Critical role of conserved histidine pairs HNXXH and HDXXH in recombinant human phosphodiesterase 4A.

Cyclic AMP-Phosphodiesterases (cAMP-PDEs) catalyse the hydrolysis cAMP to AMP and thus serve to modulate the ligand-->adenylate cyclase-->cAMP-->PKA signal transduction pathway. PDEs exist as a multigene family of enzymes that bear significant sequence homology in the catalytic domains. The sequence alignment of these domains has revealed the presence of two histidine motifs: motif I, HNXXH, and motif II, HDXXH. These amino acid sequences are canonical motifs, which act as ligands for divalent metal cations required for catalytic activity. In this paper, we report human monocyte PDE4A to be a zinc-binding protein. Substitution by site-directed mutagenesis of either histidine in motif I by serine, which is not a ligand for metals, results in complete loss of catalytic activity and loss of sensitivity to divalent metal cation activation. However, similar mutations in motif II gave proteins that retained both approximately 50% of initial activity and the ability to respond differentially to Mg2+, Mn2+ and Co2+. Moreover the motif II mutants exhibited both functional group requirements and retained their pKa values. When the inactive mutants were affinity-labelled with 8-BDB-TcAMP and probed with antibody against cAMP or antibody against PDE4A, Western blots were unaltered. These results show that the conserved histidines in motif I are an absolute requirement for catalytic activity, whereas motif II histidines are required only to achieve maximum activity.

Role of conserved histidines in catalytic activity and inhibitor binding of human recombinant phosphodiesterase 4A.

To identify critical amino acids within the central conserved region of recombinant human cAMP-specific phosphodiesterase 4 subtype A (rhPDE4A), we engineered the expression of point mutants in a fully active rhPDE4A/Met201-886. When histidine residues at positions 433, 437, 473, and 477, which are highly conserved among all PDE families, were changed independently to serine residues, cAMP hydrolyzing activities were substantially reduced or abolished. The ability of these mutants to bind prototypical PDE4 inhibitors [3H]-(R)-rolipram or [3H]RP 73401 was also decreased in parallel with the loss of catalytic activity. The parallel loss of catalytic activity and inhibitor binding suggests that these changes resulted from non-localized perturbations in the structure of the enzyme. More interesting results were obtained when histidine residues at positions 505 and 506 were changed independently to aspar agines. The K(m) value for cAMP increased 3-fold in H505N (K(m) = 11 +/- 3 microM) and 11-fold in H506N (K(m) = 44 +/- 6 microM) compared with the wild-type protein (K(m) = 4 +/- 1 microM). These mutant proteins bound [3H]-(R)-rolipram and [3H]RP 73401 with K(d) values of 1.8 +/- 0.4 and 0.3 +/- 0.1 nM, respectively, for H505N, and 3.9 +/- 0.9 and 0.5 +/- 0.1 nM, respectively, for H506N. These values are nearly identical to those obtained with the wild-type rhPDE4A/Met201-886. In contrast, the IC50 values for cAMP competition with either [3H]-(R)-rolipram or [3H]RP 73401 binding increased approximately 2-fold in H505N and approximately 13-fold in H506N compared with the wild type protein. These increases are virtually identical to the changes in the K(m) value for cAMP in these mutants. We conclude that His506 and, perhaps, His505 are involved in binding of cAMP to PDE4A/Met201-886 but not in binding of PDE4-selective inhibitors.

Inactivation of recombinant monocyte cAMP-specific phosphodiesterase by cAMP analog, 8-[(4-bromo-2,3-dioxobutyl)thio]adenosine 3',5'-cyclic monophosphate.

