- Email: [email protected]
Protein Structure: Why have six-fold symmetry?
Z. KELMAN, J. FINKELSTEIN
Z.
KELMAN,
J.
AND M. O'DONNELL
FINKELSTEIN AND
M.
O'DONNELL
PROTEIN STRUCTURE
PROTEIN STRUCTURE
Why have six-fold symmetry? Why do proteins that encircle DNA have six-fold symmetry? One important factor may be the economy in protein mass with which DNA can be encircled by six globular subunits arranged in a ring. Ring-shaped oligomeric proteins are involved in a diverse range of cellular processes. A subset of these protein oligomers encircle DNA, providing them with two seemingly opposed properties: the ability to grip DNA tightly, and the ability to move along the DNA. This group of proteins includes processivity factors of the replicative DNA polymerases, DNA helicases and a protein that resolves Holliday junction recombination intermediates. So far, all proteins examined that can encircle DNA and translocate along the double helix have been found to have either true or pseudo six-fold symmetry. This symmetry is achieved in different ways by proteins with different oligomeric states - for example, some are simple hexamers with true six-fold symmetry, whereas others are trimers of two-domain subunits or dimers of three-domain subunits with pseudo six-fold symmetry. Why do proteins that encircle DNA have six-fold symmetry, when the DNA they surround does not have a six-fold axis of symmetry? We offer a possible explanation based on the economy of the protein mass needed to encircle DNA. The first protein found to have a ring shape was the 3 subunit of Escherichia coli DNA polymerase III holoenzyme, which acts as a DNA 'sliding clamp' during chromosomal replication. The crystal structure of 13[1] showed that it forms a ring with a central cavity 35 A in diameter for encircling DNA (Fig. la). The 3-subunit ring tethers DNA polymerase III to DNA, ensuring that its enzymatic activity has the requisite high degree of processivity (in other words, ensuring that the enzyme moves along the DNA catalyzing multiple polymerization reactions without falling off its template) [1,2]. Although is only a dimer with a two-fold axis of
symmetry, it has a high degree of internal symmetry that gives it a six-fold appearance [1]. The symmetry derives from the way each 13 subunit is constructed from three globular domains, each of which has the same polypeptide chain fold, suggesting that 13 evolved via gene fusion events from a protein the size of one domain. Eukaryotes also use a circular sliding-clamp protein for chromosome replication - proliferating cell nuclear antigen (PCNA), the processivity factor of the DNA polymerase 8 holoenzyme. PCNA also forms a ring with pseudo six-fold symmetry [3], and the inner and outer diameters of the PCNA ring are similar to those of the 3 ring (Fig. lb). Furthermore, each PCNA subunit is made up of domains with chain topologies very similar to those of 3, suggesting that the proteins are evolutionarily related and their basic structure has been conserved since the divergence of prokaryotes and eukaryotes. The pseudo six-fold symmetry of PCNA has, however, a somewhat different basis from that of 3: rather than being a dimer like 3, PCNA is a trimer, each subunit of which has two globular domains. Another class of proteins with members that have been shown to encircle DNA is the helicases. The E. coli branch-migration protein RuvB has DNA helicase activity and catalyzes the movement of Holliday junctions during general genetic recombination. Electron micrographs show that RuvB encircles the DNA double helix, and three-dimensional image reconstruction analysis shows that it is organized into two hexameric rings (Fig. 2a) [4]. Gene protein 4 of bacteriophage T7 functions as a helicase and a primase during replication of the T7 genome. Gene 4 protein is expressed in two forms: one,
1. Crystal structures of sliding clamps of DNA polymerases. Ribbon representations of the polypeptide backbones of (a) a dimer of the E. coli DNA polymerase subunit [1], and (b) a trimer of yeast PCNA [3]. Strands of sheet are shown as flat ribbons, and helices are shown as spirals. The subunits within each ring are distinguished by different colors. Fig.
