Guanylate Cyclase and the Cno/Cgmp Signaling Pathway

Home , Guanylate cyclase, Soluble guanylyl cyclase

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Biochimica et Biophysica Acta 1411 (1999) 334^350

Review Guanylate cyclase and the cNO/cGMP signaling pathway

John W. Denninger a, Michael A. Marletta a;b;c;*

a 5315A Medical Sciences I, Department of Biological Chemistry, University of Michigan Medical School, 1150 West Medical Center Drive Ann Arbor, MI 48109-0606, USA b Howard Hughes Medical Institute, University of Michigan, Ann Arbor, MI 48109-0606, USA c Interdepartmental Program in Medicinal Chemistry, College of Pharmacy, University of Michigan, Ann Arbor, MI 48109-1065, USA

Received 13 August 1998; received in revised form 30 November 1998; accepted 16 December 1998

Abstract

Signal transduction with the diatomic radical nitric oxide (NO) is involved in a number of important physiological processes, including smooth muscle relaxation and neurotransmission. Soluble guanylate cyclase (sGC), a heterodimeric enzyme that converts guanosine triphosphate to cyclic guanosine monophosphate, is a critical component of this signaling pathway. sGC is a hemoprotein; it is through the specific interaction of NO with the sGC heme that sGC is activated. Over the last decade, much has been learned about the unique heme environment of sGC and its interaction with ligands like NO and carbon monoxide. This review will focus on the role of sGC in signaling, its relationship to the other nucleotide cyclases, and on what is known about sGC genetics, heme environment and catalysis. The latest understanding in regard to sGC will be incorporated to build a model of sGC structure, activation, catalytic mechanism and deactivation. ß 1999 Elsevier Science B.V. All rights reserved.

Keywords: Guanylate cyclase; Cyclic guanosine monophosphate; Nitric oxide; Carbon monoxide; Heme

Contents

1. Introduction ...... 335

2. Historical perspective ...... 335

3. Signaling with the cNO/cGMP pathway ...... 336

4. The nucleotide cyclase family: sGC, pGC and AC ...... 337

Abbreviations: AC, adenylate cyclase; ATP, adenosine 5P-triphosphate; L11385,NH2-terminal heme binding domain of the L1 subunit of soluble guanylate cyclase; cAMP, cyclic adenosine 3P,5P-monophosphate; cGMP, cyclic guanosine 3P,5P-monophosphate; CO, carbon monoxide; CO-sGC, soluble guanylate cyclase with nitric oxide bound in the distal heme coordination site; EDRF, endothelium derived relaxing factor; EPR, electron paramagnetic resonance spectroscopy; FTIR, Fourier transform infrared spectroscopy; GTP, guanosine c 3 3 5P-triphosphate; NO, nitric oxide; NO2 , nitrite; NO3 , nitrate; NOS, nitric oxide synthase; NO-sGC, soluble guanylate cyclase with nitric oxide bound in the distal heme coordination site; PDE, phosphodiesterase; pGC, particulate guanylate cyclase; sGC, soluble guanylate cyclase; RR, resonance Raman spectroscopy * Corresponding author, at address a. Fax: +1 (734) 6475687; E-mail: [email protected]

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 335

5. Soluble guanylate cyclase genetics and isoforms ...... 339

6. Puri¢cation from mammalian tissue and expression systems ...... 341

7. Heme environment ...... 342

8. Catalysis ...... 343

9. Model of sGC structure and function ...... 344

10. Conclusion ...... 346

Acknowledgements ...... 346

References ...... 346

1. Introduction protein kinase, cGMP-gated cation channels and cGMP-regulated phosphodiesterase. As will be de- Soluble guanylate1 cyclase (sGC) is the only con- scribed below, nearly 30 years of study of sGC clusively proven receptor for nitric oxide (cNO), a have given us an outline of sGC structure and func- signaling agent produced by the enzyme nitric oxide tion; however, many facets of sGC structure and synthase (NOS) from the amino acid L-arginine. As function, especially at the molecular level, remain shown in Fig. 1, sGC catalyzes the conversion of to be discovered. guanosine 5P-triphosphate (GTP) to cyclic guanosine 3P,5P-monophosphate (cGMP). Because it is one of two enzymes (the other is the membrane-bound pep- 2. Historical perspective tide-receptor guanylate cyclase) that produce cGMP, and because it is the only de¢nitive receptor for cNO, Although both sGC and its product, cGMP, were sGC is intimately involved in many signal transduc- identi¢ed in the 1960s, it has been only within the tion pathways, most notably in the cardiovascular last 10 years that the complete outline of the cNO/ system (e.g., in the regulation of vascular tone and cGMP signaling pathway has been understood. After platelet function) and in the nervous system (e.g. in the discovery of cAMP in 1958 [1] and the initial neurotransmission and, possibly, long-term potentia- elucidation of its role in physiology (for early reviews tion and depression). Soluble guanylate cyclase is a see [2,3]), the search for other physiologically rele- heterodimeric protein composed of K and L subunits; vant cyclic nucleotides began. In 1963, cGMP was in addition, it is a hemoprotein (see Fig. 2). cNO detected in the urine of rats [4]; several years later, binds to the sGC heme; this binding event activates guanylate cyclase was found in mammalian tissues the enzyme. The activation of sGC and the subse- [5^9]. In 1977, cNO 9 both as cNO gas and as cNO quent rise in cGMP concentration are what allow released from vasodilators like nitroglycerin and sGC to transmit an cNO signal to the downstream nitroprusside 9 was shown to activate guanylate elements of the signaling cascade-cGMP-dependent cyclase [10^13]. At that point, though, cNO was unknown in biological systems, except when exoge- nously supplied by drugs like nitroglycerin, so it was 1 Although we prefer `guanylate' cyclase (as does the Index assumed that cNO activation of sGC was a non- Medicus and the Oxford Dictionary of Biochemistry and Molec- physiological, if useful, phenomenon. ular Biology), `guanylyl' cyclase is also used. Neither of the com- The story soon became clear, however. In 1980, mon names for this enzyme is technically correct: `guanylate' refers to the anion of guanylic acid (guanosine monophosphate) Furchgott et al. were able to demonstrate the exis- and `guanylyl' refers to the acyl group derived from guanylic tence of a substance produced by the endothelium acid. that was required to mediate relaxation of blood

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 336 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 vessels [14]. They designated this substance endothe- di¡uses freely across cell membranes, obviating the lium-derived relaxing factor (EDRF). In 1986, Ignar- need for a transport system. cNO is relatively unreac- ro and Furchgott independently suggested that tive for a radical species, so that it can di¡use from a EDRF was cNO (reviewed in [15]); the next year, it cell in which it is produced to neighboring cells with- was demonstrated that this was the case [16,17]. At out being destroyed before it reaches its target. Be- c 3 3 the same time, our laboratory and others were un- cause NO is oxidized to NO2 and NO3 under phys- 3 3 raveling the story of nitrite (NO2 ) and nitrate (NO3 ) iological conditions, no separate mechanism for its production in mammals. After initial observations by destruction is required [29^31]; however, at low con- Tannenbaum and coworkers that mammals produce centrations of cNO and in the absence of proteins 3 3 NO3 and that NO3 synthesis is increased following and metals which contribute to its destruction, the 3 c immune stimulation [18^20], it was shown that NO2 decomposition of NO will be slow [29]. Most impor- 3 c and NO3 were produced by immunostimulated mac- tantly, the electronic structure of NO makes it an rophages from the amino acid arginine (reviewed in excellent ligand for heme, allowing it to bind to the [21,22]). Shortly thereafter, our laboratory and sGC heme at a low, non-toxic concentration. others demonstrated that a soluble enzyme in murine The one well-de¢ned target for cNO, of course, is macrophages, later puri¢ed and named nitric oxide sGC; however, several paths for cNO signal trans- synthase, was producing cNO from the amino acid duction which do not rely on the production of arginine [23^25]. At this point, the major compo- cGMP by sGC have been proposed, though they nents of the cNO/cGMP signaling pathway had are less clearly de¢ned than the cNO/cGMP pathway. been identi¢ed: nitric oxide synthase produces cNO, First, cNO was shown to inhibit mitochondrial cyto- which activates soluble guanylate cyclase, leading to chrome oxidase at concentrations of cNO found a subsequent increase in the concentration of cGMP. physiologically [32,33]. Second, cNO was reported to activate calcium-dependent potassium channels in vascular smooth muscle directly, though it is pos- 3. Signaling with the cNO/cGMP pathway sible that this only happens with levels of cNO higher than typically seen in vivo [34]. Third, cNO was Since the initial elucidation of the cNO/cGMP sys- shown to inhibit activation of NF-UB [35,36]. tem, many good examples of signal transduction by Fourth, cNO was found to activate G-proteins this system have come to light, some better charac- (such as protooncogene p21ras) by enhancing the terized than others. They include: smooth muscle rate of nucleotide exchange, leading to an increase relaxation and blood pressure regulation [26]; plate- in the GTP-bound form. In fact, a p21ras cysteine let aggregation and disaggregation [27]; and neuro- residue (C118) was shown to be S-nitrosylated in the transmission both peripherally, in non-adrenergic, presence of cNO [37^39]. Fifth, caspase, a cysteine non-cholinergic (NANC) nerves [26], and centrally, protease involved in apoptosis, was shown to be in- in the processes of long term potentiation and de- hibited by S-nitrosylation of its active-site cysteine pression [28]. Understanding the cNO/cGMP path- [40]. Finally, both hemoglobin and the cardiac calci- way, however, requires an understanding of cNO's um release channel (i.e., the ryanodine receptor) were role as a signaling agent, the alternative targets for shown to be S-nitrosylated in vivo; moreover, this cNO, the alternative activators of sGC and, natu- change resulted in a modi¢cation of their function rally, the targets for cGMP. [41^43]. As these alternative cNO targets become bet- It was initially a surprise that cNO was produced ter de¢ned, it will be important to ensure that they physiologically: prior to the demonstration of cNO are physiologically relevant. cNO, especially at high synthesis by nitric oxide synthase, the biological rele- concentration, can react with O2 through a series of c vance of NO was as a pollutant; for example, as a reactions to form N2O3, which is a potent nitrosating component of automobile exhaust. Given the phys- agent reacting capable of reacting with nucleophiles ical properties of cNO, however, it is now clear that such as the sulfur of cysteine residues. As a result, in cNO is uniquely well suited to its role as a signaling vitro experiments with high cNO concentrations may agent. Because it is small and carries no charge, cNO lead to higher than normal levels of S-nitrosocysteine