Two cAMP analogs, 8- and 2- [(4-bromo-2,3-dioxobutyl)thio]adenosine 3',5'-cyclic monophosphate (8- and 2-BDB-TcAMP) have been used in probing the catalytic site of recombinant monocyte cAMP-specific phosphodiesterase (PDE4a). 2-BDB-TcAMP is a reversible and competitive inhibitor (Ki = 5.5 mumol/L) of cAMP hydrolysis by PDE4a, 8-BDB-TcAMP irreversibly inactivates the enzyme in a time- and concentration-dependent manner with a second order rate constant of 0.022 mmol/L-1 min-1. The rate of inactivation of PDE4a is reduced by the presence of the substrate cAMP and specific inhibitors, rolipram and denbufylline, but not by cGMP or AMP. Reduction of the enzyme-inhibitor complex with sodium [3H]borohydride shows that 1.2 mol of the affinity label/mol of enzyme was incorporated. The radiolabeled peptide is composed of 10 amino acid residues (697 to 706) located near the carboxyl end of the proposed catalytic domain. The peptide (GPGHPPLPDK) has seven nonpolar and aliphatic residues, of which four are proline, giving the peptide a highly structured conformation. This peptide is the first to be identified in the putative catalytic domain involved in substrate recognition.

Mapping the functional domains of human recombinant phosphodiesterase 4A: structural requirements for catalytic activity and rolipram binding.

To identify functional domains of the 886-amino acid human recombinant cAMP-specific phosphodiesterase (PDE) subtype A (rhPDE4A), we engineered the expression of seven mutant proteins containing both NH2- and COOH-terminal truncations. The level of rhPDE4A protein expression in yeast was monitored by immunoblotting using enzyme-specific antisera. Biochemical profiles of the mutant proteins were compared with those of the full-length protein or a fully active truncated form of the enzyme (rhPDE4A Met265-886), lacking the first 264 amino acids. The smallest catalytically active fragment generated was Met332-722, which at 45 kDa is less than half the mass of the full-length enzyme (approximately 110 kDa) but spans the most highly conserved region of the PDE superfamily. Two prototypical PDE4 inhibitors, rolipram and RP 73401, inhibited cAMP hydrolyzing activity of all truncated forms of the enzyme, with IC50 values of 70-2000 nM and 0.2-0.6 nM, respectively. [3H](R)-Rolipram bound to two sites on Met265-886, a high affinity site (Kd1 = 0.7 +/- 0.3 nM) and a low affinity site (Kd2 = 34 +/- 10 nM). Interestingly, [3H](R)-rolipram failed to bind to Met332-886 with high affinity, indicating that high affinity binding is not required for inhibition of enzyme activity. Low affinity rolipram binding was still present in Met332-886 (Kd = 101 +/- 7 nM). In contrast to [3H](R)-rolipram, [3H]RP 73401 bound to a single class of high affinity sites on Met265-886 (Kd = 0.4 +/- 0.1 nM). Further truncation of the enzyme to Met332-886 had no effect on [3H]RP 73401 binding (Kd = 0.2 +/- 0.03 nM). We conclude that the catalytic center of rhPDE4A lies between amino acids 332 and 722. Furthermore, amino acids 265-332 may form a high affinity binding site for rolipram that is outside of the catalytic domain. As a more likely alternative, these amino acids may not form a distinct binding site but instead may be required for the recombinant enzyme to assume a conformation that binds rolipram at the catalytic domain with a high affinity.

Phenotypic characterization of a tungsten-tolerant mutant of Azotobacter vinelandii.

