- Email: [email protected]
A model for initiation at origins of DNA replication
Cell, Vol. 54, 915-918,
September
23, 1998, Copyright
0 1988 by Cell Press
A Model for Initiation at Origins of DNA Replication David Bramhill and Arthur Kornberg Department of Biochemistry Stanford University School of Medicine Stanford, California 94305
The initiation events that establish replication forks at the chromosomal origin of Escherichia coli (oriC) have recently been elucidated (Bramhill and Kornberg, Cell 52, 743-755,1988). After the initiator protein (dnaA) binds oriC and melts three AT-rich tandem repeats, the dnaB helicase enters the “bubble:’ unwinds the template, and directs priming of DNA replication. In prokaryotes, a number of duplex DNA origins resemble oriC in size (200-250 bp) and structure (Zyskind and Smith, Cell 46, 489-490, 1986). We have discovered similar iterated sequences in the AT-rich regions of other replication origins, suggesting that the model proposed for oriC may also apply to initiation at other origins. A Model for Initiation at oriC Our model for initiation at oriC postulates that dnaA protein performs three crucial roles (Figure 1): l
l
l
it binds tightly to four 9-mer repeats (called dnaA boxes) to form an initial complex; it successively melts three AT-rich 19mer repeats to form an open complex; and it guides the dnaB-dnaC complex into this melted region to form aprepriming complex, which marks the future forks of bidirectional DNA replication.
Stepwise melting of the 13-mer repeats provides a mechanism for exposing a large defined region of DNA (-45 bp) without requiring that the entire sequence be recognized by a single protein molecule. In addition, it is energetically more favorable than single-step melting since the energy is required in smaller increments. The evidence for this initiation model is provided by our own recent work (Bramhill and Kornberg, op. cit.) as well
Minireview
as a number of earlier insights (see reviews by Zyskind and Smith, op. cit.; Messer, J. Bacterial. 769, 3395-3399, 1987). The 9-mers and 13-mers are highly conserved among enterobacterial origins and coincide precisely with the functionally defined limits of oriC (Zyskind et al., PNAS 80, 1184-1168, 1983). Mutational analysis has shown that both the 13-mer and 9-mer sequences are essential for origin function, as is the orientation and spacing of the rightmost 13-mer (Oka et al., JMB 776, 443-458, 1984). The structure of the initial complex has been deduced from electron microscopy and protection from DNAase I digestion. It consists of negatively supercoiled oriC DNA wrapped around a central core of 20-40 dnaA protein monomers (Funnell et al., JBC 262, 10327-10334, 1987; Fuller and Kornberg, PNAS 80,5817-5821, 1983). Interaction of the dnaA protein subunits in this higher-order complex enables the precise relative spacing and orientation of the four 9-mers to be recognized. Subunits of dnaA protein within the initial complex can then act to melt the ATrich region of oriC to form an open complex. Sensitivity to the single-strand-specific Pl nuclease locates the opening in the 13-mers. Presumably, starting with the nearest (rightmost) 13-mer, successive opening of the middle and then the leftmost 13-mer follows. This melting requires ATP and dnaA protein acting at a temperature above 22%. Mutants of oriC that retain only a single 13-mer (the rightmost) can form an open complex; templates lacking all three 13-mers are inactive. All three 13-mers are required to form a prepriming complex. They provide a 45 bp region for the entry of the dnaB-dnaC complex, which supplies the dnaf3 helicase function. The single-stranded structure of the 13-mer DNA, rather than its sequence, is likely recognized by dnaB. The prepriming complex remains stable even at 0% and exhibits a broader susceptibility to Pl endonuclease cleavage than does the open complex (Bramhill and Kornberg, op. cit.). In addition to its role in opening the duplex, dnaA protein may also direct dnaB into the complex. At tempera-
HU ATP*dnaA
Figure
38’ dnaB dnaC ATP
SUPERCOILED TEMPLATE
PRIMING AND REPLICATION PREPRIMING
1. A Scheme
for Initiation
at oriC
The dnaA protein binds the four 9-mers, organizing oriC around a protein core to form the initial complex. The three 15mers are then melted serially by dnaA protein to create the open complex. The dnaS-dnaC complex can now be directed to the 15mer region to extend the duplex opening and generate a prepriming complex, which unwinds the template for priming and replication. Details are given in Sramhill and Kornberg (Cell 52, 743-755, 1988).
