Xanthine nucleotide-specific G-protein alpha-subunits: a novel approach for the analysis of G-protein-mediated signal transduction

From canonical to non-canonical cyclic nucleotides as second messengers: pharmacological implications

This review summarizes our knowledge on the non-canonical cyclic nucleotides cCMP, cUMP, cIMP, cXMP and cTMP. We place the field into a historic context and discuss unresolved questions and future directions of research. We discuss the implications of non-canonical cyclic nucleotides for experimental and clinical pharmacology, focusing on bacterial infections, cardiovascular and neuropsychiatric disorders and reproduction medicine. The canonical cyclic purine nucleotides cAMP and cGMP fulfill the criteria of second messengers. (i) cAMP and cGMP are synthesized by specific generators, i.e. adenylyl and guanylyl cyclases, respectively. (ii) cAMP and cGMP activate specific effector proteins, e.g. protein kinases. (iii) cAMP and cGMP exert specific biological effects. (iv) The biological effects of cAMP and cGMP are terminated by phosphodiesterases and export. The effects of cAMP and cGMP are mimicked by (v) membrane-permeable cyclic nucleotide analogs and (vi) bacterial toxins. For decades, the existence and relevance of cCMP and cUMP have been controversial. Modern mass-spectrometric methods have unequivocally demonstrated the existence of cCMP and cUMP in mammalian cells. For both, cCMP and cUMP, the criteria for second messenger molecules are now fulfilled as well. There are specific patterns by which nucleotidyl cyclases generate cNMPs and how they are degraded and exported, resulting in unique cNMP signatures in biological systems. cNMP signaling systems, specifically at the level of soluble guanylyl cyclase, soluble adenylyl cyclase and ExoY from Pseudomonas aeruginosa are more promiscuous than previously appreciated. cUMP and cCMP are evolutionary new molecules, probably reflecting an adaption to signaling requirements in higher organisms.

Structural basis for the high-affinity inhibition of mammalian membranous adenylyl cyclase by 2',3'-o-(N-methylanthraniloyl)-inosine 5'-triphosphate

2',3'-O-(N-Methylanthraniloyl)-ITP (MANT-ITP) is the most potent inhibitor of mammalian membranous adenylyl cyclase (mAC) 5 (AC5, K(i), 1 nM) yet discovered and surpasses the potency of MANT-GTP by 55-fold (J Pharmacol Exp Ther 329:1156-1165, 2009). AC5 inhibitors may be valuable drugs for treatment of heart failure. The aim of this study was to elucidate the structural basis for the high-affinity inhibition of mAC by MANT-ITP. MANT-ITP was a considerably more potent inhibitor of the purified catalytic domains VC1 and IIC2 of mAC than MANT-GTP (K(i), 0.7 versus 18 nM). Moreover, there was considerably more efficient fluorescence resonance energy transfer between Trp1020 of IIC2 and the MANT group of MANT-ITP compared with MANT-GTP, indicating optimal interaction of the MANT group of MANT-ITP with the hydrophobic pocket. The crystal structure of MANT-ITP in complex with the G(s)α- and forskolin-activated catalytic domains VC1:IIC2 compared with the existing MANT-GTP crystal structure revealed only subtle differences in binding mode. The higher affinity of MANT-ITP to mAC compared with MANT-GTP is probably due to fewer stereochemical constraints upon the nucleotide base in the purine binding pocket, allowing a stronger interaction with the hydrophobic regions of IIC2 domain, as assessed by fluorescence spectroscopy. Stronger interaction is also achieved in the phosphate-binding site. The triphosphate group of MANT-ITP exhibits better metal coordination than the triphosphate group of MANT-GTP, as confirmed by molecular dynamics simulations. Collectively, the subtle differences in ligand structure have profound effects on affinity for mAC.

A new tool for G protein analysis

G proteins play a pivotal role in cellular signaling by acting as molecular switches that undergo conformational changes upon binding GTP. The primary sequence constituting the binding cleft among the >160 G proteins in the human genome is highly conserved, consistent with the fact that these proteins share similar guanine nucleotide-binding characteristics. Recent work has demonstrated the feasibility of designing new analogs of GTP that can specifically activate G proteins whose nucleotide-binding sites have been remodeled through mutagenesis. This strategy has the potential to provide new insights into how G proteins act as molecular switches that engage their downstream target/effector proteins to generate specific signaling outputs.

Differential interactions of G-proteins and adenylyl cyclase with nucleoside 5'-triphosphates, nucleoside 5'-[gamma-thio]triphosphates and nucleoside 5'-[beta,gamma-imido]triphosphates

The regulatory G-proteins of adenylyl cyclase (AC), G(i) and G(s), are not only activated by GTP and the stable GTP analogs, guanosine 5'-[gamma-thio]triphosphate (GTPgammaS) and guanosine 5'-[beta,gamma-imido]triphosphate (GppNHp), but also by hypoxanthine, xanthine, uracil and cytidine nucleotides. The latter nucleotides were previously used to analyze distinct active G-protein states. Surprisingly, recent studies have shown that inosine 5'-[gamma-thio]triphosphate and uridine 5'-[gamma-thio]triphosphate can also inhibit AC directly. Therefore, we systematically compared the interactions of nucleoside 5'-triphosphates (NTPs), nucleoside 5'-[gamma-thio]triphosphates (NTPgammaSs) and nucleoside 5'-[beta,gamma-imido]triphosphates (NppNHps) with G(i), G(s) and AC. NTPgammaSs exhibited up to 26,000-fold higher affinity for G-proteins than NTPs and NppNHps. NTPgammaSs were up to 150-fold more potent direct AC inhibitors than NTPs and NppNHps. G-proteins exhibited striking preference for guanine nucleotides compared to other purine and pyrimidine nucleotides, whereas base-selectivity of various ACs, particularly the purified catalytic subunits C1.C2, was rather poor. GTP, GTPgammaS and GppNHp exhibited much higher selectivity for G-proteins relative to AC than all other purine and pyrimidine nucleotides. We have energetically characterized the interactions of purine and pyrimidine nucleotides with AC in silico, constructing pharmacophore models that correlate well with experimental affinities and have elucidated specific amino acid residues with greatest influence on nucleotide binding. Collectively, both G-proteins and ACs bind purine and pyrimidine nucleotides, with G-proteins showing much higher base-selectivity than AC. Thus, direct inhibitory effects of nucleotides on AC should be understood and considered when probing distinct active G-protein states with non-guanine nucleotides.