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 337 formation. Therefore, while there is no doubt that S- e¡ect. Moreover, the diversity of targets allows nitrosylation occurs in vivo, we believe its physiolog- cGMP to have wide-ranging e¡ects that can di¡er ical signi¢cance will not be established until two ob- by cell and tissue type. There are three known targets servations are made: ¢rst, that S-nitrosylation 9 at for cGMP which mediate the transmission of the in vivo levels of cNO 9 modi¢es the function of cNO/cGMP pathway signal downstream from guany- target proteins and, second, that in vivo concentra- late cyclase: cGMP-dependent protein kinase [51], tions of S-nitrosylated proteins can elicit a physio- cGMP-regulated phosphodiesterase [52,53] and logical response. cGMP-gated ion channels [54]. The ¢rst of these, In addition to other targets for cNO, there are cGMP-dependent protein kinase, phosphorylates tar- other potential physiological activators of sGC. get proteins in response to an increase in cGMP con- These fall into two categories 9 non-cNO nitroge- centration. For example, in smooth muscle cells, nous compounds and carbon monoxide (CO). cGMP-dependent protein kinase phosphorylates the Although most investigators are now convinced inositol 1,4,5-triphosphate receptor, leading to a de- that cNO itself is the signaling agent which activates crease in Ca2 concentration and, ultimately, smooth sGC, there are still some who believe that there are muscle relaxation. The second, cyclic nucleotide other nitrogenous products (not cNO) which acti- phosphodiesterase, catalyzes the hydrolysis of the vates sGC. Many alternatives have been proposed, 3P-phosphodiester bond of cAMP and cGMP to yield such as nitrosothiols. It is not clear, however, that AMP and GMP, respectively. Of the seven known these agents are present in high enough concentra- families, three (PDE2, PDE5 and PDE6) are allos- tion in vivo to perform any signi¢cant signaling func- terically regulated by cGMP and one (PDE3) is in- tion. Likewise, it has been hypothesized that CO, hibited by the binding of cGMP in its substrate site. which binds to the sGC heme and weakly activates Phosphodiesterases allow crosstalk between cGMP the enzyme, may act as a signaling agent [44^46]. In and cAMP signaling pathways because they cause contrast to cNO, which activates sGC approx. 400- the concentration of one cyclic nucleotide to in£u- fold, CO only activates known isoforms of sGC ap- ence the degradation of the other. The third, cyclic prox. 5-fold [47]. Additionally, CO activation re- nucleotide-gated ion channels, are non-speci¢c cation quires saturating concentrations of CO, a clearly channels found in a variety of tissues. Most notably, non-physiological alternative. Although CO is pro- these channels are found in the retina and in olfac- duced by the enzyme heme oxygenase during the tory epithelium where they are involved, respectively, breakdown of heme, it is di¤cult to imagine that in visual phototransduction and in olfaction. cells can provide an adequate supply of free heme (which is toxic to cells) for signaling with CO. In the absence of a more CO-sensitive sGC isoform or 4. The nucleotide cyclase family: sGC, pGC and AC a mechanism to increase the response of sGC to CO, we will have to reserve judgment on this potential Just as sGC is a component of the larger cNO/ pathway. Such a mechanism may have been found, cGMP signaling pathway, so is it a member of a however. It has been shown recently that CO and the larger group of related enzymes: the nucleotide cy- synthetic compound YC-1 ([48]; see Fig. 4 for the clases. This family of enzymes converts nucleotide structure) synergistically activate sGC to levels sim- triphosphates (GTP and ATP) to cyclic nucleotide ilar that of the cNO-stimulated enzyme [49,50]. If a monophosphates (cGMP and cAMP) and includes physiological counterpart for YC-1 is identi¢ed, this adenylate cyclase (AC), particulate guanylate cyclase could be a mechanism by which CO could act as a (pGC) and soluble guanylate cyclase. As a family, signaling agent. these enzymes are involved in a broad array of signal Regardless of the species that activates sGC, all transduction pathways mediated by the cyclic nucleo- signal transduction through sGC (and pGC as well) tides cAMP and cGMP-vision, blood pressure regu- takes place through an increased concentration of lation, response to hormones and many others. An cGMP. Cyclic GMP, like cAMP, is a second messen- evolutionarily conserved catalytic domain de¢nes ger; as such, it requires target proteins to have its these three groups of enzymes as a single family; in

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 338 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

Fig. 1. The guanylate cyclase reaction and cNO signal transduction. other words, the genes that code for the catalytic ti¢ed. As shown in Fig. 2, AC contains two groups portions of these enzymes are derived from a com- of six transmembrane helices and two intracellular mon ancestor. This is unsurprising, given that they loops. The intracellular loops each contain a region catalyze virtually identical reactions. Moreover, be- 9 C1 and C2, respectively 9 now known to come cause of their relatedness and similar function, in- together in a head-to-tail fashion to form a function- sights into the structure, catalytic mechanism and al catalytic domain. (Because the domain is made up possibly even regulation of one member of this fam- of two regions from a single polypeptide chain, it is ily may be generalizable to other members. sometimes referred to as `pseudo-heterodimeric'.) Adenylate cyclase, for example, has been studied Two crystal structures of catalytic domains have, in for nearly four decades. As a result, there is a rich fact, been solved recently. The ¢rst was a structure of literature of ¢ndings with AC that may, because of aC2C2 homodimer [59], the second of a C1C2 heter- the similarity of AC to sGC, prove useful in under- odimer [60]; both were engineered constructs which standing sGC. Unlike the guanylate cyclases, which lacked the transmembrane domains and were, thus, contain both receptor and catalytic functions within soluble. The crystal structure of C1C2 supports the the same protein, ACs are stimulated by hormone conclusion that this pseudo-heterodimeric catalytic receptors acting through heterotrimeric G-proteins domain has two pseudo-symmetric binding sites: [55]. AC is an integral membrane protein and is as- one binds substrate, the other forskolin. In addition, sumed to be monomeric. All of the known mamma- GsK binds to a site in the plane of the catalytic do- lian isoforms are activated by GsK; other activators main opposite the substrate binding site. are isoform speci¢c [56]. AC is also activated by the Particulate guanylate cyclase has also provided in- small molecule forskolin, a diterpene derivative iso- sight into the structure and function of sGC. After lated from the Indian plant Coleus forskohli. Forsko- the puri¢cation of cGMP and guanylate cyclase, it lin was originally shown to have a variety of cardio- was observed that guanylate cyclase activity could be vascular e¡ects [57] and was later shown to activate separated into a membrane-bound (particulate) and AC [58]; it has served as a useful tool in the study of cytosolic (soluble) fractions. As its name implies, AC, but no mammalian counterpart has been iden- particulate guanylate cyclase is the membrane-bound

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 339

Fig. 2. Schematic representation of sGC, pGC and AC. Transmembrane helices are shown in purple, the conserved sGC, pGC and

AC catalytic regions are shown in orange and the conserved NH2-terminal domains of the sGC K1 and L1 subunits are shown in green. The sGC L1 subunit also binds prosthetic heme, which is ligated by histidine 105. guanylate cyclase. In a sense, pGC is a prototypic mechanism of activation of pGC will be similar to cell-surface receptor enzyme: it contains an extracel- that of sGC. lular peptide receptor domain and an intracellular catalytic domain separated by a single transmembrane domain (Fig. 2). pGC is now generally thought 5. Soluble guanylate cyclase genetics and isoforms to be a homodimer of subunits of approx. 120 kDa [61]. Seven subclasses of mammalian pGC (GC-A Since the original identi¢cation of sGC, a great through GC-G) have been identi¢ed. Three of these deal of information about sGC genetics has become have known peptide ligands (GC-A, atrial natriuretic available. Thus far, a total of 12 sGC cDNAs have peptide and brain natriuretic peptide; GC-B, C-type been cloned from ¢ve di¡erent species (for identity natriuretic peptide; GC-C, the heat stable enterotox- and similarity comparisons see Fig. 3). The ¢rst were ins of Escherichia coli and guanylin); the rest (GC-D, cloned on the basis of protein sequence from puri¢ed olfactory epithelium; GC-E and GC-F, retina; bovine and rat K1L1 heterodimer. The cDNA se- GC-G lung, intestine, skeletal muscle) are so-called quences for bovine and rat L1 subunits, respectively, `orphan' receptors because their activating ligands were reported in 1988 [64,65], followed by the bovine have not been determined [55,61,62]. For the pur- and rat K1 subunits in 1990 [66,67]. Subsequently, poses of understanding sGC, pGC is useful because K1L1 pairs were cloned from Homo sapiens (origi- it helps to de¢ne the characteristics that are speci¢c nally called K3 and L3) ([68,69], localized to chromo- to guanylate cyclases, for example, the identity of some 4q32, [70]), Drosophila melanogaster [71^73] the amino acids which allow the di¡erent members and the ¢sh Oryzias latipes [74]. Putative sGC se- of the nucleotide cyclase family to distinguish GTP quences exist in the Caenorhabditis elegans genome, from ATP [63]. Also, given the similarity of their but these have yet to be con¢rmed with any empiri- catalytic domains, it is possible that the general cal studies beyond the sequencing itself.

