Structure of the replicative form of bacteriophage φX174

Structure of the replicative form of bacteriophage φX174

J. Md. Biol. (1969) 49, 379-399 Structure of the Replicative Form of Bacteriophage $X174 VII.? Renaturation P. H. PO-M, of Denatured Double-stranded...

898KB Sizes 0 Downloads 52 Views

J. Md. Biol. (1969) 49, 379-399

Structure of the Replicative Form of Bacteriophage $X174 VII.? Renaturation P. H. PO-M,

of Denatured Double-stranded J. vm

ROTTERDAM

AND

4X DNA

J. A. COHEN

Medical Biological Laboratory of the National Defence Research Organization TNO, 139 Lange Kleiweg, Rijswijk (Z.H.) The Netherlands (Received 6 June 1968, and in revised form 25 November 1968) Denatured double-stranded 4X DNA (RF DNA) is formed after exposure of component I (covalently closed circular duplex) to alkali and reneutralization of the alkali. Under conditions which are optimal for the renaturation of linear phage DNA, denatured double-stranded 4X DNA cannot be renatured. Renaturation will proceed very rapidly, however, in alkaline solution (pH 12) at 5O’C. Breakage of a phosphodiester bond in denatured double-stranded 4X DNA will lead to one of the following products, irrespective of the means by which the break is achieved (enzymic hydrolysis, r-irradiation or incubation with reducing agents) : (a) native double-stranded DNA (component II; nicked circular duplex) is formed in buffers of low ionic strength (0.01 M-phosphate-O.001 M-sodium citrate, pH 7) at 0 to 30°C; (b) single-stranded DNA is formed in buffers of low ionic strength at temperatures exceeding 30°C ; (c) a heterogeneous product, presumably consisting of partially renatured DNA, is formed in buffers of high ionic strength (0.1 M-Tris or M-NaCl). These results suggest that the nature of the product of denatured double-stranded +X DNA, in which a phosphodiester bond is broken, is primarily determined by environmental conditions such as temperature and ionic strength. Cross-linking of the two strands in component I by means of nitrogen mustard prevents the conversion by alkali into denatured RF DNA. On the other hand, interstrand cross-links in denatured RF DNA impede the molecules from renaturing to form component II after chain breakage. These results lend support to a model for the structure of denatured RF DNA which had been proposed earlier. This means that a shift of the two strands with respect to each other and a high number of crossings over of both strands have a major role in the irreversibility of denaturation of component I.

1. Introduction Double-stranded DNA of bacteriophage $X174 (replicative form DNA or RF DNA) isolated from Escherichia wli C infected with phage in the presence of chloramphenicol exhibits the properties of a covalently closed circular duplex (Jansz & Pouwels, 1965 ; Burton t Sinsheimer, 1965). Exposure of the closed circular duplex to alkali converts the molecule into a denatured, coiled double-stranded structure, which after neutralit Paper VI in this series is Pouwels,

Knijnenburg, 379

van Rotterdam,

Cohen & Jansz, 1968.

380

P. H.

POUWELS

ET AL.

zation of the alkali behaves as denatured double-stranded DNA (Pouwels & Jansz, 1964; Pouwels, Knijnenburg, van Rotter&m, Cohen & Jansz, 1968). The term “denatured RF DNA” will in this report refer to denatured double-stranded +X DNA at neutral pH. Denatured RF DNA in buffers of low ionic strength will renature to form nicked circular duplex molecules (component II), when a single-strand scission is introduced at room temperature (Jansz, Baas, Pouwels, van Bruggen & Olclenziel, 1968; Pouwels et al., 1968). The present report describes experiments which indicate that three different products (nicked circular duplex DNA, single-stranded DNA or partially renaturecl DNA) can be formed after breakage of a phosphodiester bond in denatured RF DNA, depending on the temperature and the ionic strength. Under specified conditions the nature of the product merely depends on the temperature and ionic strength, and not on the manner by which a chain breakage is achieved (y-irradiation, enzymic hydrolysis or reducing agents). It will be shown in this report that under appropriate conditions renaturation can be accomplished without the introduction of single-strand breaks; the conditions for renaturation include exposure of denatured RF DNA in alkaline solution to elevated temperatures. The structure of denatured double-stranded +X DNA has been further investigated using bifunctional alkylating agents in order to cross-link the two strands of doublestranded +X DNA. The results of these experiments lend support to a model which was presented earlier (Pouwels et al,, 1968), according to which the two strands are shifted with respect to each other along the longitudinal axis of the helix, when the hydrogen bonds have been broken.