tures too low to allow open or prepriming complex formation (Funnell et al., op. cit.), a dnaC-dependent association of dnaB with dnaA at oriC can be observed by electron microscopy. The dnaB protein, once established in the duplex, can then perform its two major roles: first, as a helicase to unwind the duplex and expose a template for priming (with single-strand binding protein stabilizing the exposed single strands); and second, as a guide to primase to lay down the RNA primers for DNA synthesis by DNA polymerase III holoenzyme. A General Model for Initiation When we examined the sequences of a wide variety of other prokaryotic replication origins, we discovered that many contain AT-rich repeats (Figure 2). Moreover, from the available footprinting data we deduced that these origins share a common motif for binding proteins. In each case there is evidence that an initiator protein binds tightly to one portion of the origin, usually via specific repeated sequences. A nearby region, which encompasses the ATrich repeats, appears to bind weakly to the initiator protein. For the most part, the AT-rich repeats differ between origins; however, in those instances where the repeats are similar, the initator proteins are also similar or even identical. We propose that many of these origins employ a mechanism of initiation similar to that described for oriC. According to this general model, an origin-specific protein acts at its cognate origin in place of, or in concert with, dnaA protein. This initiator protein acts sequentially on two types of reiterated sequences. One sequence localizes the protein’s site of action; the other sequence, which is AT-rich, facilitates duplex opening. In addition, protein-protein interactions may be required to deliver dnaB helicase efficiently to the melted region. Possible Initiation Events at Other Origins Bacillus subtilis: The chromosomal origin of B. subtilis contains a segment similar to the oriC of Enterobacteriaceae. Within a 200 bp sequence a series of 9-mer dnaA boxes lie adjacent to three AT-rich 16-mer repeats in tandem. These Id-mers match eight of the eleven conserved positions in the 13mer consensus of E. coli oriC. This high degree of conservation suggests a functional role for these sequences, and is consistent with the strong homology between E. coli dnaA protein and the putative dnaA protein of B. subtilis (Moriya et al., NAR 73, 2251-2258, 1985). Thus, as in E. coli oriC, initiation at this origin most likely employs dnaA protein to melt the 16-mers and guide helicase entry. Plasmid pSC101: Here we propose the concerted action of dnaA protein and a plasmid-specific protein. Initiation at the origin of pSC101 requires both dnaA and the plasmid-encoded rep protein. Rep protein binds tightly to repeated sequences in one portion of the origin and wraps the DNA around itself. The AT-rich segment of the origin, which does not appear to bind rep protein, contains both a 9-mer dnaA box and two tandem repeats of the oriC 13mer. Plausibly, following the binding of rep protein, the dnaA protein directs the melting of the 13-mers and guides dnaB entry. Plasmids F, Pl, Rl, and R6K: At these origins, the prin-
cipal role in opening the duplex does not belong to dnaA protein, but rather to a plasmid-encoded initiator protein. Each origin possesses a series of tandem repeats within its AT-rich region and one or two dnaA boxes. However, these AT-rich repeats differ from the oriC 13-mers and differ among plasmids. Each plasmid encodes its own specific initiator protein, which binds tightly to one portion of its origin. Since the AT-rich repeats are specific to each origin, it is likely that the specific initiator factor recognizes and opens them. The dnaA protein is required for replication, probably to guide dnaB helicase entry to the opened region. Supporting the involvement of these AT-rich repeat sequences, Pi plasmid replication requires methylation by the dam-encoded methyl transferase (Abeles and Austin, EMBO J. 6318%3189,1987) of the adenine residue in the 5’-GATC-3’ sequence in the AT-rich repeats of Pl. Lambdoid Phages: Tandemly repeated motifs are present in the AT-rich portions of these origins. The phageencoded 0 protein binds a series of “iterons:’ which make up part of the origin (stippled in Figure 2). We propose that 