Modeling DNA Replication Intermediates

Debjani Roy*, Brian E. Hingerty*, Robert Shapiro# and Suse Broyde*a

=X-

*13iology Department, New York University,New York, NY 10003-5181

$Health Sciences Division, Oak Ridge National Laboratory,Oak Ridge, TN 37831-8077

#chemistry Department, New York University,

New York, NY 10003-5180

ORNL/CP-101136

a Corresponding Author: (212) 998-8231, Fax (2 12) 995-4015, E-mail broyde@nyu.edu

The submitted manuscript has been authored by a contractor of the U. S. Government under contract number DE-AC05-960R22464.Accordingly, the U.S. Government retains a non-exclusive royalty free license to publish or reproduce the published form of this contribution,or allow others to do so for U. S. Government purposes.

DISCLAIMER

This report was prepared as an account of work sponsoredbyanagency of the United States Government. Neither theUnited States Government nor any agency thereof, nor anyof their employees, make any warranty, express or implied,or assumes any legal liability or responsibility for theaccuracy, completeness, or usefulness of any information,apparatus, product, or process disclosed, or represents thatits use would not infringe privately owned rights. Referenceherein to any specific commercial product, process, orservice by trade name, trademark, manufacturer, orotherwise does not necessarily constitute or imply itsendorsement, recommendation, or favoring by the UnitedStates Government or any agency thereof. The views andopinions of authors expressed herein do not necessarilystate or reflect those of the United States Government orany agency thereof.

DISCLAIMER

Portions of this document may be illegiblein electronic image products. Images areproduced from the best available originaldocument.

b ABSTR4CT

While there is now available a great deal of itiorrnation on double stranded DNA horn

X-ray crystallography, high resolution NMR and computer modeling, very little is known about

structures that are representative of the DNA core of replication intermediates. DNA replication

occurs at a single strand/double strand junction and bulged out intermediates near the junction

can lead to frameshifl mutations. The single stranded domains are particularly challenging. Our

interest is focused on strategies for modeling the DNA of these types of replication

intermediates. Modeling such structures presents special problems in addressing the multiple

minimum problem and in treating the electrostatic component of the force field. We are testing a

number of search strategies for locating low energy structures of these types and we are also

investigating two different distance dependent dielectric functions in the coulombic term of the

force field. We are studying both unmodified DNA and DNA darnaged by aromatic arnines,

carcinogens present in the environment in tobacco smoke, barbecued meats and automobile

exhaust. The nature of the structure adopted by the carcinogen modified DNA at the replication

fork plays a key role in determining whether the carcinogen will cause a mutation during

replication that can initiate the carcinogenic process. In the present work results are presented

for unmodified DNA.

1

.

INTRODUCTION

DNA replication takes place at a fork containing DNA single strand-double strand

junctions, together with polymerase and a number of other proteins. Little is known about the

confirmational preferences of the DNA in such structures. In this work, we are modeling

intermediates representing the DNA core of the replication structures present at one arm of the

replication fork.

It has long been proposed that slippage mechanisms are responsible for frameshill

mutations such as double base (-2) deletions during replication of DNA (1,2,3). The sequence

5’G1GzC~GdC~Cb3’ is a NarI restriction enzyme recognition site (4,5) and contains three G’s

(G1G&JG4C~CJ. The sequence is a mutational hotspot for -2 deletions induced by chemical

carcinogens, notably 2-(acetyl) aminofluorene (WF). Covalent modification by &4.F induces -2

frameshift mutations only when the modification is at GJ (6) even though each of the G’s shows

equal reactivity towards the reactive precursor.

A specific slippage mechanism has been proposed to explain these double base

frameshifts at the NarI sequence (7). According to this model the GC double base deletion

results from a slippage mechanism during replication (8). In this model the replication apparatus

first incorporates a cytosine opposite the modified guanine. Then template-primer slippage

occurs and a misaligned intermediate forms wherein the two terminal bases in the primer (3 ‘-CG-

5’) hydrogen bond with a repeated downstream complementary 5’-GC-3’ dinucleotide in the

template. Continued synthesis from thk intermediate now containing an unpaired GC

dinucleotide leads to a complementary strand two bases shorter than the template strand. This

model predicts that -2 deletion mutagenesis should occur at the G~ but not the G1 or G~ positions,

because, given the sequence at the NarI site, only the former would allow formation of a

misaligned intermediate stabilized by two correct terminal base pairs.