2. Materials and Methods (a) Preparation Double-stranded described previously

+X DNA labelled (Jansz, Pouwels

of double-stranded 4X DNA with 32P or non-labelled & Schiphorst, 1966).

4X DNA

was prepared

as

in a sucrose gradient

(b) Zone sedimentation

Sedimentation in a sucrose gradient was performed according to the procedure of Britten & Roberts (1960) in an International model B60 preparative ultracentrifuge at 4°C with an I.E.C. rotor SB405. After centrifugation the tubes were punctured and (45 + 1) fractions were collected on planchets and assayed for radioactivity.

(c) Sedi7nentatio-n analysis Velocity sedimentation was performed in a Spinco model E analytical ultracentrifuge equipped with an ultraviolet optical system with photo-electric scanner. Concentrations of DNA used for sedimentation were within the range of 15 to 30 pg/ml. The sedimentation coefficients have been corrected in the usual way to obtain Sz,,, values. (d) Enzymic

hydrolysis

of DNA

by pancreatic

DNase

or endonuclease

from Neurospora

crassa Pancreatic DNase (Nutritional Biochemical Corporation, twice crystallized) was dissolved in 0.01 M-phosphate buffer (pH 7.0) containing O*l”/o bovine serum albumin (Armour) at a concentration of 100 pg/ml. and stored at -35°C. A small volume (usually 10 ~1.) of DNase in an appropriate dilution, was added to a solution of DNA in 0.0 lophosphate-l mM-sodium citrate-4 mM-MgCl,, pH 7.0 at 25°C. The reaction was terminated by the addition of the sodium citrate or EDTA to a final concentration of 0.01 M. Endonuclease from Neurospora craSSa (Linn & Lehman, 1965) was a generous gift of

STRUCTURE

OF 4X

RF

DNA

381

Dr Linn. The enzyme was added in an appropriate dilution to a solution of DNA in 0.1 MTris-0.02 M-MgCl,, pH 8.4 at 37°C. The reaction was terminated by the addition of sodium citrate to a final concentration of 0.1 M and immersion of the reaction vessel in ice. (e) y-Irradiation

DNA in aqueous solutions was irradiated with y-rays using a Gamma-cell 100 6oCosource (dose rate 4.0 krad/min). The dose rate was measured with a ferrosulphate dosimeter according to Fricke. (f) Alkylation of DNA Cross-linking of double-stranded DNA was performed using the method described by Kohn, Spears & Doty (1966). Double-stranded $X DNA (component I) or denatured RF DNA in 26 m&r-triethanolamine-Hcl, 1 a-sodium citrate, pH 7.2, was treated with HNzi at 26°C. Concentration of HNs and time of incubation are given in the text. The reaction was stopped by addition of NaJ&Oe (10 mM) and cooling in ice. After alkylation RF DNA was denatured with 0.15 M-NaOH and after 3 min at room temperature the alkali was neutralized with HCI. Subsequent to alkylation denatured RF DNA was dialysed against 10 ma6-phosphate-l mM-sodium citrate buffer, pH 7.0, and incubated with pancreatic DNase in the presence of 4 mlrr-MgSO*. Control samples were treated likewise except that no nitrogen mustard was present. When treatment with formaldehyde was required, the pH of the solution was raised to 9.85 (measured in the presence of formaldehyde) by addition of 0.05 M-Na2C03 and the material was kept at 20°C for 5 min. Under these conditions the reaction of formaldehyde with denatured DNA is complete, whereas native double-stranded DNA does not react at all.

