Hidehito Urata,* Yoshiaki Ueda and Masao Akagi*

Osaka University of Pharmaceutical Sciences,
4-20-1 Nasahara, Takatsuki, Osaka 569-1094, Japan
FAX: 072-690-1005; E-mail:urata@gly.oups.ac.jp

(Received 20 October 2003 Accepted 31 October 2003)


The duplex structure of several heterochiral dodecadeoxynucleotides containing an unnatural L-enantiomer of D-deoxyribose was investigated by using nuclear magnetic resonance, ultraviolet, circular dichroism spectroscopy and restriction endonuclease Eco RI digestion. The UV melting study suggested that the L-nucleotide residue of the heterochiral 12mers retains the base pairing with the complementary natural D-counterpart. This was also supported by 1H NMR analysis of L-4, which also revealed the molecular mechanisms for the L-nucleotide to form the Watson-Crick base pairing. The CD experiments suggested that the duplex structure of the heterochiral 12mers is not significantly different from that of the parental one, and some of the 12mers were recognized by restriction endonuclease EcoRI. These results suggested that the breaking homochirality of the 12mer does not induce significant conformational changes or helical distortion, although the duplex stability is somewhat decreased. The destabilization of the duplex structure for heterochiral DNA might have been advantageous for natural selection of homochiral DNA in the processes of the chemical evolution of DNA.

(Keywords) DNA, Homochirality, Heterochiral DNA, Duplex structure, Duplex stability, Nuclear magnetic resonance, Circular dichroism, Restriction endonuclease EcoRI,


DNA forms various double-stranded structures such as A-, B- and Z-form conformations. It is very likely that the polymorphism of the DNA structure is associated with the DNA-protein recognition and the regulation of gene expression. The chirality of D-deoxyribose, which is the only chiral unit of DNA, would be owing to the formation of the highly ordered structure of DNA. The establishment of the homochirality of nucleic acids is thought to have been closely related to the origin of life and the chemical evolution of biomolecules on the primitive earth [1]. During the chemical evolution of nucleic acids, heterochiral nucleic acids might have existed, since mononucleotides would be non-enzymatically synthesized as a racemic mixture on the prebiotic earth. If heterochiral oligomers and polymers did not exist during the chemical evolution of nucleic acids, chiral selection at the monomer level must have been achieved by the beginning of the oligomerization process. Strict chiral selection at the early stage of the chemical evolution where molecules would have existed as only simple monomers, is thought to be rather less probable. Thus, the chemical and physicochemical characteristics of heterochiral nucleic acids might have been associated with the homochiral selection processes. There are few reports for the effects of the substitution of natural D-deoxyribose with L-deoxyribose on the duplex structure and function of DNA [2-4], although many groups have reported the nuclease resistance and hybridization properties towards natural DNA and RNA sequences of L-homochiral oligonucleotides [5-10]. In a preliminary report, we have reported that an oligonucleotide containing an L-deoxyribose residue, namely heterochiral DNA, maintains the stable Watson-Crick base-pairing between natural D- and unnatural L-nucleotide residues by using nuclear magnetic resonance (NMR) spectroscopy [3].
      In this paper, we report on the details of the NMR analysis, the effects of breaking homochirality at a different site on the structure of self-complementary 12mer DNA fragment [11] by using circular dichroism (CD) spectroscopy and restriction endonuclease Eco RI digestion experiments.