Timsit et al. (9) have solved the crystal structure of a DNA dodecamer duplex containing

the NarI sequence. They found a novel and intriguing structural theme in the central NarI

sequence portion of the 12-mer which suggested an intrinsic propensity for the NarI Gdhotspot to

undergo slippage. Specifically, the G1.C1, base pair is disrupted; the partner Cl, is unstacked and

the pair is highly propeller twisted. In qddhion, the neighboring C~ .GIOis highly propeller

2

twisted and C~ also participates in a three centered hydrogen bond to Gg (from N4 of C~ to 06 of

G9).

The objective of the present work is to.model the DNA of replication intermediates using

the NarI sequence in unmodified DNA. Little experimentation has been done on the structures of

these intermediates, and how to search for plausible models computationally is a challenging

problem. In the present study we present results of an effort to model computationally such

intermediates, including tests of two different dielectric functions in the electrostatic component

of the force field. We are currently exploring various treatments for this function.

We studied the following intermediates which model a fragment of DNA undergoing

replication at one arm of a replication f~rk.

.x-

Intermediate I. 5’ G1 Gz C~ G,C~ C, A,3’

3’ G,0G,T85’

In Intermediate I replication has proceeded to the base 3‘ to the Gd hotspot.

Intermediate 11. 5’ G, G2 C~ G, C5 C, A, 3’

3’,

In Intermediate II the replication machinery incorporates the correct base cytosine opposite to the

G, hotspot, which may lead to a misaligned intermediate, shown as Intermediate III.

Intermediate 111. G, C,5’ G1 Gz Cq CG A,3’

3’ C,, G,O & T*5’

In Intermediate III a misaligned intermediate has formed where the terminal two bases of the

primer (3’ C11G105’) hydrogen bond with a repeated downstream complementary 5’G2 Cq3’ in the

template, from which extension gives rise to the double-base frarneshift mutation

(3’C,1G,0G,T85’).

3

●

METHODS

We employ the torsion angle space molecular mechanics program DUPLEX (1O) to

compute energy minimized structures of the replication intermediates.

A feature of DUPLEX that played a major role in the current work is the hydrogen bond

penalty function which provided an important tool for searching for the different types of

structures. This penalty fimction or pseudopotential energy, F, is added to the energy, and is

employed to locate minimum energy conformations of any designated hydrogen bonding pattern,

or a denatured site if the function is not employed at a given position. The penalty function has

the form.

F= Wi~~[(d-dO)2 +(l+cos(z))’+~)].,-

It can be applied to a chosen Donor-Hydrogen-Acceptor pair in Watson-Crick, Hoogsteen or

any other type of selected hydrogen bonding scheme. If the fiction is not applied to a given

base pair, a denatured site can result. W is an adjustable weight and values of W are in the range

of 5-50 (kcal/mole-A2) depending on n, the number of targeted hydrogen bonds. d is the value of

the current Donor-Acceptor distance and dOis the value of the ideal target distance. The current

angle around the hydrogen atom, the Donor-Hydrogen-Acceptor angle is ~ . p = \ c1 - C212where

c1 is a unit vector perpendicular to the plane of one base and Cz is a unit vector perpendicular to

the plane of its partner. In an ideal hydrogen bond, d equals dO,~ = 180°, c1 is parallel to c1, and

F, which ‘is summed over all n hydrogen bonds at all residues, equals O. The hydrogen bond ~

penalty function F is employed in the first stage. A second minimization is then carried out in

which the fimction is released so that final structures are unrestrained energy minima.

