PLASMID 15,93-103 (1986)

The Nucleotide Sequence of pUB110: Some Salient Features in Relation to Replication and Its Regulation

TIMOTHY MCKENZIE, TARAYURI HOSHINO, TERUO TANAKA,’ AND NOBORU SUEORA

Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, Colorado 80309

Received March 26, 1985; revised November 18, 1985

For the study of DNA-membrane interaction and the regulation of replication initiation we have determined the total nucleotide sequence of pUBll0. As previously reported, this plasmid replicates in B. subtilis at a copy number of 30-50 per cell, with a majority of plasmids (60-80s) bound to the membrane (type-I binding). The type-1 membrane binding is apparently necessary for pUBll0 initiation of replication in vivo, but the membrane binding site is not known. Fur- thermore, four areas of the plasmid specifically bind to Bacillus subtilis membrane in an in vitro binding reaction (type-11 binding). These two types of membrane binding of pUB I 10 are different in that the in vivo binding (type-I) requires one (dnaBI) of the host initiation genes and is high- salt resistant, whereas the in vitro binding (type-II) does not require the dnaB1 gene product and is high-salt sensitive. A 7-mer double-strand sequence, TCAGCAA/AGTCGTT, or one-base de- rivatives of this sequence are frequently (17 of 23 of the 7-mer sequences) found in or close to the type-II binding areas. One of them is found at a restriction enzyme recognition site of a binding area that destroys the type-II membrane binding. These sequences may or may not have significance in type-11 membrane binding. In addition to the neomycin resistance gene, the sequence data indicate two sizable open reading frames, ORF LY and ORF 0, and two small ORF, y, and 6. All of these reading frames are in the same direction, which coincides with the direction of the replication. The open reading frame a! (ORF a) corresponding to 334 amino acids close to the replication origin may be essential for the initiation of replication of pUBll0. The putative protein 01 corresponding to this open reading frame contains a consensus sequence of the DNA binding sites which are found in a number of known DNA-binding proteins. The consensus DNA binding site of protein (Y is flanked by two hydrophobic areas. These two observations suggest that the corresponding protein may have both an affinity to a specific site in pUBll0, and an affinity to the membrane. 0 1986 Academic &u, Inc

Using temperature-sensitive initiation mu- tants, dnaBZ(White and Sueoka, 1973; Imada et al., 1980) and dnaBZZ (Karamata and Gross, 1970; Imada et al., 1980), we have shown that upon shifting from permissive to nonpermis- sive temperature the initiation of Bacillus subtilis chromosome replication stops, and the replication origin area is selectively released from the membrane both in vivo and in vitro (Winston and Sueoka, 1980). This indicates that dnaBZ and dnaBZZ gene products are re- quired for both initiation of replication and membrane binding of the B. subtilis chro- mosome. In contrast, the chimeric plasmid

’ Present address: Mitsubishi-kasei Life Science Institute Korn et al. (1983) have reported a specific Machida-shi, Tokyo, Japan in vitro DNA-membrane binding using iso-

pSL103, composed of PUB 110 from Staphy- lococcus aureus (for replication function) and a trpCt containing fragment from B. pumilus (Keggins et al., 1978), is able to replicate in the dnaBZZ mutant but not in the dnaBZ mu- tant at the nonpermissive temperature (Win- ston and Sueoka, 1980). The plasmid is re- leased from B. subtilis membrane only in the dnaBZ mutant and not in the dnaBZZ mutant (Imada et al., 1980; Winston and Sueoka, 1980; Korn et al., 1983) at the nonpermissive temperature. Thus, the dnaBZ gene product but not the dnaBZZ gene product is necessary for both membrane binding and replication of pSL 103 and PUB 110.

93 0147-619X/86 $3.00 Copyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.

94 MC KENZIE ET AL.

lated membrane fraction and purified pSL 103. Based upon competition experiments they have shown that this in vitro membrane bind- ing (type-II binding) is specific for the PUB 110 portion of pSL 103. In contrast, the trpC+ car- rying fragment of pSL103, ColEl, and rat DNA do not compete in membrane binding with pUBll0 (Korn et al., 1983). This type- II membrane binding of pUBll0 was later shown by Tanaka and Sueoka (1983) to be different from the in vivo membrane binding of PUB 110 and pSL 103 (type-I binding). Thus, the type-II binding is high-salt sensitive and is independent of the functional product of the initiation gene &z&Z, whereas the type-1 membrane binding of PUB 110 is high-salt re- sistant, as is the binding of the chromosomal origin area of B. subtilis (Sueoka and Ham- mers, 1974). The functional dnaBZ gene prod- uct is necessary for initiation of pUBll0 (Winston and Sueoka, 1980). Tanaka and Sueoka (1983) have described four separate binding areas of pUBI which are able to bind B. subtilis membrane in vitro (Fig. 1). Two of the four type-II binding areas flank the origin of replication of pUBl10, whereas the remaining two areas are further away.

