312 Biochimica et Btoplivsi¢'a .,1 eta 838 ( 1985 ) 312 32(t F, Ise~ier

BBA 21976

The circular dichroism properties of ~W-14 DNA containing a-putrescinylthymine

Susan Spetter a, Catherine Chen a, R.A.J. Warren ~ and Sue Hanlon "*

" Department of Biological ChemistQ,, College of Medicine, Universi(v of Illinois, Chicago, IL 60612 (U. S.A .), and b Department of Microbiology, University of British Columbia, Vancouver, B.C., V6 T 1 W5 (Canada)

(Received August 31st, 1984)

Key words: Circular dichroism; DNA structure; Putrescyhhymine: (Phage ,#W-14)

The circular dichroism properties of q~W-14 DNA containing a-putrescinylthymine and its acetylated derivative have been examined in a number of aqueous solvents. Native q~W-14 DNA exhibits a B-type CD spectrum whose characteristics do not entirely conform to what would be expected for its GC content (51%). The conformationally sensitive positive band above 260 nm has a rotational strength greater than that normally found in prokaryotic DNAs of comparable GC content, such as Escherichia coil DNA. The rotational strength of this band in the spectrum of the heat-denatured form of q~W-14 DNA, however, is similar to that of heat denatured E. coil D N A . Abolition of the positive charge on the putrescine residues of native q~W-14 DNA by reaction with C H 2 0 or by acetylatiun reduces the rotational strength to a level appropriate for its GC content. Increases in the electrolyte content of the solvent have the same effect, although the rotational strenght of this band in q~W-14 DNA does not become comparable to that of E. coli

DNA until 6 - 7 M LiCI. Titration to pH 10.6 in solvents of modest electrolyte content, however, fails to appreciably affect the CD spectral properties of either native q~W-14 DNA or the derivative in which half of the secondary and all of the primary amino groups have been acetylated. On the basis of these results we have concluded that the enhanced rotational strength of the positive band above 260 nm in the CD spectrum of q,W-14 D N A is due to a conformational difference caused by an ion-pair interaction of the positively charged primary amino groups of putrescine with the phosphate backbone. The CD spectral properties, however, reveal that these differences, averaged over the entire basepair population, appear to be relatively small. The average conformation, at least in dilute aqueous solutions, seems to be an unexceptional B variant with conformational properties which would be more appropriate for a DNA of higher CG content.

In tro duc t ion pyrimidine, a-putrescinylthymine (putThy) [1]:

The DNA of the bacteriophage, ~W-14, con- H O

\N ~ tains a unique substitution in one of its bases. L--_CH 2 _ . Approximately half of the thymine residues of this O ~ N ~ D N A are replaced with the hypermodified I

H

* To whom correspondence should be addressed. Abbrevation: putThy, a-putrescinylthymine.

H I N-(CH2)4-NH2

PutThy is synthesized at the polynucleotide level [2,3]. It is distributed non-randomly in ~W-14

0304-4165/85/$03.30 © 1985 Elsevier Science Publishers B.V.

DNA because the sequence purine-putThy-purine occurs more than twice as frequently as the se- quence, purine-thymine-purine [4]. The presence of the amine favors a very high packing density, which permits a molecule whose size is comparable to that of T4 DNA to be packaged within a head having only 72% of the volume of the T4 head [5].

Because of the hyper modified base, ~W-14 DNA manifests a number of unusual properties in solution. Although its base composition has been chemically determined to be 51% G + C (with the remaining 49% in A + putT + T) [3], its T m in the standard saline citrate solvent is considerably elevated above what would be anticipated for 51% G + C. By the same token, its buoyant density in CsC1 is much lower than would normally be ex- pected for this GC content. Both of these effects are directly attributable to the covalently bound putrescine. Putrescine's lower density reduces the density of the average nucleotide in ~W-14 DNA and its positive charges, as long as they are not ion paired within the DNA structure, should favor greater hydration in concentrated CsC1, thus accounting for the lower buoyant density of the DNA in that solvent. In more dilute electrolyte solvents, the positively charged putrescine residues would be expected to affect the T m in the observed manner, by any one of several mechanisms appli- cable to DNA as a polyelectrolyte [6-8], all of which would result in an increase in T m in ~W-14 DNA.