The search strategy employed was as follows: starting structures were an ideal B form

DNA (11) except that both anti (~=240°) and syn, (%=60°) orientations were employed for the Gd

mutational hotspot. The hydrogen bond penalty fimction was then employed in the following

way: at residues with normal Watson-Crick pairing we employed the fiction to seek out that

type of bonding. In bulged structures the function was employed only for the paired bases, and

the non-paired bases were not targeted for any hydrogen bonding. In Intermediate 11where there

is a C (Cll) opposite to the Go hotspot, five different hydrogen bonding schemes were sought for

the GA.C1lbase pair. Three of them are Watson-Crick (Gt anti), I-Ioogsteen (Gi syn), and Wedge

4

.

(G, syn) (Figure 1). The other two involved syn G, opposite anti Cl,, and anti G, opposite anti

Cl], with no hydrogen bonding between Gq and Cl,.

The energy minimizations were done in stages. The hydrogen bond penalty function

with weight of 15 was employed in all first stage minimizations to guide in the location of the

selected hydrogen bonding patterns. A second minimization was performed for each resulting

structure from the frost minimization without the hydrogen bond penalty function. Those

structures that employed protonated bases (Hoogsteen and Wedge hydrogen bonded structures,

Figure 1) were deprotonated and minimized again to permit comparison of their energies with

those of the unprotonated conformers (12). The energy minimized structures resulting from each

intermediate were ranked according to energy. It should be emphasized that the hydrogen

bonding scheme which is targeted by the hydrogen bond penalty function may or may not be

achieved during the stages of minimization, and a final unrestrained structure may have. a

different hydrogen bonding pattern than the one that was initially searched for.

DUPLEX employs a consistent force field for nucleic acids based on one devised in the

Olson laboratory (13, 14) which contains the usual Van der Wards, electrostatic and torsional

contributions to the energy, E,~, E~l, EtO,,as well as a number of special terms. These are needed

to properly model sugar conformations(15) and phosphodiester rotational preferences (16). The

original treatment for hydrogen bonding devised by Olson(17) is also employed.

E~l is an especially important term in the force field because electrostatics play a central

role in determining the molecular structure of the DNA. To model the interpenetration of

solvent as an atom pair separates, the dielectric constant is allowed to increase from its value

within DNA (believed to be in the range of 2 to 4 (13)). However, there is considerable

uncertainty as to the form the increasing dielectric function should take with increasing

interatomic distance. In the present work we employ two different treatments. In the first

treatment the E.l term is the standard coulomb potential, modified with an exponential screening

term P (13,14).

E,l = ZZ (332 qi~ e-srij) / s’rij Screened Coulomb’s Law

i<j

~ = ~l/e -Brij

We term this Dielectric Function A.

5

The choice for the value of ~ in the screening term of the dielectric function allows any selected

concentration of counterion to be represented (13). The term &’/e-Drij produces a distance

dependent dielectric function which increases as rti increases. ~ is assigned a value of 0.1 in the

present work which corresponds to screening by a monovalent salt concentration of about 0.1 M

(18,19). The choice of d, a variable parameter, governs the exact shape of the function. In the

current work we employ a value of 1 (Function A) on the basis of earlier experience in our

laboratory (20). In the second treatment we tested the dielectric function devised by Hingerty et

al. (21):

E(r) =78 - 77( r/2.5 )2 [er’2s/( er’25-1 )2]

-r We term this Dielectric Function B.

The dktance dependence of these two functions is given in Figure 2. In addition, reduced partial

charges on the pendant phosphate oxygens are employed to model charge neutralization by

counterion condensation (22).

For each of the replication intermediates (Intermediates I-III) energy minimization was

carried out using Functions A and B separately. The energy minimized structures obtained from

Function A were subsequently minimized again using Function B, and energy minimized

structures obtained from Function B were subsequently minimized again using Function A. In

addition, in several cases structures containing N-2-aminofluorene (AF) modification obtained

by energy minhnization searches or by computer graphics modeling (23) were used as starting

structures for another set of minimizations following removal of the AF. The total number of

trials for each intermediate is given in Figure 3.

Computations were carried out on a Cray C90 Supercomputer at the National Science

Foundation San Diego Supercomputer Center and on a Cray C90 at the Department of Energy

National Energy Research Supercomputer Center. The computer graphics program INSIGHT II

by Molecular Simulations Inc. was used for visualization.

6

.

