Minor groove hydration is critical to the stability of DNA duplexes [13]

Full text of DOI:10.1021/ja000686v

6512
J. Am. Chem. Soc. 2000, 122, 6512-6513
Minor Groove Hydration Is Critical to the Stability
of DNA Duplexes
Tao Lan and Larry W. McLaughlin*
Department of Chemistry, Merkert Chemistry Center
Boston College, Chestnut Hill, Massachusetts 02467
ReceiVed February 25, 2000
Since the original observation of an ordered set of water
molecules,1,2 termed the “spine of hydration,” within the primary
hydration sphere of the minor groove of duplex B-form DNA,
numerous studies have concentrated upon the function and role-
(s) of minor groove hydration. In the described spine of hydration,
individual water molecules bridge the N3-nitrogens of adenines
and the O2-carbonyls of thymines at adjacent base pairs. A second
layer of water molecules bridges the oxygen atoms of the
underlying first layer. Theoretical analyses,3,4 NMR studies5 and
additional crystallography6-11 have all largely confirmed the
presence of this chain of ordered water molecules within the minor
groove, but its importance to the structural integrity of B-form
DNA remains to be clarified.
To probe the importance of minor groove hydration for duplex
DNA, we designed an analogue dA-dT like base pair that lacks
both the N3-nitrogen of the adenine as well as the O2-carbonyl of
the thymine (Figure 1). When present in duplex DNA, this base
pair would disrupt the minor groove spine of hydration, in fact it
should eliminate the presence of any ordered water molecules or
metal ions at these sequence positions since both sites used for
hydration in the minor groove are absent. The analogue base pair
used in this study contained as a purine element, 3-deaza-2′-
deoxyadenosine (dc3A) (Figure 1), a known12 analogue in which
the N3-nitrogen is replaced by a C-H residue. Its pyrimidine-
like partner was a 3-methyl-2-pyridone (dm32P) (Figure 1). The
latter derivative was prepared by a palladium-mediated coupling13
between the iodo derivative of the 2-pyridone and the corre-
sponding protected glycal. Further protection resulted in the
phosphoramidite derivative suitable for DNA synthesis.
Figure 1. (a) Watson-Crick dA-dT base pair. (b) Watson-Crick-like
base pair between dc3A and dm32P.
in TM (<2 °C, entries 2 and 3, Table 1). With the introduction of
a modified base pair (entry 4), roughly a 3 °C change in TM is
present, and with two or four analogue base pairs (entries 5 and
6), the TM values decrease by about 10 and 26 °C, respectively.
With four analogue residues distributed over the core sequence
(entry 7) the TM value is about the same as that obtained with
two analogue base pairs (entry 5). Thermodynamic analyses have
assisted to characterize those complexes containing modified base
pairs. The introduction of a single analogue base pair results in
a 1.8 kcal/mol change in ∆G (entry 4), while the addition of a
second modified base pair (entry 5) only results in a further 0.8
kcal/mol change. The presence of all four modified base pairs
(entry 6) results in an additional loss of 1.4 kcal/mol in
stabilization energy, an average of 0.7 kcal/mol per added base
pair.
For comparison, we prepared the self-complementary Dickerson
dodecamer containing a block of four dc3A‚dm32P base pairs.
At 4 µΜ total strand concentration this sequence exhibited a single
transition at 70 °C. At increased concentration (16 µM) two
transitions were present; one at 23 °C with moderate hyperchro-
micity and a second at 70 °C, moderately higher than that
observed for the unmodified duplex. A biphasic transition, with
a concentration dependence for the early transition, is a typical
observation15 for hairpin loops that can also adopt duplex
structures with a central core devoid of Watson-Crick (W-C)
base pairing. Using non-denaturing PAGE, we confirmed that the
self-complementary dodecamer with four analogue base pairs
preferentially adopts the hairpin conformation (data not shown).
We additionally prepared a related self-complementary sequence
in which the central A-T and T-A base pairs were replaced by
G-C and C-G, respectively. This complex exhibited a single
transition at all concentrations examined. When the remaining
two T residues were replaced by the dm32P residues the TM was
largely unchanged. Thermodynamic parameters for the two
sequences were nearly identical (entries 8 and 9 in Table 1).
With the analogue phosphoramidites in hand, we prepared a
series of double-stranded sequences similar to that first crystallized
by Dickerson14 but substituted the analogue base pair for one or
more native base pairs within the central core d(AATT)2 sequence
(Table 1). Introduction of either a single dc3A or dm32P residue
into the dodecamer sequence results in a relatively minor effect
* To whom correspondence should be addressed. Telephone: (617) 552-
3622. Fax: (617) 552-2705. E-mail: larry.mclaughlin@bc.edu.
(1) Drew, H. R.; Dickerson, R. E. J. Mol. Biol. 1981, 151, 535-556.
(2) Kopka, M. L.; Fratini, A. V.; Drew, H. R.; Dickerson, R. E. J. Mol.
Biol. 1983, 163, 129-146.
