Silver Sulfadiazine: Interaction with Isolated Deoxyribonucleic Acid

Silver sulfadiazine (AgSu) was found to interact with isolated deoxyribonucleic acid (DNA) to form nondissociable complexes. These complexes differ in physical and chemical properties from those that are established when silver nitrate is added to DNA. The reaction between AgSu and DNA is visualized as occurring in two stages: (i) a weak and reversible interaction (intercalation) between DNA and the sulfadiazine moiety and (ii) a tight binding involving the silver atom. In the first stage, sodium sulfadiazine competes with AgSu for the DNA.

Silver sulfadiazine (AgSu) was synthesized by Fox, who also established its effectiveness for the prevention and treatment of burn infections due to pseudomonads (8)(9)(10)28). Even though AgSu is a sulfonamide drug, its antibacterial activity is not a function of the sulfonamide moiety because its effectiveness is not reversed by p-aminobenzoic acid (10). Silver nitrate (AgNO3) reacts with deoxyribonucleic acid (DNA) in vitro (14, 31; see also 3 and 33); hence, it has been suggested (10) that AgSu reacts with the cellular DNA and that this forms the basis of its antibacterial action. It has been proposed further (10) that when AgSu reacts with the cellular DNA it dissociates and that only the silver ion becomes associated with DNA while the sulfadiazine portion is released. However, data on both the in vivo and in vitro effects of AgSu are scarce, and therefore an examination of its mode of action was undertaken. The present study shows that in vitro AgSu forms a complex with DNA which is unlike that formed when DNA is mixed with AgNO3; but it is also shown (25; J. E. Coward, H. S. Carr, and H. S. Rosenkranz, in preparation) that this in vitro reaction between DNA and AgSu is not the basis of the latter's antibacterial action. The biological activity of AgSu appears to derive from its effect on the external bacterial structures, as documented in the accompanying report (25; also Coward, Carr, and Rosenkranz, in preparation).
MATERIALS AND METHODS Deoxyribonucleates. Calf thymus DNA was obtained either from Nutritional Biochemicals Corp. or from Calbiochem. Pseudomonas aeruginosa DNA was prepared by the procedure of Marmur (20), and bac-teriophage DNA was isolated from Klebsiella phage E-1, currently under investigation in this laboratory. 3H-labeled DNA derived from chicken embryos was prepared as described previously (13). All DNA specimens were dissolved in 0.005 M NaNO3 and extensively dialyzed against 0.005 M NaNO3 to remove all traces of chloride (capable of reacting with AgNO3 to form AgCl).
Chemical analysis. The phosphorus content of DNA specimens was determined by the procedure of Fiske and Subbarow (7).
Spectral analysis. The spectra of DNA solutions were taken with a Carey model 11 recording spectrophotometer, and the absorbance at selected wavelengths was checked with a Beckman DU-2 spectrophotometer.
Centrifugal procedures. Sedimentation velocity studies were carried out on dilute solutions in 12-mm Kel-F centerpiece centrifuge cells. A Spinco model E analytical ultracentrifuge equipped with an ultraviolet optical system was operated at 50,740 rev/min, and pictures were taken at 2-min intervals. The photographs were traced with a Joyce-Loebl Mark III B microdensitometer, and sedimentation coefficients (S5".) and distributions of sedimentation coefficients were calculated by a modification (H. S. Rosenkranz, Ph.D. thesis, Cornell Univ., Ithaca, N.Y., 1959) of the procedure of Schumaker and Schachman (27). The ultracentrifugal heterogeneity of DNA specimens was expressed as the value of one standard deviation divided by the 5"So value (24; Rosenkranz, Ph.D. thesis). The amount of ultraviolet-absorbing nonsedimenting material was calculated from the microdensitometer tracings as previously described (13). 373 The banding properties of DNA in gradients of CsCl were determined as described by Schildkraut et al. (26). Portions of the DNA, together with a reference sample (Micrococcus lysodeikticus DNA, 1.731 g/cm3), were placed in a CsCl solution (density, 1.70 g/cm3) and centrifuged at 44,770 rev/min for 24 hr. The bands formed by the specimens at their equilibrium positions were photographed and traced as above.