0 protein sequentially melts the AT-rich repeats and then directs dnaB helicase entry via the phage-encoded P protein, a dnaC analog. Replication of these phages is independent of dnaA function, and the characteristic dnaA box sequence is absent from their origins. The near identity of the 1%mer repeats in the AT-rich regions of phages h and (~82 fits the close similarity of their 0 proteins. Both the N-terminal and the C-terminal domains of the 0 proteins of the two phages are identical; of the central 140 amino acids, 88 are identical (Moore et al., Gene 74, 91-101, 1981). Phage (~80, which is more distantly related, has a different repeat motif and also a distinctive N-terminal sequence in its 0 protein (Grosschedl and Hobom, Nature 277, 621-627, 1979). Functional studies reveal that the AT-rich region of the h origin is indeed the target for the entry and initial action of dnaB helicase (Dodson et al., PNAS 83, 7638-7642, 1986). Point mutations or deletions located in any one of the three 11-mer repeats abolish origin function. By contrast, a deletion of sequences to within 2 bp of the end of the leftmost repeat does not interfere with origin function (Wickner and McKenney, JBC 262, 13163-13167, 1987). Strong evidence for the generality of our model is found in a recent report on the 0 protein by Schnos et al. (Cell 52,385-395,1988). These authors show that tight binding of the 0 protein to the h origin induces a conformational change that specifically sensitizes the adjacent AT-rich region to the single-strand-specific nuclease Sl. The 1 I-mer motif we have identified can explain the behavior of h ori- mutants. The pattern observed by Schnos et al. (op. cit.) is similar to that for oriC: Deletion of the left and middle 11-mers abolishes dnaB helicase entry and replication but still allows opening by 0 protein of the intact right 11-mer. Mutations that delete or alter the right llmer prevent both replication and duplex opening, presumably by blocking the first melting event. These later findings also suggest that, as in E. coli OK, the spacing and orientation of the first repeat are critical for opening. Plasmids P4 and RK2: These broad-host-range repli-
Minireview 917
CONSENSUS -AT-RICH-
ORIGIN
13
Erofi
-TIGHTLY
13
13
A
BOUNDdnaA protein A
SEQUENCE A
ori c
B.subfilis
on’ C A
13
pSClO1
13
A
NUMBER
OF MATCHES
5
GATCTnTTnTTTT
12,12,11
CATACCTTAnTTTTTC
14,14,15
GATCTnTTnTTTT
13,ll
repA
psclol
AGATCC;
Pl
TTTTTA
F
3
7,
6,
7,
7,
a,
a,
a,
8
9
Rl
TTTAAAnGA
9,
9,
R6K ori Y
TATTnATTTT
9,
a,10
TTnTCTTTTGT
9,11,10
TTGTCTTTTGT TCTTGT
6 80
11,ll
5,
6,
6
RK2
GGTT;fAAAA
9,
a,
9,
P4
CACTTAAAG
9,
8,
9
Figure
2. Comparison
of AT-rich
Repeats
in Prokaryotic
7
8
Origins
The arrows Indicate the locations of the AT-rich repeats in each origin, with the length (in bp) marked above each; the same styles of arrows are used for identical and near-identical repeats. Stippling identifies the sequences that bind tightly to the origin-specific initiator protein; in many cases these sequences contain tandem repeats. The stippled boxes marked “A match the 9-mer DNA box. Assignments are based on footprinting or filter binding studies, or are inferred from the close structural similarity to an origin defined by footprinting. In the consensus sequences for the AT-rich repeats, the nonspecific positions are marked “n”; the number of matches to the entire consensus (nonspecific positions included) is given for each repeat, going from left to right. Sequences are from the following sources. E. coli oriC: Meijer et al., PNAS 76, 580-584, 1979; Oka et al., Mol. Gen. Genet. 178, 9-20, 1980. 6. subtilis oriC: Moriya et al., NAR 13,2251-2258,1985. pSC101: Vocke and Bastia, PNAS 80,6557-6561, 1983; Churchward et al., NAR 17, 5645-5659, 1983; Armstrong et al., JMB 175. 331-347, 1984; Yamaguchi and Yamaguchi, Gene 29, 211-219, 1984. PI: Abeles et al., JMB 173, 307-324, 1984. F: Murotsu et al., Gene 15. 257-271, 1981. RI: Ryder et al., in hitiation of DNA Replication, ed. D. S. Ray, Academic Press, pp. 91-111, 1981; Masai et al., PNAS 80, 6814-6818, 1983. R6K: Stalker et al., PNAS 76, 1150-1154, 1979. 1 and (~82: Moore et al., Gene 14, 91-101, 1981. (~80: Grosschedl and Hobom, Nature 277,621-627, 1979. RK2: Stalker et al., Mol. Gen. Genet. 781,8-12, 1981. P4: Flensburg and Calendar, JMB 195,439-445, 1987.