3. Results (a) Renaturation

without chain breakage

Denatured linear DNA of viral origin will easily renature in buffers of high ionic strength (0.5 iw-NaCl--@o5 M-sodium citrate) when heated at 60 to 70°C for several

hours and subsequent slow cooling to room temperature (Marmur, Rownd & Schildkraut, 1963). Under these conditions denatured RF DNA does not renature. Renaturation will not take place either if the pH and/or the temperature are varied between pH 7 and pH 11.6 and between 20 and 7O”C, respectively. Denatured RF DNA renatures very rapidly, however, in iv-Nacl, pH 12 at elevated temperatures. Figure 1 shows the result of a sucrose gradient analysis after heat treatment of denatured RF DNA in alkaline solution. Inspection of Figure 1 reveals that denatured RF DNA is completely renatured within 10 minutes by heating at 50°C and is partially renatured at 30°C or 40°C. The addition of a 25fold excess of cold denatured RF DNA to 3aP-labelled DNA did not affect the renaturation of 32P-labelled denatured RF DNA, suggesting that renaturation is concentration independent. In buffers of lower ionic strength (5- to 50 m&r-sodium ions) a higher pH was required for renaturation. For example, renaturation was complete after heating for one hour at 50°C in 0.05 MNaOH, pH 12.6, but was only partially complete after heating for one hour at 50°C in 0.04 M-NaOH (pH 12.5). In order to be able to correlate the renaturation conditions with the conformation of the molecule, the sedimentation coefficient of denatured RF DNA in M-NaCl10 nnvr-Na,CO, of increasing pH was determined by boundary sedimentation using an analytical ultracentrifuge. The results of such an experiment are presented in Figure 2. In the range between pH 7 and pH 10 the sedimentation coeficient of denatured t Abbreviations chloroethylamine-HCl.

used: HN2, N-methyl-bis

(2.ohloro-ethyl) amine-HCl ; HN, N,N-dimethyl-2-

382

P. H.

POUWELS

ET

AL.

600 200

FICA 1. Renaturation of denatured double-stranded +X DNA in buffers of alkaline pII. A mixture of denatured double-stranded and single-stranded 4X DNA (32P-labelled) in 1 MNaCl, pH 11.95 was kept at various temperatures during 10 min. After neutralization of the alkali by addition of a small volume of 1 M-KH,PO* (0.2 ml.) samples were layered on top of 3.8 ml. of sucrose solution (5 to 28% sucrose in 1 M-NaCl-0.01 M-phosphate-O.005 M-sodium citrate, pH 7) and spun for 3 hr at 60,000 rev./mm. The tubes were punctured and fractions of 0.1 ml. were collected. The radioactivity in the various fractions was determined. (a) Denatured double-stranded $X DNA, 10 min 20°C at pH 11.95; (b) denatured doublestranded 4X DNA, 10 min 30°C at pH 11.95; (c) denatured double-stranded 4X DNA, 10 min 40°C at pH 11.95; (d) denatured double-stranded +X DNA, 10 min 50°C at pH 11.95.

%F4P+ 78910

II

1213

PH

FICA 2. Sedimentation coefficient of denatured double-stranded +X DNA and single-stranded +X DNA in 1 M-NaCl-0.01 M-Na,CO,-1 mM-EDTA as a function of the solvent pH. A radiometer pH meter (electrode GH 2021, type B) was used to measure the pH. The pH values reported are uncorrected for sodium ion ooncentration. (0) Denatured double-stranded 4X DNA; (0) single-stranded $X DNA. RF DNA remains

constant. and levels off at pH 12 .

Beyond pH 10 the sedimentation

coefficient risesgradually

STRUCTURE

OF +X

RF

383

DNA

The conversion of the compact form of single-stranded (5X DNA (at the neutral side of the transition) to the extended form (at the alkaline side of the transition) was recorded from the same experiment and was found to occur between pH 10.0 and pH 10.75. No accurate S-values could be determined between pH 10.0 and pH 10.75 due to heterogeneity of the material in the transition phase. Both the conversion from neutral denatured RF DNA to alkaline denatured RF DNA and from neutral single-stranded DNA to alkaline single-stranded DNA are completely reversible. It will be noted that the midpoint of the transition from neutral to alkaline denatured RF DNA is at a pH which is significantly higher than the midpoint of the transition for single-stranded DNA. The data in Figure 2 do also indicate that the renaturation of denatured RF DNA is favoured by a pH at which the molecule is fully titrated and consequently a great majority of the base pairs have been disrupted. (b) Renaturation

