impact of DNA polymerase interactions with the 2'-deoxy­ ribose moiety and the participation of these interactions in processes which contribute to fidelity. Crystal structures of DNA polymerases together with enzyme mutation studies suggest that the sugar moiety of the incoming triphosphate is fully embedded in the nucleotide binding pocket and under­ goes essential interactions with the enzyme)la.2] Here we report a functional strategy to monitor steric constraints in DNA polymerases that act on the sugar moiety of an incoming nucleoside triphosphate within the nucleotide binding pocket. We found that novel modified nucleotide probes are sub­ strates for a DNA polymerase with significantly increased selectivity compared to their natural counterpart. Through use of these high-fidelity nucleotides in functional investiga­ tions we could show that enzyme-sugar interactions are involved in DNA-polymerase fidelity mechanisms. In order to sense interactions of DNA polymerases with the sugar moiety of incoming triphosphates we introduced alkyl labels at the 4'-position in the 2'-deoxyribose moiety in such a way that they do not interfere with hydrogen bonding, nucleobase pairing, and stacking. We designed steric probes la-d by substituting the hydrogen atom at the 4'-position of thymidine triphosphate (TIP) with alkyl groups of different DNA Polymerase Selectivity: Sugar size (Scheme 1). Interactions Monitored with High-Fidelity NucIeotides** Daniel Summerer and Andreas Marx* T T PPP:~ PPP:~ 1a:R=CH3 The essential prerequisite of any organism is to keep its 1b: R = CH2CH3 OH OH genome intact and to accurately duplicate it before cell 1 c: R = CH(CH3)2 TIP 1a-d division. All DNA synthesis required for DNA repair, 1 d: R = CH2CH(CH3)2 000 11 It 11 recombination, and replication depends on the ability of PPP: -P-O-P-O-p-O-. . . T: thymine DNA polymerases to recognize the template and correctly 0- 0- 0- insert the complementary nucleotide. In a current model, Scheme I. 1l1ymidine-5' -triphosphate (TIP) and the steric probes 1 a-d. fidelity is achieved by the ability of DNA polymerases to edit nucleobase pair shape and size.PI This model is supported by crystal structures of DNA polymerases that suggest the formation of nucleotide binding pockets, which exclusively The synthesis of nucleosides 4a and 4b with different accommodate Watson - Crick base pairs.(2] Nevertheless, synthetic strategies (see Scheme 2) has been reported pre­ structural data of DNA polymerases complexed with the viouslyP' 4) However, these methods are not suitable for the DNA substrates and a noncanonical triphosphate, which synthesis of all the targeted compounds and the formation of would be very helpful for gaining insights into the causes of undesired by-products was observed in the synthesis of 4a./31 error-prone DNA synthesis, have at present not been Thus, we developed a more suitable access to the target reported. Thus, functional studies of DNA polymerases with compounds. Our synthesis started with the known alcohol 2, nucleotide analogues have been shown to be extremely which is easily accessible as described recently.fS) Compound 2 valuable)l) Since most functional studies have focused on was converted into the methylated thymidine 4a by func­ nucleobase recognition processes,P] little is known about the tional-group interconversions (Scheme 2). Alkylated nucleo­ sides 4b-d were synthesized from known compounds(5) by employing Wittig reaction, desilylation, and subsequent reduction of the aliphatic double bond. Next, nucleosides [*J Dr. A. Marx, Dipl.-Chem. D. Summerer Kckulc-Tnstitut flir Organischc Chcmic und Biochcmic 4a-d were converted into the desired triphosphates 1 a-d UnivcrsitiH Bonn following standard procedures)61 In order to gain insights into Gcrhard-Domagk-Strassc 1,53121 Bonn (Germany) potential effects of the modifications on the sugar conforma­ Fax: (+ 49) 228-73-5388 tion we performed conformation analysis based on coupling E-mail: [email protected] constants deduced from the lH NMR data by employing [**J 1l,is work was supported hy a grant from the Deutsche Forschungs­ described methodsP) We found only small differences in la­ gemeinschaft. We thank Professor Dr. Michael Famulok for his continuing support. d compared to natural TIP, which indicates that similar sugar conformations are present in solution (see Supporting In­ formation ).