In this paper, we report the entire nucleotide sequence of the PUB 110 plasmid. Translation of the sequence in all possible directions and reading frames revealed several features of the plasmid with regard to replication and mem- brane bindings. An open reading frame a (ORF a) for a protein close to the origin of replication has become evident. From a num- ber of criteria, the protein corresponding to ORF (Y is a candidate for a replication protein of the plasmid. The putative “protein (Y” is highly basic and includes a sequence of 20 amino acids, which satisfies the consensus DNA binding sequence (review by Pabo and Sauer, 1984). The binding sequence is flanked by two hydrophobic regions.

MATERIALS AND METHODS

Strains. B. subtilis 168 trp thy has been de- scribed previously (White and Sueoka, 1973). The Escherichia coli strain JM103 was in-

eluded in the Ml3 cloning kit obtained from Bethesda Research Laboratories. PUB 110 was isolated by the procedure of Tanaka and Ka- wano ( 1980).

Materials. The Ml 3 cloning kit, pUC8, pUC9, Polymerase I (Klenow fragment), and all restriction enzymes (except BstNl) were obtained from Bethesda Research Laborato- ries. BstNI was obtained from New England Biolabs. [a-32P]dTTP, [cy-32P]dCTP (specific activities were between 2000 and 3000 Ci/ mmol), and polynucleotide kinase were from New England Nuclear Corporation [y- 32P]ATP with a specific activity of 9000 Ci/ mmol was obtained from ICN. 7-deaza-2’- deoxyguanosine triphosphate was a gift from Dr. Frank Seela, Laboratory of Bioorganic Chemistry, Department of Chemistry, Uni- versity of Paderbom, West Germany.

Labeling of plasmid DNA. Digestion of PUB 110 with restriction enzymes and Y-~*P labeling have been described previously (Tan- aka and Sueoka, 1983). 3’- and 5’-32P labeling were performed according to the method de- scribed by Maxam and Gilbert (1980).

Sequencing of plasmid DNA. Chemical se- quencing was performed using the procedure of Maxam and Gilbert ( 1980) with the follow- ing modifications. The G = A reaction was initiated by the addition of 25 ~1 of 88% formic acid to 10 ~1 of PUB 110 fragments (in distilled water). All other conditions were the same as described by Maxam and Gilbert ( 1980). San- ger et al. (1977) sequencing was performed following the method that Bethesda Research Laboratories included in their Ml3 cloning kit (as outlined by Messing et al., 198 l), using either M 13 or the pUC plasmids. 7-deazu-2’- deoxyguanosine triphosphate was also used to correct some intrinsic ambiguities found by using deoxyguanosine triphosphate (S. Mi- zusawa, S. Nishimura, and F. Seela, personal communication).

Binding of [j*P]DNA to membrane frac- tion. Preparation of the bacterial lysate, bind- ing of radioactive DNA to the membrane fraction and analysis of the membrane bound [32P]DNA fragments have been described pre- viously (Tanaka and Sueoka, 1983).

pUBl10 NUCLEOTIDE SEQUENCE 95

RESULTS

The four areas of type-11 membrane binding of PUB 110 have been previously identified (Tanaka and Sueoka, 1983) and a slightly modified figure as explained below is shown in Fig. 1. The nucleotide sequence of PUB 110 is shown in Fig. 2 and the sequencing strategy is shown in Fig. 3. Major restriction sites are indicated. Two of the four binding areas (BA 1 and BA2) flank the replication origin region. The replication origin region has been previ- ously assigned by electron microscopic anal- ysis of replicating molecules (Scheer-Abra- mowitz et al., 1980). The other two binding areas (BA3 and BA4) are located opposite of the origin region (Fig. 1).

When the Mb01 C fragment which includes BAl (Tanaka and Sueoka, 1983) (Fig. 1) is

further cleaved with NcuI (Fig. l), the 558-bp fragment still binds to membrane in vitro, whereas the 320-bp fragment does not (data not shown). This indicates that the type-II binding site in this area should be within the 558-bp fragment; this 558-bp area is now des- ignated as BAl (Fig. 1). When binding area 1 is cleaved with TaqI the whole area loses type- II binding (Tanaka and Sueoka, 1983). These results suggest that the actual binding site within BAl is at or near the TuqI cleavage site.

