Single-stranded DNA Structure and DNA Polymerase Activity in the Presence of Nucleic Acid Helix-unwinding Proteins from Calf Thymus”

In the preceding articles we have described the isolation and some of the properties of two calf thymus proteins which bind selectively to single-stranded DNA and which appear analogous to previously isolated prokaryotic DNA-unwinding proteins. In the present work we demonstrate two further points of analogy. First, both the calf UP1 and the high salt eluting proteins form protein-rich complexes with single-stranded DNA, and hold this DNA in a rigid, extended conformation. Second, these proteins stimulate the calf thymus DNA polymerase-a; phage T4 gene 32-protein does not. The stimulation of a homologous DNA polymerase is characteristic of several prokaryotic DNA-unwinding proteins and is assumed to reflect their in uiuo role in DNA synthesis. Bacteriophage known genetic vitro experiments show that this DNA-unwinding specifi-cally stimulates DNA polymerase, and that is a weak interaction between 32-protein and polymerase in the Similar


Bacteriophage
T4 gene 32-protein is known from genetic analysis to be required for T4 DNA replication (1). In vitro experiments show that this DNA-unwinding protein specifically stimulates the T4 DNA polymerase, and that there is a weak interaction between 32-protein and polymerase in the absence of DNA (2). Similar stimulations of homologous DNA polymerases have been demonstrated with the Escherichia coli and T7 . The 32-protein and E. coli unwinding proteins have been shown to form tight complexes with single-stranded DNA in which all sequences of the DNA are bound and the DNA is held in an extended, regular conformation; however, the two complexes differ in contour length per nucleotide, and in composition (3,7). The specific relationship between homologous polymerases and unwinding proteins could reflect differences in the conformation of the template strand associated with unwinding protein, further specificity being derived from direct unwinding proteinpolymerase interactions.
In this report, we present properties of the single-stranded DNA complexes which are formed with the calf thymus helix-unwinding proteins described in the two preceding articles (8,9). Both UPl' and the high salt eluting protein fraction * This research was supported by grants from the National Institutes  (13,14).

MATERIALS AND METHODS
The calf thymus single-strand-specific, dextran sulfate-resistant DNA-binding proteins were isolated as described in an accompanying article (8  7). Shown in Fig. 1 are analogous electron micrographs of single-stranded fd bacteriophage DNA circles, associated with an excess of either the calf high salt eluting proteins (Fig. 1, a and b) or the calf UP1 (Fig. 1~). The fixed complexes are extended circles, apparently uniformly coated with protein. These protein-covered DNA strands are not as thick as those seen with prokaryotic DNA-unwinding proteins, and thus are more difficult to distinguish from the naked single-stranded DNA. The fd DNA contour length in the complexes formed with either of these proteins is about 2.4 pm, about 0.4 pm longer than that of the naked DNA spread in the same conditions. For comparison, the contour length of fd DNA associated with T4 32-protein was found to be about 1.1 pm longer, while the E. coli protein complex was about 0.6 pm shorter, than the naked DNA (3, 7).
Note that the fixed complexes in Fig. 1, b and c remain extended even when spread in the presence of 10 mM Mgz+ or 1 mM spermine, conditions which induce a drastic intrastrand folding in naked, single-stranded DNA (7). (The complexes contract slightly and adopt a somewhat irregular shape, possibly as a consequence of incomplete fixation.) Thus, as in the case of the prokaryotic DNA-unwinding proteins, the associated DNA appears to be rigidly held in a conformation that does not permit intrastrand DNA hairpin helix formation. This is consistent with the disappearance of ultraviolet hypochromicity, normally associated with single-stranded DNA secondary structure, upon complex formation (9).