(3) Chen, Y. Z.; Prohofsky, E. W. Biophys. J. 1993, 64, 1385-1393.
(4) Young, M. A.; Jayaram, B.; Beveridge, D. L. J. Am. Chem. Soc. 1997,
119, 59-69.
(5) Liepinsh, E.; Otting, G.; Wuthrich, K. Nucleic Acids Res. 1992, 20,
6549-6553.
(6) Shui, X. Q.; Sines, C. C.; McFail-Isom, L.; VanDerveer, D.; Williams,
L. D. Biochemistry 1998, 37, 16877-16887.
(7) Shui, X. Q.; McFailIsom, L.; Hu, G. G.; Williams, L. D. Biochemistry
1998, 37, 8341-8355.
The experimental results strongly suggest that in the presence
of the central core of analogue residues the dodecamer sequence
results in decreased helix stability, and when present in a self-
complementary sequence results in a strong preference for the
hairpin conformation rather than the duplex. At least three
parameters could be responsible for this preference, (i) changes
in interstrand hydrogen bonding, (ii) changes in base stacking,
or (iii) changes in hydration/metal ion binding.
(8) Tereshko, V.; Minasov, G.; Egli, M. J. Am. Chem. Soc. 1999, 121,
470-471.
(9) Chiu, T. K.; Kaczor-Grzeskowiak, M.; Dickerson, R. E. J. Mol. Biol.
1999, 292, 589-608.
(10) Tereshko, V.; Minasov, G.; Egli, M. J. Am. Chem. Soc. 1999, 121,
3590-3595.
(11) Woods, K. K.; McFail-Isom, L.; Sines, C. C.; Howerton, S. B.;
Stephens, R. K.; Williams, L. D. J. Am. Chem. Soc. 2000, 122, 1546-1547.
(12) Seela, F.; Grein, T. Nucleic Acids Res. 1992, 20, 2297-2306.
(13) Hsieh, H. P.; McLaughlin, L. W. J. Org. Chem. 1995, 60, 5356-
5359.
(14) Wing, R.; Drew, H.; Takano, T.; Broka, C.; Tanaka, S.; Itakura, K.;
Dickerson, R. E. Nature 1980, 287, 755-758.
(15) Xodo, L. E.; Manzini, G.; Quadrifoglio, F.; van der Marel, G.; van
Boom, J. Biochemistry 1988, 27, 6321-6326.
10.1021/ja000686v CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/23/2000

Communications to the Editor
J. Am. Chem. Soc., Vol. 122, No. 27, 2000 6513
Table 1. Effects of dc3A and dm32P Substitutions on the Stability
SM325 methods. By AM1 analysis, the dipole magnitude increases
from 2.35 D (dA) to 3.46 D (dc3A) with the analogue, but in the
second case the magnitude decreases from 4.49 D (dT) to 3.73
D (dm32P). The average for the base pair remains largely
unchanged, and a similar phenomenon is observed using PM3-
SM3 methodology. Hydrophobicity, as least as measured by
mobility during reversed-phase HPLC is only moderately affected.
With more significant detrimental base stacking effects for the
analogue, the TM values for entries 8 and 9 in Table 1 would
differ. These observations suggest that while some alteration in
the nature of the base stacking interactions due to changes in
dipolar or hydrophobicity effects may be present, the magnitude
of these differences are not responsible for signifcant losses in
duplex stabilization.
(iii) Hydration/Metal Ion Binding. Sequences containing
adjacent dc3A residues (opposite dT's) have been characterized
has having moderately reduced TM values,12 but appear to still
exist primarily as duplex structures. In those sequences the O2-
carbonyls of the partner dT residues are still available as sites
for hydration/metal ion binding in the minor groove. Hydration
in those sequences may not be as effective as in native dA-dT
containing sequences and consequently the helix is destabilized.
In the present study, the complete absence of both the N3-nitrogens
and the O2-carbonyls at as many as four sites within the center
of a dodecamer eliminates all such possible interactions, and
dramatically alters helix stability. Thermodynamic analyses
indicate that the presence of one analogue base pair results in a
1.8 kcal/mol loss in stabilization energy, while the presence of
each additional analogue results in an additional 0.7-0.8 kcal/
mol loss. This observation is consistent with a cooperative effect
relating to the spine of hydration. The initial disruption of the
ordered water molecules arising from the introduction of a single
analogue base pair may propagate along the minor groove and
disrupt additional ordered water molecules at adjacent sites.
Adjacent analogue base pairs then have a correspondingly
moderate effect. These observations are consistent with theoretical
calculations3 suggesting that in dA-dT sequences the spine of
hydration should be considered an integral part of the helix and
enhances the stability of the helix against base-pair opening events.