3693 T T T thymidine analogue adjacent at the primer 3'-end,[9] TBDPS~ a) TBDP~~ b), c). HO~ to monitor polymerase function. Figure 1 A shows HO .. H3C the pattern of insertion of thymidine-5' -triphosphate OTBS OTBS OH (TIP) and ta-d catalyzed by KF-. The results 2 3 4a obtained reveal that 1a-d are functionally compe- tent, although with varied efficiencies. T T In order to quantify these observations we deter- d'J TBDP~ f), g) .. mined the insertion efficiencies (vma) KM; vmax = the KM OTBS ~OH maximum rate of the enzyme reaction, = the T ~ TBDPS~ 6 4b Michaelis constant) of the thymidine analogues by 0- investigation of single-nucleotide insertion at vari- OTBS ous nucleotide concentrations under single complet- ~ T T 5 TBD~ ed hit and steady-state conditions as recently de- i),j) .. scribed.l8,10-12] The quantitative analyses revealed OTBS ~OH that KF- incorporates 1 a and 1 b with virtually the k'J 7 4d same efficiency as unmodified TTP (Table 1). Thus,

T T T TBDP~~ TBDP~ Table 1. Steady-state analyses for canonical nucleoside triphos- I) • m),n). phate insertion, The data presented arc averages of duplicate or OTBS OTBS ~OH triplicate experiments, Further experimental details are de- scribed in Supporting Information. 8 9 4c Nucleoside KM vmax vmaxlKM triphosphate [ILM] [min-' x 10-'] [M-'min-'] 0 0 0 T 11 11 11 T TIP 0.11 ±0.05 20±3 180000 -O-P-O-P-O-P-O~ I I I 0 la OAO± 0.01 58±4 150000 0) _0 _0 _0 H:~ .. R 1b 0.21 ±0.01 29±5 140000 OH OH le 47±4 85±5 1800 4a-d 1a-d Id 241 ±22 76±1 320

g Scheme 2. Syntheses of la-d. a) Ph,P. I" imidazole, C6H6 , 50 C, 85 %; b) Pd/C, H2 , EtOH, EtOAc. NEt,; c) TBAF, THF, 79% (two steps); d) oxidation;!5] e) CH3PPh3Br, nBuLi, nH'~ -78 to 25°C, 99%; f) TBAF, THF; g) PdlC, H" CH,oH, 88% (two steps); h)(CH')2CHPPh,l, nBuLi, THF, -78 to 25°C, 83%; i)TBAF, THF; DPd/C, H 2, CH,OH, 94% (two steps); k) alkylation and oxidation;!5] I) CH3PPh3Br, tBuOK, THF, it appears that 4'-methyl and -ethyl groups are nicely 25°C, 91 %; m) TBAF, THF; n) PdlC, H2, CH,OH. 84% (two steps); 0) POCI,. 1,8- accommodated within the nucleotide binding pocket bis(dimethylamino)naphthalene. (CH,O),PO, O°C, then (nBu3NH),H2P,O" nBu,N, of KF-. Further increase in the bulk of the probes DMF, then O,lM aqueous (Et,NH)HC03, 23-68%, TBDPS= tert-butyldiphenylsilyl, caused a marked decrease in the insertion efficiency TBS = tert-butyldimethylsilyI. TBAF = tetrabutylammonium fluoride, THF = tetra­ hydrofuran, DMF = N,N-dimethylfonnamide, of le and Id, presumably due to a steric clash since the modifications do not cause significant perturba­ tion of the sugar pucker conformation. We tested the effect of probes 1 a - d on the Klenow Since the ratio of nucleotide-insertion efficiencies opposite fragment (KF-) of E. coli DNA polymerase I (exo--mutant), a complementary to a noncomplementary l1ucleobase is a an enzyme extensively used as a model for investigations into measurement of DNA-polymerase fidelity,[8,1O.1t] we next intrinsic DNA-polymerase mechanisms and function.P] We investigated the ability of KF- to catalyze nOI1-Watson - Crick used a gel-based single nucleotide insertion assay,IS] where an nucleobase pair formation. We focused on la and lb, since adenine in the template strand codes for the insertion of a only these are inserted as efficiently as TTP by KF- in