after chain breakage in buffers of low or high ionic strength

Previous studies on the reactivity of denatured RF DNA towards pancreatic have shown that single-st’rand chain breaks promote renaturation, giving rise to nicked circular duplex molecules (Jansz et al., 1968; Pouwels et aE., 1968). In the following section experiments will be described which indicate that chain breakage of denatured double-stranded $X DNA may also lead to the formation of single-stranded DNA or a partially renatured molecule. The nature of the product formed is merely dependent on the ionic strength and on the temperature of the DNA solution, but is independent on the means by which the strand scission is produced. When denatured RF DNA was incubated at 37°C with reducing agents like ascorbic acid, cysteine or Na2S03, a homogeneous product was obtained, which sedimented in ivr-NaCl at 25 s (Fig. 3). Reducing agents that are auto-oxidable, like ascorbic acid, DNase

IBOOI Denatured

RF1

1400~

?\

si DNA7~

:



:

lOOf) -

;oJ

5 5

1000’ -

:i’

j;

lip] ’ -

$- 600 ; 200,~~j -0d” & -- E+--.I

7----‘--~1 ‘\

u

Ld’

i

10@0’1 \

600;

(c)j

200:L+-_,_,.---: ,i 1 5 13 17 iI 25 29 33 37 41 Tube no.

FIG. 3. Sucrose gradient centrifugation of denatured double-stranded 4X DNA after incubation with ascorbinate. Denatured double-stranded +X DNA (32P-labelled) in 0.01 n-phosphate-1 mn-sodium citrate W&S incubated at 37’C with 10m5 M-ascorbic acid for 0 (a), 15 (b) and 4.5 min (e). Samples (0.2 ml.) were layered on top of 3.8 ml. sucrose solution (4 to 20% sucrose in 1 r&-NaCl-0.006 M-sodium citrate, pH 7) and spun for 2 hr at 60,000 rev./min. The tubes were>unctured and fractions of 0.1 ml. were collected. The radioaotivity kIthe_various fractions was determined,

384

P. H.

POUWELS

ET

AL.

produce chain breaks in DNA only in the presence of oxygen. The reaction is promoted by the presence of trace amounts of metal ions (Cu). From sedimentation experiments in neutral and alkaline sucrose gradients and also from sedimentation equilibrium experiments in CsCl it was concluded that the reaction product was single-stranded DNA. The kinetics of inactivation of the biological activity by ultraviolet light also supported this conclusion. In order to explain the difference in results obtained with pancreatic DNase and

$-Jijj I

5

9

13 17 21 25 29 33 37 Tube no.

FIG. 4. Sucrose gradient centrifugation of denatured double-stranded +X DNA after incubation with ascorbinate at various temperatures. Denatured double-stranded 4X DNA (sap-labelled) in 0.01 an-phosphate-l mx-sodium citrate was incubated with 10V5 M-ascorbic acid + IO-5 MCuS04 at 20°C (30 min); 30°C (15 min); 4O’C (10 min). Details of centrifugation as under Fig. 3. Time of centrifugation 2 hr. (a) Denatured double-stranded 4X DNA; (b) denatured double-stranded +X DNA, after incubation with ascorbinate-CuSO, at 20%; (c) denatured double-stranded 4X DNA, after incubation with ascorbinate-CuSO, at 30°C; (d) denatured double-stranded +X DNA, after incubation with ascorbinate-&SO, at 40°C.

with reducing agents, the influence of temperature on the reaction was studied. Denatured RF DNA was incubated with 1O-5 M-ascorbic acid in the presence of 10m5M-C!uS04 at various temperatures. From the sedimentation analysis presented in Figure 4 it can be seen that at 20°C component II is formed, but at 30 or 40°C predominantly single-stranded DNA is formed, showing that at least at 20°C the results are in agreement with those obtained with pancreatic DNase. The possibility was considered that divalent copper ions (which decrease the T, of DNA) may influence the reaction (Eichhorn & Clark, 1965). However, in the absence of CuSO, similar results were obtained, arguing against an effect of copper ions on the nature of the reaction product. Since divalent cations (like magnesium) increase the melting temperature (T,) of double-stranded DNA (Eichhorn, 1962) the influence of divalent cations was also studied. When magnesium ions are added to the reaction mixture at a concentration of 4 mM, denatured RF DNA was converted to component TI both at 25°C and at 37°C. Similarly, when denatured RF DNA was incubated