A tungsten-tolerant mutant strain (CA6) of Azotobacter vinelandii first described in 1980 (P. E. Bishop, D. M. L. Jarlenski, and D. R. Hetherington, Proc. Natl. Acad. Sci. USA 77:7342-7346, 1980) has been further characterized. Results from growth experiments suggest that both nitrogenases 1 and 3 are utilized when CA6 grows in N-free medium containing Na2MoO4. Strain CA6.1.71, which lacks both nitrogenases 2 and 3, grew as well as strain CA in N-free medium containing Na2MoO4 after an initial lag. This indicates that nitrogenase 1 is fully functional in strain CA6. nifH-lacZ and anfH-lacZ transcriptional fusions were expressed in CA6 in the presence of Na2MoO4. Thus, in contrast to wild-type strain CA, transcription of the anfHDGK gene cluster in strain CA6 is not repressed by Mo. Expression of the vnfD-lacZ fusion was the same in both strains CA and CA6. In agreement with the results obtained with lac fusions, subunits of both nitrogenases 1 and 3 were found in protein extracts of CA6 cells grown in N-free medium containing Na2MoO4. However, CA6 cells, cultured in the presence of Na2WO4, accumulated nitrogenase 3 proteins without detectable amounts of nitrogenase 1 proteins. This indicates that expression of Mo-independent nitrogenase 3 is the basis for the tungsten tolerance phenotype of strain CA6. A measure of Mo accumulation as a function of time showed that accumulation by strain CA6 was slower than that for strain CA. When Mo accumulation was studied as a function of Na2MoO4 concentration, the two strains accumulated similar amounts of Mo in the concentration range of 0 to 1 microM Na2MoO4 during a 2-h period. Within the range of 1 to 5 microM Na2MoO4, Mo accumulation by strain CA increased linearly with increasing concentration whereas no further increases were observed for strain CA6. These results are consistent with the possibility that the tungsten tolerance mutation carried by CA6 is in a Mo transport system.

Regulation of nitrogenase-2 in Azotobacter vinelandii by ammonium, molybdenum, and vanadium.

Under diazotrophic conditions in the absence of molybdenum and in the presence of vanadium, Azotobacter vinelandii reduces N2 to NH4+ by using nitrogenase-2, a V-containing enzyme complex encoded by vnfH (the gene for dinitrogenase reductase-2), and vnfDGK (the genes for dinitrogenase-2 subunits). Accumulation of the vnfHorfFd and vnfDGK transcripts occurred under Mo-deficient conditions in the presence and absence of V; however, in the case of vnfDGK, the protein products only accumulated in the presence of V. This suggests that V is required for translation of the vnfDGK transcripts. In addition, expression of vnfH-lacZ and vnfD-lacZ transcriptional fusions was only partially repressed in the presence of NH4+. Transcripts hybridizing with vnfH (1.4 and 1.0 kb), vnfDG (3.4 and 1.8 kb), and vnfK (3.4 kb) were detected in RNA extracted from wild-type cells cultured with NH4+ in the presence or absence of V. However, nitrogenase-2 subunits were not detected in extracts of cells derepressed for nitrogenase-2 in the presence of NH4+. These results indicate that this nitrogen source acts at the posttranscriptional level as well as at the transcriptional level. vnf transcripts were not detected in the presence of Mo (with or without NH4+).

Removal of CO dehydrogenase from Pseudomonas carboxydovorans cytoplasmic membranes, rebinding of CO dehydrogenase to depleted membranes, and restoration of respiratory activities.

In Pseudomonas carboxydovorans, CO dehydrogenase and hydrogenase were found in association with the cytoplasmic membrane in a weakly bound and a tightly bound pool. The pools could be experimentally distinguished on the basis of resistance to removal by washes in low-ionic-strength buffer. The tightly bound pool of the enzymes could be differentially solubilized under conditions leaving the electron transport system intact and with the nondenaturing zwitterionic detergent 3-(3-cholamidopropyl) dimethylammonio 1-propane-sulfonic acid (CHAPS) and the nonionic detergent dodecyl beta-D-maltoside. In vitro reconstitution of depleted membranes with the corresponding supernatants containing CO dehydrogenase led to binding of the enzyme and to reactivation of respiratory activities with CO. The reconstitution reaction required cations with effectiveness which increased with increasing ionic charge: monovalent (Li+), divalent (Mg2+, Mn2+), or trivalent (Cr3+, La3+). Reconstitution of depleted membranes with CO dehydrogenase was specific for CO-grown bacteria. Cytoplasmic membranes from H2- or heterotrophically grown Pseudomonas carboxydovorans had no affinity for CO dehydrogenase at all, indicating the absence of the physiological electron acceptor of the enzyme, which presumably is cytochrome b561, or another membrane anchor.