cons also possess tandem repeats in the AT-rich region of their origins. Conceivably, their cognate initiator proteins open these iterons after binding to adjacent sequences. As is true for E. coli DNA, in vitro replication of P4 requires a supercoiled template (Krevolin and Calendar, JMB 782, 509-517, 1985). Plasmid ColEl: The absence of any AT-rich tandem repeats near the origins of ColEl and related plasmids is in keeping with their different initiation mechanism, which depends on RNA polymerase for both origin recognition and for site-specific helicase activity (Masukata et al., Cell 51, 1123-1130, 1987; Minden and Marians, JBC 260, 9316-9325, 1985). Eukaryotes: A role for AT-rich sequences in opening of the duplex at the replication origin seems likely for the DNA tumor virus SV40. The origin is tightly bound by T antigen, the viral-encoded initiator protein. A short AT-rich region nearby, which is not protected in footprinting
studies, is nevertheless essential for replication (Stillman et al., EMS0 J. 4,2933-2939,1985). The T antigen serves a dual role as helicase and initiator protein (Stahl et al., EMS0 J. 5, 1939-1944, 1986) and unwinds the duplex from the origin (Dodson et al., Science 238, 964-967, 1987). Regulatory Factors ATP: Nucleotide binding by dnaA protein has a profound effect on the activity of dnaA protein at oriC (Sekimizu et al., Cell 50, 259-265, 1987). An extraordinarily tight and stable ATP-dnaA complex is formed, within which the ATP is hydrolyzed very slowly in a DNA-dependent manner. Although the ADP-form of dnaA protein binds to oriC with an affinity similar to that of the ATP-form, only the latter is active for initiation. The ADP-form is incapable of duplex opening to generate an open complex. Rl replication can be supported by the ADP-form of the protein, implying that dnaA protein is not acting in duplex melting in
Cell 918
Rl initiation (Masai and Arai, PNAS 84,4781-4785, 1987). DnaC Protein: The presence of even one dnaA-box is accompanied by a requirement for dnaC protein for initiation. Phages y, (~80, and (~82 and the promiscuous P4 replicon lack a dnaA box and show no dependence on dnaC. Instead, the lambdoid phages encode a dnaC analog that interacts with dnaB and 0 proteins directly (LeBowitz et al., PNAS 82, 3988-3992, 1985); P4 is independent not only of dnaC but also of dna6 and primase. Although the replication of plasmids Pl, F, and RI is less impaired in a dnaAtS strain than replication of either pSC101 or E. coli chromosomal DNA, these plasmids nevertheless require dnaA protein-presumably for the entry of dnaB helicase. Transcriptional Activation of the Origin: Although primase provides the priming function in vitro and very likely in vivo, RNA polymerase action is still required for initiation under most circumstances. With a slightly lowered temperature or some relaxation of the supercoiled template (by topoisomerase or by the histone-like HU protein), dnaA protein fails to open oriC unless aided by the introduction of a transcript creating an R-loop near the origin (Baker and Kornberg, Cell 55, in press, 1988). Inasmuch as transcriptional activation requires an intact OriC, including the three 13-mers (Bramhill and Kornberg, op. cit.) as well as the ATP-form of dnaA protein (Sekimizu et al., op. cit.) it seems likely that the RNA transcript assists dnaA in melting the 13-mers. This mechanism for altering the topological State of the origin probably accounts for the transcriptional activation first observed in the initiation
of phage li replication (Further and Wickner, in Lambda II, eds. R. W. Hendrix et al., Cold Spring Harbor Laboratory, pp. 145-173, 1983) and may apply to many genomes, prokaryotic and eukaryotic. Thus, the level and nature of transcriptional activity near the origin may be crucial to the function of the initiator protein. In some cases (e.g., ColEl), the transcript may provide both activating and priming functions (Masukata et al., op. cit.). Membrane Attachment: Binding of dnaA protein to the head groups of acidic phospholipids in a fluid bilayer destabilizes a firmly bound nucleotide and can rejuvenate the inactive ADP-form to the ATP-form (Sekimizu and Kornberg, JBC 263,7131-71351988; Yung and Kornberg, PNAS, in press, 1988). The critical importance of membrane fluidity for chromosome initiation in vivo (Fralick and Lark, JMB 80,459-475, 1973) can now be explained by its influence on dnaA protein behavior in vitro, thus providing strong evidence that the location and orientation of the initiation protein may regulate its function. Summary Many prokaryotic origins resemble E. coli oriC in possessing essential AT-rich sequences, tandemly repeated. The role of these repeats may be in the initial opening of the duplex by the initiator protein, as has been found for the 13mers in oriC and is implied for the 11-mers of the h origin. Regulatory influences on the effective action of the initiator protein of E. coli (dnaA protein) include transcriptional activation of the origin, nucleotide binding and membrane attachment of the protein, and interactions leading to the introduction of helicases to start replication forks.