STRUCTURE

OF +X

RF

38.5

DNA

with pancreatic DNase in a reaction mixture containing magnesium ions, component II was formed both at 25°C and at 37°C. A similar influence of the temperature on the renaturation was found when singlestrand scissions were produced by y-irradiation of denatured RF DNA in O-01 Mphosphate-l mM-sodium citrate (Freifelder, 1965). At temperatures below 30°C the molecules renature when a chain break is introduced, but at higher temperatures single-stranded DNA was formed. Addition of 4 mM-magnesium had a profound effect on the renaturation conditions. At temperatures up to 50 to 60°C during yirradiation, component II was formed and only at still higher temperatures singlestranded DNA was found. Magnesium could be replaced by barium or calcium, resulting in the same increase in renaturation temperature. When denatured RF DNA was incubated at 25°C with pancreatic DNase or Neurospora crasSaendonucleasc Linn & Lehman, 1965) in buffers of high ionic strength (0.1 M-Tris), a heterogeneous product was formed sedimenting at 12 to 40 S. A similar heterogeneous product was obtained after y-irradiation of denatured RF DNA in M-NaCl. Buffers of high ionic strengt’h stabilize single-strand regions in denatured DNA and thus mismatching regions may occur along the two DNA strands, which require annealing conditions before an exact registration can proceed. Therefore we conclude that the heterogeneous product represents partially renatured molecules. High ionic strength also favours aggregation, which may in part explain the heterogeneity of the product. Since chain breakage at 25°C in buffers of low ionic strength invariably leads to the formation of component II; this result confirms the hypothesis that the nature of the product formed is merely dependent on external conditions and is independent of t’he means by which the chain is broken. (c) Effect of interstrand

cross-links

on the renaturation

of denatured

RF DNA

In a previous communication (Pouwels et al., 1968) a model has been presented for denatured RF DNA which accounts for the irreversibility of denaturation at a critical pH. According to this model the two strands in denatured RF DNA are shifted with respect to each other along the longitudinal axis of the helix, when all hydrogen bonds have been broken. Because of the continuity of the individual strands they cannot be separated. After neutralization of the alkali, non-specific base pairs arc formed. Since the number of cross-overs in covalently closed circular duplexes is constant, the two strands in denatured RF DNA are wound around each other several hundred times, which drastically reduces the mobi1it.y of the st,rands. This, taken together with the presence of non-specific base pairs, prevents the strands from shifting back to the original position, which is required for proper renaturation. The formation of inter-strand cross-links within DNA molecules does not only prevent strand separation under denaturing conditions, but also permits rapid renaturation when the conditions for denaturation have been removed (Geiduschek, 1961; Kahn et al., 1966; Lawley & Brookes, 1967). In order to verify our model experimentally we have introduced cross-links into RF DNA and have followed the properties of the material upon denaturation. The rationale behind these experiments is the following. If it is assumed that our model is correct, introduction of cross-links into RF DNA should prevent denat,uration with alkali since the two strands are prevented from shifting with respect to each other. Alternatively, the introduction of cross-links into denatured RF DNA will prevent its renaturation by a single &and-scission for the same reason. When one of 37

386

P. H.

ET AL.

POUWELS

the strands is cleaved with DNase the molecule will unwind, but since the two strands, which are shifted with respect to each other, are held together by a cross-link, they cannot come into proper register. In order to link the two strands of RF DNA covalently, the method described by Kohn et al. (1966) was followed. Figure 5 shows a sedimentation pattern in a sucrose gradient of cross linked RF DNA after denaturation with alkali. It is seen from this figure that HN, treatment does not alter the sedimentation behaviour of component I.

;;I __, j, ‘\,\_‘c;/ I 5 lo i?20

25 30 35 40

Tube no

FIG. 5. Denaturation of double-stranded $X DNA after treatment with nitrogen mustard. Double-stranded 4X DNA labelled with 32P in 25 mu-triethanol aminsHCl-1 mM-sodium citrate, pH 7.2, was incubated with 5 mM-HN, at 25°C during 30 mm. After cooling in ice, part of the material was denatured with alkali (0.15 M) and after 3 min neutralized with HCl. The control sample was treated in the same way except that no HNz was used. Details of centrifugation as under Fig. 1. Time of centrifugation 2.5 hr. (a) A mixture of component I and denatured component I; (b) component I treated with HNs for 30 min at 26%; (c) as (b), followed by denaturation with alkali.