© Current Biology 1995, Vol 5 No 11
1239
1240
Current Biology 1995, Vol 5 No 11 Fig. 2. Three-dimensional reconstruction of two DNA helicases: (a) a RuvB dodecamer [4], and (b) a gp4B hexamer [6].
gp4A, has both helicase and primase activities; the other, gp4B, is truncated at the amino terminus relative to gp4A, and has only helicase activity. Both gp4A and gp4B appear to form hexamers that can encircle single-stranded DNA (Fig. 2b) [5,6]. The large-T antigen of simian virus 40 (SV40), a DNA tumor virus, binds the viral replication origin and unwinds the DNA locally. Electron microscopy showed [7] that large-T antigen binds to the SV40 origin as two units, each with a mass equivalent to six protein molecules, which were interpreted as being two DNA-encircling hexameric rings [8]. It seems reasonable to expect that other proteins involved in DNA metabolism will also turn out to form rings that surround their DNA substrate. Good candidates are gene 45 protein, the DNA sliding clamp of phage T4 replicase; the E. coli replicative helicase DnaB, which has been shown to be active as a hexamer; components of the transcription machinery in eukaryotes and prokaryotes; and subunits of the eukaryotic origin-recognition complex. The recently determined crystal structure [9] of another type of DNA-binding protein, Gal6, a bleomycin hydrolase, shows that it too has six-fold symmetry, but as yet there is no direct evidence that Gal6 binds DNA by encircling it. It would seem that most, and perhaps all, proteins that encircle DNA have at least pseudo six-fold symmetry. Why should this be so? We suggest that the explanation may lie in economy of protein mass - in particular, the way the symmetry and aggregation state of an oligomer determines the size its subunits have to be for it to encircle DNA. Assuming a spherical shape, the subunit size needed for an oligomer to surround DNA decreases as the aggregation state of the oligomer increases (see box). In the box, the mass values for spherical subunits that oligomerize to form rings with a central cavity 34 A in diameter (the size of the central cavity of PCNA) are calculated for oligomers with from three to eight subunits. The subunit size required for a trimer to encircle DNA is over 4 MDa, and therefore over 12 MDa of protein mass
would be needed to form a trimer in the shape of a ring. PCNA, of course, is a trimer with a total mass of only 86.7 kDa but, as explained above, its component subunit has two very similar domains, so for our purposes the protein may be more appropriately classified as a hexamer. A tetrameric ring would require subunits of 235 kDa to encircle DNA, giving an aggregate size of nearly 1 MDa. Subunits of a pentameric ring need only be 48.5 kDa to encircle DNA, and those of a hexamer ring only 16.7 kDa, close to the size of the domains in P and PCNA. 16.7 kDa is also close to the observed average size of single domains in protein structures [10], which may be another reason why the hexamer arrangement appears to be common. For yet-higher-order oligomers, even less protein mass would be needed to form rings with an inner diameter of 34 A. Perhaps subunits smaller than a typical domain would lack sufficient structure to produce a stable oligomer. It is important to note that the subunits of DNA helicase rings are larger than the size of a typical domain. Indeed, three-dimensional image reconstitution shows that both RuvB and T7 gp4B consist of at least two domains; presumably one domain contains the helicase activity. Protein complexes that form rings but do not encircle, or do not translocate along, DNA may exist in other oligomeric states. Most of such complexes have a central cavity that is larger or smaller than that needed to accommodate DNA comfortably. Examples of rings with oligomeric states greater than six include the proteasomes, which form rings with seven-fold symmetry (see [11], for example), and the chaperonins, which have been shown to form rings of seven, eight or nine subunits (reviewed in [12]). The diameters of the central cavity of these higher-order rings are 45 A or greater [13-15], in keeping with a general correlation between number of ring subunits and the inner diameter of the ring (see box). Other ring-shaped protein oligomers have oligomeric states of less than six. For example, several bacterial ADP-ribosylating toxins, including pertussis toxin and cholera toxin, are pentamers of subunits with molecular
DISPATCH