September
23, 1998, Copyright
0 1988 by Cell Press
A Model for Initiation at Origins of DNA Replication David Bramhill and Arthur Kornberg Department of Biochemistry Stanford University School of Medicine Stanford, California 94305
The initiation events that establish replication forks at the chromosomal origin of Escherichia coli (oriC) have recently been elucidated (Bramhill and Kornberg, Cell 52, 743-755,1988). After the initiator protein (dnaA) binds oriC and melts three AT-rich tandem repeats, the dnaB helicase enters the “bubble:’ unwinds the template, and directs priming of DNA replication. In prokaryotes, a number of duplex DNA origins resemble oriC in size (200-250 bp) and structure (Zyskind and Smith, Cell 46, 489-490, 1986). We have discovered similar iterated sequences in the AT-rich regions of other replication origins, suggesting that the model proposed for oriC may also apply to initiation at other origins. A Model for Initiation at oriC Our model for initiation at oriC postulates that dnaA protein performs three crucial roles (Figure 1): l
l
l
it binds tightly to four 9-mer repeats (called dnaA boxes) to form an initial complex; it successively melts three AT-rich 19mer repeats to form an open complex; and it guides the dnaB-dnaC complex into this melted region to form aprepriming complex, which marks the future forks of bidirectional DNA replication.
Stepwise melting of the 13-mer repeats provides a mechanism for exposing a large defined region of DNA (-45 bp) without requiring that the entire sequence be recognized by a single protein molecule. In addition, it is energetically more favorable than single-step melting since the energy is required in smaller increments. The evidence for this initiation model is provided by our own recent work (Bramhill and Kornberg, op. cit.) as well
Minireview
as a number of earlier insights (see reviews by Zyskind and Smith, op. cit.; Messer, J. Bacterial. 769, 3395-3399, 1987). The 9-mers and 13-mers are highly conserved among enterobacterial origins and coincide precisely with the functionally defined limits of oriC (Zyskind et al., PNAS 80, 1184-1168, 1983). Mutational analysis has shown that both the 13-mer and 9-mer sequences are essential for origin function, as is the orientation and spacing of the rightmost 13-mer (Oka et al., JMB 776, 443-458, 1984). The structure of the initial complex has been deduced from electron microscopy and protection from DNAase I digestion. It consists of negatively supercoiled oriC DNA wrapped around a central core of 20-40 dnaA protein monomers (Funnell et al., JBC 262, 10327-10334, 1987; Fuller and Kornberg, PNAS 80,5817-5821, 1983). Interaction of the dnaA protein subunits in this higher-order complex enables the precise relative spacing and orientation of the four 9-mers to be recognized. Subunits of dnaA protein within the initial complex can then act to melt the ATrich region of oriC to form an open complex. Sensitivity to the single-strand-specific Pl nuclease locates the opening in the 13-mers. Presumably, starting with the nearest (rightmost) 13-mer, successive opening of the middle and then the leftmost 13-mer follows. This melting requires ATP and dnaA protein acting at a temperature above 22%. Mutants of oriC that retain only a single 13-mer (the rightmost) can form an open complex; templates lacking all three 13-mers are inactive. All three 13-mers are required to form a prepriming complex. They provide a 45 bp region for the entry of the dnaB-dnaC complex, which supplies the dnaf3 helicase function. The single-stranded structure of the 13-mer DNA, rather than its sequence, is likely recognized by dnaB. The prepriming complex remains stable even at 0% and exhibits a broader susceptibility to Pl endonuclease cleavage than does the open complex (Bramhill and Kornberg, op. cit.). In addition to its role in opening the duplex, dnaA protein may also direct dnaB into the complex. At tempera-
HU ATP*dnaA
Figure
38’ dnaB dnaC ATP
SUPERCOILED TEMPLATE
PRIMING AND REPLICATION PREPRIMING
1. A Scheme
for Initiation
at oriC
The dnaA protein binds the four 9-mers, organizing oriC around a protein core to form the initial complex. The three 15mers are then melted serially by dnaA protein to create the open complex. The dnaS-dnaC complex can now be directed to the 15mer region to extend the duplex opening and generate a prepriming complex, which unwinds the template for priming and replication. Details are given in Sramhill and Kornberg (Cell 52, 743-755, 1988).