EcoRI was purchased from Takara Shuzo Co. (Kyoto, Japan). L-Deoxynucleosides and their phosphoramidites were synthesized by the previously reported procedure [12]. Dodecadeoxynucleotides were synthesized on an Applied Biosystems model 392 DNA/RNA synthesizer. Reagents for the DNA/RNA synthesizer were purchased from Applied Biosystems Japan (Tokyo, Japan). Cleavage from the support and deprotection were effected by treatment with concentrated aqueous ammonia (6 h at 55°C) and 80% acetic acid (15 min at room temperature), purification was carried out on a reversed-phase HPLC. HPLC analyses were performed on a Shimadzu LC-10A system. A mBondasphere C185µm 100Åcolumn (3.9 x 150 mm, Waters Corporation, Milford, MA, USA) was used with a linear gradient of acetonitrile in 50 mM triethylammonium acetate (TEAA, pH 7.0).
1H NMR measurements Samples (240 OD units) of the 12mers were dissolved (2.5 mM) in a buffer containing 0.1 M NaCl, 0.1 mM EDTA, 10 mM sodium phosphate, pH 7.5 in 20% D2O for measurements of exchangeable protons or 100% D2O for mesurements of non-exchangeable protons. 1H NMR spectra were measured on a JEOL GX400 or Bruker AM 500 spectrometer. Chemical shifts were measured relative to internal 2-methyl-2-propanol (1.23 ppm from sodium 2,2-dimethyl-2-silapentane-5-sulfonate). Two-dimensional NOESY and DQF-COSY spectra in D2O were recorded at 25°C with 2048 points in 2 and 256 points in t1 (spectral width, 5000 Hz each). The NOESY and DQF-COSY data were collected in the phase-sensitive mode by the method of States et al [13]. In the NOESY experiments, a mixing time of 150 ms or 200 ms was used. The time domain data were zerofilled to 1024 points in the 1 dimension before Fourier transformation.
Measurements of melting curves The concentrations of oligonucleotide solutions were calculated by using equation and coefficients described by Borer [14]. The coefficients of heterochiral 12mers were calculated on the assumption that modified dimers have the same hypochromicity as the corresponding unmodified dimers. Molar coefficients (ε260) of L-dG, L-dA, L-dC and L-dT were assumed to be the same as natural one. The 12mers were dissolved in a buffer containing 1 M NaCl, 10 mM sodium phosphate (pH 7.5) at single strand concentration of 8.5 µM. After annealing, the solution was transferred to a quartz cell (1 cm path length), and melting curves were measured at least twice at 275 nm on a JASCO Ubest-55 spectrophotometer. The temperature was raised at a rate of 0.5°C/min, and the Tm values were obtained by the first-derivative plots of the melting curves.
Measurements of CD spectra The same samples (12mer single strand concentration; 8.5 µM in 1 M NaCl, 10 mM sodium phosphate, pH 7.5) as used for UV melting experiments were employed for CD experiments. Measurements were carried out on a JASCO J-500 spectropolarimeter.
Digestion of the 12mers with EcoRI Dodecanucleotides (0.1 A260) were incubated at 20°C in 50 mM Tris-HCl (pH 8.0), 0.7 mM 2-mercaptoethanol, 10 mM MgCl2 and 100 mM NaCl with Eco RI (250 units) in a final volume of 100 µl. Aliquots of the reaction mixture were removed and heated at 95°C for 5 min, then analyzed by a reversed phase HPLC with a linear gradient of acetonitrile (2.5-12.5 %) containing 50 mM ammonium acetate.

Results and Discussion

We synthesized heterochiral 12mers and the parental homochiral one, which have a self-complementary sequence, shown below:
Native d(C-G-C-G-A-A-T-T-C-G-C-G)
L-3 d(C-G-(L-C)-G-A-A-T-T-C-G-C-G)
L-4 d(C-G-C-(L-G)-A-A-T-T-C-G-C-G)
L-5 d(C-G-C-G-(L-A)-A-T-T-C-G-C-G)
L-9 d(C-G-C-G-A-A-T-T-(L-C)-G-C-G)
L-10 d(C-G-C-G-A-A-T-T-C-(L-G)-C-G)
To investigate their duplex stability, UV melting experiments were conducted. Figure 1 shows the melting curves. Although the homochiral native 12mer showed a sigmoidal curve, suggesting the two-state transition, L-5 showed a non-cooperative melting curve. Therefore, L-5 would have a more complex melting pathway (e.g. duplex ⇔ sigle strand ⇔ hairpin) [15]. The destabilization of the core sequence (-AATT-) by the introduction of L-deoxyadenosine might cause the preferential formation of the intramolecular base pairing between the two CGCG sequences. In contrast, the heterochiral 12mers other than L-5 showed cooperative sigmoidal curves as well as the native 12mer. The some decreases of the melting temperatures (Tms) of the heterochiral 12mers were observed compared with the native 12mer, however, the extent of the Tm decreases per substitution (ΔTm 10°C) is much less than that by introducing a base-pair mismatch (ΔTm 20°C) [16]. Thus, the L-nucleotide residue in the 12mers other than L-5 is likely to form the base pairing with the complementary natural D-residue. Figure 2 shows circular dichroism (CD) spectra of the 12mers. The CD spectrum of native 12-mer at 20 °C has positive and negative cotton bands at 283 nm and 252 nm, respectively, indicating that native 12-mer adopts the B-form conformation. The other 12mers showed the similar spectra with the native 12mer, suggesting them to retain the right-handed B-form duplex, and the 12mers other than L-5 showed isoelliptic points at around 275 nm and 227 nm. This result indicates that the structural transition of these 12mers induced by heating is the simple two-state transition as suggested by the UV melting experiments. However, the temperature dependence of the CD spectrum of L-10 is considerably different from those of other 12mers, suggesting the duplex structure of L-10 to be somewhat different from the others.