The effect on the packing density, however, is not as easy to rationalize. In principle, one might anticipate higher densities of fiber forms for DNA whose negative charge is partly neutralized by an internally bound cation. Our recent studies, how- ever, on DNA-monoamine complexes prepared by covalently coupling n-butylamine to exocyclic amino groups of calf thymus DNA indicate that the center-to-center distance (as measured by the equatorial reflection) for DNA amine complexes in B-form fibers at 98% r.h. is only marginally smaller than that of the amine-free controls. Fur- thermore, there seems to be no systematic trend as the content of the amine in the calf thymus DNA complexes increases from 0.06 to 0.14 mol amine/mol nucleotide [9]. It thus seems unlikely that a putrescine content of 0.11 mol /mol nucleo- tide or a positive charge content at pH 7 of 0.22

313

mol /mol nucleotide should, by itself, give rise to such a dramatic increase in packing density. An increase in density could result, however, if the putrescine was ion paired to DNA phosphates of the backbone. Alternatively, the conformation of the @W-14 DNA in aqueous solution may be ex- ceptional and perhaps would permit a closer center-to-center approach (such as observed for the D form of DNA [10]) a n d / o r a shorter length at the level of hydration of the DNA in the phage head. This current study was undertaken in order to evaluate these possibilities.

Materials and Methods

The ~W-14 DNA was prepared as described previously [1]. It was acetylated with acetic anhydride and the extent and partitioning of acetyl residues between the primary and secondary amino groups of putrescine was determined as described in Gerhard and Warren [11]. (The specific deriva- tive employed in these studies had 100% of the primary and 40-50% of the secondary amino groups modified.) E. coli DNA was a commercial preparation obtained from Cal-Biochem. It was further purified by ethanol precipitation from a 2 M NaC1 solution followed by exhaustive dialysis against 20 mM NaC1 (pH 7).

The concentration of DNA in all solutions was determined spectrophotometrically from absorb- ances at the maximum (258.5 nm) and the extinc- tion coefficients at the same wavelength. The latter, in units of M-1 . cm-1 were determined for several preparations using the phosphate determination of Ames and Dubin [12]. The values for ~W-14, acetylated q~W-14 and E. coli DNA were 6700, 7200 and 6454 M -1 .cm -1, respectively. The agreement with the previous reported value of 6800 M -1. cm -~ for ~W-14 DNA [1] and 6500 M -1- cm -1 for E. coli DNA [13] was quite good.

Absorption spectra were obtained with a Cary Model 14 CMR at 1.0 cm path lengths over the wavelength range of 210 to 400 nm. Except for the solution of ~W-14 DNA in 500 mM NaC1 which had been titrated from pH 10.6 to 7, there was no spectral evidence of light scattering due to aggre- gation in any of the solutions examined in this study.

The thermal melting profiles were obtained in

314

the manner previously described [14] on solutions which had been exhaustively dialyzed against 1 mM NaCI (pH 7). Dialysis fluid was used as the reference solvent, and both sample and reference compartments were heated to the same tempera- ture. The temperature of the sample cell contents was measured by a Yellow Springs Bridge and thermistor assembly. Values of the absorbance, A~, at a given temperature (T) and wavelength (X) were corrected for thermal expansion effects by multiplying by the factor Pw,25/Pw,7, where p,,. rep- resents the density of water at 25°C and tempera- ture T, respectively. The hyperchromic increase upon melting is defined as the ratio of the cor- rected value of a~°°c ,~:5°('

~ ' 2 5 9 t o ' ' 2 5 9 •

Circular dichroism spectra were obtained at 27°C with a Cary 60 spectropolarimeter equipped with a 6001 CD unit, in the manner previously described [15]. Path lengths were 1000 cm and concentrations of DNA ranged between 0.1 and 0.27 mM with no concentration dependency noted in this range. All spectral data are reported in mean residue ellipticity, [O]x, in deg-cm2/dmol , at wavelength k (in nanometers).