When the Mb01 F fragment which includes BA3 (Tanaka and Sueoka, 1983) (Fig. 1) is further cut with the enzyme Thai (Fig. I), the larger 186-bp fragment binds to the mem- brane, whereas the smaller 75-bp fragment does not (data not shown). However, when this Mb01 fragment is cut with Hinf I, neither of the two subfragments is bound to the B.

FIG. 1. A membrane binding map of restriction fragments of pUB 110. The thick arcs outside of the map show the length of DNA fragments that can bind to membrane, termed binding areas (BA 1 to BA4). Detailed descriptions of these binding areas have been described previously (Tanaka and Sueoka, 1983). The shaded arc with an asterisk indicates the area that has been sequenced by J. I-Iahn and D. Dubnau (personal com- munication). The shaded arc of neo’ area with two asterisks has been sequenced by Matsumura et al. (1984). Thick solid arrows inside the map represent five open reading frames; four of them, ORF 01, ORF p, ORF y, and ORF 8, have been recognized by this work and the one (*a) for neo’ (or Km’) has been reported by Matsumura et al. (1984). The origin, and direction of replication reported by Scheer-Abramowitz et al. ( 1980) is described by the arrow outside of the map itself (ori). neo’ stands for the neomycin resistance gene.

~?~,~~AT~,,~MT?~,,~~~~~C~~~??C~~~~~~~~T~T~OTT??~C~C~~~~T~?~?~?CCG~G~CO??C~C~UT~~~?~~?~~~~C

21e1 r~nuttctrr~~~r~rirct~~*~*~ic~cc~c~k~Ac~~??AC~AT~~cTcTc~~~ickr~r~T~~~ Mwuc;tr-w~TMEcrtrrrMt 2- END nsor c -~.rl~~~v.r~v.rr1.r~r~l~l”~r~1Ycl”r~~r~~r~

FIG. 2. Nucleotide sequence of pUBl10. The relative positions of the four binding areas (as shown by thick arrows) correspond to those in Fig. 1. Three open reading frames, CZ, 8, and y, which are each equivalent to more than 80 amino acids, are shown. Although ORF 6 consists of more than 80 amino acids, the amino acid sequence of ORF 6 is not shown in this figure to avoid confusion with ORF 8. The ORF 8 sequence occupies the nucleotide positions from 1075 to 833 with a possible ATG start at position 1075, counter- clockwise, although a GTG start at position 1138 is also a possibility. The amino- and carboxy-terminals are indicated by -N and C-, respectively. The N-terminals of these open reading frames are not necessarily the best initiation sites. A plausible area for transcriptional initiation is seen around the boxed TATAAT (positions 3599 to 3594, counterclockwise). Thin arrows above the sequence indicate inverted repeats as

96

probable transcriptional terminators. A consensus sequence for DNA binding sites (reviewed by Pabo and Sauer, 1984) is underlined (positions 3854 to 3795). The conserved amino acids within this sequence are underlined twice. Major restriction sites are shown (including MboI). The ‘I-mer TCAGCAA/AGTCGTT and its one-base variants are also marked by a line between the double-strand 7-mer sequence. The area of the replication origin and the direction of replication reported by Scheer-Abramowitz et al. (1980) are indicated. Position 4035 to 4377 has also been sequenced by J. Hahn and D. Dubnau (personal commu- nication) and position 1979 to 3203 has been sequenced by Matsumura et al. (1984). The sequence of those areas is in complete agreement with ours. Some parts of our nucleotide sequence have been previously reported (Sueoka et al., 1984) and the current result includes several corrections.

97

98 MC KBNZIE ET AL

1 814 1 ria I Nrn’ I lsazj Sll I

Tag I -

r-

Mbo I c_

- /I -- r- -- Hinf I I

Bst NI - - - Em RI, .1 Tha I -

Hoe III

Avo II -

Xbo I -

Ava I -

Nco I -

Hpo II L-U L I I L I

0 I 2 3 4 Kb

FIG. 3. Sequencing strategy of pUBll0. The DNA se- quence has been analyzed at least twice by the method of Maxam and Gilbert ( 1980) or by Sanger’s dideoxynucleo- tide method (1977) except in the area sequenced by Mat- sumura et al. (1984). The vertical lines show the relative positions of each restriction enzyme recognition site. The arrows above the restriction site describe the direction of sequencing after either 5’ or 3’ 32P labeling for the Maxam- Gilbert technique, while the arrows below the sites show the direction of sequencing using Sanger’s method. The boxed areas on the top line show the relative positions of the binding areas and the neomycin resistance gene, neo’. kb is the length of DNA in kilobases.

subtilis membrane (Tanaka and Sueoka, 1983). The 186-bp fragment is now designated as BA3. The Thai site is separated from the Hinf I site by 108 bases (Figs. 1 and 2). These results suggest that the actual binding site within BA3 is at or near the HinfI cleavage site.