Figure 1. UV melting curves of the native and heterochiral 12mers. Samples contained 8.5 mM strand in 1 M NaCl, 10 mM sodium phosphate, pH 7.5.

Figure 2. CD spectra of the native and heterochiral 12mers. Samples contained 8.5 mM strand in 1 M NaCl, 10 mM sodium phosphate, pH 7.5.

     The macroscopic structural analyses of the 12mers provided the information for overall duplex structure of the 12mers. Our subsequent interest is how they form the duplex structure, especially at the substitution site with L-nucleotide. To obtain such information, we carried out nuclear magnetic resonance (NMR) experiments. The signal assignment was carried out by using the sequential assignment method which is widely applied to B-form DNA fragments [17]. Figure 3 shows the 1H NMR spectra of exchangeable imino resonance region of the native 12mer and L-4. Imino protons (guanine-1NH and thymine-3NH) serve as a proton donor for base pairing. Since imino protons, which are exchangeable with solvent protons, are unable to be detected when an oligonucleotide exists as a single strand. However, when the oligonucleotide forms a duplex (base pairing), the exchange rate of the imino protons dramatically reduces and the imino resonances become to be observed. In this case, the sequence of the 12mers has the twofold axis, and then only the protons in the region surrounded by a rectangle (Figure 4) are observable by 1H NMR. The native 12mer showed six imino resonances at 0°C (Figure 3, bottom), indicating all the base pairs to be formed. In the case of L-4, all imino resonances containing G4NH are also observed (Figure 3, upper), thus the unnatural G4 residue forms the base-pairing as well as other natural residues as expected by the UV melting experiments. Figure 5a and 5b shows the temperature dependence of the chemicalshifts for the imino resonances of L-4 and the native12mer, and the chemical shift differences of the non-exchangeable protons between L-4 and the native 12mer, respectively. The temperature dependence of the imino resonances is quite similar in both 12mers (Figure 5a). Imino protons in duplex DNA are positioned where are strongly affected by ring current effects of neighboring base pairs. Therefore, this result suggests that L-4 still retains the similar base pairing mode and base-base stacking geometry as parental one. In the case of the non-exchangeable protons, the large chemical shift differences between L-4 and the native 12mer are converged at the G4 residue (Figure 5b), and therefore the introduction of L-deoxyguanosine into the G4 site hardly affects the structure of the other regions. These results suggest that L-4 still retains the B-form type base pairing and base stacking, and the introduction of L-deoxyguanosine does not significantly affect the overall duplex structure.

Figure 3. 1H NMR spectra of L-4 (upper) and the native 12mer (bottom) in the exchangeable imino resonance region. Samples contained 2.5 mM duplex in 0.1 M NaCl, 0.1 mM EDTA, 10 mM sodium phosphate (pH 7.5) in 20 % D2O at 1°C.

Figure 4. Nucleotide sequence and numbering system of the 12mer duplexes. The G4 residue of L-4 is unnatural L-deoxyguanosine. Since this sequence has a dyad axis, only protons in the residues surrounded by a rectangle can be observed.