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 340 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

Fig. 3. Sequence identity and similarity among the predicted amino acid sequences of known sGC subunits. Translated cDNA sequences for the indicated sGC subunits were compared using the program Gap with gap weight = 12 and length weight = 4 (Wisconsin Package Version 9.1, Genetics Computer Group (GCG), Madison, WI, USA). The de¢nitions of the three letter species codes above are: bov, bovine, Bos taurus; hum, human, H. sapiens; rat, Rattus rattus; fsh, Medaka ¢sh O. latipes; dro, D. melanogaster. GenBank accession numbers for the cDNA sequences are as follows: bovine K1, X54014; human K1, X66534; rat K1, M57405; O. latipes K1, AB000849; Drosophila K1, U27117; human K2, X63282; bovine L1, Y00770; human L1, X66533; rat L1, M22562; O. latipes L1, AB000850; Drosophila L1, U27123; rat L2, M57507.

In addition to the cloning of K1L1 pairs, the (-CVVL) for isoprenylation/carboxyl-O-methylation cDNAs for two additional sGC subunits have been [75], but it is not known whether such a modi¢cation cloned: L2 from rat kidney [75] and K2 from human exists on the mature polypeptide. Because full-length fetal brain ([76], localized to chromosome 11q21-q22 sequences of the K2 and L2 subunits from species [77]. In transient transfection experiments, the K2 other than the ones from which they were originally subunit has been shown to form an active hetero- isolated have yet to be identi¢ed, one might question dimer with the L1 subunit [76]. Although the L2 sub- their signi¢cance. However, a partial sequence for unit has been recently reported to form an active the bovine K2 has been determined [76] and a puta- heterodimer with the K1 subunit in transient expres- tive L2 subunit partial sequence exists in a mouse sion experiments [78], similar experiments in our lab- placenta EST database (GenBank accession No. oratory (unpublished results) and others (unpub- AA023957, from the Soares mouse placenta library). lished results, see [79,80]) were unable to generate These subunits, then, may well be common to all an active heterodimer containing the L2 subunit. mammals and perhaps to other sGC-containing spe- Thus, the issue of the L2 subunit partner remains cies. to be conclusively proven. A unique feature of the The potential existence of novel sGC isoforms (i.e., L2 subunit is that its predicted amino acid sequence in addition to the K1L1) raises an important issue contains a COOH-terminal consensus sequence about the distinction between sGC subunits and

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 341

existence) of sGC isoforms other than the K1L1is poorly understood, our discussion of sGC characteristics will, by necessity, focus on the K1L1 isoform. An additional curiosity about the heterodimeric nature of sGC is the apparent requirement for co- expression of sGC subunits. For example, sGC K1 and L1 subunits, when expressed separately in transient mammalian expression systems, have no guanylate cyclase activity [67,73,81]. In fact, there is still no detectable guanylate cyclase activity when the soluble fractions from cells expressing the K1 and L1 subunits individually are mixed [73,81]. Only when the K1 and L1 subunits are co-expressed (i.e., produced in Fig. 4. Activators and inhibitors of sGC. Shown in panel A are the same cell) does basal and cNO-stimulated activity compounds known to interact directly with sGC. As described result [67,73,81]. The sGC requirement for co-expres- in the text, YC-1 is an activator of sGC that acts synergistically sion is similar to that of bacterial luciferase, which with CO to yield activity comparable to that observed with has been investigated extensively in the Baldwin lab- cNO. ODQ is a relatively potent inhibitor of the enzyme, although the mechanism of action is not clearly established. It oratory ([82], for a review of co-translational protein has been reported to oxidize the ferrous heme moiety [124]. folding, see [83]). It is possible that sGC, like bacte- Shown in panel B are compounds that have been reported to rial luciferase, has one subunit that can, in addition a¡ect catalytic activity. Methylene blue and LY83583 are re- to normal heterodimerization, form a kinetically ported to inhibit the enzyme and isoliquiritigenin has been re- trapped homodimer or fold to a dimerization-incom- ported to activate the enzyme. Studies with these compounds have typically been carried out with crude preparations of the petent monomer. enzyme and so the exact mechanism of action is not clear.

6. Puri¢cation from mammalian tissue and expression sGC isoforms. Our de¢nition of an sGC isoform is a systems catalytically competent heterodimer (in principle, however, it could be a homodimer). For AC (which Although the existence of other isoforms of sGC exists as a monomer) and pGC (which exists as a besides the K1L1 has been suggested by cloning and homodimer), the identi¢cation of di¡erent isoforms expression experiments, only the K1L1 has been pu- 9 with di¡erent tissue or developmental distribution, ri¢ed from mammalian tissue. In the past, a number di¡erent regulation and/or di¡erent kinetic parame- of di¡erent sources for sGC, including lung, liver, ters 9 has been relatively simple: each distinct gene and brain, were used for puri¢cations. Early puri¢- product is a distinct isoform. For sGC, the picture is cations (see, for example, [84,85]) produced enzyme somewhat more complicated. Given that four distinct with little or no bound heme [86]. Because of the sGC subunits have been cloned, there are 16 possible high sGC content and ready availability, bovine homodimeric and heterodimeric combinations of lung is now used as a source for puri¢cation from sGC subunits. Each of these combinations, if shown mammalian tissue. The groups currently purifying to exist in vivo, could be rightly called an isoform. sGC from bovine lung use a variety of puri¢cation The only isoform, therefore, that has been shown to procedures [47,87^92]. Most of the more recent pu- exist at the protein level in animal tissues is the K1L1. ri¢cations have been modi¢ed to increase the yield of Again, as mentioned above, putative K2L1 and K1L2 heme-containing sGC. In addition, yields have been isoforms have been demonstrated in transient mam- increased: puri¢cation from tissue can yield up to 3^ malian expression experiments [76,78], but until these 5 mg of pure protein (S. Kim, M.A. Marletta, un- subunits are shown to exist in vivo and, ideally, pu- published data; [92]). ri¢ed from tissue sources, they cannot be considered Because isolation of sGC from bovine lung is typ- bona ¢de sGC isoforms. Because the role (even the ically arduous, there has been strong motivation for