tures too low to allow open or prepriming complex formation (Funnell et al., op. cit.), a dnaC-dependent association of dnaB with dnaA at oriC can be observed by electron microscopy. The dnaB protein, once established in the duplex, can then perform its two major roles: first, as a helicase to unwind the duplex and expose a template for priming (with single-strand binding protein stabilizing the exposed single strands); and second, as a guide to primase to lay down the RNA primers for DNA synthesis by DNA polymerase III holoenzyme. A General Model for Initiation When we examined the sequences of a wide variety of other prokaryotic replication origins, we discovered that many contain AT-rich repeats (Figure 2). Moreover, from the available footprinting data we deduced that these origins share a common motif for binding proteins. In each case there is evidence that an initiator protein binds tightly to one portion of the origin, usually via specific repeated sequences. A nearby region, which encompasses the ATrich repeats, appears to bind weakly to the initiator protein. For the most part, the AT-rich repeats differ between origins; however, in those instances where the repeats are similar, the initator proteins are also similar or even identical. We propose that many of these origins employ a mechanism of initiation similar to that described for oriC. According to this general model, an origin-specific protein acts at its cognate origin in place of, or in concert with, dnaA protein. This initiator protein acts sequentially on two types of reiterated sequences. One sequence localizes the protein’s site of action; the other sequence, which is AT-rich, facilitates duplex opening. In addition, protein-protein interactions may be required to deliver dnaB helicase efficiently to the melted region. Possible Initiation Events at Other Origins Bacillus subtilis: The chromosomal origin of B. subtilis contains a segment similar to the oriC of Enterobacteriaceae. Within a 200 bp sequence a series of 9-mer dnaA boxes lie adjacent to three AT-rich 16-mer repeats in tandem. These Id-mers match eight of the eleven conserved positions in the 13mer consensus of E. coli oriC. This high degree of conservation suggests a functional role for these sequences, and is consistent with the strong homology between E. coli dnaA protein and the putative dnaA protein of B. subtilis (Moriya et al., NAR 73, 2251-2258, 1985). Thus, as in E. coli oriC, initiation at this origin most likely employs dnaA protein to melt the 16-mers and guide helicase entry. Plasmid pSC101: Here we propose the concerted action of dnaA protein and a plasmid-specific protein. Initiation at the origin of pSC101 requires both dnaA and the plasmid-encoded rep protein. Rep protein binds tightly to repeated sequences in one portion of the origin and wraps the DNA around itself. The AT-rich segment of the origin, which does not appear to bind rep protein, contains both a 9-mer dnaA box and two tandem repeats of the oriC 13mer. Plausibly, following the binding of rep protein, the dnaA protein directs the melting of the 13-mers and guides dnaB entry. Plasmids F, Pl, Rl, and R6K: At these origins, the prin-
cipal role in opening the duplex does not belong to dnaA protein, but rather to a plasmid-encoded initiator protein. Each origin possesses a series of tandem repeats within its AT-rich region and one or two dnaA boxes. However, these AT-rich repeats differ from the oriC 13-mers and differ among plasmids. Each plasmid encodes its own specific initiator protein, which binds tightly to one portion of its origin. Since the AT-rich repeats are specific to each origin, it is likely that the specific initiator factor recognizes and opens them. The dnaA protein is required for replication, probably to guide dnaB helicase entry to the opened region. Supporting the involvement of these AT-rich repeat sequences, Pi plasmid replication requires methylation by the dam-encoded methyl transferase (Abeles and Austin, EMBO J. 6318%3189,1987) of the adenine residue in the 5’-GATC-3’ sequence in the AT-rich repeats of Pl. Lambdoid Phages: Tandemly repeated motifs are present in the AT-rich portions of these origins. The phageencoded 0 protein binds a series of “iterons:’ which make up part of the origin (stippled in Figure 2). We propose that 