As a first approach for identifying the type- II binding sequence, the binding areas have been examined for sequence homology with the aid of a computer. The search for the larg- est sequence similarity in all four binding areas (excluding sequences composed almost en- tirely of A and T) gave the result that the se- quence, TCAGCAA/AGTCGTT, is found seven times in PUB 110; once in BA 1 (position 3802 to 3808, Fig. 2) and twice in BA3 (po- sitions 1439 to 1445 and 1540 to 1546, Fig. 2). Three others (1289 to 1295, 1328 to 1334, and 1563 to 1557) are within 60 bases from BA3 and one outside of any binding area (position 4255 to 4261). The sequence, TCAGCAA/AGTCGTT, located at 1439 to 1445 of BA3, is adjacent to the Hinf I site that,

if cut, results in the loss of membrane binding of BA3. Among 16 one-base variants of the 7- mer, 7 are found in the type-II binding areas and 4 are within 80 bases of some of the bind- ing areas. As there are no identical sequences of more than 6 bp in the four membrane bind- ing areas, identification of the consensus se- quence responsible for type-II binding is cur- rently not possible using this method.

Recently, Matsumura et al. ( 1984) reported the nucleotide sequence of the area covering the kanamycin (neomycin) resistance gene (neo’). The sequence reported by them is ex- actly the same as our results shown in Fig. 2 (position 1979 to 3203). The sequence is pre- sented in Fig. 2 in the opposite orientation of that of Matsumura et al. (1984). In addition, J. Hahn and D. Dubnau (personal commu- nication) sequenced the area between BA 1 and BA4 (Fig. 2).

Five open reading frames (neo’, ORF (Y, ORF fi, ORF y, and ORF 6) can be recognized, all of which have the same transcriptional ori- entation (counterclockwise in Fig. 1). p-inde- pendent transcriptional terminator structures are found immediately after the translational stop codons of ORF (Y and ORF /3 (Fig. 2). Thus these two ORFs seem to be transcribed. We believe that ORF a is the prime candidate for a plasmid replication gene (rep) for the fol- lowing reasons. ORF (Y is in the ori region. Plasmids lacking ORF & ORF y, or ORF 6 regions can replicate, but no reconstructed PUB 110 molecules lacking the ORF (Y region were capable of replication (Scheer-Abra- mowitz et al., 1980; Tanaka and Sueoka, 1983). Also, there is no additional room for other long open reading frames in pUBll0. Several properties of the putative protein (Y, deduced from the nucleotide sequence of ORF (Y according to Chou and Fasman (1978) are presented in Fig. 4.

We have also recognized a typical, 20-amino acid consensus sequence for a DNA binding site in protein (Y which is similar to those ob- served in several procaryotic DNA-binding proteins (reviewed by Pabo and Sauer, 1984). The sequence occupies the location between nucleotide positions 3854 to 3795 of the se- quence presented in Fig. 2. This sequence also

pUBI NUCLEOTIDE SEQUENCE 99

HYDROPHOBICITY

HYDROPHILICITY

H CHARGE

RLPHA HELICITY L BETA SHEET

B 50 IW 150 280 258 388 N C

fWIN0 RCID POSITION

HYDROPHOBICITY

HYDROPHILICITY

ALPHR HELICITY b BETR SHEET . . . . . . . . . . . . .

L 1

0.5 I . . . . I . . ..I.... I.

BEl 100 120

AMINO RCID POSITION

FIG. 4. Some physical properties of the putative protein (Y. Hydropathy, (Y helicity and j3 sheet potential were calculated by using the parameters assigned by Chou and Fasman (1978). For Hydropathy, average values for 7 consecutive amino acids are plotted for each amino acid position. For (x helicity and fl sheet potential, average values for 4 consecutive amino acids are plotted. The stretch of the 2O-arnino acid sequence (position 3854 to 3795 in Fig. 2) that corresponds to the consensus DNA-binding amino acid sequence (Pabo and Sauer, 1984) is indicated as the thick bar. A: the total area; B: extended picture of the putative DNA-binding area.

favors two a-helical regions (Fig. 4) as in most The ATG at position 4136 (Fig. 2) as the of the other DNA binding sequences compiled true initiation codon of “protein CY” is doubt- by Pabo and Sauer (1984). ful, because of the lack of a sequence for a

100 MC KENZIE ET AL.

tibosome binding site in front of it (reviewed by Murray and Rabinowitz, 1982). In this sense, GTG at position 4067 seems more likely to be the starting codon. It is likely that the 3’-terminal region of ORF (Y covers the repli- cation origin of the plasmid reported by Scheer-Abramowitz et al. (1980). There is no sizable open reading frame in BA3.