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 342 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 the development of alternative sources of sGC, forms, it became clear that the COOH-terminal por- namely, the development of expression systems. tion of both sGC subunits was most likely catalytic Both transient and stable expression of sGC in mam- in nature. In addition, as our laboratory noted malian expression systems using COS cells (an Afri- [47,87], the NH2-terminal portions 9 upstream can green monkey kidney cell line) and HEK293 cells from the conserved catalytic regions 9 of both the (a human embryonic kidney cell line) has successfully K and L subunits were homologous to each other produced active cyclase in several laboratories (i.e., 30% identical and 39% similar), though less ho- [67,73,76,78,81]. Unfortunately, while these systems mologous at the extreme NH2 terminus. This obser- have been useful for exploring the catalytic activity vation led us to hypothesize that both subunits might of sGC isoforms and mutants, the low level of pro- bind heme. Initially, our heme stoichiometry deter- tein produced has made them unsuitable as sources mination for sGC supported that hypothesis [87]; of puri¢ed protein. In order to take advantage of an however, recent work in our laboratory has demon- expression system that can generate amounts of sGC strated that there had been an error in the quanti- suitable for puri¢cation, several labs (including ours) tative amino acid analysis used to determine sGC have used expression and puri¢cation of sGC in the concentration. As a result, we have de¢nitively deter- baculovirus/Sf9 system [93^96]. In general, puri¢ca- mined the heme stoichiometry to be one heme per tion of sGC from baculovirus/Sf9 expression systems heterodimer ([93]; S. Kim, P.E. Brandish, M.A. Mar- is more straightforward because it requires a lower letta, unpublished data). puri¢cation factor. Obviously, the ability to express The hypothesis that the sGC heme is ligated by a heterodimeric sGC at high levels in E. coli would be histidine residue, much like the well-studied heme a tremendous boon to the ¢eld. Many laboratories proteins hemoglobin and myoglobin, is not new; have doubtless tried to do just that and have been, to however, it has been only within the last several this point, unsuccessful 9 and here we can judge years that that suspicion has been con¢rmed and only by our own experience, given the lack of pub- the ligand identi¢ed as histidine 105 of the L1 sub- lished reports 9 because sGC tends to form insolu- unit (the numbering applies to bovine, human and ble inclusion bodies when expressed in E. coli.To rat L1). Early observations of the sGC heme Soret our knowledge, the only sGC construct that has (i.e., the maximal heme absorbance peak) as isolated been successfully overexpressed in E. coli is the L1 [102] hinted that sGC was likely to bind heme in a subunit heme binding region that has been expressed way similar to deoxyhemoglobin and deoxymyoglo- and puri¢ed in our laboratory [97,98]. bin, namely, as a ¢ve-coordinate histidyl complex. It was not until two relatively recent observations were made that this picture began to become clear. First, a 7. Heme environment mutant of sGC in which histidine 105 of the bovine L1 subunit had been mutated to phenylalanine Although there has been some debate in the past, (H105F) and transiently expressed in COS cells was it is clear now that sGC binds one protoporphyrin found to have lost stimulation by cNO and, in addi- IX type heme per heterodimer. The early history of tion, to lack any heme absorbance [94]. Second, elec- the determination that sGC is a hemoprotein is re- tronic absorption spectroscopy on pure sGC revealed viewed in [99]. After initial observations that heme a spectrum that was entirely consistent with a histi- was needed to allow stimulation of sGC by cNO in dine-ligated heme, namely, a Soret peak at 431 nm crude preparations [100,101], sGC was shown to con- and a broad K/L band at 562 nm [47,105]. Additional tain a prosthetic heme moiety that was important in observations with electronic absorption spectroscopy activation by cNO and cNO donors [85,102^104]. Be- [106], magnetic circular dichroism spectroscopy [106], cause the early puri¢cations yielded sGC with heme resonance Raman spectroscopy (RR) [107,108] and stoichiometries less than one (often far less than one) Fourier transform infrared spectroscopy (FTIR) [86], it was assumed that the heme stoichiometry was, [109] con¢rmed that the heme ligand was histidine. in fact, one. After the cloning of both K and L sGC Subsequent observations in our laboratory de¢ni- subunits and the cloning of some of the pGC iso- tively identi¢ed L1-His105 as the heme ligand: the

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 343

NH2-terminal portion of the L1 subunit (L11385) ex- 8. Catalysis pressed and puri¢ed from E. coli was capable of binding heme with spectral characteristics virtually Several lines of evidence have converged to unam- identical to the heterodimeric enzyme [97]; mutants biguously de¢ne the catalytic domain of sGC: se- of His105 did not bind heme [98]; and heme binding quence comparisons to AC and pGC to de¢ne an to the H105G mutant could be rescued by including evolutionarily conserved region, site-directed muta- imidazole during growth and puri¢cation [98]. genesis to produce inactive enzyme, the construction The majority of groups studying sGC purify the of deletion mutants that retain catalytic function, enzyme in a ¢ve-coordinate, high-spin histidyl com- and the creation of mutants with altered substrate plex that resembles deoxyhemoglobin and deoxy- speci¢city. The ¢rst method that was used to de¢ne myoglobin in terms of ligation, oxidation state and the catalytic regions of AC, pGC and sGC was align- spectral characteristics. Unlike hemoglobin and my- ment of their predicted amino acid sequences (once oglobin, however, sGC has an extremely low a¤nity the cDNA sequences became available). These align- for oxygen, which fails to bind to the sGC heme even ments revealed a highly homologous region approx. under 100% O2 [47]. The molecular basis for this 200 amino acids long that was common to all three property is still under investigation. The failure of enzymes. Two regions were found in the AC se- sGC to bind oxygen enables sGC to act as a receptor quence (C1 and C2), one region in pGC monomers for cNO by allowing cNO to bind directly to the and one region in each half of the sGC heterodimer ferrous-deoxy heme; in contrast, cNO will react (K1 and L1). Several additional pieces of evidence with ferrous-oxy complexes (like oxyhemoglobin) to showed that these regions in the K1 and L1 subunits give nitrate and ferric heme [110]. The Soret position of sGC came together to form the catalytic domain. for sGC heme (431 nm) coupled with a broad K/L First, the mutation of two aspartic acid residues in peak at 562 nm [47,105] is consistent with ¢ve-coor- the sGC K1 subunit each abolished catalytic activity dinate high-spin histidyl-ligated proteins and model in co-expressed heterodimers [114]. Second, Wedel compounds [111]. Despite this seeming consensus, and coworkers demonstrated that co-expressed however, there have been reports that sGC, as iso- NH2-terminal deletion mutants of K1 (deletion of lated, is a mixture of ¢ve-coordinate high-spin histid- the ¢rst 366 amino acids) and L1 (deletion of the ¢rst yl heme and six-coordinate low-spin bis-histidyl 305 amino acids) retained catalytic activity [115]. Fi- heme [95,106,108,112]. It is not clear whether heme nally, any remaining doubt that these regions formed reconstitution, or some other di¡erence in procedure, the catalytic domain was removed by the recent ¢nd- is the cause of these multiple populations of enzyme. ing that sGC can be made into an cNO-responsive Extensive study of the sGC heme in its cNO-bound adenylate cyclase with changes to these regions (NO-sGC) and CO-bound (CO-sGC) forms has giv- alone: by mutating one residue in the K1 subunit en us a clear picture of the characteristics of these and two in the L1 subunit in the putative catalytic species. NO-sGC is a ¢ve-coordinate nitrosyl heme, regions that were predicted 9 on the basis of the AC in which the Fe-His bond has been broken [47,92,107,108,112,113]. A combination of electronic absorption spectroscopy, electron paramagnetic res- Table 1 Peak positions and extinction coe¤cients from electronic ab- onance spectroscopy (EPR) and RR has been used to sorption spectroscopy of sGC show this. When CO is bound to the sGC heme, the a Fe-His bond apparently remains intact, forming a sGC form Soret LKPor-Fe six-coordinate heme [47,92,106,108,112]. Observa- Ferrous 431 (111) 562 (12) 770 (0.2) tions by electronic absorption spectroscopy and RR CO complex 423 (145) 541 (12) 567 (12) NO complex 399 (85) 537 (11) 572 (11) have demonstrated this to be the case. Table 1 shows Ferric 391 (110) 511 (14) 642 (4) a summary of electronic absorption spectroscopy Values for this table were drawn from [47,105]. Extinction coef- maxima for sGC in its as isolated, CO-bound and ¢cients (mM31 cm31) are given in parentheses. c NO-bound forms; Table 2 shows RR and FTIR aPor-Fe refers to the putative porphyrin to iron charge transfer vibrations for the heme ligands of these forms. band.

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 344 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 catalytic domain crystal structures 9 to be important stereochemistry, which indicates that the reaction in nucleotide base binding [59,60], Sunahara and proceeds by a single displacement reaction [118]. coworkers were able to change the nucleotide specif- icity of sGC [116]. Many aspects of the catalytic mechanism of sGC 9. Model of sGC structure and function are poorly understood. The type of reaction catalyzed by sGC 9 the attack of an activated hydroxyl In order to synthesize some of the latest sGC de- on the K-phosphate of a nucleotide triphosphate 9 is velopments into a model 9 albeit a working model a common one. This is, for example, the same reac- 9 we have divided sGC enzymology into four sub- tion catalyzed in the polymerization of deoxynucleo- areas that represent what are, in our view, the most tide triphosphates in the synthesis of DNA; however, important open questions: structure, activation, cat- in the case of DNA it is an intermolecular attack of alytic mechanism and deactivation. the 3P-hydroxyl on an K-phosphate. In fact, after No X-ray crystal structure of sGC or of any sGC publication of the AC C2C2 structure, it was sug- domain has been solved; as a result, any model of gested that AC catalytic domain contains a fold sim- sGC structure must draw heavily on related proteins: ilar to the `palm' domain in prokaryotic DNA poly- The high degree of homology (approx. 30% identity merase I [117]. It is possible that some of the features and 40% similarity) between the catalytic domain of of the DNA polymerase mechanism will hold true AC and the catalytic domain of sGC, in addition to for the nucleotide cyclases. Thus, the outline of the their almost identical catalytic function, suggests that reaction is clear: the 3P-hydroxyl is activated in some these catalytic domains are descended from a com- way to become more nucleophilic and attacks the K- mon ancestor and, as such, are likely to share a phosphate, causing the release of the L- and Q-phos- common three-dimensional fold. In fact, the Hurley phates as inorganic pyrophosphate. What is not clear group has used the AC catalytic domain structure is how the 3P-hydroxyl is activated, whether pyro- determined in his laboratory to create a homology phosphate release is concerted or proceeds through model of the catalytic domain of sGC [63]. Such a pentavalent phosphorous transition state and what models have been used to correctly predict the amino role divalent metal cations play in the mechanism. acid residues that control the substrate speci¢city of Despite these questions, however, some possibilities both ACs and GCs. For example, the substrate spe- for the sGC mechanism have been eliminated. For ci¢city of retinal guanylate cyclase (a pGC) was example, the displacement of PPi by the 3P-hydroxyl changed from GTP to ATP [119], sGC was con- of GTP has been shown to occur with inversion of verted into an cNO-activated AC and AC was con- verted into a non-speci¢c (but still GsK-activated) purine nucleotide cyclase [116]. As a result, the Table 2 X-ray crystal structures of the AC C1C2 heterodimer Resonance Raman and FTIR vibrations for sGC heme ligands [60] and C2C2 homodimer [59] appear to be excellent models for the sGC catalytic domain. We would con- sGC form cm31 Description cur, then, with the prediction made by others [63], Ferrous that the COOH-terminal portions of the K and L X(Fe-His) 204 Fe-His stretch CO complex sGC subunits form a head-to-tail catalytic domain. X(Fe-CO) 472 Fe-CO stretch If the pseudo-heterodimeric AC catalytic domain is N(Fe-C-O) 562 Fe-C-O bending any guide, sGC may well have only one substrate- X(C-O) 1987 C-O stretch binding site. NO complex As for the rest of the protein, the lack of direct X(Fe-NO) 525 Fe-NO stretch X(N-O) 1677 N-O stretch evidence (even by analogy) makes it more di¤cult to choose among the possibilities; still, there are ¢nd- The values in this table were drawn from [107,109]. Similar values have been reported by other laboratories [91,92,108,112]; ings upon which we can continue to build the model. some di¡erences have been reported for reconstituted sGC The homology between the K and L subunits of sGC [91,108,112]. (33% identity and 42% similarity overall, although