weights in the range of 8-22 kDa, which surround a central pore (reviewed in [16,17]). The central cavity formed by the pentamer of heat-labile enterotoxin [18] has a maximal diameter of 15 A, too small to accommodate double-stranded DNA (which in the B-form has a diameter of about 20 A). There are two known cases of proteins that probably surround DNA but do not translocate along the helix axis, and in both instances the proteins lack six-fold symmetry. One of these two proteins is topoisomerase I: the crystal structure of an amino-terminal fragment of topoisomerase I [19] shows that the polypeptide chain folds to produce a cavity with an average diameter of 27 A, which probably accommodates DNA. The other is the B subunit of DNA gyrase: the crystal structure of an amino-terminal fragment of this DNA gyrase subunit [20] shows that it forms a homodimer with a central cavity of 20 A diameter. As these enzymes are not thought to translocate along DNA, perhaps the higher-order
symmetry observed for sliding clamps and some helicases is required for motion along the DNA. One possible reason why higher-order rings are not used by proteins that encircle DNA - at least no examples are known at this time - is that the stability ol;the oligomer may decrease as the number of protomers increases, because of the greater probability of a subunits dissociating from the structure. This is expected to be especially true for a ring-shaped oligomer. In ringshaped oligomers, each subunit contacts only two adjacent neighbors, so upon loss of one subunit, the adjacent subunits will only contact one other subunit, leading to further dissociation and cooperative disassembly of the entire ring. Not only will dissociation of a subunit from the ring be more probable as the number of subunits in the oligomer increases, but the rate of ring reassembly from individual subunits will decrease with oligomer complexity, as more subunits must collide in a small time frame to assemble and lock the
1241
1242
Current Biology 1995, Vol 5 No 11 oligomer into a stable, closed ring structure (in which each subunit contacts two other subunits). 7.
Why did the six globular subunits of DNA sliding clamps fuse during evolution so that they assemble into rings of two or three subunits, instead of simply remaining as six individual subunits like some of the helicases? One major difference between helicases and the DNA sliding clamps is that the latter do not assemble onto DNA by themselves, but require a second protein factor, and the energy of ATP, to get on and off DNA [21,22]. In contrast, the helicases RuvB and gp4 appear capable of assembling onto DNA by themselves [4,6]. The need for assistance in getting on and off DNA implies that the interfaces of the sliding clamps are relatively tight. Why sliding clamps should require other proteins is an open question. Perhaps this requirement serves a regulatory role, as processive replication can not commence until the clamp is properly assembled onto DNA. Whatever the reason, one way to ensure that the sliding clamps do not get on and offDNA by themselves is to have a stable ring structure, and as discussed above, the more subunits needed to complete a ring, the more it will tend towards instability. Hence, the increased stability of a ring with fewer subunits may have driven the fusion events that appear to have occurred during the evolution of DNA polymerase sliding clamps.
8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Acknowledgments: We are grateful to John Kuriyan for assistance in
providing Figure 1, and for Edward Egelman and Steve West for
19.
assistance in preparation of Figure 2. 20.
References 1. Kong X-P, Onrust R, O'Donnell M, Kuriyan, J: Three dimensional structure of the P subunit of Escherichia coli DNA polymerase III holoenzyme: a sliding DNA clamp. Cell 1992, 69:425-437. 2. Stukenberg PT, Studwell-Vaughan PS, O'Donnell M: Mechanism of Biol the sliding clamp of DNA polymerase III holoenzyme. Chem 1991, 266:11328-11334. 3. Krishna TSR, Kong X-P, Gary S, Burgers PM, Kuriyan J: Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell 1994, 79:1233-1243. 4. Stasiak A, Tsaneva IR, West SC, Benson CJB, Yu X, Egelman EH: The Escherichia coli RuvB branch migration protein forms double hexamer rings around DNA. Proc Natl Acad Sci USA 1994, 91:7618-7622. 5. Patel SS, Hingorani MM: Oligomeric structure of bacteriophage Biol Chem 1993, T7 DNA primase/helicase proteins. 268:10668-10675. 6. Egelman EH, Yu X, Wild R, Hingorani MM, Patel SS: T7
21. 22.