DISCUSSION

Although the function of type-11 membrane binding in chromosome replication is pres- ently not known, the clear sequence specificity for this binding suggests its importance (Tan- aka and Sueoka, 1983). Of the four binding areas of pUBl10, the two areas flanking the origin (BA 1 and BA2) seem indispensable for replication in vivo. No reconstructed pUB 110 molecules made from Mb01 fragments Iacking BAl and BA2 were capable of replication (Tanaka and Sueoka, 1983). On the other hand plasmids lacking in BA3 (MboI-F) or BA4 (MboI-B) have been isolated (Scheer- Abramowitz et al., 1980; Tanaka and Sueoka, 1983). As for the possible importance of type- II binding, we have proposed as a working hy- pothesis that the type-II binding is necessary for the plasmid to bind to a specific site or sites of the membrane first, and, subsequently, the type-I membrane complex, the putative ini- tiation and replication complex, may be built around the replication origin (Sueoka et al., 1984). According to this hypothesis, type-II binding is an essential first step in the repli- cation (initiation) of the plasmid. Since type- II binding is sequence specific and approxi- mately the same number of plasmids per cell can be bound to the membrane in vitro as the number in vivo, the type-II binding step may be the basis for host-plasmid specificity, for compatibility between different plasmids, and for copy number control (Sueoka et al., 1984).

Plasmid pBC 16, which replicates both in B. subtilis and in B. cereus has an extensive ho- mology with pUBl10, as reported by Polak and Novick (1982); the homology between the two plasmids covers the region of replication origin and type-11 binding areas, BAl, BA2, and BA4. BA3 of pUBll0 does not exist in

pBC 16. In place of the neomycin-kanamycin resistance gene of pUBl10 a tetracycline re- sistance gene is in pBC16. Both plasmids co- exist when the two plasmids are put into the same cell under the condition selective for tet- racycline resistance. However, when kana- mycin is the sole selective agent or under non- selective conditions, pBC16 is lost from the culture, maintaining only 2 to 20% of the pop- ulation with pBC 16 after 20 to 30 generations (Polak and Novick, 1982). Either of these plasmids by themselves is entirely stable under nonselective conditions in B. subtilis (Polak and Novick, 1982; Scheer-Abramowitz et al., 1980). It is an interesting possibility that the apparent incompatibility between PUB 110 and pBC16 may come from the fact that the two plasmids compete for the same type-II binding sites on the membrane and that PUB 110 overtakes pCB 16 by a more efficient type-II binding of pUBI with four binding areas over the three binding areas of pBC 16. Recently, Hoshino et al. (1985b) reported that pTHT 15, a tetracycline-resistant plasmid originally isolated from thermophilic soil Ba- cillus (Hoshino et al., 1985a), shows strong homology to PUB 110 in the replication origin and BA3. Thus, pTHT 15 seemed to have sim- ilar membrane binding areas to BAl, BA2, and BA3 of pUBll0. It has recently been shown that the nucleotide sequence of the replication region of pTHTl5 is identical to that of pUBll0 presented here (T. Hoshino and K. Furukawa, unpublished data). They also examined the incompatibility between pTHT 15 and pUB I 10 under the same exper- imental conditions as described by Polak and Novick (1982). pTHTl5 and pUBll0 are semicompatible in B. subtilis host cell: only about 10 to 20% of the total population lost pTHT 15 under nonselective conditions for 20 to 30 generations or when kanamycin was the sole selective agent (Hoshino et al., 1985b). The difference in compatibilities between pBC16 and pTHT 15 with pUBll0 suggests that BA3 may contribute significantly in the efficiency of plasmid maintenance or incom- patibility, although BA3 is not essential for replication (Scheer-Abramowitz et al., 1980; Tanaka and Sueoka, 1983).