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 345 higher in the COOH-terminal catalytic region than in Fe-His bond which occurs with cNO binding may be the NH2-terminal regulatory region) suggests that correlated with the molecular event that brings about the fold of the two subunits may be pseudo-symmet- activation, but may not directly cause it. In other rical. The K and L subunits of sGC have a predicted words, it is possible that the breaking of the Fe-His amphipathic helical region just NH2-terminal to the bond can be eliminated as a step in the activation of predicted catalytic domain. These helices may be sGC. There may be, then, some other e¡ect of cNO evolutionarily related to the predicted amphipathic binding to the heme that leads to activation. We are helices NH2-terminal to the catalytic region of currently pursuing this possibility in our laboratory. pGC, which have been demonstrated to mediate di- Our model of the sGC catalytic mechanism is merization [120]. This conclusion is supported by the based largely on proteins that catalyze similar reac- homodimerization of the L11385 [97] and by the di- tions. In addition to this, we have incorporated what merization of K1 and L1 subunit NH2-terminal dele- is known about the nucleotide cyclase catalytic mech- tion mutants [115], which retain the putative dimeri- anisms through crystal structures and mutational zation helices. The conformational change which studies. The active site functions/characteristics that shifts the AC catalytic domain to its active state need to be elucidated are as follows: the number of 2 upon binding of GsK may be similar to the confor- substrate binding sites; the number of Mg ions mational change which shifts the sGC catalytic do- participating in substrate binding and/or catalysis; main to its active state when cNO binds to the sGC the amino acid residues important for substrate bind- heme. It is possible, then, that the sGC heme domain ing; the amino acid residues involved in Mg2 bind- (the NH2-terminal portion of the L subunit) abuts ing; and the amino acid residues involved in the the catalytic domain in the same way that GsK abuts catalytic mechanism (e.g., activation of the 3P-OH the AC catalytic domain [60]. This would put both and transition-state stabilization). There are almost the K and L subunit NH2-terminal domains in the certainly no more than two substrate-binding sites in plane of the catalytic domain (which is roughly sGC; given that the catalytic domain of sGC, like disk-shaped) 9 the heme domain opposite the sub- that of AC, is made up of homologous but non-iden- strate site (in the same way GsK is opposite the AC tical components, we hypothesize that there is only substrate site) and the K subunit NH2-terminal por- one substrate-binding site, as in AC. This hypothesis tion next to the substrate site (opposite the site anal- also ¢ts in well with the obvious asymmetry of the ogous to where forskolin binds in the AC structure). sGC regulatory region: the L1 subunit binds heme, Our model of sGC activation by cNO does not the K1 does not. As for important residues, the mu- di¡er substantially from the following simpli¢ed pic- tation of rat K1D513 and K1D529 each abolished ture. Binding of cNO to the sGC heme severs the catalytic activity when transiently expressed with bond between the heme Fe and the ligating histidine, wild-type L1 subunit [114], but it is di¤cult to pin- presumably puckers the heme toward the distal nitro- point the role that these residues may be playing syl ligand and, through some combination of without further study. changes in heme geometry, causes a conformational The ¢nal component of our model of sGC struc- change that activates the enzyme. Recent work, how- ture and function is the question of deactivation, ever, has re¢ned our view of the heme environment which remains an open question because of the fol- of sGC in a way that allows us to re¢ne our working lowing observation: the deactivation of sGC in or- model of activation. As mentioned above, YC-1 is an gan bath studies has suggested that the half-life of activator of sGC that has the interesting property of sGC activation is on the order of minutes [17], synergistically activating sGC in the presence of CO whereas the half-lives of most nitrosyl-heme com- (see Fig. 4 for the structure of YC-1 and other sGC plexes are longer. Two hypotheses have been ad- activators and inhibitors). Electronic absorption vanced to explain this discrepancy. First, the sGC spectroscopy studies have shown that sGC activated heme may simply have di¡erent characteristics that by YC-1 and CO has a six-coordinate heme, even allow cNO to dissociate more rapidly. Second, there though its level of activity is similar to the cNO-acti- may be some extrinsic factor 9 a small molecule or vated enzyme [49,121]. Therefore, the breaking of the protein 9 which assists in the deactivation either by

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 346 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 directly interacting with the ferrous-nitrosyl heme or (2) understanding the catalytic mechanism and (3) via an allosteric e¡ect that increases the rate of cNO piecing together the molecular details of activation dissociation from the heme. Studies of the sGC and deactivation. As these questions are answered, heme, for example, showed that its properties were a complete view of sGC structure and function will di¡erent from most histidine-ligated heme proteins emerge. and model compounds: the half-life (extrapolated to 37³C) for cNO dissociation was on the order of 2 min [122]. Recent work in our laboratory has Acknowledgements shown that thiol compounds like glutathione, at concentrations similar to that found in vivo, increase the We would like to thank all of the Marletta lab for rate of cNO release from the heme [93]. Compounds many thought-provoking discussions and for their that trap cNO, like oxyhemoglobin, also increase the insightful comments on this manuscript. rate of cNO release from the sGC heme [93,122]. There has been a report that the presence of Mg2- GTP increases the rate of deactivation [123]; however, we have found that a precipitate which forms References under these conditions complicates the determination of such rates [93]. We believe, therefore, that the [1] T.W. Rall, E.W. Sutherland, Formation of a cyclic adenine ribonucleotide by tissue particles, J. Biol. Chem. 232 (1958) c relatively fast NO o¡-rate for the sGC heme [122] 1065^1076. and contribution of thiol compounds like glutathione [2] J.G. Hardman, G.A. Robison, E.W. Sutherland, Cyclic nu- account for most, if not all, of the in vivo rate of cleotides, Annu. Rev. Physiol. 33 (1971) 311^336. deactivation. [3] G.A. Robison, R.W. Butcher, E.W. Sutherland, Cyclic AMP, Academic Press, New York, 1971. [4] D.F. Ashman, R. Lipton, M.M. Melicow, T.D. Price, Iso- lation of adenosine 3P,5P-monophosphate and guanosine 10. Conclusion 3P,5P-monophosphate from rat urine, Biochem. Biophys. Res. Commun. 11 (1963) 330^334. We have attempted in this review to provide some [5] A.A. White, G.D. Aurbach, Detection of guanyl cyclase in insight into the structure and function of soluble mammalian tissues, Biochim. Biophys. Acta 191 (1969) 686^ guanylate cyclase. In doing so, we have placed sGC 697. [6] G. Schultz, E. Bo«hme, K. Munske, Guanyl cyclase. Deter- in its larger context by discussing the history of the mination of enzyme activity, Life Sci. 8 (1969) 1323^1332. sGC ¢eld, the physiological importance of the cNO/ [7] N.D. Goldberg, S.B. Dietz, A.G. O'Toole, Cyclic guanosine cGMP pathway and its components, and the nucleo- 3P,5P-monophosphate in mammalian tissues and urine, tide cyclase family to which sGC belongs. We have J. Biol. Chem. 244 (1969) 4458^4466. delved into the current understanding of sGC by [8] E. Ishikawa, S. Ishikawa, J.W. Davis, E.W. Sutherland, De- termination of guanosine 3P,5P-monophosphate in tissues discussing what is known about sGC isoforms, ex- and of guanyl cyclase in rat intestine, J. Biol. Chem. 244 pression and puri¢cation, heme environment and cat- (1969) 6371^6376. alysis. And ¢nally, we have presented a working [9] J.G. Hardman, E.W. Sutherland, Guanyl cyclase, an enzyme model of sGC structure, activation, catalytic mecha- catalyzing the formation of guanosine 3P,5P-monophosphate nism and deactivation. Naturally, a great deal of from guanosine triphosphate, J. Biol. Chem. 244 (1969) work remains to be done to complete our under- 6363^6370. [10] S. Katsuki, W. Arnold, C. Mittal, F. Murad, Stimulation of standing of sGC. Two important milestones in guanylate cyclase by sodium nitroprusside, nitroglycerin and achieving that progress will be the identi¢cation nitric oxide in various tissue preparations and comparison to and characterization of sGC subunits and isoforms the e¡ects of sodium azide and hydroxylamine, J. Cyclic from several species and the development of an sGC Nucleotide Res. 3 (1977) 23^35. expression system capable of producing large quanti- [11] N. Miki, Y. Kawabe, K. Kuriyama, Activation of cerebral guanylate cyclase by nitric oxide, Biochem. Biophys. Res. ties of the heterodimeric enzyme. With or without Commun. 75 (1977) 851^856. these breakthroughs, the most pressing problems to [12] W.P. Arnold, C.K. Mittal, S. Katsuki, F. Murad, Nitric address for sGC are: (1) determining the structure, oxide activates guanylate cyclase and increases guanosine