RESULTSIntermediate I: 5’ G, G2 C~ GdC~ CG A, 3’

3’

..*

GIOGgT85’

Results are summarized in Tables 1 and 4. All minima obtained are included in Table 1

without an energy cutoff. The two most important structures for this intermediate (Table 1, #1,

Figure 4A and Table 1, #2, Figure 4B) are very similar. The double stranded region is normally

paired. The unpaired G1 is normal anti and stacked with the adjacent Cj.GIO base pair on its 3’

side; GAis unstacked on its 5’ side. There is a bend between Gi and C~. C~, G2 and Gl are nicely

stacked. tie single stranded region of the template strand interacts with residues of the duplex

region of the primer. The sugar-phosphate between G, and Gz interacts with the Gg and T8.-.

residues. Higher energy structures are summarized in Table 4 and illustrated in FigWe 4.

Intermediate 11: 5’ G1 G2 CJ Gd C~ CGA, 3’

3’

Results are summarized in Tables 2 and 5. Computed structures to 25 kcal/mol are

included in Table 2. The most important structure (Table 2, #l, Figure 5A) has the following

features: All four base pairs in the double stranded region are intact. The G1 hotspot is normal

anti and nicely stacked with the adjacent C~.GIObase pair on its 3‘ side. The C~ is not stacked

with the GOC,l base pair. The Gz and C~ in the single stranded region are nicely stacked with

each other but G, and Gz are not stacked, instead, G1is parallel to C, ~and partly stacked with it.

Higher energy structures are summarized in Table 5 and illustrated in Figure 5.

G, C5Intermediate 111: 5’ G1 G2 C~ Cb A, 3’

. .

3’ C1l GIO “ “Gg T* 5’

Results are summarized in Tables 3 and 6. All minima obtained are included in Table 3

without an energy cutoff.’l%e two most important structures for this intermediate (Table 3, #1,

Figures 6A and Table 3, #2, Figure 6B) have the following features:

7

G} is only partly stacked with Gz in both structures. Gz is properly paired with its partner

C,l. CJ is properly paired with GIOand the two base pairs are well stacked. In structure #1 Gt is

anti and protrudes into the major groove ; in structure #2 G1 is syn and also protrudes into the

major groove. The unpartnered C~ is stacked with Gg on one face; the other face is mainly

exposed. The Cb.G~and A,.Tg base pairs are normal. Stacking between the Cc.Gg base pair and

A7.T8 base pair is fair in both structures. The helix axis between the two base paired regions is

shifted in both cases, resulting in a bend. Descriptions of a number of higher energy structures

are summarized in Table 6 and illustrated in Figure 6. Figure 7 shows low energy structures of

Intermediates 1, H, and III

-DISCUSSION

Search Strategies: Reasonable sets of starting structures were employed for our energy

minimizations, includlng canonical B forms, structures built by computer graphics modeling and

structures obtained from carcinogen-DNA studies with removal of the carcinogen. In addition

various hydrogen bondhg possibilities were considered for Intermediate II at the GJ hotspot.

The syn as well as anti conformations were considered for Gd in all intermediates. Moreover,

structures found in the search employing Dielectric Function A were used as starting structures in

Dielectric Function B minimizations and vice versa. A total of 44 trials were made for the three

intermediates.

Dielectric Function: A gratifying aspect of our results is that the most important

structures for each category of intermediate were similar irrespective of which dielectric fimction

was employed in the searches, although the energy rankings of similar structures diverged

considerably in the higher energy forms. It is possible that different dielectric treatments might

be more suitable for differing structural types (normal duplexes, single strand/double strand

junctions, bulges) but it will be necessary to have experimental benchmarks in the same sequence

and length context available to permit such discrimination. Moreover other dielectric function

variants may need to be tested (13,24,25,26).

8

Relevance of Higher Energy Structures: We have included among our results

structures of fairly high energy. We feel that in the presence of relevant proteins including

polymerases, single strand binding proteins, topoisomerases and others in normal replication, and

repair enzymes with modified DNA, structures that are higher energy in their absence could

become favored. Examples of protein induced alterations to DNA structure are accumulating in

the crystallographic literature (27,28). We note that among our structures of Intermediate H , for

example, the structure most plausible for normal replication with normal alignment of the

template is not the lowest energy form but has an energy of about 16 kcal/mol. In addition,

minima for the small fragments treated here may be more or less favored at an authentic

replication fork where the DNA is essentially infinitely long.