For Cs2SO4 density gradient analyses, mixtures containing 0.3 ml of radioactive DNA solution, 1.7 ml of 0.005 M NaNO3, and 1.33 ml of saturated Cs2SO4 (23 C) solution were spun at 30,000 rev/min in the SW-50 rotor of a Spinco model L-2 ultracentrifuge for 39 hr. Fractions were collected as previously described (5) and were assayed for acid-precipitable radioactivity. The refractive indices of selected fractions were determined. These values were used to calculate buoyant densities (30).
For zonal centrifugations in gradients of sucrose, radioactive DNA (0.3 ml) was layered on top of 4.5 ml of a gradient (5 to 20%) of sucrose dissolved in 0.01 M tris(hydroxymethyl)aminomethane (Tris) buffer, pH 7.0. The solutions were spun at 30,000 rev/min in the SW-50 rotor of a Spinco model L-2 ultracentrifuge for 8 hr. Fractions were collected and processed as described above.
Thermal transition profiles. Thermal transition profiles were measured with a Beckman DU spectrophotometer in the manner described by Marmur and Doty (21). All DNA specimens were dissolved in 0.005 M NaNO3. RESULTS Effect of AgSu on spectral properties. Addition of AgNO3 and AgSu to DNA resulted in a shift of the absorption maxima from 257.5 to 263 nm (Fig. 1). Such an effect of AgNO3 has been described previously (14). Further analysis indicated, however, that the effects of the two silver compounds were not identical, as evidenced by differences in hyperchromic shifts and ratios of absorbancies at selected wavelengths ( Table 1). The lack of effect of AgSu (and AgNO3) on the absorbance in the visible range (300 to 450 nm) indicates that no precipitation of DNA occurs when these substances interact with DNA. Contrariwise, when polyamines or basic polypeptides are complexed with DNA, this results in an absorbance in the visible range due to aggregation (light-scattering effect; 2, 4, 6,11,12).
Effect on the sedimentation coefficient of DNA. The addition of increasing amounts of AgNO3 to calf thymus DNA resulted in increases in the sedimentation coefficients ( Table 2, experiment V). On the other hand, no such concentration dependence was observed when AgSu was added to DNA (Table 2). Thus, a maximal effect was observed when the molar ratio of AgSu to DNAphosphorus (AgSu/DNA-P) was 0.5. Moreover, it was consistently found that the maximal effect   Fig. 1. The absorbance at selected wavelengths was checked with a Beckman DU-2 spectrophotometer.
of AgSu was always smaller than that produced by the addition of equimolar amounts of AgNO3 (Table 2). (In no instance was there a decrease in the sedimentation coefficient of DNA upon exposure to either AgSu or AgNO3 ; i.e., DNA was not degraded.) Analysis of the amount of ultraviolet-absorbing material sedimenting in mixtures of AgSu and DNA revealed that such mixtures contained ultraviolet-absorbing nonsedimenting species, presumably unbound AgSu (AgSu absorbs in the ultraviolet). However, as the time of exposure of  (Fig. 3) upon addition ). Solutions of AgSu, it stiUl exhibited the same distribution of ialyzed in a sedimentation coefficients ( Table 4). The sediuge.
mentation coefficients of high-molecular-weight DNA species (Klebsiella bacteriophage and P. AgSu in aeruginosa) also increased upon addition of AgSu ( Fig. 3). In these cases, too, there was no sig-Per cent nificant change in the distribution of sedimentalonsedimenting tion coefficients (Table 4).
These data indicate that DNA species of dissim-0 ilar molecular weights, different sources (mam- a The sedimenting boundaries of mixtures of DNA + AgSu and DNA + AgNO3 (in each case the ratio of DNA-phosphorus to silver was 1.0) were analyzed as a function of time. Calculation of the percentage of nonsedimenting material was performed as described previously (13). the DNA to AgSu increased, the amount of unbound ultraviolet-absorbing material (AgSu) decreased (Table 3). This finding suggests that the reaction between DNA and AgSu is a slow one and that it will be difficult to determine the exact molar ratio of AgSu bound per DNA-P   Helix-to-coil thermal transitions. The presence of AgNO3 had a profound effect on the thermal denaturation profile of calf thymus DNA (Fig.  4). Thus, low levels of AgNO3, i.e., AgNO3/ DNA-P of 0.02 and 0.21, increased the Tm, the midpoint of the thermal transition curve, by 3.1 and 15.2 C, respectively. Under these conditions, however, complete denaturation was achieved, as evidenced by the extent of the hyperchromic shift (Fig. 4). At an AgNO3/DNA-P ratio of 0.5, denaturation required higher temperatures and was incomplete even at 95 C. When the molarity of AgNO3 equaled that of the nucleotides (i.e., AgNO3/DNA-P = 1.0), denaturation was completely prevented (Fig. 4).