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 347

3P:5P-cyclic monophosphate levels in various tissue prepara- [30] J.M. Fukuto, Chemistry of nitric oxide: biologically relevant tions, Proc. Natl. Acad. Sci. USA 74 (1977) 3203^3207. aspects, Adv. Pharmacol. 34 (1995) 1^15. [13] K. Schultz, G. Schultz, Sodium nitroprusside and other [31] J.F. Kerwin Jr., J.R. Lancaster Jr., P.L. Feldman, Nitric smooth muscle-relaxants increase cyclic GMP levels in rat oxide: a new paradigm for second messengers, J. Med. ductus deferens, Nature 265 (1977) 750^751. Chem. 38 (1995) 4343^4362. [14] R.F. Furchgott, J.V. Zawadzki, The obligatory role of en- [32] C. Giulivi, Functional implications of nitric oxide produced dothelial cells in the relaxation of arterial smooth muscle by by mitochondria in mitochondrial metabolism, Biochem. J. acetylcholine, Nature 288 (1980) 373^376. 332 (1998) 673^679. [15] S. Moncada, R.M. Palmer, E.A. Higgs, The discovery of [33] G.C. Brown, Nitric oxide regulates mitochondrial respira- nitric oxide as the endogenous nitrovasodilator, Hyperten- tion and cell functions by inhibiting cytochrome oxidase, sion 12 (1988) 365^372. FEBS Lett. 369 (1995) 136^139. [16] L.J. Ignarro, G.M. Buga, K.S. Wood, R.E. Byrns, G. [34] V.M. Bolotina, S. Najibi, J.J. Palacino, P.J. Pagano, R.A. Chaudhuri, Endothelium-derived relaxing factor produced Cohen, Nitric oxide directly activates calcium-dependent po- and released from artery and vein is nitric oxide, Proc. tassium channels in vascular smooth muscle, Nature 368 Natl. Acad. Sci. USA 84 (1987) 9265^9269. (1994) 850^853. [17] R.M. Palmer, A.G. Ferrige, S. Moncada, Nitric oxide release [35] D. Sekkai, F. Aillet, N. Israel, M. Lepoivre, Inhibition of accounts for the biological activity of endothelium-derived NF-kappaB and HIV-1 long terminal repeat transcriptional relaxing factor, Nature 327 (1987) 524^526. activation by inducible nitric oxide synthase 2 activity, [18] D.A. Wagner, V.R. Young, S.R. Tannenbaum, Mammalian J. Biol. Chem. 273 (1998) 3895^3900. 15 nitrate biosynthesis: incorporation of NH3 into nitrate is [36] M. Spiecker, H.B. Peng, J.K. Liao, Inhibition of endothelial enhanced by endotoxin treatment, Proc. Natl. Acad. Sci. vascular cell adhesion molecule-1 expression by nitric oxide USA 80 (1983) 4518^4521. involves the induction and nuclear translocation of Ikappa- [19] L.C. Green, K. Ruiz de Luzuriaga, D.A. Wagner, W. Rand, Balpha, J. Biol. Chem. 272 (1997) 30969^30974. N. Istfan, V.R. Young, S.R. Tannenbaum, Nitrate biosyn- [37] H.M. Lander, P.K. Sehajpal, A. Novogrodsky, Nitric oxide thesis in man, Proc. Natl. Acad. Sci. USA 78 (1981) 7764^ signaling: a possible role for G proteins, J. Immunol. 151 7768. (1993) 7182^7187. [20] L.C. Green, S.R. Tannenbaum, P. Goldman, Nitrate syn- [38] H.M. Lander, J.S. Ogiste, S.F. Pearce, R. Levi, A. Novog- thesis in the germfree and conventional rat, Science 212 rodsky, Nitric oxide-stimulated guanine nucleotide exchange (1981) 56^58. on p21ras, J. Biol. Chem. 270 (1995) 7017^7020. [21] M.A. Marletta, M.A. Tayeh, J.M. Hevel, Unraveling the [39] H.M. Lander, D.P. Hajjar, B.L. Hempstead, U.A. Mirza, biological signi¢cance of nitric oxide, Biofactors 2 (1990) B.T. Chait, S. Campbell, L.A. Quilliam, A molecular redox 219^225. switch on p21(ras). Structural basis for the nitric oxide- [22] M.A. Marletta, Nitric oxide: biosynthesis and biological sig- p21(ras) interaction, J. Biol. Chem. 272 (1997) 4323^4326. ni¢cance, Trends Biochem. Sci. 14 (1989) 488^492. [40] J. Li, T.R. Billiar, R.V. Talanian, Y.M. Kim, Nitric oxide [23] J.B. Hibbs Jr., R.R. Taintor, Z. Vavrin, E.M. Rachlin, Ni- reversibly inhibits seven members of the caspase family via tric oxide: a cytotoxic activated macrophage e¡ector mole- S-nitrosylation, Biochem. Biophys. Res. Commun. 240 cule, Biochem. Biophys. Res. Commun. 157 (1988) 87^94. (1997) 419^424. [24] M.A. Marletta, Mammalian synthesis of nitrite, nitrate, ni- [41] J.S. Stamler, L. Jia, J.P. Eu, T.J. McMahon, I.T. Demchen- tric oxide, and N-nitrosating agents, Chem. Res. Toxicol. 1 ko, J. Bonaventura, K. Gernert, C.A. Piantadosi, Blood £ow (1988) 249^257. regulation by S-nitrosohemoglobin in the physiological oxy- [25] J.M. Hevel, K.A. White, M.A. Marletta, Puri¢cation of the gen gradient, Science 276 (1997) 2034^2037. inducible murine macrophage nitric oxide synthase. Identi¢- [42] L. Xu, J.P. Eu, G. Meissner, J.S. Stamler, Activation of the cation as a £avoprotein, J. Biol. Chem. 266 (1991) 22789^ cardiac calcium release channel (ryanodine receptor) by 22791. poly-S-nitrosylation, Science 279 (1998) 234^237. [26] T.D. Warner, J.A. Mitchell, H. Sheng, F. Murad, E¡ects of [43] A.J. Gow, J.S. Stamler, Reactions between nitric oxide and cyclic GMP on smooth muscle relaxation, Adv. Pharmacol. haemoglobin under physiological conditions, Nature 391 26 (1994) 171^194. (1998) 169^173. [27] W.A. Buechler, K. Ivanova, G. Wolfram, C. Drummer, J.M. [44] T. Ingi, J. Cheng, G.V. Ronnett, Carbon monoxide: an en- Heim, R. Gerzer, Soluble guanylyl cyclase and platelet func- dogenous modulator of the nitric oxide-cyclic GMP signal- tion, Ann. NY Acad. Sci. 714 (1994) 151^157. ing system, Neuron 16 (1996) 835^842. [28] S.R. Ja¡rey, S.H. Snyder, Nitric oxide: a neural messenger, [45] A. Verma, D.J. Hirsch, C.E. Glatt, G.V. Ronnett, S.H. Annu. Rev. Cell Dev. Biol. 11 (1995) 417^440. Snyder, Carbon monoxide: a putative neural messenger, Sci- [29] R.S. Lewis, W.M. Deen, Kinetics of the reaction of nitric ence 259 (1993) 381^384. oxide with oxygen in aqueous solutions, Chem. Res. Toxicol. [46] R. Zakhary, K.D. Poss, S.R. Ja¡rey, C.D. Ferris, S. Tone- 7 (1994) 568^574. gawa, S.H. Snyder, Targeted gene deletion of heme oxygen-

