Nature Structural Biology
8, 290 - 293 (2001)
doi:10.1038/86149
Making the most of metal ionsNancy C. Horton
& John J. PeronaNancy C. Horton and John J. Perona are in the Department of Chemistry & Biochemistry, University of California at Santa Barbara, Santa Barbara, California 93106-9510, USA.
Correspondence should be addressed to Nancy C. Horton horton@chem.ucsb.edu or John J. Perona perona@chem.ucsb.eduCrystal structures of the homing endonuclease I-CreI bound to substrate DNA and divalent metals show that one metal ion is shared between the two active sites of the enzyme. This arrangement appears uniquely suited to the formation of double-stranded DNA breaks via a concerted reaction.Many group I introns possess open reading frames which encode DNA endonucleases that promote the mobility of the intron in a process known as homing1. The introns exhibit widespread phylogenetic diversity, and have now been found in all three biological domains. In the homing mechanism, a donor allele possessing the intron (I+) is paired with a recipient cognate allele that lacks the element (I-). The intron-encoded homing endonuclease cleaves both strands of the I- allele at the splice site, initiating a duplication of the intron followed by resolution via a double-stranded break repair pathway (Fig. 1). Thus, the homing endonuclease gene functions in a selfish manner, ensuring propagation by virtue of the ability of its protein product to initiate recombination events.
 | | |
The homing endonucleases must be highly specific for their target sites in order to avoid generating multiple double-stranded breaks within the host genome. On the basis of conserved sequence motifs, group I homing enzymes fall into four structural families. The largest of these, the LAGLIDADG family, is the best studied and is now represented by crystal structures of the I-CreI, PI-SceI, I-DmoI and PI-PfuI enzymes2,
3,
4,
5. However, the catalytic mechanisms of these homing endonucleases have remained obscure. A significant step in understanding the basis for rate enhancement is now reported on page page 312 of this issue of Nature Structural Biology. Based on the high-resolution structures of two ternary enzyme−DNA− metal ion complexes, Chevalier et al.6 propose a novel DNA cleavage mechanism for I-CreI in which two catalytic active sites share a total of three metal ions.
Metal ions in biological reactions
The importance of divalent metal cations in promoting biologically important phosphoryl transfer reactions has been appreciated for some time. Recognition of the central role of ATP hydrolysis in biochemistry led to early emphasis on understanding the mechanisms for accelerating hydrolysis of phosphomonoesters7,
8,
9. Model reactions catalyzed by small organic compounds, together with enzymatic studies and theoretical approaches, have shown that metal ions can play important roles in NTP cleavage reactions10,
11,
12,
13,
14,
15. These reactions are necessary for energy conversion, regulation of protein activity and cell signaling processes. By contrast, enzymes that hydrolyze phosphodiesters instead are central to replication, transcription, recombination, and DNA repair through their manipulation of DNA and RNA.
Several important model systems have established and quantified the contributions of divalent metal ions in phosphodiester hydrolysis16,
17. For these reactions, the mechanism involves in-line attack of the nucleophile on the tetrahedral ground state phosphorus, followed by passage through a trigonal bipyramidal pentavalent transition state and departure of the leaving oxygen anion18. Three ways in which metal ions can directly assist this reaction are: (i) Lewis acid activation by coordination to nonbridging phosphate oxygens (those oxygens bound only to phosphorus); (ii) nucleophile activation by direct coordination, thereby reducing the pKa and inducing deprotonation; and (iii) leaving group activation by direct coordination to the oxygen anion. Rate acceleration can also be provided via mechanisms that do not require direct metal ion contact with substrate. These include: (i) general base catalysis, by which a metal ion-bound hydroxide accepts a proton; (ii) general acid catalysis, in which a metal ion-bound water molecule donates a proton; and (iii) electrostatic stabilization by location of the positively charged metal ion in the vicinity of the negatively charged transition state. In general, it is believed that the direct catalytic functions are capable of facilitating much better rate enhancement than are the indirect16,
19. The three direct roles can increase the cleavage rate by two, eight and six orders of magnitude, respectively, thus giving the necessary 1016-fold rate enhancement needed to hydrolyze DNA on the biological time scale16. Ligation of phosphate oxygens to the metal ions may also induce strain in the O-P-O angle for preferential stabilization of the transition state. This provides another mechanism by which direct metal coordination may accelerate the difficult phosphodiester cleavage chemistry.