The pH of all solutions were determined with a Corning pH meter and a combination electrode small enough to permit measurements in the spec- trophotometer cells. The titrations of q,W-14 DNA samples to pH 10.6 were conducted as rapidly as possible in these cells in order to avoid the alkaline degradation known to ensue at pH 11 [16].

All salts employed in these experiments were reagent grade products. The bovine serum albumin and pronase preparations (nuclease free) used to check protein contamination were products of Sigma and Cal Biochem, respectively.

Results

In order to assess the conformational effects of the putrescine attachment to DNA, the CD spec- tra should be compared to that of a putrescine-free D N A of identical sequence under the same solvent conditions. Such a control was unavailable and hence we selected a procaryotic DNA of compara- ble GC base content to serve in that capacity. The CD spectra of procaryotic DNA's show a good linear correlation with the average base composi- tion [17] as long as unmodified bases predominate.

This is attributable to the fact that CD spectra of nonrepetitive DNA can be adequately accounted for as a linear combination of first neighbor con- tributions [18]. In most unique sequence pro- caryotic DNA species, the latter are more or less a linear function of the overall base composition. We thus selected E. colt DNA whose unmodified base content (50% GC; [19]) is comparable to that of ,fW-14 DNA to serve in this capacity. The CD spectrum of E. colt DNA (beaded curve) is shown together with the spectrum of 4,W-14 DNA (solid line curve) in Fig. 1. As is obvious, the spectral characteristics of ~W-14 DNA differ significantly from those of E. colt all through the spectral region. The enhancement in the rotational strength of the conformationally sensitive positive band above 258 nm in ~W-14 DNA, as well as the spectral effects seen below this wavelength are more comparable to the CD properties of DNA of a higher GC content, such as, Micrococcus lysodeikticus (shown as the dashed line), whose GC content is 72% [19]. The effect on the positive band above 258 nm is especially interesting, since

I0 I i : i

5 o ol _~ o ~

0 011

o c,~ oo ~ o o o o ° o

o L ° I - -

o ; - o ° E

..of '~o j :,,/"-'h',, ° '

o " o

_,0

2 5 25 0 275 500

X (nm)

Fig. l . Comparison of the C D spectrum of native ~pW-14 D N A with that o f other native prokaryot ic D N A s in 20 m M NaCI (pH 7), 27°C. - - , ,#W-14, 51% GC; O O O O , E. coil, 50% GC; - , M . l v s o d e i k t i c u s , 72% GC.

315

the rotational strength of this band is particularly sensitive to changes in winding angle, tilt, twist and other helical parameters [20]. An elevated rotational strength of the magnitude observed for q~W-14 DNA is normally found only at substan- tially elevated temperatures (approx. 60°C) in the premelting region for this GC content [21,14]. All three CD spectra, however, are of the conservative type typically assigned to conformational variants of DNA with B backbone geometry.

Apart from differences in secondary structure, however, the differences observed between the CD spectra of ~W-14 and E. coli DNA could be due to (1) variations in protein content, (2) differences between the intrinsic chromophore properties of thymine and putThy or (3) changes in the chromo- phore properties of other bases in the vicinity of putThy due to the interaction with the positive charges of the diamine. Effects of protein con- tamination could be ruled out by subjecting q~W-14 DNA to prolonged pronase digestion. Under con- ditions where the negative rotational strength of the peptide band at 222 nm in the spectrum of an aqueous solution of bovine serum albumin was completely abolished by proteolysis, the negative bands in the same region in the spectrum of q~W-14 DNA as well as the rotational strength of the major positive band were unaffected.

Substitution of putrescine might shift the ab- sorption spectrum of thymine to the red and simi- lar effects should be observed if a positive charge is placed in the vicinity of a base. One would not anticipate, however, that either of these would have a dramatic effect on the rotational strength of thymine or other bases which would, in turn, lead to differences in the CD spectral properties of ,~W-14 and E. coli DNA of the type seen in Fig. 1. It is also unlikely that a specific interaction of the positively charged primary amino group of the diamine with other bases in the near vicinity of putThy would alter the intrinsic chromophore properties in such a way as to lead to a greater contribution to the rotational strength of the mononucleotide. Such charge interactions should produce electronic effects equivalent to those which arise in protonation of the bases. The most likely base interaction sites accessible to the the putres- cine in the major groove are the N-7 positions of the purines. The predicted spectral changes due to

