -----_-

TTTTTATTTC -

ATCGACATAA -

ACCGACTATC - - -

AATGGAAATT -

Fig. 3. Comparison of the 50 terminal nucleotides of the phage genomes (PRDI, 429 and Cp-1) replicating in a linear form. $29L and

$29R are the left- and right-hand termini of the 429 genome. The sequences conserved within each phage group are underlined (in the

$29 phage group, the phage GA-I is omitted). The dashed line indicates the homologies between the phage groups found in the conserved

areas. The data for 429 is from Yoshikawa et al. (1985) and for Cp-1 from Escarmis et al. (1985). For PRDl, see Fig. 2.

204

this sequence is a homologous sequence: CCCCZCCC. The same sequence is also found in the Cp-1 termini a few nucleotides further from the genome end. The right-hand end of $29 does not have this sequence. The fact that the left-hand terminus of $29 enters the phage head first during packaging (Bjomsti et al., 1983) suggests that this sequence of $29, PRD 1 and Cp-1 might be required for DNA packaging.

The conserved areas of the right-hand end of $29 and Cp-1 termini also share another common sequence, TTCCCC(C)A, beginning at positions 31 and 34, respectively. In the PRDl group of ITRs, there are four conserved C’s, beginning at nt 32. These might be related to the $29 and Cp- 1 sequence above. This type of sequence similarity between the different viruses might indicate a common origin, even though they infect a number of both Gram- positive and Gram-negative hosts. We are currently studying the function of the different parts of the ITRs in PRDl replication.

NOTE

During the processing of this manuscript we became aware of the sequencing of the termini of bacteriophage PRDl by D.D. Gerendasy and J. Ito (The genome of lipid-containing bacteriophage PRD 1, which infects Gram-negative bacteria, contains long, inverted, terminal repeats. Submitted for publication). The sequence is identical to that of the PRDl ITR presented in this paper, with the exception that they found five G’s at the extreme termini compared to four G’s suggested by us. The reason for this discrepancy is most probably the difficulty in determining the very last nucleotides with the Maxam-Gilbert technique.

ACKNOWLEDGEMENTS

This investigation was supported by a research grant to D.B. from the Academy of Finland. H.S. is the recipient of a fellowship from the Emil Aaltonen foundation.

REFERENCES

Alestram, P., Stenlund, A., Li, L. and Pettersson U.: A common sequence in the inverted terminal repetitions of human and avian adenoviruses. Gene 18 (1982) 193-197.

Bamford, D.H. and Mindich, L.: Structure ofthe lipid-containing bacteriophage PRD 1: disruption of wild type and nonsense mutant phage particles with guanidine hydrochloride. J. Virol. 44 (1982) 1031-1038.

Bamford, D.H. and Mindich, L.: Characterization of the DNA-protein complex at the termini of the bacteriophage PRDI genome. J. Virol. 50 (1984) 309-315.

Bamford, D.H., Rouhiainen, L., Takkinen, K. and Sjderlund, H.: Comparison of the lipid-containing bacteriophages PRDl, PR3, PR4, PRS and L17. J. Gen. Virol. 57 (1981) 365-373.

Bamford, D.H., McGraw, T., Mackenzie, G. and Mindich, L.: Identification of a protein covalently bound to the termini of bacteriophage PRDl DNA. J. Virol. (1983) 311-316.

Bjomsti, M-A., Reilly, B.E. and Anderson, D.L.: Morphogenesis of bacteriophage $29 of Bacillus subtih: oriented and quantized in vitro packaging of DNA protein gp3. J. Virol. 45 (1983) 383-396.

Challberg, M.D. and Kelly Jr., T.J.: Adenovirus DNA replication in vitro. Proc. Natl. Acad. Sci. USA 76 (1979) 655-659.

Coetzee, W.F. and Bekker, P.J.: Pilus-specific, lipid-containing bacteriophages PR4 and PR722: comparison of physical characteristics ofgenomes. J. Gen. Viral. 45 (1979) 195-200.

Davis, T.N., Muller, E.D. and Cronan Jr., J.E.: The virion of the lipid-containing bacteriophage PR4. Virology 120 (1982) 287-306.

Davis, T.N. and Cronan Jr., J.E.: Nonsense mutants of the lipid-containing bacteriophage PR4. Virology 126 (1983) 600-613.

Desiderio, S.V. and Kelly Jr., T.J.: Structure of the linkage between adenovirus DNA and the 55,000 molecular weight terminal protein. .I. Mol. Biol. 145 (1981) 319-337.

Escarmis, C. and Salas M.: Nucleotide sequence at the termini of the DNA of Bacillus subtilk phage $29. Proc. Natl. Acad. Sci. USA 78 (1981) 1446-1450.

Escarmis, C., Garcia, P., Mkndez, E., Ldpez, E., Salas, M. and Garcia, E.: Inverted terminal repeats and terminal proteins of the genomes of pneumococcal phages. Gene 36 (1985) 341-348.

Fuller, R.S., Funnel, B.E. and Kornberg, A.: The dnaA protein complex with the E. coli chromosomal replication origin (oriC) and other DNA sites. Cell 38 (1984) 889-900.

Garcia, E., Gomez, A., Ronda, A., Escarmis, C. and L6pez, R.: Pneumococcal bacteriophage Cp-1 contains a protein bound to the 5’-termini of its DNA. Virology 128 (1983) 92-104.

Garcia, P., Hermoso, J.M., Garcia, J.A., Garcia, E., L6pez, R. and Salas, M.: Formation of a covalent complex between the terminal protein of pneumococcal bacteriophage Cp-1 and 5’-dAMP. J.&irol. 58 (1986) 31-35.