Figure 5. (a) Temperature dependence of chemical shifts of imino resonances of L-4 (solid line) and the native 12mer (dotted line). (b) Chemical shift differences between L-4 and the native 12mer at 25°C. The differences are defined by Δδd = δ (L-4)δ (native 12mer).
      Figure 6a shows the sequential NOE connectivities between the base H8 (purines) or H6 (pyrimidines) protons and sugar H1' protons of L-4. As shown in Figure 6a, the NOE cross-peaks could be completely connected. This means that the duplex structure of the 12mer does not change dramatically by the substitution of the G4 residue with L-deoxyguanosine. Figure 6b shows the sequential NOE connectivities between the base H8 or H6 protons and sugar H3' protons of L-4. In this case, the NOEs derived from G4H8-G4H3' and A5H8-G4H3' are missing, instead, unusual NOEs derived from G4H8-G4H4' and A5H8-G4H4' were observed. It was found that all the residues in L-4 including the G4 residue adopt the S-type sugar puckering by the analysis of the DQF-COSY spectrum (data not shown). When the deoxyguanosie residue with the S-type sugar conformation adopts a usual anti glycosyl conformation, strong NOEs between base H8 and the sugar H2', H3' are observed. In the case of the G4 residue of L-4, the NOE between H6 and H3' could not be observed as described above, but the weak NOE between H6 and H4' was observed. This means that H6 is close to rather H4' than H3' [3]. This result indicates that the unnatural G4 residue of L-4 adopts an unusual low anti glycosyl conformation (χ = 180° 30°) (χ is defined by a torsion angle of C4-N9-C1'-O4' for prines and C2-N1-C1'-O4' for pyrimidines), although each residue of regular B-form DNA fragments adopts an anti glycosyl conformation (χ = 135° 30°) (Figure 7).
Based on the above experimental results, the model structure of L-4 was constructed. As shown in Figure 8, the G4 residue forms the Watson-Crick base-pairing with the complementary C9 residue. Each O4' atom of the sugar moiety in the regular B-form DNA is oriented toward the 5'-end direction like as the C3 and A5 residues in L-4, however, that of the G4 residue is oriented toward the opposite direction. This characteristic conformation would make the L-deoxyguanosine residue enable to form the usual Watson-Crick base pairing.

Figre6. NOESY spectrum (150 ms mixing time) of L-4 in D2O at 25°C. Duplex concentration is 2.5 mM in 0.1 M NaCl, 0.1 mM EDTA, 10 mM sodium phosphate, pD 7.5. (a) Expanded base proton and sugar H1' region of the spectrum. (b) Expanded base proton and sugar H3' region of the spectrum. Arrowhead represents the chemical shift of the missing G4H8-G4H3'.

Figure 7. Structures of L-deoxyguanosine 5'-monophosphate with the anti (χ = 135°30°, left) and low anti (χ = 180°30°, right) glycosyl conformations. χ is defined by a torsion angle of C4-N9-C1'-O4'.

Figure 8. Model structure of L-4. Protons and the residues other than C3, L-G4, A5 are omitted for clarity. The O4' atoms of the C3, L-G4, A5 residues are highlighted with a red sphere.

     To further evaluate the duplex structure of the 12mers, digestion experiments of the 12mers with restriction endonuclease EcoRI was carried out. This enzyme strictly recognizes the sequence of 5'-GAATTC-3' in double-stranded B-form DNA and cleaves a phosphodiester bond at the GpA step in the recognition sequence. The specificity of the DNA recognition by this enzyme is achieved by both direct and indirect read-out mechanisms [18,19], which involve protein-base contacts through hydrogen-bonding and van der Waals contacts (direct read-out), and recognition of the overall conformation of DNA (indirect read-out). Therefore, the susceptibility of DNA fragments for the restriction endonuclease digestion affords structural information of the substrate DNAs. The time-course of the cleavage reaction of the 12mers with EcoRI is shown in Figure 9. The reaction was performed at 20°C, since these 12mers form the stable duplex at this temperature from the CD and the UV melting experiments. Under the conditions that native 12-mer is completely hydrolyzed (48 hr), L-4 and L-5 show complete resistance to digestion with this enzyme. In contrast, L-3, L-10 and even L-9 that contains the L-nucleotide in the recognition sequence were hydrolyzed, although the cleavage rate of L-3 and L-9 is significantly reduced. Surprisingly, the cleavage rate of L-10 is about 20-fold enhanced compared with that of the native 12mer. From the CD experiments, the overall duplex structure of L-10 has significantly different characteristics from that of the other 12mers. The enhanced hydrolysis rate of L-10 would be due to the indirect effect such as conformational change, since the direct interaction between EcoRI and the 3'-flanking residue of the recognition sequence, corresponding to the L-deoxyguanosine residue in L-10, was not observed in the co-crystal of the EcoRI-DNA fragment complex [19]. Thus, L-10 might have the kinked structure to easily accommodate the binding with EcoRI as seen in the EcoRI-DNA fragment complex. The results obtained for L-4 and L-5 may suggest that the introduction of L-nucleotide at the catalytic site is critical for the reaction (catalysis). Thus, heterochiral 12mers retains the B-form type duplex structure that can be recognized by EcoRI.