helicase/primase proteins form rings around single-stranded DNA that suggest a general structure for hexameric helicases. Proc Natl Acad Sci USA 1995, 92:3869-3873. Mastrangelo IA, Hough PVC, Wall JS, Dodson M, Dean F, Hurwitz J: ATP-dependent assembly of double hexamer of SV40 T antigen at the viral origin of DNA replication. Nature 1989, 338:658-662. Dean F, Borowiec A, Eki T, Hurwitz J: The simian virus 40 T antigen double hexamer assembles around the DNA at the replication origin. I Biol Chem 1992, 267:14129-14137. Joshua-Tor L, Xu HE, Johnston SA, Rees DC: The crystal structure of Gal6/bleomycin hydrolase: a conserved protease that binds DNA. Science 1995, 269 945-950. Berman AL, Kolker E, Trifonov EN: Underlying order in protein sequence organization. Proc Natl Acad Sci USA 1994, 91:4044-4047. Lbwe J, Stock D, Jap B, Zwickl P, Baumeister, W, Huber R: Crystal structure of the 20S proteasome from the archaeon T. acidophilum at 3.4 A resolution. Science 1995, 268:533-539. Kim S, Willison KR, Horwich AL: Cystosolic chaperonin subunits have a conserved ATPase domain but diverged polypeptide-binding domains; Trends Biochem Sci 1994, 19:543-548. Braig K, Otwinowski Z, Hegde R, Boisvert DC, Joachimiak A, Horwich AL, Sigler PB: The crystal structure of the bacterial chaperonin GroEL at 2.8 A. Nature 1994, 371:578-686. Marco S, Carrascosa JL, Valpuesta JM: Reversible interaction of 1actin along the channel of the TCP-1 cytoplasmic chaperonin. Biophys 1994, 67:364-368. Knapp S, Schmidt-Krey I, Hebert H, Bergman T, Jrrivall H, Ladenstein R: The molecular chaperonin TF55 from the thermophilic archaeon Sulfolobussolfataricus. JMol Biol 1994, 242:397-407. Spangler BD: Structure and function of cholera toxin and related Escherichia coli heat-labile enterotoxin. Microbiol Rev 1992, 56:622-647. Burnette WN: AB5 ADP-ribosylating toxins: comparative anatomy and physiology. Structure 1994, 2:151-158. Sixma TK, Pronk SE, Kalk KH, Wartna ES, van Zanten BAM, Witholt B, Hol WGJ: Crystal structure of a cholera toxin-related heat-labile enterotoxin from E. coil. Nature 1991, 351:371-377. Lima CD, Wang JC, Mondragon A: Three-dimensional structure of the 67K N-terminal fragment of E. coli DNA topoisomerase I. Nature 1994, 367:138-145. Wigley BD, Davies GJ, Dodson EJ,Maxwell A, Dodson G: Crystal structure of an N-terminal fragment of the DNA gyrase B protein. Nature 1991, 351:624-629. Kelman Z, O'Donnell M: DNA replication - enzymology and mechanisms. Curr Opin Genet Dev 1994, 4:185-195. Stukenberg PT, Turner J, O'Donnell M: An explanation for lagging strand replication: polymerase hopping among DNA sliding clamps. Cell 1994, 78:877-887.
Zvi Kelman*, Jeff Finkelsteint and Mike O'Donnellt, Microbiology Department, Hearst Research Foundation and tHoward Hughes Medical Institute, Cornell University Medical College, 1300 York Avenue, New York, New York 10021, USA. *Present address: Department of Molecular Biology and Genetics, The John Hopkins Medical School, Baltimore, Maryland 21205, USA.