pUB I 10 NUCLEOTIDE SEQUENCE 101

Binding areas BA 1 and BA2 contain the or- igin area of pUBll0 as Sheer-Abramowitz et al. (1980) have assigned through electron mi- croscopic analysis of partially replicating mol- ecules of pUBll0 (Fig. 1). They determined that replication proceeds counterclockwise, as indicated in Fig. 1. The nucleotide sequence presented in BAl and BA2 covers the origin area. From sequence analysis, some potential promoter sites (e.g., TATAAT at position 3599) of transcription by RNA polymerase are recognized. However, there does not appear to be a ribosome binding sequence that matches any of the known sequences in B. subtilis (Murray and Rabinowitz, 1982; Moran et al., 1982) near this region. This po- tential promoter exists in the same orientation as the direction of replication (counterclock- wise). Therefore, transcription is likely to occur in this region, but translation may not. More- over, the direction of transcription is consistent with the direction of DNA replication from the origin area, as reported by Scheer-Abra- mowitz et al. ( 1980). These results give support to the scheme that the initiation of pUB 110 replication might start as RNA-polymerase- mediated transcription in this area, and the RNA might serve as the primer for the DNA strand elongation. The presence of the repli- cation origin within the rep gene (repC) has been reported by Novick et al. ( 1982) in pT 18 1 and by Shon et al. ( 1982) in R6K. These cases may present a novel way of regulating the ini- tiation rate of plasmid replication.

The sequence around the pUBll0 origin shows none of the long, inverted complemen- tary repeat sequences which have been re- ported around the replication origins of two other S. aureus plasmids that replicate in B. subtilis, pC194 and pE194 (Ehrlich, 1977; Gryczan et al., 1978; Horinouchi and Weis- blum, 1982a,b; Scheer-Abramowitz et al., 1980). There is no region of multiple loops as seen in the ColE 1 initiation area, nor are there found direct or inverted repeats of 19- to 22- bp sequences as observed in the replication origin area of plasmids, F, P 1, and R6K (re- viewed by Scott, 1984), and in the origin area of the genomes of phage lambda, 80, phage 82 (reviewed by Friedman et al., 1984), E. coli

(Zyskind et al., 1983) and B. subtilis (Moriya et al., 1985). In pUBl10, there is not a single sequence of the 9-mer, TTAT(C/A)CACA, that is repeated several times in the origin area of E. coli and B. subtilis chromosomes. It is possible that the dnaBZZ gene product that is necessary for the initiation and membrane binding of the B. subtilis chromosome, but is not necessary for PUB 110 (Winston and Sueoka, 1980), might be involved in the rec- ognition of the repeated 9-mer of the B. subtilis replication origin.

The putative protein LY of PUB 110 corre- sponding to the ORF (Y is highly basic (46 basic amino acids vs 23 acidic amino acids) and contains about five short hydrophobic areas (Fig. 4). Particularly interesting is the existence of a consensus amino acid sequence which binds to a specific DNA sequence (Fig. 5; re- viewed by Pabo and Sauer, 1984). This se- quence in the putative protein ar is flanked by two hydrophobic areas (Fig. 4). This suggests the possibility that the protein (Y may bind both to a specific DNA sequence of PUB 110 and to the membrane. This binding site is not far upstream of the replication origin assigned by Scheer-Abramowitz et al. (1980).

The replication of pUB 110 requires the

pUB!lB ‘Protein alpha’

RtnTrpRrgRrgRl~~tLy~nl*Glyll~GlnLy*v.lv~l~l,Gl”v~lll~ . . . . . . . . . . . .

Lambda cm Protein GlnThrLytThrRldyrRpLeuGlyVllTyrGl”~~~l,Il~~~”Ly,~l~Il.Hl,

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

Lambda Repressor GlnGl~SsrV~lfll~R~pLy=tttGl~tGIyGln~rG~~~~tGlyRI~.uPheR~~

. . . . . . . . . . . . E. coli CRP Protein

A~gGl~Gl~I1~GlyGl~ll~V~lGlyCyoSerRrgGl~Th~~~lGl~~SIl~L~~Ly~ . . . . . . . . . . . .

E. coli Lac R L~UT~~R~~V~IRI~G~~T~~RIIG~~V~~~~T~~G~~T~~~SV~~V~~R~~

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

E. coli Gal R II~Ly~R~pV~IAI~ArSLsuAIIGlyV~l~~V.lfl~.Th~V~l~r~~SV.lIl~fl~~

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

Mat Rlpha LyrGl”Gl”V.IRldy,Ly*Cy,G,yI,~Th~P~~L~”G,”V.,~~SV,,T~pCy,~=”

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

1 Helix 3

I

FIG. 5. A consensus DNA binding sequence in “protein ~2’ of PUB1 10 (position 3854 to 3795 in Fig. 2), in com- parison with some other known proteins. Asterisks indicate the conserved amino acid positions. Proposed a helical structures and DNA-binding sequences of proteins other than pUBI “protein a” are cited from the review by Pabo and Sauer (1984).