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 348 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

ase 2 reveals neural role for carbon monoxide, Proc. Natl. [63] Y. Liu, A.E. Ruoho, V.D. Rao, J.H. Hurley, Catalytic Acad. Sci. USA 94 (1997) 14848^14853. mechanism of the adenylyl and guanylyl cyclases: modeling [47] J.R. Stone, M.A. Marletta, Soluble guanylate cyclase from and mutational analysis, Proc. Natl. Acad. Sci. USA 94 bovine lung: activation with nitric oxide and carbon mon- (1997) 13414^13419. oxide and spectral characterization of the ferrous and ferric [64] D. Koesling, J. Herz, H. Gausepohl, F. Niroomand, K.D. states, Biochemistry 33 (1994) 5636^5640. Hinsch, A. Mu«lsch, E. Bo«hme, G. Schultz, R. Frank, The [48] S.-C. Kuo, F.Y. Lee, C.-M. Teng, United States patent No. primary structure of the 70 kDa subunit of bovine soluble 5,574,168, 1996. guanylate cyclase, FEBS Lett. 239 (1988) 29^34. [49] J.R. Stone, M.A. Marletta, Synergistic activation of soluble [65] M. Nakane, S. Saheki, T. Kuno, K. Ishii, F. Murad, Mo- guanylate cyclase by YC-1 and carbon monoxide: implica- lecular cloning of a cDNA coding for 70 kilodalton subunit tions for the role of cleavage of the iron-histidine bond dur- of soluble guanylate cyclase from rat lung, Biochem. Bio- ing activation by nitric oxide, Chem. Biol. 5 (1998) 255^ phys. Res. Commun. 157 (1988) 1139^1147. 261. [66] D. Koesling, C. Harteneck, P. Humbert, A. Bosserho¡, R. [50] A. Friebe, G. Schultz, D. Koesling, Sensitizing soluble gua- Frank, G. Schultz, E. Bo«hme, The primary structure of the nylyl cyclase to become a highly CO-sensitive enzyme, larger subunit of soluble guanylyl cyclase from bovine lung. EMBO J. 15 (1996) 6863^6868. Homology between the two subunits of the enzyme, FEBS [51] S.M. Lohmann, A.B. Vaandrager, A. Smolenski, U. Walter, Lett. 266 (1990) 128^132. H.R. De Jonge, Distinct and speci¢c functions of cGMP- [67] M. Nakane, K. Arai, S. Saheki, T. Kuno, W. Buechler, F. dependent protein kinases, Trends Biochem. Sci. 22 (1997) Murad, Molecular cloning and expression of cDNAs coding 307^312. for soluble guanylate cyclase from rat lung, J. Biol. Chem. [52] E. Degerman, P. Belfrage, V.C. Manganiello, Structure, lo- 265 (1990) 16841^16845. calization, and regulation of cGMP-inhibited phosphodies- [68] G. Giuili, U. Scholl, F. Bulle, G. Guellaen, Molecular clon- terase (PDE3), J. Biol. Chem. 272 (1997) 6823^6826. ing of the cDNAs coding for the two subunits of soluble [53] M.D. Houslay, G. Milligan, Tailoring cAMP-signalling re- guanylyl cyclase from human brain, FEBS Lett. 304 (1992) sponses through isoform multiplicity, Trends Biochem. Sci. 83^88. 22 (1997) 217^224. [69] V. Chhajlani, P.A. Frandberg, J. Ahlner, K.L. Axelsson, J.E. [54] W.N. Zagotta, S.A. Siegelbaum, Structure and function of Wikberg, Heterogeneity in human soluble guanylate cyclase cyclic nucleotide-gated channels, Annu. Rev. Neurosci. 19 due to alternative splicing, FEBS Lett. 290 (1991) 157^ (1996) 235^263. 158. [55] B.J. Wedel, D.L. Garbers, New insights on the functions of [70] G. Giuili, N. Roechel, U. Scholl, M.G. Mattei, G. Guellaen, the guanylyl cyclase receptors, FEBS Lett. 410 (1997) 29^33. Colocalization of the genes coding for the K3 and L3 sub- [56] J. Hanoune, Y. Pouille, E. Tzavara, T. Shen, L. Lipskaya, units of soluble guanylyl cyclase to human chromosome 4 at N. Miyamoto, Y. Suzuki, N. Defer, Adenylyl cyclases: struc- q31.3-q33, Hum. Genet. 91 (1993) 257^260. ture, regulation and function in an enzyme superfamily, Mol. [71] S. Yoshikawa, I. Miyamoto, J. Aruga, T. Furuichi, H. Oka- Cell. Endocrinol. 128 (1997) 179^194. no, K. Mikoshiba, Isolation of a Drosophila gene encoding a [57] E. Lindner, A.N. Dohadwalla, B.K. Bhattacharya, Positive head-speci¢c guanylyl cyclase, J. Neurochem. 60 (1993) inotropic and blood pressure lowering activity of a diterpene 1570^1573. derivative isolated from Coleus forskohli: forskolin, Arznei- [72] W. Liu, J. Yoon, M. Burg, L. Chen, W.L. Pak, Molecular mittelforschung 28 (1978) 284^289. characterization of two Drosophila guanylate cyclases ex- [58] K.B. Seamon, W. Padgett, J.W. Daly, Forskolin: unique pressed in the nervous system, J. Biol. Chem. 270 (1995) diterpene activator of adenylate cyclase in membranes and 12418^12427. in intact cells, Proc. Natl. Acad. Sci. USA 78 (1981) 3363^ [73] S. Shah, D.R. Hyde, Two Drosophila genes that encode the 3367. K and L subunits of the brain soluble guanylyl cyclase, [59] G. Zhang, Y. Liu, A.E. Ruoho, J.H. Hurley, Structure of the J. Biol. Chem. 270 (1995) 15368^15376. adenylyl cyclase catalytic core, Nature 386 (1997) 247^ [74] T. Mikami, T. Kusakabe, N. Suzuki, Molecular cloning of 253. cDNAs and expression of mRNAs encoding K and L sub- [60] J.J. Tesmer, R.K. Sunahara, A.G. Gilman, S.R. Sprang, units of soluble guanylyl cyclase from medaka ¢sh Oryzias Crystal structure of the catalytic domains of adenylyl cyclase latipes, Eur. J. Biochem. 253 (1998) 42^48. in a complex with GsKWGTPQS, Science 278 (1997) 1907^ [75] P.S. Yuen, L.R. Potter, D.L. Garbers, A new form of gua- 1916. nylyl cyclase is preferentially expressed in rat kidney, Bio- [61] H.J. Fulle, D.L. Garbers, Guanylyl cyclases: a family of chemistry 29 (1990) 10872^10878. receptor-linked enzymes, Cell. Biochem. Funct. 12 (1994) [76] C. Harteneck, B. Wedel, D. Koesling, J. Malkewitz, E. 157^165. Bo«hme, G. Schultz, Molecular cloning and expression of a [62] S. Schulz, B.J. Wedel, A. Matthews, D.L. Garbers, The clon- new K-subunit of soluble guanylyl cyclase. Interchangeability ing and expression of a new guanylyl cyclase orphan recep- of the K-subunits of the enzyme, FEBS Lett. 292 (1991) 217^ tor, J. Biol. Chem. 273 (1998) 1032^1037. 222.

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 349

[77] F. Yu, D. Warburton, S. Wellington, R.S. Danziger, Assign- [93] P.E. Brandish, W. Buechler, M.A. Marletta, Regeneration ment of GUCIA2, the gene coding for the K2 subunit of of the ferrous heme of soluble guanylate cyclase from the soluble guanylyl cyclase, to position 11q21-q22 on human nitric oxide complex: acceleration by thiols and oxyhemo- chromosome 11, Genomics 33 (1996) 334^336. globin, Biochemistry (1998) in press. [78] G. Gupta, M. Azam, L. Yang, R.S. Danziger, The L2 sub- [94] B. Wedel, P. Humbert, C. Harteneck, J. Foerster, J. Mal- unit inhibits stimulation of the K1/L1 form of soluble gua- kewitz, E. Bo«hme, G. Schultz, D. Koesling, Mutation of nylyl cyclase by nitric oxide. Potential relevance to regula- His-105 in the L1 subunit yields a nitric oxide-insensitive tion of blood pressure, J. Clin. Invest. 100 (1997) 1488^1492. form of soluble guanylyl cyclase, Proc. Natl. Acad. Sci. [79] B. Mayer, D. Koesling, E. Bo«hme, Characterization of nitric USA 91 (1994) 2592^2596. oxide synthase, soluble guanylyl cyclase, and Ca2/calmodu- [95] G. Gupta, J. Kim, L. Yang, S.L. Sturley, R.S. Danziger, lin-stimulated cGMP phosphodiesterase as components of Expression and puri¢cation of soluble, active heterodimeric neuronal signal transduction, Adv. Second Messenger Phos- guanylyl cyclase from baculovirus, Protein Expr. Purif. 10 phoprotein Res. 28 (1993) 111^119. (1997) 325^330. [80] D. Koesling, P. Humbert, G. Schultz, in: S.R. Vincent (Ed.), [96] W.A. Buechler, S. Singh, J. Aktas, S. Muller, F. Murad, R. Nitric Oxide in the Nervous System, Academic Press, San Gerzer, High-level expression of biologically active soluble Diego, CA, 1995, pp. 43^50. guanylate cyclase using the baculovirus system is strongly [81] C. Harteneck, D. Koesling, A. So«ling, G. Schultz, E. Bo«hme, heme-dependent, Adv. Pharmacol. 34 (1995) 293^303. Expression of soluble guanylyl cyclase. Catalytic activity re- [97] Y. Zhao, M.A. Marletta, Localization of the heme binding quires two enzyme subunits, FEBS Lett. 272 (1990) 221^223. region in soluble guanylate cyclase, Biochemistry 36 (1997) [82] A.C. Clark, S.W. Raso, J.F. Sinclair, M.M. Ziegler, A.F. 15959^15964. Cha¡otte, T.O. Baldwin, Kinetic mechanism of luciferase [98] Y. Zhao, J.P. Schelvis, G.T. Babcock, M.A. Marletta, Iden- subunit folding and assembly, Biochemistry 36 (1997) ti¢cation of histidine 105 in the L1 subunit of soluble gua- 1891^1899. nylate cyclase as the heme proximal ligand, Biochemistry 37 [83] A.N. Fedorov, T.O. Baldwin, Cotranslational protein fold- (1998) 4502^4509. ing, J. Biol. Chem. 272 (1997) 32715^32718. [99] L.J. Ignarro, Haem-dependent activation of guanylate cy- [84] D.L. Garbers, Puri¢cation of soluble guanylate cyclase from clase and cyclic GMP formation by endogenous nitric ox- rat lung, J. Biol. Chem. 254 (1979) 240^243. ide: a unique transduction mechanism for transcellular sig- [85] M.S. Wolin, K.S. Wood, L.J. Ignarro, Guanylate cyclase naling, Pharmacol. Toxicol. 67 (1990) 1^7. from bovine lung. A kinetic analysis of the regulation of [100] P.A. Craven, F.R. DeRubertis, D.W. Pratt, Electron spin the puri¢ed soluble enzyme by protoporphyrin IX, heme, resonance study of the role of NOc catalase in the activa-