Z.
KELMAN,
J.
AND M. O'DONNELL
FINKELSTEIN AND
M.
O'DONNELL
PROTEIN STRUCTURE
PROTEIN STRUCTURE
Why have six-fold symmetry? Why do proteins that encircle DNA have six-fold symmetry? One important factor may be the economy in protein mass with which DNA can be encircled by six globular subunits arranged in a ring. Ring-shaped oligomeric proteins are involved in a diverse range of cellular processes. A subset of these protein oligomers encircle DNA, providing them with two seemingly opposed properties: the ability to grip DNA tightly, and the ability to move along the DNA. This group of proteins includes processivity factors of the replicative DNA polymerases, DNA helicases and a protein that resolves Holliday junction recombination intermediates. So far, all proteins examined that can encircle DNA and translocate along the double helix have been found to have either true or pseudo six-fold symmetry. This symmetry is achieved in different ways by proteins with different oligomeric states - for example, some are simple hexamers with true six-fold symmetry, whereas others are trimers of two-domain subunits or dimers of three-domain subunits with pseudo six-fold symmetry. Why do proteins that encircle DNA have six-fold symmetry, when the DNA they surround does not have a six-fold axis of symmetry? We offer a possible explanation based on the economy of the protein mass needed to encircle DNA. The first protein found to have a ring shape was the 3 subunit of Escherichia coli DNA polymerase III holoenzyme, which acts as a DNA 'sliding clamp' during chromosomal replication. The crystal structure of 13[1] showed that it forms a ring with a central cavity 35 A in diameter for encircling DNA (Fig. la). The 3-subunit ring tethers DNA polymerase III to DNA, ensuring that its enzymatic activity has the requisite high degree of processivity (in other words, ensuring that the enzyme moves along the DNA catalyzing multiple polymerization reactions without falling off its template) [1,2]. Although is only a dimer with a two-fold axis of
symmetry, it has a high degree of internal symmetry that gives it a six-fold appearance [1]. The symmetry derives from the way each 13 subunit is constructed from three globular domains, each of which has the same polypeptide chain fold, suggesting that 13 evolved via gene fusion events from a protein the size of one domain. Eukaryotes also use a circular sliding-clamp protein for chromosome replication - proliferating cell nuclear antigen (PCNA), the processivity factor of the DNA polymerase 8 holoenzyme. PCNA also forms a ring with pseudo six-fold symmetry [3], and the inner and outer diameters of the PCNA ring are similar to those of the 3 ring (Fig. lb). Furthermore, each PCNA subunit is made up of domains with chain topologies very similar to those of 3, suggesting that the proteins are evolutionarily related and their basic structure has been conserved since the divergence of prokaryotes and eukaryotes. The pseudo six-fold symmetry of PCNA has, however, a somewhat different basis from that of 3: rather than being a dimer like 3, PCNA is a trimer, each subunit of which has two globular domains. Another class of proteins with members that have been shown to encircle DNA is the helicases. The E. coli branch-migration protein RuvB has DNA helicase activity and catalyzes the movement of Holliday junctions during general genetic recombination. Electron micrographs show that RuvB encircles the DNA double helix, and three-dimensional image reconstruction analysis shows that it is organized into two hexameric rings (Fig. 2a) [4]. Gene protein 4 of bacteriophage T7 functions as a helicase and a primase during replication of the T7 genome. Gene 4 protein is expressed in two forms: one,
1. Crystal structures of sliding clamps of DNA polymerases. Ribbon representations of the polypeptide backbones of (a) a dimer of the E. coli DNA polymerase subunit [1], and (b) a trimer of yeast PCNA [3]. Strands of sheet are shown as flat ribbons, and helices are shown as spirals. The subunits within each ring are distinguished by different colors. Fig.