. .

Bending. An interesting feature of all the important structures of the intermediates

modeled in this work is that the DNA is bent. In the cases of Intermediates I and II there is a

strong bend near the next to be replicated templating base. This is in line with a model based on

the crystal structure of human DNA polymerase ~ complexed with DNA (29) which contains a

90° bend in the single stranded template. It was proposed by Pelletier et al. (29) that such a bend

of the template at the active site could increase the enzyme’s sensitivity to mismatch prior to.

catalysis, by minimizing &e importance of indiscriminate base stacking between the primer

terminus and an incoming nucleotide until a check has been made against the template for proper

base pairing . In the case of the bulged Interinediate III, bending is known to be associated with

bulged structures (30-41) although the biological significance, if any, is not known.

Bulged Intermediates versus Normal Extension: Replication Fidelity and

Carcinogen Modification: Comparison of the energies of the G4 anti bulged structure

(Intermediate HI) versus its G, anti normal counterpart (Intermediate II) shows that the unbulged ~

form (Intermediate H) is strongly favored, as would be expected to maintain ftithful replication

even in sequences such as iVarI which accommodate slippage to form a bulge (Table 7).

However, such slipped structures can become stabilized in the presence of a covalently bound

carcinogen (42-47). on the other hand when the Gq is syn the energies of Intermediates II and III

are very close (Table 7). This suggests that if a carcinogen like AAF which induces the syn

9

.

conformation were covalently bound to GAthe slipped Intermediate 111may be much closer in

energy to the normally extended one than when the Gq is anti.

CONCLUSION

Our modeling studies of, replication interrnediates for the NarI mutational hotspot

sequence of E. coli has yielded some plausible structures that are in line with faithful replication

at the G~ template when this base is undamaged by carcinogen. On the other hand when the GAis

syn the energies of normally extended Intermediate 11 and slipped Intermediate III are much

closer than when GAis anti. This suggests that if a carcinogen like AAF which induces the syn

conformation were covalently bound to the Gd hotspot the slipped Intermediate 111may be close

-m energy to the normally extended one.

ACKNOWLEDGEMENTS

This research is supported by NIH Grants CA 75449, CA 28038

Grant DE-FG02-90ER60931 to S.Broyde and R. Shapiro, and DOE

960R22464 with Lockheed Martin Energy Research to B.E. Hingerty.

RR-06458 and DOE

Contract DE-AC05-

10

REFERENCES1. Streisinger, G., Okada, Y., Ernrich, J., Newton, J., Tsugita, A., Terzaghi, E., and Inouye, M.

Cold Spring Harbor Symp. Quant. Biol. 31,77-84 (1966).2. Masurekar, M., Kreuzer, K. N., and Ripley, L. S. Genetics 127,453-462 (1991).3. Kaiser, V. L., and Ripley, L. S. %oc. Nat]. Acad. Sci. USA. 92,2234-2238 (1995).4. Fuchs, R.P.P., Schwartz, N. and Daune, M. Nature 294,657-659 (198 1).5. Koffel-Schwartz. N., Verdier. J.M., Bichara, M., Freund. A.M., Daune, M. and Fuchs R.P.P.

J Mol. Biol. 177,33-51 (1984).

6. Burnouf, D., Koehl, P. and Fuchs, R. P. P. Proc. Natl. Acad Sci. USA. 86,4147-4151(1989).

7.

8.9.

10.-11.12.

Sch&per, R. M., Koffel-Schwartz, N. and Fuchs R. P. P. Carcinogenesis (Lend.)1 l,10S7-1095 (1990).Kunkel, T. A. Biodzemisz?y 29,8003 (1990).Timsit, Y., Westhof, E., Fuchs, R. P. P., Moras, D. Nature 341,459-462 (1989).