the - N H 3 . . . N~ / interaction should resemble those seen when guanosine is transferred from a pH 7 to a pH 1 medium, since the predominant protona- tion site for that base is the N-7 position. These effects include a shift of approx. 3 to 5 nm to the red with an accompanying decrease in the oscilla- tor strength throughout the wavelength range of 230 to 310 nm [19]. This leads to a small red shift and a decrease in the rotational strength of the affected monomer. The spectral effects observed for native q,W 14 DNA are the reverse. This sug- gests that the enhanced rotational strength of the positive band is due primarily to differences in secondary structure.

In order to demonstrate this point experimen- tally, we examined the CD spectral properties of q~W-14 and E. coli DNA under denaturing condi- tions where most - if not all - of their organized regular secondary structure has been lost, but the potential for base-specific or nonspecific charge interactions remained. Solutions of each of these DNA's and a derivative of q,W-14 DNA with all of the primary and approx. 40 to 50% of the sec- ondary amino group modified were dialyzed into 1 mM NaC1 (pH 7), and melting profiles were ob- tained in this solvent. Cooperative melting was observed with T m values of 47.7, 55.4 and 64.4°C for E. coli, acetylated q~W-14 and q~W-14 DNA, respectively, and it was established that at 85°C the melting process was complete for all three. Separate aliquots of each DNA were then heated to 85°C for 5 min, quenched by plunging into an ice/water bath and examined immediately in the CD spectrometer at 27°C. The spectra of these denatured samples are shown in Fig. 2 as the solid line for q~W-14 DNA, beaded for E. coli DNA and dashed for acetylated q~W-14 DNA. The dif- ferences in the rotational stength of the positive bands above 260 nm of q,W-14 and E. coli DNA have been substantially reduced by converting both DNAs to the disordered conformation. Since the amine-base interactions should still be possible in the denatured form of q~W-14, the decrease in spectral differences indicated that these interac- tions, if they exist, are not contributing signifi- cantly to differences in the chromophore prepara- tions of the monomers. This is confirmed by the close correspondence of the spectral properties of the denatured q~W-14 DNA and its acetylated de-

316

I0

5 E o

N .

u

c~ 0

,'7 ! 0 !

t ' , ,,~ ~\ ?f , iLo o\

/ N \ %

%0 2

-I0

225 250 225 500 ), (n m)

Fig. 2. Comparison of the CD spectra of denatured DNAs in 20 mM NaCI (pH 7), 27°C. - - , q~W-14; 0 0 ( 3 , E. colt;

, acetylated 4~W-14.

rivative, which totally lacks the capacity to engage in specific base-primary amino interactions. Above 258 nm, its spectral properties are indistinguisha- ble from that of the melted unacetylated ~W-14 DNA. Hence we conclude that the spectral dif- ferences observed in the rotational strength of the positive band in the native DNA species reflect mainly differences in secondary structural proper- ties. Significant differences in the CD spectra of the denatured forms are still apparent below 250 nm. This region of the spectrum may be more sensitive to either a sequence effect, differences in electronic and magnetic effects or differences in intra strand base-sugar interactions due to the substituion of putThy for thymine.

The differences in the melting temperatures of q~W-14 and E. coli DNA in the same solvent reveal that the putrescine residues, as expected, are posi- tively charged. When this positive charge is abolished, the rotational strength of the positive band above 260 nm becomes comparable to that of E. coli DNA. This point is illustrated in Fig. 3, which compares the spectra of native q~W-14 DNA (solid line) and E. coli DNA (open circles) with spectra of 4,W-14 DNA derivatives in which the putrescine groups are predominantly uncharged.

I0

g

0 E u

ID

? _o_ 5

Z

-I0

i

i

/ $ ;I ~.'\°o /~"

eQe 0 ~

260 360 ), (rim)

Fig. 3. Effects on the CD spectrum of abolishing the positive charges on the putrescine residues of ~W-14 DNA. - - q~W-14 DNA in 20 mM NaC1 (pH 7); ~ , ~W-14 DNA in 20 mM NaCI, 2% CH20 (pH 7); - , acelylated 4,W-14 DNA in 20 mM NaC1 (pH 7); O O O , E. coil DNA in 20 mM NaCI (pH 7).