Figure 9. Time-course of cleavage reaction of the 12mers with restriction endonuclease EcoRI.

      The establishment of the homochirality of nucleic acids and proteins is thought to have been essential for the origins of life. The chemical processes that biomolecules acquire the homochirality can be considered to involve three events: 1) symmetry breaking, 2) chiral amplification, and 3) chiral transfer. Symmetry breaking is explained by, for example, a predicted small energy difference between enantiomeric molecules as a result of a parity violation in the weak nuclear force [20], or a merely statistical chance [21]. Chiral amplification can make molecules possible to achieve large enantiomeric excesses from small chiral imbalance [22]. The amplified chirality of a molecule transfers to other molecules (chiral transer). Since the rate of racemization of amino acids is relatively too fast, it would not be reasonable to consider the chirality of amino acids and peptides as a source of the chirality of the biosphere. We have thus focused on the chirality of nucleic acids. Our results described here indicate that the introduction of an L-nucleotide into DNA does not induce significant helical distortions. In other words, the DNA double-helical structure permits the incorporation of a chiral antipode, although some decreases of the duplex stability are observed. This allows us to speculate that heterochiral DNA can serve as a template for the non-enzymatic oligomerization. However, the decrease of the duplex stability would serve unfavorably for the complex formation of a template with monomers and growing strands, and for the chemical stability of DNA [23]. Actually, Orgel and coworkers reported that the non-enzymatic oligomerization of activated guanosine 5'-monophosphate (2-MeImpG) on a DNA template containing one or two successive L-deoxycytidine residues proceeds to the end of the template through the L-deoxycytidine residues [24]. However, the efficiency is rather low relative to the oligomerization on an all-D-template, furthermore a template containing three successive L- deoxycytidine residues or alternating D-,L-template does not facilitate the oligomerization. The structural characteristics of heterochiral DNA shown in this study well explain the template activity of the DNA oligomers containing L-nucleotides reported by Orgel et al [24]. Overall, the introduction of L-type monomers into DNA should be unfavorable for the non-enzymatic oligomerization and replication of DNA. Such characteristic of heterochiral DNA might have in part contributed to the natural selection of homochiral DNA.


We have investigated the duplex structure of heterochiral DNA 12mers that have an unnatural L-nucleotide residue, and found that the breaking homochirality of the 12mer does not cause a drastic structural alteration. However, the breaking homochirality of the 12mer leads to the some destabilization of the duplex. Such destabilization might have been disadvantageous for the natural selection of heterochiral DNAs in the processes of the chemical evolution of DNA via oligomerization of mononucleotides that would have been abiotically synthesized as a racemate on the primitive earth.


This work was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, and Culture, Japan.