Moreover, under conditions where untreated RF DNA is converted into denatured RF DNA, cross-linked RF DNA sediments like native double-stranded $X DNA and thus has not been irreversibly denatured. Similarly denatured RF DNA was treated with HN,, followed by incubation with pancreatic DNase. The control was treated likewise except that no nitrogen mustard was present. The results presented in Figure 6 demonstrate that the sedimentation rate of denatured RF DNA in iv-NaCl had not been changed after alkylation and treatment with DNase. This is not unexpected since the sedimentation coefficient of denatured RF DNA fits the relation between S-value and molecular weight (Studier, 1965). A single-strand break would therefore presumably not lead to a change in S-value as long as the molecular weight remains unaltered (Fig. 6). After incubation of this material with formaldehyde (5%) at 65°C for 10 minutes, one expects the molecules without chain breaks to sediment approximately at the same rate as in the absence of formaldehyde whereas molecules with single-&and breaks (with or without, cross-links) should sediment much slower. Closed circles assume a compact structure in formaldehyde (Pouwels & Jansz, 1964; Crawford & Black, 1964) as in alkali. but nicked circles are converted into single-stranded DNA. Clearly two single strands

STRUCTURE

OF +X

RF

DNA

387

linked to each other will sediment more rapidly than single-stranded DNA. After sedimentation in a sucrose gradient two main peaks are visible (Fig. 6(e)). The fast’ one corresponds to denatured RF DNA with no single-strand break. The amount of material sedimenting at this position is dependent on the time of incubation with DNase. After longer exposure to the enzyme the amount of material sedimenting at this position diminishes and the bulk of the material is found as the slower moving component. The slower component sediments with a rate expected for single-stranded DNA of twice the size of single-stranded +X DNA (Hagen, 1967). This is taken to mean that cross-linking does occur in denatured RF DNA and that DNase converts this molecule into two single strands held together by covalent linkages. When singlestranded DNA produced by denaturation of component II is subjected to alkylation with HN, and treatment with formaldehyde, virtually all material sediments at the position of single-stranded DNA. In dilute solutions of single-stranded DNA apparently no inter-strand cross-linking takes place. The role of cross-links in preventing renaturation is also suggested by an experiment using the mono-functional alkylating Denatured



1

5

9

RF1

-_ RF11

/

13 17 21 25

29 33 37

Tube no

of denatured double-stranded +X DNA after treatment with nitrogen FIG. 6. Rem&nation mustard. Denatured double-stranded I$X DNA (labelled with aaP) in 25 m&f-triethauolamine-HCl1 m&r-sodium citrate, pH 7.2, was incubated with 5 miu-HNs at 25’C for 30 min. After cooling in ice the reaction was stopped by addition of Na,SaOs to a final concentration of 0.01 M and dialysis (overnight) against 10 mrd-phosphate-l mx-sodium citrate, pH 7.0. After dialysis, part of the material was incubated with pancreatic DNase (0.0005 pg/8 pg DNA/ml.) at 26’C for 5 min. The reaction was terminated by addition of sodium citrate to a final concentration of 0.01 M. Following DNase digestion a sample was heated at 65°C for 6 min in the presence of 6% formaldehyde. Details of centrifugation as under Fig. 1. Time of centrifugation 2 hr. (a) Denatured double-stranded +X DNA; (b) as (a), after incubation with DNase; (c) as (a), after incubation with HN,; (d) as (c), after incubation with DNase; (e) as (d), after heating at 65% with 6% HCHO. This sample was sedimented in sucrose containing 1 M-NaCl-50 mMphosphate buffer-l maa-sodium citrate-1.8% HCHO, pH 7.

