Binding of meso ‐tetrakis ( n ‐methylpyridiniumyl)porphyrin isomers to DNA: Quantitative comparison of the influence of charge distribution and copper(II) derivatization

Factors influencing the binding of tetracationic porphyrin derivatives to DNA have been comprehensively evaluated by equilibrium dialysis, stopped‐flow kinetics, etc., for meso ‐tetrakis (4‐ N ‐methylpyridiniumyl)porphyrin TMpyP(4). Technical difficulties have previously precluded a comprehensive study of metalloporphyrins. Since electrostatic interactions with the DNA and metal derivatization of the porphyrins have important consequences, we have investigated in greater detail two isomers of TMpyP(4) { meso ‐tetra‐kis(3‐ N ‐methylpyridiniumyl)porphyrin, [TMpyP (3)] and meso ‐tetrakis (2‐ N ‐methylpyridiniumyl)porphyrin [TMpyP(2)]} in which the position of the charged centers has been varied. A comprehensive study of the Cu(II) derivatives, e.g., CuTMpyP(4), was possible since the difficulties encountered previously with Ni (II) compounds were not a problem with Cu(II) porphyrins [J. A. Strickland, L. G. Marzilli, M. K. Gay, and W. D. Wilson (1988) Biochemistry 27 , 8870–8878]. At 25°C, the apparent equilibrium constants [ K obs ] decreased with increasing [Na + ] for all porphyrins. The K obs values were comparable for TMpyP(4) and TMpyP(3) binding to either polyd(G‐C)·polyd(G‐C) [poly[d(G‐C) 2 ]] or poly[d(A‐T)·poly[d(A‐T)]][poly[d(A‐T) 2 ]]. For the copper (II) porphyrins, the K obs values were about fivefold greater. The K obs value for CuTMpyP(2) binding to poly[d(G‐C) 2 ] was too small to measure under typical salt conditions; however, K obs for CuTMpyP(4) or CuTMpyP(3). Application of the condensation theory for polyelectrolytes suggests about three charge interactions when CuTMpyP(4), CuTMpyP(3), and TMpyP(4) bind to poly[d(G‐C) 2 ] or poly[d(A‐T) 2 ], a result comparable to that reported for TMpyP(4). At 20°C and 0.115 M [Na + ], incorporation of copper decreased the rates of dissociation from poly[d(A‐T) 2 ] by a 100‐fold compared to those reported for TMpyP(4) but had little effect on the rates of dissociation from poly[d(G‐C) 2 ]. Also, movement of the H 3 CN + group from the fourth to the third position of the pyridinium ring enhanced the rates of dissociation from poly[d(A‐T) 2 ] but decreased the rates of dissociation from poly[d(G‐C) 2 ]. From polyelectrolyte theory, the [Na + ] dependence of the dissociation rates from poly[d(G‐C) 2 ] is consistent with intercalative binding, while that for poly[d(A‐T) 2 ] is consistent with an outside binding model. For calf thymus [CT]DNA at 20°C, a greater decrease in the AT than in the GC imino 1 H‐nmr signal was observed upon addition of CuTMpyP(2), suggesting selective outside binding to the AT regions. Flow dichroism experiments with calf thymus DNA revealed a small reduced dichroism [ red D] value with CuTMpyP(2), indicative of disordered, outside binding. However, a large red D with CT DNA was found for CuTMpyP(4) and CuTMpyP(3), suggesting ordered intercalative binding. Titrations of closed circular superhelical DNA (CCS DNA) with CuTMpyP(4) and CuTMpyP(3) produced large increases in the solution reduced viscosity (SRV), indicative of unwinding. Twice the concentration of TMpyP(4) was needed for a similar effect, a result suggesting either that CuTMpyP (4) and CuTMpyP(3) intercalate into more sites, or that they produce more unwinding per site. Alternatively, CuTMpyP(4) and CuTMpyP(3) could unwind CCS DNA through a nonintercalative binding mode. Addition of CuTMpyP(4) or CuTMpyP(3) increased the SRV of poly[d(A‐T) 2 ], poly[d(G‐C) 2 ], and a variety of native linear DNAs varying in percentage GC. However, CuTMpyP(4) decreased the SRV of poly[d (A)]·poly[d(T)]. Whereas the viscosity increases with poly[d(G‐C) 2 ] probably result from intercalation, the unusual increase in the SRV of poly[d(A‐T) 2 ] could arise from conversion of a small amount of hairpin to the B form or from some other type of cooperative conformational change. Together these results suggest generally more favorable interactions with DNA of the copper porphyrins than the analogous metal‐free porphyrins. However, our evidence clearly rules out appreciably greater GC vs AT selectivity for the copper porphyrins. Electrostatic interactions with DNA are similar for the TMpyP(4) and TMpyP(3) species, but are evidently much less favorable for the TMpyP(2) porphyrins. Finally, we find no clear evidence for additional binding modes for copper porphyrins.