These effects of AgNO3 on the thermal denaturation of DNA were reproducible, and they were independent of the period of incubation which preceded the measurement.
AgSu also had a profound effect on the thermal denaturation profile of DNA (Fig. 5). This effect depended upon AgSu concentration and was also influenced by the duration of the contact between DNA and AgSu (Fig. 5). As the time of incubation increased, the effect on the thermal denaturation profile became more pronounced. A similar effect was also observed by sedimentation velocity analysis (Table 3). However, unlike the effects of silver substances on the sedimentation behavior of DNA (Table 2), AgSu was almost as potent as AgNO3 in modifying the "melting-out" behavior of DNA (compare Fig. 4 and 5). Although AgSu-DNA complexes exhibited increased resistance to thermal denaturation, once denaturation occurred, it was irreversible, as illustrated by the data summarized in Fig. 6. This indicates that AgSu did not cross-link the DNA strands and thereby cause renaturation.
Cesium chloride density gradient centrifugation. Specimens of DNA, DNA plus AgSu, and DNA plus AgNO3 were placed in CsCl and analyzed in an analytical ultracentrifuge by buoyant density gradient centrifugation. The tracings of the DNA bands at equilibrium are shown in Fig. 7. The behavior of the AgNO3-DNA complex was similar to that of the control DNA ( Fig. 7A and B). This presumably resulted from the presence of concentrated CsCl and dissociation of the AgNO3-DNA complex owing to the formation of insoluble AgCl (14). However, the banding behavior of the DNA-AgSu complex (Fig. 7C) differed from that of the others both in position and in shape. This unexpected finding suggests that, whereas the AgNO3-DNA complex is dissociated by CsCl, the AgSu-DNA complex is not. The increase in band width of the AgSu-DNA complex (Fig. 7C) indicates a heterogeneity of attachment sites and a possible preferential reaction of AgSu with certain species of DNA.
Cesium sulfate density gradient centrifugation. AgCl is insoluble, but Ag2SO4 is not; this has permitted the use of Cs2SO4 density gradient centrifugation for the analysis of the properties of AgNO3-DNA complexes (14). It was thus found (Fig. 8) that the AgNO3-DNA complex had a buoyant density higher than that of the control DNA; this confirms the earlier finding of Jensen and Davidson (14) and presumably reflects the increased molecular weight of the silver nucleate DNA and AgNO3-DNA complexes differed in their banding behavior, it should be noted that neither was dissociated by high salt concentra-1 0 tions (i.e., Cs2SO4), as evidenced by the differences between their banding behavior and that of the control DNA.
Zonal centrifugation in sucrose. The rate of migration of macromolecules in gradients of sucrose is a function of their molecular weights 0 10 20 (22). Analyses of silver-DNA complexes in such gradients showed (Fig. 9) that the rate of sedimentation of the AgNO3-DNA complex was far  slightly, but reproducibly, faster than that of the control DNA (Fig. 9).
Effect of cyanide on silver-DNA complexes.
Silver ions form soluble complexes wit cyanide (32), and indeed it has been reported tht AgNOa-DNA complexes can be dissociated by cyanide (3,14). This reversal of the AgNO3-DNA complexes was confirmed in the present study: (i) KCN caused a decrease in the sedimentation coefficients of AgNO3-DNA complexes (unpublished data) and (ii) addition of KCN to AgNO3-DNA complexes caused the thermal helix-coil transition profiles to become indistinguishable from that of untreated DNA (unpublished data). However, the effect of KCN on AgSu-DNA complexes could not be studied by this technique because KCN forms strongly ultraviolet-absorbing complexes with both AgSu and NaSu (unpublished data). To overcome this effect of KCN, advantage was taken of the availability of 3H-labeled DNA. It was thus shown that addition of KCN to AgNO3-DNA restored the rate of sedimentaion in a sucrose gradient to a level similar to that of the unmodified DNA (Fig. 10B); i.e., the complex dissociated. On the other hand, addition of KCN to the AgSu-DNA complex did not result in dissociation of the complex (Fig. 10A).