and nitrosyl-heme, J. Biol. Chem. 257 (1982) 13312^13320. tion of guanylate cyclase by NaN3 and NH2OH. Modula- [86] S.A. Waldman, F. Murad, Cyclic GMP synthesis and function of enzyme responses by heme proteins and their nitro- tion, Pharmacol. Rev. 39 (1987) 163^196. syl derivatives, J. Biol. Chem. 254 (1979) 8213^8222. [87] J.R. Stone, M.A. Marletta, Heme stoichiometry of heterodi- [101] P.A. Craven, F.R. DeRubertis, Restoration of the respon- meric soluble guanylate cyclase, Biochemistry 34 (1995) siveness of puri¢ed guanylate cyclase to nitrosoguanidine, 14668^14674. nitric oxide, and related activators by heme and hemepro- [88] A. Mu«lsch, R. Gerzer, Puri¢cation of heme-containing solu- teins. Evidence for involvement of the paramagnetic nitro- ble guanylyl cyclase, Methods Enzymol. 195 (1991) 377^ syl-heme complex in enzyme activation, J. Biol. Chem. 253 383. (1978) 8433^8443. [89] P. Humbert, F. Niroomand, G. Fischer, B. Mayer, D. Koes- [102] R. Gerzer, E. Bo«hme, F. Hofmann, G. Schultz, Soluble ling, K.D. Hinsch, H. Gausepohl, R. Frank, G. Schultz, E. guanylate cyclase puri¢ed from bovine lung contains Bo«hme, Puri¢cation of soluble guanylyl cyclase from bovine heme and copper, FEBS Lett. 132 (1981) 71^74. lung by a new immunoa¤nity chromatographic method, [103] L.J. Ignarro, J.N. Degnan, W.H. Baricos, P.J. Kadowitz, Eur. J. Biochem. 190 (1990) 273^278. M.S. Wolin, Activation of puri¢ed guanylate cyclase by [90] E.A. Dierks, J.N. Burstyn, The deactivation of soluble gua- nitric oxide requires heme. Comparison of heme-de¢cient, nylyl cyclase by redox-active agents, Arch. Biochem. Bio- heme-reconstituted and heme-containing forms of soluble phys. 351 (1998) 1^7. enzyme from bovine lung, Biochim. Biophys. Acta 718 [91] E.A. Dierks, S.Z. Hu, K.M. Vogel, A.E. Yu, T.G. Spiro, (1982) 49^59. J.N. Burstyn, Demonstration of the role of scission of the [104] R. Gerzer, F. Hofmann, G. Schultz, Puri¢cation of a solu- proximal histidine-iron bond in the activation of soluble ble, sodium-nitroprusside-stimulated guanylate cyclase guanylyl cyclase through metalloporphyrin substitution stud- from bovine lung, Eur. J. Biochem. 116 (1981) 479^486. ies, J. Am. Chem. Soc. 119 (1997) 7316^7323. [105] J.R. Stone, M.A. Marletta, Spectral and kinetic studies on [92] T. Tomita, T. Ogura, S. Tsuyama, Y. Imai, T. Kitagawa, the activation of soluble guanylate cyclase by nitric oxide, E¡ects of GTP on bound nitric oxide of soluble guanylate Biochemistry 35 (1996) 1093^1099. cyclase probed by resonance Raman spectroscopy, Biochem- [106] J.N. Burstyn, A.E. Yu, E.A. Dierks, B.K. Hawkins, J.H. istry 36 (1997) 10155^10160. Dawson, Studies of the heme coordination and ligand bind-

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart 350 J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

ing properties of soluble guanylyl cyclase (sGC): character- D. Koesling, Functional domains of soluble guanylyl cy- ization of Fe(II)sGC and Fe(II)sGC(CO) by electronic ab- clase, J. Biol. Chem. 270 (1995) 24871^24875. sorption and magnetic circular dichroism spectroscopies [116] R.K. Sunahara, A. Beuve, J.J.G. Tesmer, S.R. Sprang, and failure of CO to activate the enzyme, Biochemistry D.L. Garbers, A.G. Gilman, Exchange of substrate and 34 (1995) 5896^5903. inhibitor speci¢cities between adenylyl and guanylyl cy- [107] G. Deinum, J.R. Stone, G.T. Babcock, M.A. Marletta, clases, J. Biol. Chem. 273 (1998) 16332^16338. Binding of nitric oxide and carbon monoxide to soluble [117] P.J. Artymiuk, A.R. Poirrette, D.W. Rice, P. Willett, A guanylate cyclase as observed with resonance Raman spec- polymerase I palm in adenylyl cyclase?, Nature 388 troscopy, Biochemistry 35 (1996) 1540^1547. (1997) 33^34. [108] A.E. Yu, S.Z. Hu, T.G. Spiro, J.N. Burstyn, Resonance [118] P.D. Senter, F. Eckstein, A. Mu«lsch, E. Bo«hme, The stereo- Raman-spectroscopy of soluble guanylyl cyclase reveals dis- chemical course of the reaction catalyzed by soluble bovine placement of distal and proximal heme ligands by NO, lung guanylate cyclase, J. Biol. Chem. 258 (1983) 6741^ J. Am. Chem. Soc. 116 (1994) 4117^4118. 6745. [109] S.Y. Kim, G. Deinum, M.T. Gardner, M.A. Marletta, G.T. [119] C.L. Tucker, J.H. Hurley, T.R. Miller, J.B. Hurley, Two Babcock, Distal pocket polarity in the unusual ligand bind- amino acid substitutions convert a guanylyl cyclase, ing site of soluble guanylate cyclase: implications for the RetGC-1, into an adenylyl cyclase, Proc. Natl. Acad. Sci. control of cNO binding, J. Am. Chem. Soc. 118 (1996) USA 95 (1998) 5993^5997. 8769^8770. [120] E.M. Wilson, M. Chinkers, Identi¢cation of sequences me- [110] M.P. Doyle, J.W. Hoekstra, Oxidation of nitrogen oxides diating guanylyl cyclase dimerization, Biochemistry 34 by bound dioxygen in hemoproteins, J. Inorg. Biochem. 14 (1995) 4696^4701. (1981) 351^358. [121] A. Friebe, D. Koesling, Mechanism of YC-1-induced acti- [111] E. Antonini, M. Brunori, Hemoglobin and Myoglobin in vation of soluble guanylyl cyclase, Mol. Pharmacol. 53 their Reactions with Ligands, vol. 21, North-Holland Publ. (1998) 123^127. Co., Amsterdam, 1971. [122] V.G. Kharitonov, V.S. Sharma, D. Magde, D. Koesling, [112] B. Fan, G. Gupta, R.S. Danziger, J.M. Friedman, D.L. Kinetics of nitric oxide dissociation from ¢ve- and six-co- Rousseau, Resonance Raman characterization of soluble ordinate nitrosyl hemes and heme proteins, including solu- guanylate cyclase expressed from baculovirus, Biochemistry ble guanylate cyclase, Biochemistry 36 (1997) 6814^ 37 (1998) 1178^1184. 6818. [113] J.R. Stone, R.H. Sands, W.R. Dunham, M.A. Marletta, [123] V.G. Kharitonov, M. Russwurm, D. Magde, V.S. Sharma, Electron paramagnetic resonance spectral evidence for the D. Koesling, Dissociation of nitric oxide from soluble gua- formation of a pentacoordinate nitrosyl-heme complex on nylate cyclase, Biochem. Biophys. Res. Commun. 239 soluble guanylate cyclase, Biochem. Biophys. Res. Com- (1997) 284^286. mun. 207 (1995) 572^577. [124] J. Garthwaite, E. Southam, C.L. Boulton, E.B. Nielsen, K. [114] P.S. Yuen, L.K. Doolittle, D.L. Garbers, Dominant nega- Schmidt, B. Mayer, Potent and selective inhibition of nitric tive mutants of nitric oxide-sensitive guanylyl cyclase, oxide-sensitive guanylyl cyclase by 1H-[1,2,4]oxadiazo- J. Biol. Chem. 269 (1994) 791^793. lo[4,3-a]quinoxalin-1-one, Mol. Pharmacol. 48 (1995) 184^ [115] B. Wedel, C. Harteneck, J. Foerster, A. Friebe, G. Schultz, 188.

BBABIO 44737 23-4-99 Cyaan Magenta Geel Zwart