The closed circles represent the spectrum of ~W-14 DNA in 2% CH20, whose addition to putrescine should result in the loss of the protons from both the secondary and primary amine groups. CH20 is expected to add predominantly to the primary and secondary amine groups of putrescine in the reac- tion:

H H +1 +1

U r a - C H 2 - N - ( C H 2 ) 4 - N - H + 2CH20 I I

H H

~- Ura-CH 2 - N - ( C H 2 ) 4 - - N - H + 2H + I I

CHzOH CH2OH

(The buffer capacity absorbs the liberated protons, thus maintaining the pH of the solution at 7). Under the solvent conditions employed in this experiment, base unstacking is minimal (as as- sessed by the absence of hyperchromicity in the absorbance spectrum) and the CD spectrum of E. coli DNA under the same solvent conditions for the same period of time is unaffected.

The other spectrum (dashed line) shown in Fig. 3 is the acetylated derivative of ~W-14 DNA in its native form. (The CD spectrum of its denatured form has been displayed in Fig. 2). The spectrum of the native acetylated derivative shows a rota- tional strength of the positive band above 260 nm comparable to that found for the spectrum of ~W-14 DNA in 2% CH20 and for native E. coli DNA. Thus, the secondary structural effects giving rise to the CD spectral differences above 260 nm between ~W-14 DNA compared to its uncharged derivatives are directly attributable to the positive charge on the putrescines rather than the presence of a bulky substituent in the major groove of the DNA duplex. It might be noted, furthermore, that in the acetyiated derivative, a fraction of the sec- ondary amines (approx. 60-50%) of the putrescine residues are still, in principle, positively charged. These experimental results thus demonstrate that the conformational effect is due mainly to the positive charge of the free primary amino terminus of the putrescine residues.

A third method of diminishing the positive charge on the putrescine residues failed to yield the expected results. The intrinsic pK a values of putrescine should be approx. 10.5 at room temper- ature. Because of electrostatic effects, however, two pKa values are observed [22] which, corrected to 27°C, should be 9.2 and 10.6 [23]. When at- tached to thymine in native duplex DNA, one would anticipate some elevation of these values, due to the negatively charged phosphate environ- ment. In the absence of specific site interaction, however, the higher of the two pK a values should not exceed 11, especially in a solvent with a sub- stantial concentration of neutral electrolyte. In- deed, we have observed behavior of n-butylamine, attached as a secondary amine to one of the DNA bases, which is interpretable in terms of a pK~ of approx. 10.5 in 20 mM NaC1, 10 mM sodium phosphate buffer (pH 7) [24].

With the rationale in mind, we obtained the CD spectra of ~W-14 DNA and its acetylated deriva- tive as a function of pH over the range in which the hypochromism of the molecules was not appre- ciably altered, and the alkaline degradation known to ensue above pH 11 was not appreciable [16]. Fig. 4 shows the result of that experiment. Panels A and B display the effects of pH between 7 and

317

Ell

"E

,'_o

A B C

260 70 80 90 300 260 70 80 90 300 260 70 80 90 300

), (rim}

Fig. 4. Effects of increasing pH on the CD spectra of ~W-14 DNA and its acetylated derivative. (A) ~W-14 DNA in 20 mM NaC1 at pH 6.6-9.3 (curve 1), pH 10.2 (curve 2) and pH 10.6 (curve 3). The dashed curve was obtained upon titration back to pH 6.8 (B) ~W-14 DNA in 500 mM NaCI at pH 6.4 to 7.5 (curve 1), pH 9.3 (curve 2), pH 10.0 (curve 3) and pH 10.6 (curve 4). (C) Acetylated ,~W-14 DNA in 40 mM NaC1 at pH 7.3 (curve 1), pH 8.1 to 9.4 (curve 2) and pH 10.5 (curve 3). The dashed curve was obtained upon titration back to pH 7.5.