0 protein sequentially melts the AT-rich repeats and then directs dnaB helicase entry via the phage-encoded P protein, a dnaC analog. Replication of these phages is independent of dnaA function, and the characteristic dnaA box sequence is absent from their origins. The near identity of the 1%mer repeats in the AT-rich regions of phages h and (~82 fits the close similarity of their 0 proteins. Both the N-terminal and the C-terminal domains of the 0 proteins of the two phages are identical; of the central 140 amino acids, 88 are identical (Moore et al., Gene 74, 91-101, 1981). Phage (~80, which is more distantly related, has a different repeat motif and also a distinctive N-terminal sequence in its 0 protein (Grosschedl and Hobom, Nature 277, 621-627, 1979). Functional studies reveal that the AT-rich region of the h origin is indeed the target for the entry and initial action of dnaB helicase (Dodson et al., PNAS 83, 7638-7642, 1986). Point mutations or deletions located in any one of the three 11-mer repeats abolish origin function. By contrast, a deletion of sequences to within 2 bp of the end of the leftmost repeat does not interfere with origin function (Wickner and McKenney, JBC 262, 13163-13167, 1987). Strong evidence for the generality of our model is found in a recent report on the 0 protein by Schnos et al. (Cell 52,385-395,1988). These authors show that tight binding of the 0 protein to the h origin induces a conformational change that specifically sensitizes the adjacent AT-rich region to the single-strand-specific nuclease Sl. The 1 I-mer motif we have identified can explain the behavior of h ori- mutants. The pattern observed by Schnos et al. (op. cit.) is similar to that for oriC: Deletion of the left and middle 11-mers abolishes dnaB helicase entry and replication but still allows opening by 0 protein of the intact right 11-mer. Mutations that delete or alter the right llmer prevent both replication and duplex opening, presumably by blocking the first melting event. These later findings also suggest that, as in E. coli OK, the spacing and orientation of the first repeat are critical for opening. Plasmids P4 and RK2: These broad-host-range repli-
Minireview 917
CONSENSUS -AT-RICH-
ORIGIN
13
Erofi
-TIGHTLY
13
13
A
BOUNDdnaA protein A
SEQUENCE A
ori c
B.subfilis
on’ C A
13
pSClO1
13
A
NUMBER
OF MATCHES
5
GATCTnTTnTTTT
12,12,11
CATACCTTAnTTTTTC
14,14,15
GATCTnTTnTTTT
13,ll
repA
psclol
AGATCC;
Pl
TTTTTA
F
3
7,
6,
7,
7,
a,
a,
a,
8
9
Rl
TTTAAAnGA
9,
9,
R6K ori Y
TATTnATTTT
9,
a,10
TTnTCTTTTGT
9,11,10
TTGTCTTTTGT TCTTGT
6 80
11,ll
5,
6,
6
RK2
GGTT;fAAAA
9,
a,
9,
P4
CACTTAAAG
9,
8,
9
Figure
2. Comparison
of AT-rich
Repeats
in Prokaryotic
7
8
Origins
The arrows Indicate the locations of the AT-rich repeats in each origin, with the length (in bp) marked above each; the same styles of arrows are used for identical and near-identical repeats. Stippling identifies the sequences that bind tightly to the origin-specific initiator protein; in many cases these sequences contain tandem repeats. The stippled boxes marked “A match the 9-mer DNA box. Assignments are based on footprinting or filter binding studies, or are inferred from the close structural similarity to an origin defined by footprinting. In the consensus sequences for the AT-rich repeats, the nonspecific positions are marked “n”; the number of matches to the entire consensus (nonspecific positions included) is given for each repeat, going from left to right. Sequences are from the following sources. E. coli oriC: Meijer et al., PNAS 76, 580-584, 1979; Oka et al., Mol. Gen. Genet. 178, 9-20, 1980. 6. subtilis oriC: Moriya et al., NAR 13,2251-2258,1985. pSC101: Vocke and Bastia, PNAS 80,6557-6561, 1983; Churchward et al., NAR 17, 5645-5659, 1983; Armstrong et al., JMB 175. 331-347, 1984; Yamaguchi and Yamaguchi, Gene 29, 211-219, 1984. PI: Abeles et al., JMB 173, 307-324, 1984. F: Murotsu et al., Gene 15. 257-271, 1981. RI: Ryder et al., in hitiation of DNA Replication, ed. D. S. Ray, Academic Press, pp. 91-111, 1981; Masai et al., PNAS 80, 6814-6818, 1983. R6K: Stalker et al., PNAS 76, 1150-1154, 1979. 1 and (~82: Moore et al., Gene 14, 91-101, 1981. (~80: Grosschedl and Hobom, Nature 277,621-627, 1979. RK2: Stalker et al., Mol. Gen. Genet. 781,8-12, 1981. P4: Flensburg and Calendar, JMB 195,439-445, 1987.