Effect of NaSu in DNA. The addition of NaSu to DNA did not significantly affect the sedimenta-tion behavior of calf thymus DNA (Table 5). An analysis of the sedimentation boundaries did not reveal any NaSu co-sedimenting with the DNA.
Although NaSu had no effect on the sedimentation behavior of DNA, it greatly influenced the thermal helix-coil transition profile (Fig. 11). This effect was not due to a contribution of NaSu to the ionic strength of the solution as an equivalent amount of NaNO3 or NaCl had no such effect. Dialysis of NaSu-DNA mixtures against 0.005 M NaNO3 completely reversed this effect of NaSu. (Bound AgSu was not removable by dialysis [see below].) Exposure of DNA to mixtures of NaSu and AgSu revealed that NaSu competed with AgSu for the poiydeoxynucleotide. Thus, the maximal effect of AgSu on the sedimentation coefficient of DNA was inhibited by the simultaneous presence of NaSu (Table 5). On the other hand, addition of NaSu after the formation of the AgSu-DNA complex was without effect on the sedimentation coefficient. Thus, the 1 :1 Klebsiella phage DNA-AgSu complex had a sedimentation coefficient of 43.6S. When NaSu (NaSu/DNA-P = 1.0) was added after the formation of this complex (t = 24 hr) and the mixture was incubated for another 24     Table 3). This 0 70 80 90 100 suggested that even in the presence of NaCl, the AgSu was still bound to the DNA. When the Temperature (°C) NaCl was removed by dialysis and no further 11. Effect of sodium suliadiazinie (NaSu) on AgSu was added, the sedimentation coefficient ermal helix-coil transitiont of calf thymus DNA. increased once more (Table 6); moreover, undashed line, control DNA; solid line, NaSu + bound AgSu had been removed (Table 6).
cleates form complexes with ethidium bromide, chloroquine, and proflavine; in each of these ine sedimentation coefficient was found to be stances, there is intercalation of the chemical (i.e., unchanged within experimental error). between adjacent base pairs. These interactions ect of NaCl on silver-DNA complexes. The with DNA can be detected in the spectra of these -DNA complex appears to be undissociated substances (1,(16)(17)(18). It was found that even ther CsCl or Cs2SO4 as determined by excessive amounts of AgSu (AgSu/DNA-P = nt density centrifugation (see Fig. 7 and 8). 1.0) did not interfere in the formation of com-Lbove results with NaSu suggest, however, plexes between DNA and ethidium bromide, in addition to this "irreversible" binding, a proflavine, and chloroquine (unpublished data).   DISCUSSION The antibacterial effectiveness of AgSu has been attributed (10) to its ability to penetrate into the bacterial cell wherein it dissociates, thereby allowing the silver ion to interfere with the base pairs of DNA in a manner similar to the one shown to occur in vitro between DNA and AgNO3 (3,14,33). In the present study, it is shown that AgSu interacts with isolated DNA but that the product is different in all respects from that obtained when AgNOs is added to DNA. It would appear that one of the foremost differences between AgNO3 and AgSu is the inability of AgSu to ionize. (Thus, addition of NaCl to AgSu does not result in formation of insoluble AgCl.) AgSu must thus be viewed as a nondissociable molecule.
It is shown that AgSu reacts with DNA species of various sizes (4.9 to 38.2S), sources (mammalian, avian, bacterial, and viral), and base compositions (47 to 67% guanine plus cytosine).