© Current Biology 1995, Vol 5 No 11
1239
1240
Current Biology 1995, Vol 5 No 11 Fig. 2. Three-dimensional reconstruction of two DNA helicases: (a) a RuvB dodecamer [4], and (b) a gp4B hexamer [6].
gp4A, has both helicase and primase activities; the other, gp4B, is truncated at the amino terminus relative to gp4A, and has only helicase activity. Both gp4A and gp4B appear to form hexamers that can encircle single-stranded DNA (Fig. 2b) [5,6]. The large-T antigen of simian virus 40 (SV40), a DNA tumor virus, binds the viral replication origin and unwinds the DNA locally. Electron microscopy showed [7] that large-T antigen binds to the SV40 origin as two units, each with a mass equivalent to six protein molecules, which were interpreted as being two DNA-encircling hexameric rings [8]. It seems reasonable to expect that other proteins involved in DNA metabolism will also turn out to form rings that surround their DNA substrate. Good candidates are gene 45 protein, the DNA sliding clamp of phage T4 replicase; the E. coli replicative helicase DnaB, which has been shown to be active as a hexamer; components of the transcription machinery in eukaryotes and prokaryotes; and subunits of the eukaryotic origin-recognition complex. The recently determined crystal structure [9] of another type of DNA-binding protein, Gal6, a bleomycin hydrolase, shows that it too has six-fold symmetry, but as yet there is no direct evidence that Gal6 binds DNA by encircling it. It would seem that most, and perhaps all, proteins that encircle DNA have at least pseudo six-fold symmetry. Why should this be so? We suggest that the explanation may lie in economy of protein mass - in particular, the way the symmetry and aggregation state of an oligomer determines the size its subunits have to be for it to encircle DNA. Assuming a spherical shape, the subunit size needed for an oligomer to surround DNA decreases as the aggregation state of the oligomer increases (see box). In the box, the mass values for spherical subunits that oligomerize to form rings with a central cavity 34 A in diameter (the size of the central cavity of PCNA) are calculated for oligomers with from three to eight subunits. The subunit size required for a trimer to encircle DNA is over 4 MDa, and therefore over 12 MDa of protein mass
would be needed to form a trimer in the shape of a ring. PCNA, of course, is a trimer with a total mass of only 86.7 kDa but, as explained above, its component subunit has two very similar domains, so for our purposes the protein may be more appropriately classified as a hexamer. A tetrameric ring would require subunits of 235 kDa to encircle DNA, giving an aggregate size of nearly 1 MDa. Subunits of a pentameric ring need only be 48.5 kDa to encircle DNA, and those of a hexamer ring only 16.7 kDa, close to the size of the domains in P and PCNA. 16.7 kDa is also close to the observed average size of single domains in protein structures [10], which may be another reason why the hexamer arrangement appears to be common. For yet-higher-order oligomers, even less protein mass would be needed to form rings with an inner diameter of 34 A. Perhaps subunits smaller than a typical domain would lack sufficient structure to produce a stable oligomer. It is important to note that the subunits of DNA helicase rings are larger than the size of a typical domain. Indeed, three-dimensional image reconstitution shows that both RuvB and T7 gp4B consist of at least two domains; presumably one domain contains the helicase activity. Protein complexes that form rings but do not encircle, or do not translocate along, DNA may exist in other oligomeric states. Most of such complexes have a central cavity that is larger or smaller than that needed to accommodate DNA comfortably. Examples of rings with oligomeric states greater than six include the proteasomes, which form rings with seven-fold symmetry (see [11], for example), and the chaperonins, which have been shown to form rings of seven, eight or nine subunits (reviewed in [12]). The diameters of the central cavity of these higher-order rings are 45 A or greater [13-15], in keeping with a general correlation between number of ring subunits and the inner diameter of the ring (see box). Other ring-shaped protein oligomers have oligomeric states of less than six. For example, several bacterial ADP-ribosylating toxins, including pertussis toxin and cholera toxin, are pentamers of subunits with molecular
DISPATCH