T'rp PTP A} Primer 5·----ACA r B} C) Prllller 5""'AC~ Template 3'---TGTACT----- Template 3'----TGTACT .... -

o 1...... , 50 1...... , 50 1...... , 50 ...... ,so ...... ,50 ...... , ...... , 50...... , 500 50...... , 500 o 10...... , 500 10...... , 500 10...... , 500 TTP 1a 1b 1<: 1d dCTP TTP 1a 1b TTP 1a 1b Figure 1. Nucleoside triphosphate insertion catalyzcd by KF-. A) Insertion encoded by adenine (A) in the template strand, Conditions: Primer/template complex (50 nM), KF" (2 nM), 3rC, 5 min, B) Nucleoside triphosphate insertion encoded by guanine (G) in the template strand, Conditions: Primer/ template complex (50 nM), KF- (4 nM). 37"C, 5 min. C) Mismatch extension. Conditions: Primer/template complex (50 nM), KF- (2 nM). 3rC, 20 min, In each case. the nuckotide concentrations [ILM] are shown in the figure. dCTP =2'·deoxycytidine-S'-triphosphate, T*TP = TTP analogues, Further experimental details and the DNA sequences employed are described in the Supporting Information,

3694 canonical base pair formation and thus they should be ideally In conclusion, the selectivity of nucleotide insertion by a suited for use in the monitoring of differential enzyme DNA polymerase can be significantly increased by modified interactions between insertion and misinsertion events. As­ sugar moieties. Our results strongly implicate the involvement says of nucleotide incorporation opposite G, T, and C revealed of differential DNA-polymerase interactions with the sugar in that KF- misinserts TIP and 1 a, b opposite G with the processes that contribute to the fidelity of DNA synthesis. highest efficiency.l9, 13) Strikingly, as is already apparent from Furthermore, our studies provide a new functional and the results shown in Figure 1 B, 1 a and 1 b exhibit dramatically general method to monitor steric constraints in nucleotide decreased misinsertion efficiency. Steady-state kinetic analy­ binding pockets of DNA polymerases. Further analyses with ses revealed an approximately lOO-fold decrease in misinser­ such steric probes, both at the functional and structural level tion efficiency using the 4' -alkylated probes 1 a, b compared should reveal more insights into mechanisms of DNA to TIP (Table 2). polymerase selectivity.

Tab)e 2. Steady-state analyses for nucleoside triphosphate misinsertion and mismatch extension. The data presented are averages of duplicate or [I] Recent reviews and commentaries: a) T. A. Kunkel, K. Bebenek, triplicate experiments. Further experimental details are described in A11Illl. Rev. Biochem. 2000,69,497-529: b) E. T. Kool, 1. Morales, Supporting Information. c: K. M. Guckian, Angew. Chem. 2000,112. 1046-1068; Angew. Chel7l.