Hingerty, B.E., Figuero~ S., Hayden, T. and Broyde, S. BiopoZymers 28, 1195-1222 (1989).Arnott, S., Dover, S., and Wonacott, A. Acts Cryst. B25, 2192 (1969).Shapiro, R., Hingerty, B. E. and Broyde, S. J Biomol Struct. Dyn. 7,493-513 (1989).

13. Olson, W. K. and Srinivasan, A. R. in Landolt-Bornstein Numerical Data and FunctionalRelationships in Science and Technology, Group VII. Biophysics. Volume 1D, W. Saenger,Editor, Springer Verlag, Berlin, pp. 415-435 (1990).

14. Srinivasan, A. R. and Olson, W. K. Fed. Proc., Fed. Am. Sot. Exp. Biol. 39,2199 (1980).15. Olson, W. J. Am. Chem. Sot. 104,270-286 (1982).16. Srinivasan, A. R., Yathin@ N., Rae, V.S.R. and Trakash, S. Biopol. 19, 165-191 (1980).17. Olson, W. Biopolymers 17, 1015-1040 (1978).

18. Fenley, M.O., Manning, G.S. and Olson, W. K. Biopolymers 30, 1191-1203 (1990).19. Fenley, M. O., Olson, W. K., Tobias, I. and Manning, G.S. Biophysical Chemis@ 50,255-

271 (1994).20. Shapiro, R., Ellis, S. and Broyde, S. manuscript in preparation.21. Hingerty, B.E., Ritchie, R. H., Ferrell, T. L. and Turner, J. E. Biopolymers 24,427-439

(1985).22. Manning. G.S. Quart. Rev. Biophys. 11, 179-246 (1978).23. Roy, D., Hingerty, B. E. and Broyde, S. unpublished.24. Olson, Wilma K. and %-inivavasan, A. R. Computer Science Information Project e-book

Oak Ridge National Laboratories (1997)25. Ramstein, J., Lavery, R. Proc. NatL Acad. USA. 85,723 1-7235(1988).26. Lavery, R. Structure and expression, Volume 3, DNA Bending and Curvature, Olson, W. K.,

Sarma, M. H., Sarma, R. H., Sundaralingam, M., eds., Adenine Press, Schenectady, NZ pp.

27.

28.

29.

30.31.

191-211 (1988).Dickerson, R. E. Oxford University Press, Oxford (S. Neidle, editor) in press (1997).

Berman, H. M., W. K. Olson, D. L. Beveridge, J. Westbrook, A. R. Srinivasan, and B.Schneider. Biophys J. 63, 751-759(1992).Pelletier, H., Saway~ M. R., WoMe, W., Wilson, S. H. and Kraut, J. Biochemistry 35,

12742-12761 (1996).Woodson, S. A., and Crothers, D. M. Biochemistry 27,3130-3141 (1988).

Hsie~ C. H. and Grifl%h, J. D. Proc. Natl. Acad. USA. 86,4833-4837 (1989).

11

32.33.34.35.36.37.38.39.

40.41,

Rice, J. A. and Crothers, D. M. Biochemistry 28,4512-4516 (1989).Bhattacharyy~ A. and Lilley, D. M. J. Nucleic Acids Res. 17,6821-6840 (1989).Bhattacharyya, A. Murchie, A. I. H. and Lilley, D. M. J. Nature 343,484-487 (1990).Tang, R. S. and Draper, D. E. Biochemistry 29,5232-5237 (1990).Wang, Y. H. and Griffith, J. Biochemistry 30, 1358-1363 (1991).

Rosen, M. A., Shapiro, L. and Patel,D.J.Biochemis@31, 4015-4026 (1992).Wang, Y. H., Baker, P. and Griffith, J. J. BioL them. 267,4911-4915 (1992).

Gohlke, C., Murchie, A. I. H., Lilley, D. M. and Clegg, R. M. Proc. Natl. Acad. USA.91,11660-11664 (1994).Zacharias, M. and Hagerman, P. J. J Mol. Biol. 247,486-500 (1995).Lilley, D. M. J. Proc. Natl. Acad. USA. 92,7140-7142 (1995).