weights in the range of 8-22 kDa, which surround a central pore (reviewed in [16,17]). The central cavity formed by the pentamer of heat-labile enterotoxin [18] has a maximal diameter of 15 A, too small to accommodate double-stranded DNA (which in the B-form has a diameter of about 20 A). There are two known cases of proteins that probably surround DNA but do not translocate along the helix axis, and in both instances the proteins lack six-fold symmetry. One of these two proteins is topoisomerase I: the crystal structure of an amino-terminal fragment of topoisomerase I [19] shows that the polypeptide chain folds to produce a cavity with an average diameter of 27 A, which probably accommodates DNA. The other is the B subunit of DNA gyrase: the crystal structure of an amino-terminal fragment of this DNA gyrase subunit [20] shows that it forms a homodimer with a central cavity of 20 A diameter. As these enzymes are not thought to translocate along DNA, perhaps the higher-order
symmetry observed for sliding clamps and some helicases is required for motion along the DNA. One possible reason why higher-order rings are not used by proteins that encircle DNA - at least no examples are known at this time - is that the stability ol;the oligomer may decrease as the number of protomers increases, because of the greater probability of a subunits dissociating from the structure. This is expected to be especially true for a ring-shaped oligomer. In ringshaped oligomers, each subunit contacts only two adjacent neighbors, so upon loss of one subunit, the adjacent subunits will only contact one other subunit, leading to further dissociation and cooperative disassembly of the entire ring. Not only will dissociation of a subunit from the ring be more probable as the number of subunits in the oligomer increases, but the rate of ring reassembly from individual subunits will decrease with oligomer complexity, as more subunits must collide in a small time frame to assemble and lock the
1241
1242
Current Biology 1995, Vol 5 No 11 oligomer into a stable, closed ring structure (in which each subunit contacts two other subunits). 7.
Why did the six globular subunits of DNA sliding clamps fuse during evolution so that they assemble into rings of two or three subunits, instead of simply remaining as six individual subunits like some of the helicases? One major difference between helicases and the DNA sliding clamps is that the latter do not assemble onto DNA by themselves, but require a second protein factor, and the energy of ATP, to get on and off DNA [21,22]. In contrast, the helicases RuvB and gp4 appear capable of assembling onto DNA by themselves [4,6]. The need for assistance in getting on and off DNA implies that the interfaces of the sliding clamps are relatively tight. Why sliding clamps should require other proteins is an open question. Perhaps this requirement serves a regulatory role, as processive replication can not commence until the clamp is properly assembled onto DNA. Whatever the reason, one way to ensure that the sliding clamps do not get on and offDNA by themselves is to have a stable ring structure, and as discussed above, the more subunits needed to complete a ring, the more it will tend towards instability. Hence, the increased stability of a ring with fewer subunits may have driven the fusion events that appear to have occurred during the evolution of DNA polymerase sliding clamps.
8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Acknowledgments: We are grateful to John Kuriyan for assistance in
providing Figure 1, and for Edward Egelman and Steve West for
19.
assistance in preparation of Figure 2. 20.
References 1. Kong X-P, Onrust R, O'Donnell M, Kuriyan, J: Three dimensional structure of the P subunit of Escherichia coli DNA polymerase III holoenzyme: a sliding DNA clamp. Cell 1992, 69:425-437. 2. Stukenberg PT, Studwell-Vaughan PS, O'Donnell M: Mechanism of Biol the sliding clamp of DNA polymerase III holoenzyme. Chem 1991, 266:11328-11334. 3. Krishna TSR, Kong X-P, Gary S, Burgers PM, Kuriyan J: Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell 1994, 79:1233-1243. 4. Stasiak A, Tsaneva IR, West SC, Benson CJB, Yu X, Egelman EH: The Escherichia coli RuvB branch migration protein forms double hexamer rings around DNA. Proc Natl Acad Sci USA 1994, 91:7618-7622. 5. Patel SS, Hingorani MM: Oligomeric structure of bacteriophage Biol Chem 1993, T7 DNA primase/helicase proteins. 268:10668-10675. 6. Egelman EH, Yu X, Wild R, Hingorani MM, Patel SS: T7
21. 22.