10.6 on the positive band of ffW-14 DNA above 260 nm in 20 mM NaC1 (A) and in 500 mM NaCI (B). With increasing pH, the crossover of the posi- tive band undergoes a blue shift resulting in a slightly enhanced rotational strength below 275 nm and a slightly decreased value above 275 nm. These two effects balance with no real reduction in the total rotational strength of this band up to pH 10.6. This effect is somewhat more exaggerated at the higher ionic strength, but of the same char- acter. The behavior of the acetylated derivative displayed in Panel C showed a similar but signifi- cantly smaller effect at comparable ionic strengths.

The dashed curves in panels A and C are the CD spectra obtained upon return to pH 7. The small differences observed between the original and the return are probably attributable to minor hydrolytic effects at the elevated pH.

The smaller effect observed in the titration curves for the acetylated derivative which has all of the primary groups acetylated, but still retains approx, half of the secondary amino groups un- modified, suggests that the small transformations observed in the spectra in Fig. 4A and B as the pH increased from 7 to 10.6 are due principally to the removal of charge from the secondary amino groups of putrescine, leaving the primary amino groups still protonated. Thus, the pK a of the latter has been apparently shifted significantly above 11. Such a substantial elevation implies extensive -

318

and close - association with the negatively charged phosphate backbone. This conclusion is also one arrived at by Gerhard and Warren [11] based on the greater susceptibility of the secondary amino of putThy to acetylation. Because of the hydrolysis problems at elevated pH, however, we were unable to test this hypothesis directly by chemical titra- tion procedures.

Another mechanism of minimizing the in- fluence of the positive charge of the putrescine residues is to immerse the DNA molecule in a highly ionic environment of low water activity, such as that provided by high concentrations of neutral electrolytes. Such solvents have their own unique effect on the conformation of under- ivatized DNA [15,25]. In these solvents, the rota- tional strength of the positive band decreases with increasing electrolyte content of the solvent. This is correlated with an increase in the winding angle

of DNA [14,26]. Probably other helical parameters are changing as well, but there is no marked transformation in the geometry of the backbone which has been demonstrated to remain in the B conformation in amine dervivatives of DNA which exhibit similarly transformed spectra [9,27]. Typi- cal effects of increasing concentrations of monova- lent salt, for instance, on the positive band of the CD spectrum of DNA are demonstrated by the data shown in Fig. 5 for E. coli DNA (beaded curve). The decrease in the rotational strength of the positive band above 260 nm as the salt con- centration is increased is not accompanied by

E G

a )

? 0

×

0

I

0

° o

/o ° o k ,

I

C

I I

I

26O

i E

B

i

D

i I

F

L i I

300 260 300 X (nm) X(nrn)

Fig. 5. Effects of increasing content of LiCI on the CD spectra of E. coil DNA ( O O O ) and 4~W-14 DNA ( ). (A) 20 mM NaC1 (0 M LiC1): (B) 0.77 M LiC1 ( O O O ) and 0.86 M LiCI ( ): (C) 1.78 M LiCI, ( O O O O ) ; (D) 4.37 M LiC1 ( 0 0 0 ) and 4.43 M LiCI ( ); (E) 5.99 M LiCI ( 0 0 0 ) and 5.61 M LiC1 ( ); (F) 7.08 M LiC1 ( O O O ) and 6.99 M LiCI ( ).

Fig. 6. A Pauling-Cory-Koltun spacefilling model of DNA in the B form with putThy substituted at the position indicated by the solid line arrows. This veiw shows the pr ima~ amine group (located at the tip of the longest solid line arrow) of putrescine interacting with the oxygens of two consecutive Y-phosphate groups (located at the tip of the broken line arrows) in the same strand in the major groove. The order of increasing Van dcr Waals', radii of the atoms in the model is H, N, O and P.

319

marked changes in the other spectral bands below 258 nm.