The corollary to this suggestion is that in the complete absence
of minor groove hydration, base pair opening at dA-dT base pairs
can be expected to be enhanced and to contribute to the
destabilization of the B-form duplex.
of a Duplex Dodecamer1
1 Tm values were obtained at a concentration of 3 µM in 20 mM
NaH2PO4, 1 M NaCl, pH 7.0 buffer using an AVIV 14DS spectro-
photometer. Samples were heated at 0.5 °C/min. 2 a ) dc3A and t )
dm32P. 3 Calculated with T ) 25 °C and reported in kcal/mol for ∆G
and ∆H and in cal/mol °K for ∆S.
(i) Interstrand Hydrogen Bonding. We designed the analogue
base pair such that bidendate W-C-like hydrogen bonding would
not only be possible but probable (Figure 1). The dc3A residue is
essentially a ring-fused derivative of 2-aminopyridine, and UV
studies of 2-aminopyridines have clearly established that these
derivatives prefer the amino rather than the imino form.16 Studies
of acid dissociation constants have led to the same conclusions.17
For the pyrimidine analogue we chose to prepare the C-nucleoside
of the requisite 2-pyridone. Also as the result of ultraviolet studies,
2-pyridones have been shown to prefer the keto rather than the
enol form.18 In the absence of the O2-carbonyl, less hindered
rotation about the glycosidic bond will occur, but it seems likely
that as the hydrogen bonding face of the dc3A residue approaches
the dm32P analogue, the complementary hydrogen bonding face
will be selected. If free rotation about the glycosidic bond
represented a significant entropic disadvantage to base pairing,
then the thermodynamic parameters measured for entries 8 and
9 in Table 1 would differ. Had either analogue undergone an
unfavorable tautomeric shift, the base pair containing the analogue
residue would have a lone pair-lone pair clash or an N-H) (H-N
steric interactionsboth lead to severe helix destabilization when
present at even a single site.19,20 Other effects could be present
and related to the change in the length of the glycosidic bond for
dm32P resulting from the replacement of the N1-nitrogen by
carbon, but when present in a G-C rich sequence (entry 9 in
Table 1) no loss of stabilization was observed. Additionally,
pseudouridine is a related derivative in which the uracil base has
been rotated and attached to the carbohydrate through the C5
position. It exhibits essentially normal base pairing with adenosine
in simple duplexes.21
In addition to the absence of hydration, the inability to localize
metal ions6-8,10,11 to either the N3-nitrogens or O2-carbonyls in
the minor groove, or to water molecules hydrating these sites
(there is an alternative to this electrostatic view of the minor
groove9), may increase phosphate-phosphate repulsion across the
minor groove to destabilize the duplex, or in the case of the self-
complementary sequence, facilitate the noted conformational
change. Although bidentate hydrogen bonding base pairs with
minimally altered base stacking properties are present in the core
sequence, these effects alone do not appear to be sufficient for
duplex stabilization. The spine of hydration associated with
possible metal ion binding in the minor groove is critical for the
structural integrity of duplex DNA.
(ii) Base-Base Stacking. The base stacking interactions
between neighboring heterocycles are more difficult to quantify.
In general it seems that heterocycles exhibit better stacking
interactions than simple aromatic compounds but the reasons for
these observations are less clear. Both dipole-dipole interactions22
as well as the hydrophobic effect23 have been suggested by recent
studies as being origin for base stacking. We have calculated the
dipole moment for the two native and analogue nucleosides used
in this study in an aqueous environment both by AM124 and PM3-
Acknowledgment. We thank Tim Searls and Dr. Narendra Vaish,
Boston College, for experimental assistance. We thank Professor Loren
Williams, Georgia Institute of Technology, for helpful discussions. This
work was supported by a grant from the NSF (MCB-9723844).
(16) Anderson, J.; Seeger, I. J. Am. Chem. Soc. 1943, 71, 340-345.
(17) Angyal, H. J. Chem. Soc. 1952, 1461-1465.
(18) Katritzky, A. R.; Lagowski, J. M. AdV. Heterocyl. Chem. 1962, 1,
131-141.
(19) Rajur, S. B.; McLaughlin, L. W. Tetrahedron Lett. 1992, 33, 6081-
6084.
(20) Searls, T.; McLaughlin, L. Tetrahedron 1999, 55, 11985-11996.
(21) Hall, K. B.; McLaughlin, L. W. Nucleic Acids Res. 1992, 20, 1883-
1889.
(22) Newcomb, L. F.; Gellman, S. H. J. Am. Chem. Soc. 1994, 116, 4993-
4994.
Supporting Information Available: Synthesis procedures for the
dm32P nucleoside analogue, sample HPLC analyses of hydrolyzed
oligonucleotides, sample absorbance vs temperature plots, and gel
electrophoresis experiments (PDF). This material is available free of
charge via the Internet at http//:pubs.acs.org.
JA000686V
(23) Friedman, R. A.; Honig, B. Biophys. J. 1995, 69, 1528-1535.
(24) Dixon, R. W.; Leonard, J. M.; Hehre, W. J. Isr. J. Chem. 1993, 33,
427-434.
(25) Cramer, C. J.; Truhlar, D. G. J. Comput.-Aided Mol. Design 1992, 6,
629-666.