102 MC KENZIE ET AL.

product of one (dnaB1) of the two host initi- ation genes (Winston and Sueoka, 1980). If PUB 110 indeed requires the ORF CY encoding protein to replicate, pUB 110 needs at least two proteins (one, &zaBIgene product of the host, and the other, protein (Y of the plasmid) for its replication. We have previously shown that the product of the dnaBI gene may also be a membrane protein (Imada et al., 1976, McKenzie and Sueoka, unpublished results). Since pUBll0 is bound to the membrane, the properties of the proteins LY and dnaBI should be quite interesting from the view point of their potential interactions with the replication or- igin of pUB 110 and also with the membrane.

ACKNOWLEDGMENTS

This research has been supported by NSF Grant PCM8011549 and NIH Grant GM28133; We thank Dr. David Pribnow and Dr. Larry Klobutcher for their helpful instructions in DNA sequencing procedures. We are grateful to Dr. Jannet Hahn and David Dubnau for making the result of their sequence analysis available before pub- lication and to Dan Sonis and Peter Newton for improve- ment of the manuscript. A gift of 7-deaza-2’deoxyguano- sine triphosphate from S. Mizusawa, S. Nishimura, and F. Se& before their publication is gratefully acknowledged.

REFERENCES

CHOU, P. Y., AND FASMAN, G. D. (1978). The secondary structure of proteins from their amino acid sequence. Adv. Enzymol. 47,45-148.

EHRLICH, S. D. (1977). Replication and expression of plasmids from Staphylococcus aureus in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 74, 1680-1682.

FRIEDMAN, D. I., OLSEN, E. R., GEOGOFQULOS, C., TILLY, K., HERSKOWITZ, I., AND BANUETT, F. (1984). Inter- action of bacteriophage and host macromolecules in the growth of bacteriophage X. Microbial. Rev 48, 299- 325.

GRYCZAN, T. J., CONTENTE, S., AND DUBNAU, D. (1978). Characterization of Staphylococcus aureus plasmids in- troduced by transformation into Bacillus subtilis. J. Bacterial. 134, 3 18-329.

HORINOUCHI, S., AND WEISBLUM, B. (1982a). Nucleotide sequence and functional map of pC 194, a plasmid that specifies inducible chloramphenicol resistance. J. Bac- teriol. 150, 815-825.

HORINOUCHI, S., AND WEISBLUM, B. (1982b). Nucleotide sequence and functional map of pE 194, a plasmid that specifies inducible resistance to macrotide, lincosamide, and streptogramin type B antibiotics. J. Bacterial. 150, 804-8 14.

HOSHINO, T., IKEDA, T., NARUSHIMA, H., AND TOMI- ZUKA, N. (1985a). Isolation and characterization of an-

tibiotic-resistance plasmids in thermophilic bacilli. Canad. J. Microbial. 31, 339-345.

HOSHINO, T., IKEDA, T., FURUKAWA, K., AND TOMIZUKA, N. (1985b). Genetic relationship between pUBl IO and antibiotic-resistant plasmids obtained from thermophilic bacilli. Canad. J. Microbial. 31, 614-619.

IMADA, S., CARROL, L. E., AND SUEOKA, N. (1976). DNA- membrane complex in Bacillus subtilis. In “Micro- biology- 1976” (D. Schlessinger, ed.), pp. I 16- 122, American Society for Microbiology.

IMADA, S., CARROLL, L. E., AND SUEOKA, N. ( 1980). Ge- netic mapping of a group of temperature-sensitive dna initiation mutants in Bacillus subtilis. Genetics 94, 809- 823.

KARAMATA, D., AND GROSS, J. D. (1970). Isolation and genetic analysis of temperature sensitive mutants of B. subtilis defective in DNA synthesis. Mol. Gen. Genet. 108,277-287.

KEGGINS, K., LOVET’~, P., AND DUVALL, D. (1978). Mo- lecular cloning of genetically active fragments of Bacillus DNA in Bacillus subtilis and properties of the vector plasmid pUBll0. Proc. Natl. Acad. Sci USA 75,1423- 1427.

KORN, R., WINSTON, S., TANAKA, T., AND SUEOKA, N. (1983). Specific in vitro binding of a plasmid to a mem- brane fraction of Bacillus subtilis. Proc. Natl. Acad. Sci. USA 80,5-f&578.