The two-metal mechanism
X-ray crystallographic investigations of phosphodiester manipulating enzymes have provided the structural details of stalled ground state and product complexes, including the details of substrate and product orientation, contacts with active site residues, and the precise locations of divalent cations and water molecules20. The first well-formulated metal-dependent phosphoryl transfer mechanism for phosphodiester hydrolysis was developed from crystallographic studies of the 3'-5' exonuclease activity of E. coli DNA polymerase I21. A similar two-metal mechanism was proposed for the phosphomonoesterase alkaline phosphatase22, and has since been invoked for a number of other enzymes including ribozymes23, restriction endonucleases (refs 24,
25,
26,
27,
28; Horvath M. et al., PDB entry 1QPS), and the synthetic active site of polymerases29. In the two-metal mechanism, the metal ions are positioned parallel to the apical axis of the trigonal bipyramidal transition state, and interact with both the attacking and leaving groups found at the two apical positions (Fig. 2). The metal ion at site A orients and activates the nucleophile by lowering the pKa of a nucleophilic water, thus inducing deprotonation. Metal ion B directly ligates the leaving group atom, facilitating the bond breakage step by neutralizing the developing negative charge on this oxygen anion. Both metals also ligate a nonbridging oxygen of the phosphate to stabilize the additional incipient negative charge in the transition state, and help to position the substrate phosphate group properly in the active site. In this mechanism the metal ions are proposed to be fully capable of providing all of the required rate enhancement.
| | Figure 2. Variations of the two-metal ion mechanism. | | |  | The essential elements of the phosphoryl-transfer mechanism are depicted at the top. Variations among enzymes are noted at bottom, for the restriction endonucleases (REs): EcoRV24, EcoRI (Horvath et al., PDB entry 1QPS), PvuII25, BamHI26, BglI27 and BglII28; the homing endonucleases (HEs): I-CreI6 and I-PpoI30; Serratia nuclease31 (SN), the synthetic site of polymerases29 (P), the 3'-5' exonuclease of DNA polymerase I31 (E), alkaline phosphatase31 (AP), and purple acid phosphatase36 (PAP). The models differ with respect to the presence/absence and the identities of surrounding groups (R1, R2, R3, R4) and metal ions: 1. present: REs, I-CreI, AP, P, E, PAP, absent: no metal ion A in I-PpoI, SN; 2. present: PvuII, BglI, BamHI, BglII, EcoRI, I-CreI, AP, P, E, absent: EcoRV, PAP, no metal ion A in I-PpoI, SN; 3. present: EcoRV, PvuII, BglI, BamHI, I-CreI, I-PpoI, SN, AP, PAP, P, E, absent: no metal ion B in BglII, EcoRI; 4. present: BglI, BamHI, I-CreI, I-PpoI, SN, AP, P, E, absent: EcoRV, PvuII, PAP, no metal ion B in BglII, EcoRI, 5. present: EcoRV, PvuII, BglI, BamHI, I-PpoI, SN, absent: I-CreI, P, E, AP, PAP, no metal ion B in BglII, EcoRI. Groups B and R2 serve as general base and for orientation of the nucleophile, respectively.
 Full Figure and legend (42K) |
| |
The classical framework of the two-metal mechanism as described above can also accommodate a number of variations. These include positioning a potential proton acceptor (a general base catalyst) in contact with the attacking metal ion-ligated nucleophile24,
26, metal ligation to either of the two nonbridging oxygens, and stabilization of the leaving group by protonation from a metal ion-bound water instead of by direct ligation. The observed variations raise some interesting and unresolved mechanistic questions. For example, although a general base would appear to be required to accept the proton from water, there is sometimes no clear candidate present in the active site to fulfill this role. I-CreI provides an illustration of this: the metal-bound attacking water is hydrogen-bonded only to other waters rather than directly to an appropriate enzyme side chain (one with a pKa near the physiological range)6. Similarly, while type II restriction endonucleases often possess well-positioned Lys or Glu side chains as candidates to accept a proton24,
26, the crystal structure of BglII shows that a Gln residue occupies the equivalent position instead28. So far, there are no definitive mechanistic studies establishing the identity of the general base catalyst for any of the restriction or homing endonuclease enzymes that employ two-metal catalysis. However, a one-metal mechanism appears to function in the His-Cys box homing endonuclease I-PpoI30 (as well as in Serratia endonuclease31), where the equivalent of metal A is absent. A histidine imidazole occupies this position instead, and is an excellent candidate for the general base.
