Stable, metastable and unstable cellulose solutions.

We have characterized the dissolution state of microcrystalline cellulose (MCC) in aqueous tetrabutylammonium hydroxide, TBAH(aq), at different concentrations of TBAH, by means of turbidity and small-angle X-ray scattering. The solubility of cellulose increases with increasing TBAH concentration, which is consistent with solubilization driven by neutralization. When comparing the two polymorphs, the solubility of cellulose I is higher than that of cellulose II. This has the consequence that the dissolution of MCC (cellulose I) may create a supersaturated solution with respect to cellulose II. As for the dissolution state of cellulose, we identify three different regimes. (i) In the stable regime, corresponding to concentrations below the solubility of cellulose II, cellulose is molecularly dissolved and the solutions are thermodynamically stable. (ii) In the metastable regime, corresponding to lower supersaturations with respect to cellulose II, a minor aggregation of cellulose occurs and the solutions are kinetically stable. (iii) In the unstable regime, corresponding to larger supersaturations, there is macroscopic precipitation of cellulose II from solution. Finally, we also discuss strong alkali solvents in general and compare TBAH(aq) with the classical NaOH(aq) solvent.

Studies on the nitroreductase prodrug-activating system. Crystal structures of complexes with the inhibitor dicoumarol and dinitrobenzamide prodrugs and of the enzyme active form.

The E. coli nitroreductase enzyme (NTR) has been widely used in suicide gene therapy (GDEPT and ADEPT) applications as a activating enzyme for nitroaromatic prodrugs of the dinitrobenzamide class. NTR has been previously shown to be a homodimeric enzyme with two active sites. We present here the crystal structures of the reduced form of NTR and its complexes with the inhibitor dicoumarol and three dinitrobenzamide prodrugs. Comparison of the structures of the oxidized and reduced forms of the native enzyme shows that the principal structural changes occur in the FMN cofactor and indicate that the enzyme itself is a relatively rigid structure that primarily provides a rigid structural framework on which hydride transfer occurs. The aziridinyldinitrobenzamide prodrug CB 1954 binds in nonidentical ways in both of the two active sites of the homodimeric enzyme, employing both hydrophobic and (in active site B) a direct H-bond contact to the side chain of Lys14. In active site A the 2-nitro group stacks above the FMN, and in active site B the 4-nitro group does, explaining why reduction of either nitro group is observed. In contrast, the larger mustard group of the dinitrobenzamide mustard compound SN 23862 forces the prodrug to bind at both active sites with only the 2-nitro group able to participate in hydride transfer from the FMN, explaining why only the 2-hydroxylamine reduction product is observed. In each site, the nitro groups of the prodrug make direct H-bond contacts with the enzyme; in active Site A the 2-nitro to Ser40 and the 4-nitro to Asn71, while in active Site B the 2-nitro contacts the main chain nitrogen of Thr41 and the 4-nitro group the Lys14 side chain. The related amide-substituted mustard SN 27217 binds in a broadly similar fashion, but with the larger amide group substituent able to reach and contact the side chain of Arg107, further restricting the prodrug conformations in the binding site. The inhibitor dicoumarol appears to bind primarily by pi-stacking interactions and hydrophobic contacts, with no conformational changes in the enzyme. One of the hydroxycoumarin subunits stacks above the plane of the FMN via pi-overlap with the isoalloxazine ring, penetrating deep into the groove, with the other less well-defined. These studies suggest guidelines for further prodrug design. Steric bulk (e.g., mustard rather than aziridine) on the ring can limit the possible binding orientations, and the reducible nitro group must be located para to the mustard. Substitution on the carboxamide side chain still allows the prodrugs to bind, but also limits their orientation in the binding site. Finally, modulating substrate specificity by alteration of the structure of the enzyme rather than the prodrug might usefully focus on modifying the Phe124 residue and those surrounding it.

Sequence-selective targeting of long stretches of the DNA minor groove by a novel dimeric bis-benzimidazole.

A dimeric bis-benzimidazole molecule has been designed by computer modeling to bind to a DNA sequence via the DNA minor groove that covers a complete turn of B-DNA. A series of bis-benzimidazole dimers incorporating a -O-(CH(2))(n)()-X-CH(2))(n)()-O- linker, with n = 2 or 3 and X = O or N(+)H(Me), were screened for their capacity to fit the DNA minor groove. The modeling studies enabled an optimal linker to be devised (n = 3, X = N(+)H(Me)), and the synthesis of the predicted "best" molecule, N-methyl-N,N-bis-3,3-[4'-[5' '-(2' "-p-methoxyphenyl)-5' "-1H-benzimidazolyl]-2' '-1H-benzimidazolyl]phenoxypropylamine (5), is reported. The optimized linker permits the two symmetric bis-benzimidazole motifs to maintain hydrogen-bonded contacts with the floor of the DNA minor groove. DNase I footprinting studies have shown that this ligand binds with high affinity to sequences representing approximately a complete turn of B-DNA, represented by the [A.T](4)-[G.C]-[A.T](4) motif, and only poorly to sequences of half this site size, in accord with the computer modeling studies. Compound 5 does not show acute cellular cytotoxicity, in contrast with its monomeric bis-benzimidazole precursors, yet is rapidly taken up into cells.