MATSUMURA, M., KATAKURA, Y., IMANAKA, T., AND AIBA, S. (1984). Enzymatic and nucleotide sequence studies of a kanamycin-inactivating enzyme encoded by a plasmid from thermophilic Bacilli in comparison with that encoded by plasmid pUBl10. J. Bacterial. 160,4 13-420.

MAXAM, A. M., AND GILBERT, W. (1980). Sequencing end-labelled DNA with base-specific chemical cleavages. In “Methods in Enzymology” (L. Grossman and K. Moldave, eds.), Vol. 65, pp. 499-560.

MESSING, J., CREA, R., AND SEEBURG, P. H. (198 I). A system for shotgun DNA sequencing. Nucleic. Acids Rex ,9, 309-321.

MORIYA, S., OGASAWARA, N., AND YOSHIKAWA, H. ( 1985). Structure and function of the region of the rep l&ion origin of the Bacillus subtilis chromosome. III. Nucleotide sequence of some 10,000 base pairs in the origin region. Nucleic Acids Res. 13, 225 l-2265.

MOWN, C. P., LANG, N., LEGRICE, S. F. J., LEE, G., STEPHENS, M., SONENSHEIN, A. L., PERO, J., AND Los ICK, R. ( 1982). Nucleotide sequences that signal the ini- tiation of transcription and translation in Bacillus sub- tilis. Mol. Gen. Genet. 186, 339-346.

MURRAY, C. L., AND RABINOWITZ, J. C. (I 982). Species specific translation: Characterization of B. subtilis ri- bosome binding sites. In “Molecular Cloning and Gene Regulation in Bacilli.” (A. T. Ganesan, S. Chang, and J. A. Hoch, eds.), pp. 271-285.

NOVICK, R. P., ADLER, G. K., MAJUMDER, S., KAHN, S. A., CARLETON, S., ROSENBLUM, W. D., ANDIORDA- NESCU, S. ( 1982). Coding sequence for the pT18 1 repC product: A plasmid-ccded protein uniquely required for replication. Proc. Natl. Acad. Sci. USA 79,4108-4112.

pUBl10 NUCLEOTIDE SEQUENCE 103

PABO, C. O., ANDSAUER, R. T. (1984). Protein-DNA rec- ognition. Annu. Rev. Biochem. 53,293-32 1.

POLAK, J., AND NOVICK, R. P. (1982). Closely related plasmids from Staphylococcus aureus and soil bacilli. Plasmid 7, 152-162.

SANGER, F., NICKLEN, S., AND COULSON, A. R. (1977). DNA sequencing with chain terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463-5467.

SCHEER-ABRAMOWITZ, J., GRYCZAN, T. J., AND DUBNAU, D. (1980). Origin and mode of replication of plasmids pE194 and pUBll0. Plasmid 6,67-71.

SCOTT, J. R. (1984). Regulation of plasmid replication. Microbial. Rev. 48, l-23.

SHON, M., GERMINO, J., AND BASTIA, D. (1982). The nu- cleotide sequence of the replication origin /3 of the plas- mid R6K. J. Biol. Chem. 257, 13823-13827.

SUEOKA, N., AND HAMMERS, J. M. (1974). Isolation of DNA-membrane complex in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 71,4787-479 1.

SUEOKA, N., KORN, R., MCKENZIE, T., TANAKA, T., AND WINSTON, S. (1984). Two types of binding of PUB110 to Bacillus subtilis membrane. In “Genetics and Bio-

technology of Bacilli” (A. Ganesan and J. Hoch, eds.) pp. 79-88, Academic Press.

TANAKA, T., AND KAWANO, N. (1980). Cloning vehicles for the homologous Bacillus subtilis host-vector system. Gene 10, 131-136.

TANAKA, T., AND SUEOKA, N. (1983). Site-specific in vitro binding of plasmid PUB 110 to Bacillus subtilis mem- brane fraction. J. Bacterial. 154, 1184-l 194.

WHITE, K., AND SUEOKA, N. (1973). Temperature-sen- sitive DNA synthesis mutants of BaciNus subtihs - Appendix: Theory of density transfer for symmetric chromosome replication. Genetics 73, 185-2 14.

WINSTON, S., AND SUEOKA, N. (1980). DNA-membrane association is necessary for initiation of chromosomal and plasmid replication in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 77,2834-2838.

ZYSKIND, J. W., CLEARLY, J. M., BRUSILOW, W. S. A., HARDING, N. E., AND SMITH, D. W. (1983). Chromo- somal replication origin from the marine bacterium Vibrio haveyi functions in Escherichia coli:oriC con- sensus sequence. Proc. Natl. Acad. Sci. USA 80, 1164- 1168.