Hermoso, J.M. and Salas, M.: Protein p3 is linked to the DNA of phage +29 through a phosphoester bond between serine and 5’-dAMP. Proc. Natl. Acad. Sci. USA 77 (1980) 6425-6428.

205

Hirochica, H., Nakamura, K. and Sakaguchi, K.: A linear DNA

plasmid from Sfreptomyces rochei with an inverted terminal repetition of 614 base pairs. EMBO J. 3 (1984) 761-766.

Hishinuma, F., Nakamura, K., Hirai, K., Nishizawa, R., Gunge, N. and Maeda, T.: Cloning and nucleotide sequencies of the linear DNA killer plasmids from yeast. Nucl. Acids Res. 12

(1984) 7581-7597. Kemble, R.J. and Thompson, D.: Sl and S2, the linear

mitochondrial DNAs present in a male sterile line of maize, possess terminally attached proteins. Nucl. Acids Res. 10

(1982) 8181-8190. Lechner, R.L. and Kelly Jr., T.J.: The structure of replicating

adenovirus 2 DNA molecules. Cell 12 (1977) 1007-1020.

Levings HI, C.S. and Sederoff, R.R.: Nucleotide sequence of the S-2 mitochondrial DNA from the S cytoplasm ofmaize. Proc. Natl. Acad. Sci. USA 80 (1983) 4055-4059.

Lichy, J.H., Horwitz, M.S. and Hurwitz, J.: Formation of a covalent complex between the 80,000 dalton adenovirus terminal protein and 5’-dCMP in vitro. Proc. Natl. Acad. Sci.

USA 78 (1981) 2678-2682. Lundstrom, K.H., Bamford, D.H., Palva, E.T. and Lounatmaa,

K.: Lipid-containing bacteriophage PR4: structure and life cycle. J. Gen. Virol. 43 (1979) 583-592.

Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual, Cold Spring Harbor laboratory, Cold Spring Harbor, NY, 1982.

Matsui, M., Oka, A., Takanami, M., Yasuda, S. and Hirota, Y.: Sites of dnaA protein-binding in the replication origin of the Escherichiu coli K-12 chromosome. J. Mol. Biol. 184 (1985) 529-533.

Maxam, A.M. and Gilbert, W.: Sequencing end-labeled DNA with base-specific chemical cleavage. Methods Enzymol. 65 (1980) 499-560.

McGraw, T., Yang, H.-L. and Mindich, L.: Establishment of a physical and genetic map for bacteriophage PRDl. Mol. Gen. Genet. 190 (1983) 237-244.

Messing, J., Crea, R. and Seeburg, P.H.: A system for shotgun DNA sequencing. Nucl. Acids Res. 9 (1981) 309-321.

Mindich, L., Cohen, J. and Weisburd, M.: Isolation of nonsense suppressor mutants in Pseudomonas. J. Bacterial. 126 (1976) 177-182.

Mindich, L., Bamford, D., Goldthwaite, C., Laverty, M. and

Mackenzie, G.: Isolation of nonsense mutants of lipid containing bacteriophage PRDI. J. Virol. 44 (1982a) 1013-1020.

Mindich, L., Bamford, D., McGraw, T. and Mackenzie, G.: Assembly of bacteriophage PRDI: particle formation with wild-type and mutant viruses. J. Viral. 44 (1982b) 1021-1030.

Muller, E.D. and Cronan Jr., J.E.: The lipid-containing bacterio- phage PR4: effects of altered lipid composition on the virion. J. Mol. Biol. 165 (1983) 109-124.

Olsen, R.H., Siak, J.-S. and Shipley, P.L.: Characteristics of PRD 1, a plasmid-dependent broad host range bacteriophage. J. Virol. 14 (1974) 689-699.

Pefialva, M.A. and Salas, M.: Initiation of phage 429 DNA

replication in vitro: formation of a covalent complex between the terminal protein, p3, and 5-dAMP. Proc. Natl. Acad. Sci. USA 79 (1982) 5522-5526.

Sanger, F. and Coulson, A.R.: The use of thin acrylamide gels for DNA sequencing. FEBS Lett. 87 (1978) 107-l 10.

Sor, F., Wbsolowski, M. and Fukuhara, H.: Inverted terminal repetitions of the two linear DNA associated with the killer character of the yeast Kluyveromyces la&. Nucl. Acids Res.

11 (1983) 5037-5044. Watabe, K., Shih, M.F., Sugino, A. and Ito, J.: In vitro replication

of bacteriophage $29 DNA. Proc. Natl. Acad. Sci. USA 79 (1982) 5245-5248.

Yanisch-Perron, C., Vieira, J. and Messing, J.: Improved Ml3 phage cloning vectors and host strains: nucleotide sequences of the Ml3mp18 and pUC19 vectors. Gene 33 (1985) 103-l 19.

Yoshikawa, H. and Ito, J.: Terminal proteins and short inverted terminal repeats of small Bacillus bacteriophage genomes. Proc. Natl. Acad. Sci. USA 78 (1981) 2596-2600.

Yoshikawa, H., Garvey, K.J. and Ito, J.: Nucleotide sequence analysis of DNA replication origins of small Bacillus bacterio- phages: evolutionary relationships. Gene 37 (1985) 125-130.

Zagursky, R.J., Baumaister, K., Lomax, N. and Berman, M.L.: Rapid and easy sequencing of large linear double-stranded DNA and supercoiled plasmid DNA. Gene Anal. Tech. 2

(1985) 89-94.

Communicated by Z. HradeEna