Sharing a metal between two sites
The newest variation on the two-metal ion mechanism is reported by Chevalier et al.6 in this issue. They determined crystal structures of I-CreI bound to DNA and divalent metal ions in a new better-ordered crystal form, at 1.8 Å and 2.05 Å resolution for a Mg2+ product complex and a Ca2+ substrate complex, respectively. (As is the case for type II restriction enzymes, Ca2+ ions inhibit I-CreI function allowing trapping of uncleaved phosphodiester linkages in both strands of the DNA.) This accomplishment represents the first high-resolution view of metal ions bound in the active site of a LAGLIDADG class homing endonuclease. The striking finding is that each scissile phosphate appears to be cleaved by a conventional two-metal mechanism, but the metal ion occupying site B bridges both active sites and interacts simultaneously with the 3'-leaving oxygen of both DNA strands. This metal ion is located at the subunit interface of the homodimeric enzyme, and is ligated directly to a conserved critical aspartate residue (Asp 20) from the LAGLIDADG motif of both monomers. The bridging between subunits is possible because the scissile phosphates across the minor groove are separated by only 8 Å, which is a shorter distance compared to straight B-DNA because in this case the duplex is curved around the enzyme surface. The metal ions occupying site A are bound entirely by residues from a single subunit, and also ligate the conserved Asp 20 carboxylate group6.
The finding of a shared metal site is consistent with mutational data obtained for the homologous monomeric PI-SceI enzyme32, which carries two copies of the LAGLIDADG peptide that can be superimposed on the two motifs present within the separate subunits in the I-CreI dimer3,
4. This work predicted the existence of a single catalytic center which cleaves both DNA strands, and the finding of a shared metal ion essential to both subunits indeed reveals that the dimeric I-CreI may be considered to possess one active site capable of hydrolyzing both phosphodiester linkages. Alternatively, the structural organization may also be described as consisting of two overlapping active sites.
As suggested by Chevalier et al.6, the sharing of an essential metal ion between both active sites of I-CreI implies that the cleavage of the two DNA strands may occur nearly simultaneously. A consequence of this would be to ensure the production of double-stranded breaks in the DNA, while minimizing the potential for generating single-stranded nicks. In turn this would be expected to increase the efficiency of the homing process, since DNA ligase can repair single-stranded nicks arising from enzyme dissociation after cleavage of only one strand.
Why might a homing endonuclease require a concerted hydrolysis mechanism, rather than two sequential cleavages occurring within the lifetime of the enzyme−DNA complex? In type II restriction endonucleases, single-turnover kinetics towards plasmid substrates has revealed the existence of open-circle (nicked) DNA as a transient intermediate, showing that sequential cutting of the two strands occurs at cognate sites33. However, restriction enzymes are known to dissociate from the DNA after cleavage of only one strand at noncognate sites differing by one base pair from the specific sequence, generating a nick that can be repaired by ligase34. Thus, two well-separated active sites confer a biological advantage to the restriction enzyme, because the sequential cleavage of the two strands can permit repair of potentially lethal damage at incorrect loci. In contrast, the homing endonucleases require flexible recognition of long target sites in order to maximize their ability to cleave closely related variants. This is essential for recognition of the homing site in cognate alleles with slightly differing sequences. Without a structural organization that directly couples cleavage of the two strands to each other, the homing enzymes might be susceptible to dissociation after nicking at certain diverged targets. Indeed examination of PI-SceI activity toward pre-nicked substrates showed significant sequence-dependent cleavage rates35. Because the active sites of this monomeric enzyme are not identical, it will be important to also perform such experiments using a homodimeric endonuclease such as I-CreI.
It appears that restriction enzymes and LAGLIDADG homing endonucleases have each evolved structural mechanisms for phosphodiester hydrolysis consistent with their biological roles. Whether or not some other classes of homing endonucleases possess similar mechanisms remains as an engaging question for future research.