As expected from their DNA melting properties (9), both the high salt eluting proteins and UP1 induce denaturation bubbles in SV40 supercoiled, double-helical DNA. This result with UP1 is shown in Fig Direct determination of the stoichiometry of the saturated DNA.protein complex was carried out on sedimented complexes formed with excess UP1 (Fig. 2B, input protein:DNA weight ratios of 13:l and 24:l). The protein:DNA weight ratio in the peaks was found to be 12:l and 9.1:1, respectively. Taking the weight ratio to be lO:l, we calculate that each UP1 molecule would occupy a single-stranded DNA site about seven nucleotides long in the saturated complex. The protein:DNA weight ratio is in the same range as that obtained by similar 'Note that the Escherichia coli DNA-unwinding protein, which binds to DNA in a highly cooperative manner, gives two distinct and separated DNA peaks in such gradients if the protein is present in subsaturating amounts; one peak consists of nearly fully saturated complexes, and the other of essentially free DNA (3). However, the filamentous phage gene 5 protein, which also binds cooperatively (16), gives results of the same type as seen in Fig. 2A (18). Thus, the data of Fig. 2A do not necessarily imply that UP1 binds to DNA non-cooperatively. Electron microscopy of such subsaturated mixtures of UP1 and fd DNA also fails to show distinct regions of associated and unassociated DNA with either UP1 or the high salt eluting proteins (unpublished results); this, too, is a rather stringent test of cooperativity, and conclusions should be guarded.
techniques with prokaryotic "DNA-unwinding" proteins: 12: 1 for the phage T4 gene 32-protein (19); 7:l for the Ff gene 5 protein (16, 18), and 8:l for the E. coli DNA-unwinding protein (3). Fig. 2B also contains a sedimentation profile obtained under identical conditions for fd DNA saturated with the T4 gene 32.protein. Both sedimentation and electron microscopy (7, 20) suggest that this complex adopts a greatly extended conformation with a high frictional coefficient. The UP1 complex sediments even slower than the 32-protein complex. In fact, the frictional coefficients of the two complexes are nearly identical, the faster sedimentation of the 32-protein complex being a consequence of its somewhat larger mass. Thus, the sedimentation data provide independent evidence that the UP1 complex with single-strand DNA is highly extended in solution.
Stimulation of Calf Thymus DNA Polymerase-a-The effect of UP1 on the activity of the calf thymus DNA polymerase-a was tested, using as the template native DNA partially degraded with E. coli exonuclease III. With a quantity of UP1 approximately sufficient to cover the single-stranded DNA "tails" on this template, the rate of nucleotide incorporation was stimulated lo-fold (Fig. 3); higher amounts of UP1 inhibit the polymerase. In both respects, these results parallel those with prokaryotic unwinding proteins (2,3). Note that no stimulation of the calf DNA polymerase was obtained with additions of T4 gene 32-protein (Fig. 3), showing that mere removal of secondary structure from the template is not sufficient for the DNA polymerase stimulation.
Slight (<a-fold) stimulations of this polymerase by UP1 have been observed both using heat-denatured DNA templates, and several synthetic Chang-Bollum "hook polymers" (17). In addition, strong stimulation (>5-fold) was found for oligo(dG)-primed synthesis of poly(dG) on a poly(dC) template (15): For polymerase stimulation, a UP1 fraction composed primarily of the most basic subspecies (as judged by isoelectric focusing (8)) is the most effective. For example, in experiments using the exonuclease III-degraded DNA template, the optimal protein to single-stranded DNA weight ratio for stimulation was about 5:l for the most basic UP1 subspecies, as against 27:l for the most acidic subspecies. The basic subspecies is also the one which binds to DNA the most tightly (8).
The high salt eluting proteins behave similarly to UP1 in many respects (8,9,and Fig. 1). Preliminary studies using the poly(dC) oligo(dG) template-primer reveal this fraction to be even more efficient than the basic subspecies of UP1 in stimulating the calf DNA polymerase-a (15). Further characterization of this protein fraction must await the development of better purification procedures for its several components. Three reports of mammalian DNA polymerase-stimulating proteins have appeared (21, 22, 23), but details necessary for comparison with our results have not been published.