42. Garci~ A., Lambert, I. B. and Fuchs, R. P. P. Proc. Natl. Acad. Sci. USA. 90,5989-5993

(1993).,

43. Milhe’, C., Dhalluin, C., Fuchs R. P. P. and Lefevre, J. F. Nucleic Acids Res. 22,4646-4652

(1994).

“44. Milhe’, C., Fuchs R. P. P. and Lefevre, J. F. Eur J Biochem. 235,120-127 (1996).45. Y% N. Q., Smimov, S., Cosman, M., Bhanot, S., Ibanez, V., and Geacintov, N. E. Structural

Biolo~: The State of the Art, Proceedings of the Eighth Conversation in the DisciplineBiomoiecuIar Stereodynamics, State Universi~ of New York at Albany, June 22-26, 1993(Sarma R. H., Sarma. M., Eds.) Vol. 2. pp 349-366, Adenine Press, Schenectady, NY. (1994),

46. Mao, B., Cosman, M., Hingerty, B.E., Broyde, B., Patel, J. P. Biochemistry 34,6226-

6238 (1995).47. Mao, B., Flingerty, B. E., Broyde, B., Patel, J. P. Biochemistry 34, 1664’ -16653 (1995).

12

TABLE 1Energy minimized conformations of Intermediate ~ 5’ G1 G2 CJ G4C5 C6 A, 3’

3’. . .

GIOGgT85’

Dielectric Function AMk# x Domain Figure #

(KC:v:ol)1 0.0 anti 4A, 7A

2 2.2 anti 4B.-3 5.1 Syn 4C4 10.5 syn 4D5 11.3 syn 4E

Dielectric Function BMin# x Domain Figure #

(KcaYmol)1 0.0 anti 4B2 7.4 anti b

3 17.3 Syn 4E4 21.2 Syn 4C5 25.3 Syn 4D

‘All computed structures shown. Structures with the same figure number are virtually identical.‘TJot shown, structure is a somewhat distorted variant of 4A.

13

TABLE 2Energy minimized conformations of Intermediate 11’ 5’ G, Gz C~ Gd C~ CG& 3’

. . . .

3’ Cll GIOGyT~ 5’

Dielectric Function AMin# x Domain Figure #

(Kc~UEmol)1 0.0 anti 5A, 7B

.< 2 16.2 anti 5B

3 22.4 anti 5C

Dielectric Function BMin# AE x Domain Figure #

(Kcal/mol)1 0.0 anti 5A, 7B2 16.4 anti 5B

3 24.5 syn 5D

‘Computedidentical.

structures to 25 kcal/mol shown. Structures with the same figure number are virtually

14

TABLE 3 GdC,

Energy minimized conformations of Intermediate 11~ 5’ GI G2 C~ Cb A, 3’. . . .

3’ Cll GIO Gg Tg 5’

Dielectric Function A

Min# AE x Domain Figure #(Kcal/mol)

1 0.0 anti 6A, 7C

2 0.2 Syn 613

3 12.3 syn 6C

4 18.1 anti 6D

5 19.5 anti 6E

Dielectric Function B

Min# AE x Domain Figure.#(Kcal/mol)

1 0.0 anti b

2 5.9 syn 6C

3 9.7 syn 6B

4 23.4 anti 6D

5 26.1 anti 6E

‘All computed structures shown. Structures with the same figure number are virtually identical.!Not shown, structure is a somewhat distorted variant of 6A.

15

TABLE 4Structure characterization of Intermediate I :5’ GI G2 C~ G~C~ CG A, 3’

3’. . .

G,OGg Tg 5’

~igure ,Xof Gd Hydrogen A7.T~ Pair Single Stranded Region Stacking in# Bonding Single

in Double StrandedStranded Region Region

4A Anti All three are Normal Sharp bend between GJ and CJ CJ,G2 and G,intact are stacked

4B Anti All three are Normal Sharp bend between Gd and C~ C~ ,G2and G,intact are stacked

4C Syn All three are Normal U-turn between G1 and C~ C3 ,Gz and G1intact are stacked

4D Syn All three are Normal U-turn between C~ and Gq C~ ,Gz and G,intact are stacked

4E Syn Two are intact Disrupted Sharp bend between Gq and C~ Gz and G, arestacked

16

TABLE 5Structure characterization of Intermediate II :5’ GI G* CJ GdC~ Cb A, 3’

. . . .