One would anticipate that the effects of increas- ing the electrolyte content of the solvent would have similar effect on the CD spectrum of q~W-14 DNA, since high ion concentrations would effec- tively compete for the interactions of the primary amino group of putrescine with the phosphates of the DNA backbone. Fig. 5 demonstrates that this is indeed the case, although the rotational strength of the positive bands of E. coli and q~W-14 DNA (solid line) do not become comparable until 6 to 7 M LiCI. It might be noted that despite the high salt concentration in the q, W14-DNA solutions, the spectral changes as a function of salt con- centration are those expected for the GC content of q~W-14 DNA. The pronounced spectral distor- tion resembling the contribution of a small frac- tion of the bases in an A form sor.aetimes seen in high salt for high GC content DNA species such as Micrococcus luteus DNA [28] is absent. Nor is there evidence in the CD spectra for the formation of ~p structures or other condensed forms of DNA, at the concentrations of DNA employed in these experiments.

Discussion

The CD spectral properties of q,W-14 DNA, although possessing some unusual features, suggest that the average backbone geometry is a B form. The conformation appears to be that more usually found in higher-GC-content DNAs. The enhanced rotational strength of the positive band above 260 nm, compared to a procaryotic DNA of compara- ble base composition is attributable to conforma- tional differences between the two induced by the positively charged putrescine residues. The spec- tral difference suggests a smaller average winding angle of approx. 0.4 ° (0.2 bp) for ~kW-14 DNA. On the basis of these present results it is not possible to decide whether this conformational difference is consistently propagated as a small perturbation throughout the length of the DNA duplex structure or whether there is a more dramatic difference in the neighborhood of the putThy residues, while the remainder of the basepairs assume B-form geometry characteristic of random-sequence DNA. The primary amino group of putThy appears to be mainly responsible

for these effects. The results of the titration experi- ments suggest that this group is probably involved in intimate ion-pair formation with the phosphate groups of the backbone. This conclusion is in accord with the kinetics of acetylation of putres- cine in q,W-14 DNA [11].

The identification of putThy in q,W-14 DNA and the properties of the DNA led originally to the suggestion by Kropinski [29] that the putres- cine amino groups might form salt bridges with phosphates in the opposite strand of the helix. Inspection of a Pauling-Corey-Koltun model of DNA with putThy residues indicates that, if aver- age B geometry spacings are maintained, these interactions would be mediated via water bridges. Our model does not, however, incorporate the sequence-dependent conformational variability which has been observed in crystals of a self-com- plementary dodecamer [30]. Such variability could, in principle, facilitate a more intimate amino- phosphate interstrand interaction.

Direct ion pairs can be formed within the same strand without the mediation of water molecules between the primary amine and the phosphates attached to either the 5' and 3' position of the putThy nucleoside or to the 5' end of the putThy nucleotide and the 5' end of the adjacent nucleo- tide on the 5' side of putThy. The latter associa- tion seems to be sterically less hindered. A view of the Watson Crick B form of DNA with the primary amine group of putrescine site bound in the this manner is shown in Fig. 6. With minor adjust- ments of the backbone tortional angles and other helical parameters (which we are unable to do with this semirigid model) it would be possible to get an ion-pair association of lower energy in which the positively charged primary amino group is more symmetrically disposed between and closer to the phosphate oxygens. Transformation of the back- bone in the vicinity of the putThy residues to a more A-like geometry in which the center-to-center distance between adjacent phosphates on the same strand is shorter and the nonesterified oxygen atoms of the phosphates face inward toward the groove [31] would be particularly effective in facilitating good ion-pair formation in the case of both the intrastrand interactions as well as the interstrand ion-pair formation originally suggested by Kropinski [29].

320

The relationship between the higher packing density and the conformational properties of q~W- 14 DNA remains obscure. Ion-pair formation will almost certainly promote a higher packing density, since the backbone is likely to be less hydrated, thus permitting a closer distance of approach of the duplex structure of q,W-14 DNA in the col- lapsed form in the phage head. Whether the con- formational differences observed will enhance this effect is unclear. Whatever conformationally de- pendent mechanism is invoked to explain the packing density, however, it must accomodate the fact that the CD properties of the moecule at dilute concentrations in aqueous solution are those normally interpreted in terms of one in which the majority of the basepairs is in an unexceptional B-type structure which would normally be found in a higher-CG-content DNA. There is no spectral evidence for the presence of a dramatically differ- ent form, such as the D, which favors higher packing densities.