388

P. H.

POUWELS

ET

AL.

agent HN. Renaturation of denatured RF DNA, treated with HN, normally occnrretl after incubation with DNase. The concentration of nitrogen mustard (HN,) used in these experiments is likely to introduce several cross-links per DNA molecule. When the application of low concentrations of HN,, introducing on the average only about one cross-link per DNA molecule, is followed by DNase treatment, the following products are to be expected: (a) not cross-linked and not nicked molecules (unaltered denatured RF DNA). (b) cross-linked and not nicked molecules, (c) cross-linked and nicked molecules, (d) not cross-linked and nicked molecules (component II). Molecules of types (b) and (c) will have the same sedimentation velocity as unaltered denatured RF DNA (a). After formaldehyde treatment, (a) and (b) will show up as denatured RF DNA and (c) as single-stranded DNA of twice the size of mature phage DNA. Under these reaction conditions component II does not react with formaldehyde and therefore remains unaltered (see Materials and Methods). This expectation is borne out by the experimental data. Denatured RF DNA was treated with 0.1 m&r-HN, for 40 minutes at 25°C and incubated with pancreatic DNase (0.0005 pg DNase/4 pg DNA/ml.) for five minutes at 25°C in the presence of 4 m&r-MgSC,. The samples were then subjected to the formaldehyde treatment and centrifuged in a sucrose gradient. Approximately 200/, of the total radioactivity present was found at the position of denatured RF DNA. Of the material which had received single-strand breaks by DNase, approximately 30% sedimented at the position of DNA of twice the size of single-stranded 4X DNA. The remainder of the material had a sedimentat,ion rate corresponding to component II.

4. Discussion The results presented in this paper support, although they do not prove conclusively, the correctness of the model for denatured RF DNA which we have proposed earlier on the basis of the then available data (Results section (c) of this paper; see also Pouwels et al., 1968). Inherent in this model is that no appreciable base sequences in denatured RF DNA should be in register. As an alternative it was thought that merely a high number of crossings-over in denatured RF DNA would suffice to hinder the molecule from folding back into a helical conformation. The inability of RF DNA molecules containing cross-links to become irreversibly denatured strongly suggests that after denaturation no single sequence of bases in one strand is in register with the corresponding sequence in the other strand. Otherwise denatured RF DNA without crosslinks should also fold back after neutralization of the alkali. In an experiment recorded in Figure 6 it was shown that the presence of cross-links in denatured RF DNA prevents the renaturation after strand scission. This result also suggests that primarily a shift of the two strands is responsible for the irreversibility of denaturation of the molecules. In a series of experiments it was demonstrated that chain breakage in denatured RF DNA under certain conditions leads to products other than component II. This result supports the view that no base sequences in denatured RF DNA are in register. One would expect that denatured DNA molecules containing base sequences in register would renature after chain breakage irrespective of ionic strength and temperature of the solution (Geiduschek, 1961). We have shown that under conditions as described in this paper denatured RF DNA, unlike linear phage DNA, cannot be

STRUCTURE

OF r$X RF

DXA

389

renatured at neutral pH. However, denatured RF DNA, like linear phage DNA, can be renatured at alkaline pH. Two other differences may be noted between the rpnaturation of circular and linear DNA molecules. (1) The renaturation of denatured RF DNA was found to be concentration independent, while the renaturation of T4 DNA is concentration dependent (Vinograd, Morris, Davidson & Dove, 1963). (2) The optimum renaturation pH for circular DNA is much higher (12-O) than that for linear DNA (10.7 to 10.8). For both types of DNA the renaturation pH is close to the pH at which the molecules are irreversibly denatured. Chain breakage of denatured RF DNA will lead to the formation of component I I 1 single-stranded DNA or partially renatured RF DNA. The formation of different products could be shown to be due to changes in the conditions of the experiment (ionic strength and temperature). Therefore the nature of the product formed depends on external conditions but is independent of the cause of the chain break. For the infiuence of the temperature and ionic strength on the nature of the product after chain breakage in denatured RF DNA, the same interpretation can be applied as for the renaturation of linear DNA molecules. Renaturation of denatured DNA depends both on temperature and ionic strength (Britten & Kohne, 1966; Wetmur & Davidson, 1968). At a given ionic strength the rate of renaturation increases with temperature, reaches an optimum and then decreases because of a higher probability of dissociation than of reassociation of the strands at higher temperatures. According to Britten & Kohne, the rate of renaturation in buffers containing O-08 M-sodium ions is at a maximum at about 30°C below the melting temperature, and at lower salt concentration the maximum rate will occur further below the melting temperature. From the melting temperature of component II in 0.01 M-phosphate-l mnl-sodium citrate (72°C) and the data of Britten & Kohne (1966), it can be calculated that the rate of renaturation of 4X DNA in 0.01 M-phosphate-l mM-sodium citrate will be optimum at a temperature below 42”C, which is in agreement with the present, findings. From our experimental data it seems likely that the optimum renaturation temperature is approximately 25°C. At temperatures beyond the optimum renaturation temperature (e.g. 40°C) the rate of renaturation declines abruptly, and 1argelJ single-stranded DNA is formed at this temperature. In buffers of high ionic strength the single-strand character of the molecule is stabilized. Consequently higher temperatures are required for renaturation, and at, room temperature only a part,ially renatured molecule is found. The increase of the renaturation temperature due to the addition of bivalent ions (magnesium, barium or calcium) cannot be explained by ionic strength effects since the presence of sodium ions at the same ionic strength as in the experiment using magnesium resulted only in a very small increase in renaturation temperature. Boedtker (1960) has demonstrated that a 25,000 times higher concentration of sodium than magnesium ions is required to achieve comparable helix stability in tobacco mosaic virus RNA. The conclusion inferred by the author was that the much greater specificity of interaction between RNA (phosphate groups) and magnesium than between RNA and sodium, cannot be explained by a difference of the charge density of the two cations. Therefore the possibility should be considered that bivalent cations like magnesium specifically interact with phosphate groups of DNA, thus facilitating recombination of the strands (Eichhorn, 1962).

