The therapeutic and commercial success of phosphodiesterase 5 inhibitors such as Viagra, Levitra and Cialis has sparked renewed interest in the phosphodiesterases as drug discovery targets. Virtually all the phosphodiesterases are expressed in the CNS, making this gene family a particularly attractive source of new targets for the treatment of psychiatric and neurodegenerative disorders. Significantly, all neurons express multiple phosphodiesterases, which differ in cyclic nucleotide specificity, affinity, regulatory control and subcellular compartmentalization. Therefore, phosphodiesterase inhibition represents a mechanism through which it could be possible to precisely modulate neuronal activity. In this article, we review the current state of the art in the burgeoning field of phosphodiesterase pharmacology in the CNS.

The hydrolysis of cephalosporins containing good leaving groups at the 3'-position [those used in this study were the chromogenic cephalosporin PADAC [pyridine-2-azo-4'-(N',N'-dimethylaniline) substituted on cephalosporin], cephaloridine, and cephalothin], catalyzed by the Staphylococcus aureus PC1 beta-lactamase, proceeds in two spectrophotometrically observable phases. The first involves formation of an acyl-enzyme intermediate while the second involves partitioning of this intermediate between two pathways. One path yields the normal cephalosporoate (3) from which the 3'-leaving group is spontaneously eliminated in solution to give the 3-methylenedihydrothiazine 2, while the second involves initial elimination of the 3' substituent, thus yielding a second acyl-enzyme intermediate, which then hydrolyzes to give the same final product as from the first pathway. The second acyl-enzyme is relatively inert to hydrolysis (t1/2 congruent to 10 min at 20 degrees C), and its formation thus leads to transient inhibition of the enzyme. The partition ratio between hydrolysis and elimination at the enzyme active site could be determined either spectrophotometrically from burst experiments or from measurements of residual beta-lactamase activity as a function of cephalosporin concentration. This ratio varied with the leaving group ability of the 3' substituent (acetoxy greater than N,N-dimethylaniline-4-azo-2'-pyridinium greater than pyridinium) in the anticipated fashion. The inert acyl-enzyme intermediate was isolated by exclusion chromatography and shown to contain the cephem nucleus, but not the 3' substituent, covalently bound to the enzyme. As would be expected, PADAC, cephaloridine, and cephalothin yielded the same inert intermediate. Cephalosporins with poor or no 3'-leaving groups, e.g., dansylcephalothin and desacetoxycephalothin, neither displayed the branched pathway nor yielded the long-lived acyl-enzyme.

Since ribosomally mediated protein biosynthesis is confined to the L-amino acid pool, the presence of D-amino acids in peptides was considered for many years to be restricted to proteins of prokaryotic origin. Unicellular

Inhibition of alanine racemase from the Gram-positive bacterium Bacillus stearothermophilus by (1-aminoethyl) phosphonic acid (Ala-P) proceeds via a two-step reaction pathway in which reactivation occurs very slowly. In order to determine the mechanism of inhibition, we have recorded low-temperature, solid-state 15N NMR spectra from microcrystals of the [15N]Ala-P-enzyme complex, together with spectra of a series of model compounds that provide the requisite database for the interpretation of the 15N chemical shifts. Proton-decoupled spectra of the microcrystals exhibit a line at approximately 150 ppm, which conclusively demonstrates the presence of a protonated Ala-P-PLP aldimine and thus clarifies the structure of the enzyme-inhibitor complex. We also report the pH dependence of Ala-P binding to alanine racemase.

Alanine racemases are bacterial pyridoxal 5'-phosphate (PLP) dependent enzymes providing D-alanine as an essential building block for biosynthesis of the peptidoglycan layer of the cell wall. Two isozymic alanine racemases, encoded by the dadB gene and the alr gene, from the Gram-negative mesophilic Salmonella typhimurium and one from the Gram-positive thermophilic Bacillus stearothermophilus have been examined for the racemization mechanism. Substrate deuterium isotope effects and solvent deuterium isotope effects have been measured in both L----D and D----L directions for all three enzymes to assess the degree to which abstraction of the alpha-proton or protonation of substrate PLP carbanion is limiting in catalysis. Additionally, experiments measuring internal return of alpha-3H from substrate to product and solvent exchange/substrate conversion experiments in 3H2O have been used with each enzyme to examine the partitioning of substrate PLP carbanion intermediates and to obtain the relative heights of kinetically significant energy barriers in alanine racemase catalysis.