The sedimentation coefficients of macromolecules are determined by size and shape, whereas rates of migration in gradients of sucrose have been equated with size (22). AgNO3-DNA complexes show great increases in sedimentation coefficients and in rates of sedimentation in sucrose, which suggests that they possess vastly increased molecular weights. AgSu-DNA complexes also display increased sedimentation coefficients, but their rates of migration in sucrose gradients are nearly identical to that of unmodified DNA. These facts can be interpreted as signifying that the main effect of AgSu on DNA involves an alteration in shape with a minimal effect on the molecular weight. This is in agreement with the observation that even when the AgNOs concentration greatly exceeds the DNA-phosphorus ratio (>200:1) the sedimentation coefficient still increases (Table 2). On the other hand, AgSu exerts its maximal effect at an AgSu/DNA-P ratio of 0.5, a -d even at that concentration a large proportion of the AgSu remains unbound.
The observation that unlike AgNOa-DNA, AgSu-DNA is not dissociated by NaCl, CsCl, and KCN may be a result of the lack of ionization of the AgSu molecule, or it may indicate that the AgSu to DNA bond is very tight. The fact that the complex is not dissociated by dialysis against NaCl suggests the second possibility. The observation that after dialysis of the AgSu-DNA complex the amount of nonsedimentable material, i.e., unbound AgSu, was eliminated (Table 6) indicates that the dialysis procedure was efficient and that the "real" amount of AgSu bound is much smaller than the amount added. Moreover, the temporary return, in the presence of NaCl, to a lower sedimentation coefficient may indicate that the shape of the molecule was restored even while AgSu was still attached. This could be due to cswinging out" of the sulfadiazine moiety while silver was still attached (see below), thereby leading to a temporary restoration of shape. The most puzzling observation is that of a weak association between NaSu and DNA. This association appears to be involved also in the reaction between AgSu and DNA, as AgSu and NaSu appear to compete for the DNA. On the other hand, if AgSu is added first, NaSu no longer competes with it. This leads to the suggestion that the reaction between AgSu and DNA proceeds in two steps: a weak interaction between DNA and the sulfadiazine moiety, followed by a strong binding between the silver atom and the DNA. It is possible that the sulfadiazine moiety puts the AgSu molecule in place so that the second step can take place.
Concerning the nature of these reactions, conceivably the interaction between the sulfadiazine moiety and the DNA could involve an intercalation: indeed, the spectral hyperchromicity (Fig. 1) and the decreased buoyant density (Fig. 8) support such a suggestion (15,16,19,29). The nature of this hypothetical intercalation remains to be elucidated. AgSu does not compete with pro-flavine, ethidium bromide, or chloroquine, which may indicate either that a different sort of intercalation is involved or that the binding between DNA and these dyes is much more efficient than between it and AgSu. This last possibility is reinforced by the fact that the binding between DNA and AgSu is a slow process and that in reality only a few molecules of AgSu are bound per DNA molecule, even when the ratio of added AgSu to DNA-P is 1.0.
It has been shown that the shift in the absorption maximum that occurs when AgNO3 is added to DNA is due to binding of the Ag+ to base-pairs. It is interesting, therefore, that the AgSu-DNA complex also exhibits this shift (Fig. 1), and by analogy it can be suggested that the same bonds are involved in the binding of AgSu to DNA. If this is true, then it can be visualized that the binding of AgSu involves (i) intercalation between base pairs followed by (ii) a strong binding between AgSu and base pairs. Since the binding of silver is strongest for guanine-cytosine base pairs (3,14), this in effect would make AgSu an intercalator with a preference for GC group base pairs.
In the present study, it was shown that AgSu was capable of binding to DNA species of different base compositions but, owing to the difficulty in assessing exact binding, it was not possible to determine whether the interaction was greater for some preparations than for others. However, the finding that after addition of AgSu to calf thymus DNA there was an increased banding heterogeneity (in CsCl) suggests that there are some molecules that are preferential substrates for AgSu. The recent availability of radioactive AgSu should permit a better determination of the "binding constant" as well as the role of base composition in the binding. In addition, the nature of the binding (intercalation) may also be revealed by use of radioactive AgSu. Such studies are proceeding in this laboratory.
It is interesting as well as important to note that none of the interactions that have been detected in the present study appear to be determinants of the biological activity of AgSu. Studies on the antibacterial action of AgSu are presented in the accompanying report (25; also Coward, Carr, and Rosenkranz, in preparation).