[1] Joyce, G. F., Schwartz, A. W., Miller, S. L. and Orgel, L. E. The case for an ancestral genetic system involving simple analogs of the nucleotides, Proc. Natl. Acad. Sci. U.S.A., 84, 4398-4402 (1987).
[2] Urata, H., Shinohara, K., Ogura, E., Ueda, Y. and Akagi, M. Mirror Image DNA, J. Am. Chem. Soc. 113, 8174-8175 (1991).
[3] Urata, H., Ueda, Y., Suhara, H., Nishioka, E. and Akagi, M. NMR study of a heterochiral DNA: stable Watson-Crick-type base-pairing between the enantiomeric residues, J. Am. Chem. Soc. 115, 9852-9853 (1993).
[4] Blommers, M. J. J., Tondelli, L. and Garbesi, A. Effects of the introduction of L-nucleotides into DNA. Solution structure of the heterochiral duplex d(G-C-G-(L)T-G-C-G)d(C-G-C-A-C-G-C) studied by NMR spectroscopy, Biochemistry 33, 7886-7896 (1994). [5] Fujimori, S., Shudo, K. and Hashimoto, Y. Enantio-DNA recognizes complementary RNA but not complementary DNA, J. Am. Chem. Soc. 112, 7436-7438 (1990).
[6] Asseline, U., Hau, J.-F., Czernecki, S., Diguarher, T. L., Perlat, M.-C., Valery, J.-M. and Thuong, N. T. Synthesis and physicochemical properties of oligonucleotides built with either a-L or b-L nucleotides units and covalently linked to an acridine derivative, Nucleic Acids Res. 19, 4067-4074 (1991).
[7] Anderson, D. J., Reischer, R. J., Taylor, A. J. and Wechter, W. J. Preparation and characterization of oligonucleotides of D- and L-2'-deoxyuridine, Nucleosides Nucleotides 3, 499-512 (1984).
[8] Morvan, F., Genu, C., Rayner, B., Gosselin, G. and Imbach, J.-L. Synthesis, nuclease resistance and base pairing properties of α- and β-L-octathymidylates, Biochem. Biophys. Chem. Res. 172, 537-543 (1990).
[9] Garbesi, A., Capobianco, M. L., Colonna, F. P., Tondelli, L., Arcamone, F., Manzini, G., Hilbers, C. W., Aelen, J. M. E. and Blommers, M. J. J. L-DNAs as potential antimessenger oligonucleotides: A reassessment, Nucleic Acids Res. 21, 4159-4165 (1993).
[10] Damha, M. J., Giannaris, P. A. and Marfey, P. Antisense L/D-oligodeoxynucleotide chimeras: Nuclease stability, base-pairing properties, and activity at directing ribonuclease H, Biochemistry 33, 7877-7885 (1994).
[11] Urata, H. and Akagi, M. Sequence dependence of thermodynamic stability of heterochiral DNA, Tetrahedron Lett. 37, 5551-5554 (1996).
[12] Urata, H., Ogura, E., Shinohara, K., Ueda, Y. and Akagi, M. Synthesis and properties of mirror-image DNA, Nucleic Acids Res. 20, 3325-3332 (1992).
[13] States, D. J., Haberkorn, R. A. and Ruben, D. J. A two-dimensional nuclear Overhauser experiment with pure absorption phase in four quadrants, J. Magn. Reson. 48, 286-292 (1982).
[14] Fasman, G. D. Ed., Handbook of biochemistry and Molecular Biology, 3rd ed., Vol. 1, Nucleic Acids; pp. 589, CRC Press, Boca Raton, FL, 1975.
[15] Marky, L. A., Blumenfeld, K. S., Kozlowski, S. and Breslauer, K. J. Salt-dependent conformational transitions in the self-complementary deoxydodecanucleotide d(CGCGAATTCGCG): evidence for hairpin formation, Biopolymers 22, 1247-1257 (1983).
[16] Leonard, G. A., Booth, E. D. and Brown, T. Structural and thermodynamic studies on the adenineguanine mismatch in B-DNA, Nucleic Acids Res. 18, 5617-5623 (1990).
[17] Hare, D. R., Wemmer, D. E., Chou, S. H., Drobny, G. and Reid, B. R. Assignment of the nonexchangeable proton resonances of d(C-G-C-G-A-A-T-T-C-G-C-G) using two-dimensional nuclear magnetic resonance methods, J. Mol. Biol. 171, 319-336 (1983).
[18] Lesser, D. R., Kurpiewski, M. R. and Jen-Jacobson, L. The energetic basis of specificity in the Eco RI endonuclease-DNA interaction, Science 250, 776-786 (1990).
[19] McClarin, J. A., Frederick, C. A., Wang, B.-C., Greene, P., Boyer, H. W., Grable, J. and Rosenberg, J. M. Structure of the DNA-Eco RI endonuclease recognition complex at 3 Å resolution, Science 234, 1526-1541 (1986).
[20] Wu, C. S., Ambler, E., Hayward, R. W., Hoppes, D. D. and Hudson, R. P. Experimental test of parity conservation in β-decay, Physical Review 105, 1413-1415 (1957). [21] Thiemann, W. The origin of optical activity, Naturwissenschaften 61, 476-483 (1974).
[22] Soai, K., Shibata, T., Morioka, H. and Choji, K. Asymmetric autocatalysis and amplification of enantiomeric excess of a chiral molecule, Nature 378, 767-768 (1995).
[23] Lindahl, T. Instability and decay of the primary structure of DNA, Nature 362, 709-715 (1993).
[24] Kozlov, I. A., Pitsch, S. and Orgel, L. E. Oligomerization of activated D- and L-guanosine mononucleotides on templates containing D- and L-deoxycytidylate residues, Proc. Natl. Acad. Sci. USA 95, 13448-13452 (1988).