cons also possess tandem repeats in the AT-rich region of their origins. Conceivably, their cognate initiator proteins open these iterons after binding to adjacent sequences. As is true for E. coli DNA, in vitro replication of P4 requires a supercoiled template (Krevolin and Calendar, JMB 782, 509-517, 1985). Plasmid ColEl: The absence of any AT-rich tandem repeats near the origins of ColEl and related plasmids is in keeping with their different initiation mechanism, which depends on RNA polymerase for both origin recognition and for site-specific helicase activity (Masukata et al., Cell 51, 1123-1130, 1987; Minden and Marians, JBC 260, 9316-9325, 1985). Eukaryotes: A role for AT-rich sequences in opening of the duplex at the replication origin seems likely for the DNA tumor virus SV40. The origin is tightly bound by T antigen, the viral-encoded initiator protein. A short AT-rich region nearby, which is not protected in footprinting
studies, is nevertheless essential for replication (Stillman et al., EMS0 J. 4,2933-2939,1985). The T antigen serves a dual role as helicase and initiator protein (Stahl et al., EMS0 J. 5, 1939-1944, 1986) and unwinds the duplex from the origin (Dodson et al., Science 238, 964-967, 1987). Regulatory Factors ATP: Nucleotide binding by dnaA protein has a profound effect on the activity of dnaA protein at oriC (Sekimizu et al., Cell 50, 259-265, 1987). An extraordinarily tight and stable ATP-dnaA complex is formed, within which the ATP is hydrolyzed very slowly in a DNA-dependent manner. Although the ADP-form of dnaA protein binds to oriC with an affinity similar to that of the ATP-form, only the latter is active for initiation. The ADP-form is incapable of duplex opening to generate an open complex. Rl replication can be supported by the ADP-form of the protein, implying that dnaA protein is not acting in duplex melting in
Cell 918
Rl initiation (Masai and Arai, PNAS 84,4781-4785, 1987). DnaC Protein: The presence of even one dnaA-box is accompanied by a requirement for dnaC protein for initiation. Phages y, (~80, and (~82 and the promiscuous P4 replicon lack a dnaA box and show no dependence on dnaC. Instead, the lambdoid phages encode a dnaC analog that interacts with dnaB and 0 proteins directly (LeBowitz et al., PNAS 82, 3988-3992, 1985); P4 is independent not only of dnaC but also of dna6 and primase. Although the replication of plasmids Pl, F, and RI is less impaired in a dnaAtS strain than replication of either pSC101 or E. coli chromosomal DNA, these plasmids nevertheless require dnaA protein-presumably for the entry of dnaB helicase. Transcriptional Activation of the Origin: Although primase provides the priming function in vitro and very likely in vivo, RNA polymerase action is still required for initiation under most circumstances. With a slightly lowered temperature or some relaxation of the supercoiled template (by topoisomerase or by the histone-like HU protein), dnaA protein fails to open oriC unless aided by the introduction of a transcript creating an R-loop near the origin (Baker and Kornberg, Cell 55, in press, 1988). Inasmuch as transcriptional activation requires an intact OriC, including the three 13-mers (Bramhill and Kornberg, op. cit.) as well as the ATP-form of dnaA protein (Sekimizu et al., op. cit.) it seems likely that the RNA transcript assists dnaA in melting the 13-mers. This mechanism for altering the topological State of the origin probably accounts for the transcriptional activation first observed in the initiation
of phage li replication (Further and Wickner, in Lambda II, eds. R. W. Hendrix et al., Cold Spring Harbor Laboratory, pp. 145-173, 1983) and may apply to many genomes, prokaryotic and eukaryotic. Thus, the level and nature of transcriptional activity near the origin may be crucial to the function of the initiator protein. In some cases (e.g., ColEl), the transcript may provide both activating and priming functions (Masukata et al., op. cit.). Membrane Attachment: Binding of dnaA protein to the head groups of acidic phospholipids in a fluid bilayer destabilizes a firmly bound nucleotide and can rejuvenate the inactive ADP-form to the ATP-form (Sekimizu and Kornberg, JBC 263,7131-71351988; Yung and Kornberg, PNAS, in press, 1988). The critical importance of membrane fluidity for chromosome initiation in vivo (Fralick and Lark, JMB 80,459-475, 1973) can now be explained by its influence on dnaA protein behavior in vitro, thus providing strong evidence that the location and orientation of the initiation protein may regulate its function. Summary Many prokaryotic origins resemble E. coli oriC in possessing essential AT-rich sequences, tandemly repeated. The role of these repeats may be in the initial opening of the duplex by the initiator protein, as has been found for the 13mers in oriC and is implied for the 11-mers of the h origin. Regulatory influences on the effective action of the initiator protein of E. coli (dnaA protein) include transcriptional activation of the origin, nucleotide binding and membrane attachment of the protein, and interactions leading to the introduction of helicases to start replication forks.