Acknowledgements

This work was supported by grants from the National Science Foundation (PCM 7724879-A01) and the National Institutes of Health (GM 30284) to S.H. and Operating Grant A3686 from the Natural Sciences and Engineering Research Council of Canada to R.A.J.W.

References

1 Kropinski, A.M.B., Bose, R.J. and Warren, R.A.J. (1973) Biochem. 12, 151-157

2 Kelln, R.A. and Warren, R.A.J. (1973) J. Virol. 12, 1427-1433

3 Maltman, K.L., Neuhard, J., Lewis, H.A. and Warren, R.A.J. (1980) J. Virol. 34, 354-359

4 Warren, R.A.J. (1981) Curr. Microbiol. 6, 185-188 5 Scraba, D.G., Bradley, R.D., Leyritz-Wills, M. and Warren,

R.A.J. (1983) J. Virol. 124, 152-160 6 Manning, G.S. (1978) Q. Rev. Biophys. 11, 179-246 7 Record, Jr., M.T., Anderson, C. and Lehman, T. (1978) Q.

Rev. Biophys. 11, 103-178

8 McGhee, J.D. and Felsenfeld, G. (1980) Nucleic Acids Res. 8, 2751-2769

9 Chen, C., Pheiffer, B.H., Zimmerman, S.B. and Hanlon, S. (1983) Biochem. 22, 4746-4751

10 Arnott, S., Chandrasekaren, R., Hukins, D., Smith, P. and Watts, L. (1974) J. Mol. Biol. 88, 523 533

11 Gerhard, B. and Warren, R.A.J. (1982) Biochem. 21, 5458-5462

12 Ames, B. and Dubin, D. (1960) J. Biol. Chem. 235, 769-775 13 Felsenfeld, G. and Hirschman, S.Z. (1965) J. Mol. Biol. 13,

407 427 14 Chan, A., Kilkuskie, R. and Hanlon, S. (1979) Biochem. 18,

85-91 15 Hanlon, S., Brudno, S., Wu, T.T. and Wolf, B. (1975)

Biochem. 14, 1648-1660 16 Lewis, H.A., Miller, R.C., Jr., Stone, J.C. and Warren,

R.A.J. (1975) J. Virol. 16, 1375-1379 17 Gratzer, W.B., Hill, L.R. and Owen, R.J. (1970) Biochem.

15, 209-214 18 (}ray, D.M., Hamlton, F.D. and Vaughan (1978) Biopoly-

mers 17, 85-106 19 Sober, H. (ed.) (1968) Handbook of Biochemistry, Chemical

Rubber Company, Cleveland 20 Johnson, B., Dahl, K., Tinoco, Jr., 1., lvanov, V. and

Zhurkin, V. (1981) Biochem. 20, 73 78 21 Gennis, R.B. and Cantor, C.R. (1972) J. Mol. Biol. 65,

381-399 22 Weast, R. and Astle, M. (eds,) (1982) CRC Handbook of

Chemistry and Physics, 63rd Edition, Chemical Rubber Company Press, Bocca Raton

23 Robinson, R.A. and Stokes, R.H. (1959) Electrolyte Solu- tions, Butterwordth, London

24 Chen, C., Kilkuskie, R. and Hanlon, S. (1981) Biochemistry 20, 1987-1995

25 Wolf, B. and Hanlon, S. (1975) Biochemistry 14, 1661 1670 26 Baase, W.A. and Johnson, Jr., W.C. (1979) Nucleic Acid

Res. 6, 797-814 27 Fish, S., Chen, C., Thomas, G., Jr. and Hanlon, S. (1983)

Biochemistry 22, 4751-4756 28 Ringquist, S., Cheng, S. and Hanlon, S. {1984) Fed. Proc.

43, 1884a 29 Kropinski, A.M.B. (1973) The Physico-Chemical Properties

of Bacteriophage 4,W-14 Deoxyribonucleic Acid, Ph.D. Thesis, University of British Columbia

30 Drew, H., Wing, R., Takano, T., Broka, C., Tanaka, S.. itakura, K. and Dickerson, R. (1981) Proc. Natl. Acad. Sci. USA 78, 2179-2183

31 Zhurkin, V., Lysor, Y. and Ivanov, V. (1980) Biopolymers 19, 1415-1434