REFERENCES
- Jurica, M.S. & Stoddard, B.L. Cellular and Molecular Life Sciences 55, 1304−1326 (1999). | Article | PubMed | ISI | ChemPort |
- Heath, P.J., Stephens, K.M., Monnat, R.J. & Stoddard, B. Nature Struct. Biol. 4, 468−476 (1997). | Article | PubMed | ISI | ChemPort |
- Duan, X., Gimble, F.S. & Quiocho, F.A. Cell 89, 555−564 (1997). | Article | PubMed | ISI | ChemPort |
- Silva, G.H., Dalgaard, J.Z., Belfort, M. & van Roey, P. J. Mol. Biol. 286, 1123−1126 (1999). | Article | PubMed | ISI | ChemPort |
- Ichiyanagi, K., Ishino, Y. Ariyoshi, M., Komori, K. & Morikawa, K. J. Mol. Biol. 300, 889−901 (2000). | Article | PubMed | ISI | ChemPort |
- Chavelier, B.S., Monnat, R.J.Jr., & Stoddard, B.L. Nature Struct. Biol. 8, 312−316 (2001). | Article | PubMed | ISI | ChemPort |
- Meyerhof, O. & Lohman, K. Biochem. Z. 253, 431−461 (1932). | ChemPort |
- Sigel, H. Coordination Chemistry Reviews 100, 453−539 (1990). | Article | ISI | ChemPort |
- Mildvan, A.S. Advances in Enzymology and Related Areas of Molecular Biology 49, 103−126 (1979). | ChemPort |
- Butcher, W.W. & Westheimer, F.H. J. Am. Chem. Soc. 77, 2420−2424 (1955). | Article | ISI | ChemPort |
- Bunton, C.A., Llewellyn, D.R., Oldham, K.G. & Vernon, C.A. J. Chem. Soc., 3574−3587 (1958). | Article | ChemPort |
- Kirby, A.J. & Varvoglis, A.G. J. Am. Chem. Soc. 89, 415−423 (1967). | Article | ISI | ChemPort |
- Herschlag, D. & Jencks, W.P. Biochemistry 29, 5172−5179 (1990). | Article | PubMed | ISI | ChemPort |
- Aqvist, J., Kolmodin, K., Florian, J. & Warshel, A. Chemistry & Biology 6, R71−R80 (1999). | Article | PubMed | ChemPort |
- Maegley, K.A., Admiraal, S.J. & Herschlag, D. Proc. Natl. Acad. Sci. USA 93, 8160−8166 (1996). | Article | PubMed | ChemPort |
- Williams, N.H., Takasaki, B., Wall, M. & Chin, J. Accts. Chem. Res. 32, 485−493 (1999). | Article | ChemPort |
- Blasko, A. & Bruice, T.C. Accts. Chem. Res. 32, 475−484 (1999). | Article | ChemPort |
- Benkovic, S.J. & Schray, K.J. In The Enzymes . (ed. Boyer, P. D.) 201−238 (Academic Press, New York; 1973). | ChemPort |
- Browne, K.A. & Bruice, T.C. J. Am. Chem. Soc. 114, 4951−4958 (1992). | Article | ISI | ChemPort |
- Strater, N., Lipscomb, W. N., Klabunde, T. & Krebs, B. Angewandte Chemie 35, 2024−2055 (1996). | ISI |
- Beese, L.S. & Steitz, T.A. EMBO J. 10, 25−33 (1991). | PubMed | ISI | ChemPort |
- Kim, E.E. & Wyckoff, H.W. J. Mol. Biol. 218, 449−464 (1991). | Article | PubMed | ISI | ChemPort |
- Steitz, T.A. & Steitz, J.A. Proc. Natl. Acad. Sci. USA 90, 6498−6502 (1993). | PubMed | ChemPort |
- Horton, N.C., Newberry, K.J. & Perona, J.J. Proc. Natl. Acad. Sci. USA 95, 13489−13494 (1998). | Article | PubMed | ChemPort |
- Horton, J.R. & Cheng, X.D. J. Mol. Biol. 300, 1049−1056 (2000). | Article | PubMed | ISI | ChemPort |
- Viadiu, H. & Aggarwal, A.K. Nature Struct. Biol. 5, 910−916 (1998). | Article | PubMed | ISI | ChemPort |
- Newman, M. et al. EMBO J. 17, 5466−5476 (1998). | Article | PubMed | ChemPort |
- Lukacs, C., Kucera, R., Schidkraut, I. & Aggarwal, A.K. Nature Struct. Biol. 7, 134−140 (2000). | Article | PubMed | ISI | ChemPort |
- Steitz, T.A. Nature 391, 231−232 (1998). | Article | PubMed | ISI | ChemPort |
- Galburt, E.A. et al. Nature Struct. Biol. 6, 1096−1099 (1999). | Article | PubMed | ISI | ChemPort |
- Miller, M.D., Cai, J. W. & Krause, K. L. J. Mol. Biol. 288, 975−987 (1999). | Article | PubMed | ISI | ChemPort |
- Gimble, F.S. & Stephens, B.W. J. Biol. Chem. 270, 5849−5855 (1995). | Article | PubMed | ISI | ChemPort |
- Sasnauskas, G., Jeltsch, A., Pingoud, A. & Siksnys, V. Biochemistry 38, 4028−4036 (1999). | Article | PubMed | ISI | ChemPort |
- Taylor, J.D. & Halford, S.E. Biochemistry 28, 6198−6207 (1989). | Article | PubMed | ISI | ChemPort |
- Christ, F. et al. EMBO J. 18, 6908−6916 (1999). | Article | PubMed | ChemPort |
- Klabunde, T., Strater, N., Frohlich, R., Witzel, H. & Krebs, B. J. Mol. Biol. 259, 737−748 (1996). | Article | PubMed | ISI | ChemPort |
Acknowledgments
We are grateful to F. Gimble for valuable discussion and comments on the manuscript. Work in our laboratory in this area of research is funded by the National Institutes of Health and by the National Science Foundation POWRE program (N.C.H.).
|