The third major calf DNA-binding protein fraction obtained by our isolation procedures does not appear to be a helixunwinding protein by several criteria. Designated as the "33,000 dalton, low salt eluting protein" (8), it formed no detectable complex with fd DNA as judged by the electron microscopic criterion applied to the two other calf protein fractions (Fig. 1). Moreover, it is not capable of strong induction of DNA melting (9). Over a range of concentrations it has a marked inhibitory effect on the calf DNA polymerase-a S. Imada, G. Herrick, and F. Bollum, unpublished results. high salt eluting proteins. The fd DNA (7 &ml) was incubated with single-strand-specific, dextran sulfate-resistant DNA-binding proteins the protein (160 &ml) for 10 min at 37" in 10 mM potassium from calf thymus. a, complex between fd single-stranded DNA and the phosphate/l mM Na&DTA, pH 7. As detailed in Ref. was between 65 and 80%. Inset, the peak position in each gradient is plotted against the amount of protein added (0); the same type of data, obtained from the experiment in b, are also plotted (0). b, comparison with T4 gene 32-protein.
Performed as in a, except that the sucrose gradients contained 5 mM Tris-HCl, pH 7.4 at 20", 1 mM Na,EDTA, 50 pg/ml of bovine serum albumin, and 10% glycerol. Two samples contained 7.6 and 14.3 pg, respectively, of UPl, while a third sample contained 18 pg of bacteriophage T4 gene 32-protein.
Each sample received 0.60 pg of fd [sH]DNA, as in a. In a parallel run, a sample containing 4.8 pg of UP1 was sedimented without DNA; gel electrophoresis of gradient fractions showed that all of this protein remained near the top of the gradient. As expected, the gradient with 14.3 pg of UP1 had considerable free protein left near the top, while the gradient with 7.6 Gg of UP1 had very little. No protein was detected in either of these gradients between the position of free protein and that of the DNA-protein complex. Iri determining the protein to DNA ratio of complexes, the UP1 concentration in each gradient fraction was assayed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by staining with Coomassie brilliant PROTEIN ADDED (i/g) blue (8). Band intensity was quantitated by integration of densitometer tracings with a Du Pont curve resolver, compared with parallel gels run with known amounts of the protein.
In controls, dye intensity was proportional to the amount of UP1 applied to the gel between 0.5 and 10 pg. The presence of DNA does not interfere with quantitation of UP1 by this method.
The concentration of the DNA was calculated from the aH label and the known specific activity of the DNA. FIG. 3 (right) [26][27][28][29][30][31][32] suggest that this DNA polymerase is involved in nuclear DNA replication, and it is thus possible that one or both of the two calf thymus helix-unwinding proteins are also involved in DNA replication.
Note that there is a close analogy between the properties of these two proteins and the properties of the T4 gene 32-protein and the E. coli "DNA-unwinding protein," both of which appear to be involved in prokaryotic DNA synthesis. (a) All four proteins can be isolated as single-strandspecific, dextran sulfate-resistant DNA-binding proteins; (b) while all have nonbasic isoelectric points, their DNA binding is salt-sensitive and thus relies in part on ionic interactions with the DSA phosphates; (c) these proteins all hold singlestranded DNA in an extended complex which is unineme and protein-rich; (d) these complexes show nearly complete DNA hyperchromicity, indicating that the bases are held in a nonstacked conformation; (e) by virtue of the above properties and their relative lack of affinity for native DNA, these proteins all have the ability to denature native DNA; (f) all of these proteins stimulate a homologous DNA polymerase.
On the other hand, there are two major differences between 32-protein and the calf proteins.
First, as is true for the E. coli protein (3), no catalysis of DNA renaturation has been observed with the calf proteins (9). Second, we have failed to demonstrate cooperativity of binding of the calf proteins, either by sedimentation (Fig. 2) or by electron microscopic analysis (15). These two negative results may reflect insensitive or inappropriate testing conditions, or they may be real, suggesting that these properties may not be essential to all facets of unwinding protein function. For instance, the renaturation-catalyzing ability of 32-protein may only be essential to its role in genetic recombination (33). 20. 21. 22.