helicase/primase proteins form rings around single-stranded DNA that suggest a general structure for hexameric helicases. Proc Natl Acad Sci USA 1995, 92:3869-3873. Mastrangelo IA, Hough PVC, Wall JS, Dodson M, Dean F, Hurwitz J: ATP-dependent assembly of double hexamer of SV40 T antigen at the viral origin of DNA replication. Nature 1989, 338:658-662. Dean F, Borowiec A, Eki T, Hurwitz J: The simian virus 40 T antigen double hexamer assembles around the DNA at the replication origin. I Biol Chem 1992, 267:14129-14137. Joshua-Tor L, Xu HE, Johnston SA, Rees DC: The crystal structure of Gal6/bleomycin hydrolase: a conserved protease that binds DNA. Science 1995, 269 945-950. Berman AL, Kolker E, Trifonov EN: Underlying order in protein sequence organization. Proc Natl Acad Sci USA 1994, 91:4044-4047. Lbwe J, Stock D, Jap B, Zwickl P, Baumeister, W, Huber R: Crystal structure of the 20S proteasome from the archaeon T. acidophilum at 3.4 A resolution. Science 1995, 268:533-539. Kim S, Willison KR, Horwich AL: Cystosolic chaperonin subunits have a conserved ATPase domain but diverged polypeptide-binding domains; Trends Biochem Sci 1994, 19:543-548. Braig K, Otwinowski Z, Hegde R, Boisvert DC, Joachimiak A, Horwich AL, Sigler PB: The crystal structure of the bacterial chaperonin GroEL at 2.8 A. Nature 1994, 371:578-686. Marco S, Carrascosa JL, Valpuesta JM: Reversible interaction of 1actin along the channel of the TCP-1 cytoplasmic chaperonin. Biophys 1994, 67:364-368. Knapp S, Schmidt-Krey I, Hebert H, Bergman T, Jrrivall H, Ladenstein R: The molecular chaperonin TF55 from the thermophilic archaeon Sulfolobussolfataricus. JMol Biol 1994, 242:397-407. Spangler BD: Structure and function of cholera toxin and related Escherichia coli heat-labile enterotoxin. Microbiol Rev 1992, 56:622-647. Burnette WN: AB5 ADP-ribosylating toxins: comparative anatomy and physiology. Structure 1994, 2:151-158. Sixma TK, Pronk SE, Kalk KH, Wartna ES, van Zanten BAM, Witholt B, Hol WGJ: Crystal structure of a cholera toxin-related heat-labile enterotoxin from E. coil. Nature 1991, 351:371-377. Lima CD, Wang JC, Mondragon A: Three-dimensional structure of the 67K N-terminal fragment of E. coli DNA topoisomerase I. Nature 1994, 367:138-145. Wigley BD, Davies GJ, Dodson EJ,Maxwell A, Dodson G: Crystal structure of an N-terminal fragment of the DNA gyrase B protein. Nature 1991, 351:624-629. Kelman Z, O'Donnell M: DNA replication - enzymology and mechanisms. Curr Opin Genet Dev 1994, 4:185-195. Stukenberg PT, Turner J, O'Donnell M: An explanation for lagging strand replication: polymerase hopping among DNA sliding clamps. Cell 1994, 78:877-887.
Zvi Kelman*, Jeff Finkelsteint and Mike O'Donnellt, Microbiology Department, Hearst Research Foundation and tHoward Hughes Medical Institute, Cornell University Medical College, 1300 York Avenue, New York, New York 10021, USA. *Present address: Department of Molecular Biology and Genetics, The John Hopkins Medical School, Baltimore, Maryland 21205, USA.