Nucleoside KM Vmllx vmaxlKM 1nt. Ed. 2000,39,991-1009; c) T. A. Kunkel, S. H. Wilson, Nat. Slrucl. triphosphate [fLMJ [min-' x 1O-'J [M-'min-'J Bioi. 1998,5, 95-99; d) U. Diederichsen, Angew. Chem. 1998, no, 1745-1747: Angew. Che111. 1nl. Ed. 1998. 37, 1655-1657: e) M. F. misinserlion: Goodman, Proc. NaIl. Aead. Sri. USA 1997,94,10493-10495. TIP 22± 0.2 16± 1 730 [2] a) S. Douhlie, S. Tabor, A. M. Long, c: c: Richardson, T. Ellenherger, la 65±5 0.34±0.02 5 Nalure 1998,397,251-258; b) Y. Li, S. KoroJev, G. Waksman, EMBO 1b 228±5 1.7 ± 0.1 7 1. 1998, 77, 7514-7525; c) 1. R. Kiefer, c: Mao, J. c: Braman, L. S. mismatch extension: Beese, Nalure 1998, 397, 304-307; d) H. F. Huang, R. Chopra, G. L. TIP 19±1 36±2 1900 Verdine, S. c: Harrison, Science 1998, 282, 1669 - 1675; e) H. Pelletier, la 80± 17 2.9±0.1 40 M. R. Sawaya, A. Kumar, S. H. Wilson, J. Kraut, Science 1994, 264, 1b 40±5 4.0 ±0.3 100 1891-1903. [3J c: O.-Yang, W. KUIZ, E. M. Eugui, M. J. McRoberts, J. P. H. Verhey­ den, L. 1. Kurz, K. A. M. Walker, Telrahedron Left. 1992,33,41-44. [4] I. Sugimoto, S. Shuto, S. Mori, S. Shigeta, A. Matsuda, Bioorg. Med. Further experiments also showed that the analogues la and Chem. Lett. 1999, 9, 385-388. [5] A. Marx, P. Erdmann, M. SemI, S. Korner, T. Jungo, M. Petretla, P. 1 b were not as well inserted opposite T or C as unmodified Imwinkelried, A. Dussy, K. 1. Kulicke, L. Macko, M. Zehnder, B. TIP was (see Supporting Information). These results clearly Gicsc, He/v. Chim. Acla 1996, 79, 1980-1994. show that nucleotide-insertion selectivity is increased by [6] T. KOV8CS, L. btvos. Telrahedron Lell. 1988,29,4525-4528: recent substitution of the hydrogen atom at the 4'-position of the review: K. Burgess, D. Cook, Chem. Rev. 20()(). lOO. 2047 -2059. sugar with bulkier alkyl groups. [7] L. 1. Rinkel, C. Altona . .I. Biomol. Slruct. Dyll. 1987, 4, 621-648. [8] S. Creighton, L. B. Bloom, M. F. Goodman, Methods Enzymol. 1995, The second critical determinant of intrinsic DNA-polymer­ 262,232-256. ase fidelity is the capacity to extend from mismatched primerl [9] Detailed experimental procedures as well as DNA sequences applied template ends.l1 a 8,11) Here again, la and 1 b should be ideally in the in vitro replication assays are provided in the Supporting suited to monitor differential DNA-polymerase interactions Information. [10J M. S. Boosalis, 1. Petruska, M. F. Goodman, 1. Bioi. Chell1. 1987,262, with the sugar moiety in mismatch extension events. We 14689-14696. investigated extension from template-T/primer-G, TIT, and [I1J M. F. Goodman, S. Creighton, L. B. Bloom, 1. Petruska, Crit. Rev. TIC by using KF- to extend from a mismatched base at the Biochem. Mol. BioI. 1993,28,83-126. primer 3'-end by incorporation of TIP or la, b. The most [12] Quantitative investigations towards the action of 4'-modified triphos· efficient extension was from TIG for all thymidine triphos­ phates on HIV·I reverse transcriptase and cellular DNA polymerases have been reported recently. 4'C-Azidothymidine triphosphate: M. S. 9 phate derivatives (data shown for TIG in Figure 1 C).f , 13) The Chcn, R. T. Sutlmann, E. Papp, P. D. Cannon, M. J. McRoberts, c: results of quantitative analyses of TIG extensions are pre­ Bach, W. C. Copc1and. T. S.-F. Wang. Biochemistry 1993,32,6002- sented in Table 2. KF- extends a TIG mismatched primer end 6010; 4'C-acctylated thymidine triphosphate: a) A. Marx, M. Spichty, with probes la, b significantly less efficiently than with TTP. M. Amacker, U. Schwitter, U. Htibscher, T. A. Bickle, G. Maga, B. Giese, Chem. Bioi. 1999,6, 111 -116: b) A. Marx, M, Amacker, M. Again, this is in contrast to the results in Table 1 where TIP Stucki, U. Htibscher, T. A. Bickle, B. Giese, Nucleic Acids Res. 1998, and 1 a, b extension of a perfectly matched primer were 26, 4063 -4067. equivalent. Our results indicate that nucleobase pair mis­ [13] Results obtained with TIP agree with recently published data: matches at the 3'-end of a primerltemplate complex trigger a) C. M. Joyce, X. C. Sun, N. D. F. Grindley, 1. Bioi. Chelll. 1992,267, unfavorable enzyme interactions with the sugar moiety of an 24485-24500; b) S. S. Carroll, M. Cowart, S.1. Benkovic, Biochem· istry 1991,30,804-813: c) K. Bebcl1ck. C. M. Joycc, M. P. Fitzgerald, incoming nucleoside triphosphate and prevent inadvertent T. A. Kunkel, 1. Bioi. Chenl. 1990,265, 13878 -13 887. sealing of a base substitution in the nascent DNA strand. Indeed, this mechanism may be considered essential for proofreading, and may allow a pause for other functions such as intrinsic exonuclease activity or dissociation of the DNA­ replication complex.

3695