3’ C1lGIOG9T8 5’

‘igmx#

5A

5B

5C

5D

Anti

Anti

Anti

Syn

Iydrogen BondingrnDouble Stranded

Region

All four are intact

All four are intact

All four are intact

“Ihree are intact

Normal

Normal

Normal

Only onehydrogen

bond isformed

stacking

No stacking

Stacking

No stacking

SingleStranded

Region

Sharp bendbetween G,

andC~

Sharp bendbetween Gq and

C3

Shallow bendbetween Gd and

c,

Sharp bendbetween Gi and

C3

31is parallel to C1,md stacked with it;

C~ and Gzarestacked

C~, Gzand G1arestacked

G2and G1are notstacked; C~ and Gz

are stacked

Cg,Gzand G1arestacked. G~is

parallel to Cll butnot stacked with it

17

.

TABLE 6 G, C,Structure characterization of Intermediate III: 5' GlG2 CJ

?igure til Stacking G2.C1l C3”G3 Unpartnered Gq Unpartnered C~ Cb.Gg A7.T8# with GI Pair Pair Pair Pair

6A Poor Normal Normal Anti One face stacked Normal NormalMajor groove with Gg

6B Poor Normal Normal Syn One face Normal Normal

Major groove stacked with Gg

6C Partly Normal Normal Syn Stacked with Distorted DistortedStacked Major groove CG.G9

6D None Normal Normal Anti, Stacked with C~ Stacked with Normal DistortedCGand G,

6E Poor Normal Normal Anti, Stacked with C~ Stacked with Gd Normal Distorted

18

.

TABLE7 Energies of Intemediate II@xtended) versus Intermediate III@ulged)in lowest energy anti and syn conformations

Energies (kcalhnol)’

Intermediate II Intermediate HI Intermediate II Intermediate IIIAnti Anti SYn Syn

Dielectric Function A -327.5 (0.0) -295.6 (0.0) -294.0 (33.5) -295.~ (0.2)Dielectric Function B -281.7 (0.0) -~55.o (().0) -257.2 (24.5) -249.1 (5.9)

-..

aAbsolute energies are followed by AE, the energy of the given structure relative to thelowest energy form forthat intermediate, inparentheses. See Tables 2and3.

19

FIGURE

Figure 1.(a)

Figure 2.

Figure 3.

Figure 4.

Figure 6.

Figure 7

LEGEND

Protonated hydrogen bonding patterns investigated.Hoogsteen and (b) Wedge

Dielectric functions studied.

Total number of trials employed for energy minimization.

Energy minimized conformations of Intermediate I (See Tables #1 and #4).Stereo views.

Energy minimized conformations of Intermediate II (See Tables #2 and #5).Stereo views.

Energy minimized conformations of Intermediate III (See Tables #3 and #6).Stereo views.

Space-filling models for low energy conformations ofA. Intermediate I B. Intermediate 11 C. Intermediate III (See Tables 1-3)Stereo views.Color code: template strand - yellow, primer strand - blue; Gt hotspot - red;

unpaired C~ - green (Intermediate III)

20

.

CYTeSLmE WANME

(al

Susan

CTTOSSME

SIJsAa

CUANIRE

SUGAR

Cx●

SU6AR

(b)

.

.

/,/

Dielectric Functions

.~.~-x-x-x-x”x-x-x-x”x-~”~-------

70 .

60 .

50 .

&413 ..

30 .

20

..z-

10

0

\

UAS W@ law“w! UAS UeeJs “qq l]UW

UAs J]uv ON 4)6POM ‘600H ON ~“m UAS jylv

,

A.

B.

c.

. D.

Ii w

A.

B.

c.

D.

E

A.

B.

c.

ew-

FD.

= w

A.

B.

c.

FD. v

-.

.-

A.

B.

A.

B.

c. c.

D.

%

E.

D.

E.

?

?

r-

-ii