390

P. H. POUWELS

ET

AL.

It is a pleasure to thank Drs H. S. Jansz and P. Borst for their interest and critical reading of the manuscript. We also gratefully acknowledge the assistance of Mr J. Meyer in the analytical ultracentrifugation experiments. REFERENCES Boedtker, H. (1960). J. Mol. BioZ. 2, 171. Britten, R. J. & Kohne, D. E. (1966). Yearb. Carnegie In&n, 1965, p. 78. Britten, R. J. & Roberts, R. B. (1960). Science, 131, 32. Burton, A. & Sinsheimer, R. L. (1965). J. Mol. BioZ. 14, 327. Crawford, L. V. & Black, P. H. (1964). l’ilirology, 24, 388. Eichhorn, G. L. (1962). Nature, 194, 474. Eichhorn, G. L. & Clark, P. (1965). Proc. Nat. Acad. Sci., Wash. 53, 586. Freifelder, D. (1965). Proc. Nat. Acad. Sci., Wash. 54, 118. Geiduschek, E. P. (1961). Proc. Nat. Acud. Sci., Wash. 47, 950. Hagen, U. (1967). Biochim. biophys. Acta, 134, 45. Jansz, H. S., Baas, P. D., Pouwels, P. H., van Bruggen, E. F. J. & Oldenziel, H. J. (1968). J. Mol. Biol. 32, 159. Jansz, H. S. & Pouwels, P. H. (1965). Biochem. Biophys. Res. Comm. 18, 589. Jansz, H. S., Pouwels, P. H. & Schiphorst, J. (1966). Biochim. biophys. Acta, 123, 626. Kohn, K. W., Spears, C. L. I%Doty, P. (1966). J. Mol. BioZ. 19, 266. Lawley, P. D. and Brookes, P. (1967). J. Mol. BioZ. 25, 143. Linn, S. & Lehman, I. R. (1965). J. BioZ. Chem. 240, 1287. Marmur, J., Rownd, R. & Schildkraut, C. L. (1963). In Progress in Nucleic Acid Research, ed. by J. N. Davidson & W. E. Cohn, vol. 1, p. 231. New York: Academic Press. Pouwels, P. H. & Jansz, H. S. (1964). Biochim. biophys. Acta, 91, 117. Pouwels, P. H., Knijnenburg, C. M., Rotterdam, J. van, Cohen, J. A. & Jansz, H. S. (1968). J. Mol. BioZ. 32, 169. Studier, F. W. (1965). J. Mol. Biol. 11, 373. Vinograd, J., Morris, J., Davidson, N. & Dove, W. F. (1963). Proc. Nat. Acad. Sci., Wash. 49, 12. Wetmur, J. G. & Davidson, N. (1968). J. Mol. BioZ. 31, 349.