Biochimica et Biophysica Acta 1411 (1999) 334^350

Review

Guanylate cyclase and the c NO/cGMP signaling pathway

John W. Denninger a , Michael A. Marletta
a

aYbYc Y

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 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. 1999 Elsevier Science B.V. All rights reserved.
Keywords: Guanylate cyclase; Cyclic guanosine monophosphate; Nitric oxide; Carbon monoxide; Heme

Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Historical perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Signaling with the c NO/cGMP pathway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. The nucleotide cyclase family: sGC, pGC and AC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335 335 336 337

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 3 5P-triphosphate; c NO, nitric oxide; NO3 2 , 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 : marle@umich.edu 0005-2728 / 99 / $ ^ see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 5 - 2 7 2 8 ( 9 9 ) 0 0 0 2 4 - 9

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350
5. Soluble guanylate cyclase genetics and isoforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Purication from mammalian tissue and expression systems . . . . . . . . . . . . . . . . . . . . . . . 7. Heme environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9. Model of sGC structure and function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 341 342 343 344 346 346 346

335

1. Introduction Soluble guanylate1 cyclase (sGC) is the only conclusively proven receptor for nitric oxide (c NO), a signaling agent produced by the enzyme nitric oxide synthase (NOS) from the amino acid L-arginine. As shown in Fig. 1, sGC catalyzes the conversion of guanosine 5P-triphosphate (GTP) to cyclic guanosine 3P,5P-monophosphate (cGMP). Because it is one of two enzymes (the other is the membrane-bound peptide-receptor guanylate cyclase) that produce cGMP, and because it is the only denitive receptor for c NO, sGC is intimately involved in many signal transduction pathways, most notably in the cardiovascular system (e.g., in the regulation of vascular tone and platelet function) and in the nervous system (e.g. in neurotransmission and, possibly, long-term potentiation and depression). Soluble guanylate cyclase is a heterodimeric protein composed of K and L subunits; in addition, it is a hemoprotein (see Fig. 2). c NO binds to the sGC heme; this binding event activates the enzyme. The activation of sGC and the subsequent rise in cGMP concentration are what allow sGC to transmit an c NO signal to the downstream elements of the signaling cascade-cGMP-dependent
1 Although we prefer `guanylate' cyclase (as does the Index Medicus and the Oxford Dictionary of Biochemistry and Molecular Biology), `guanylyl' cyclase is also used. Neither of the common names for this enzyme is technically correct: `guanylate' refers to the anion of guanylic acid (guanosine monophosphate) and `guanylyl' refers to the acyl group derived from guanylic acid.

protein kinase, cGMP-gated cation channels and cGMP-regulated phosphodiesterase. As will be described below, nearly 30 years of study of sGC have given us an outline of sGC structure and function; however, many facets of sGC structure and function, especially at the molecular level, remain to be discovered. 2. Historical perspective Although both sGC and its product, cGMP, were identied in the 1960s, it has been only within the last 10 years that the complete outline of the c NO/ cGMP signaling pathway has been understood. After the discovery of cAMP in 1958 [1] and the initial elucidation of its role in physiology (for early reviews see [2,3]), the search for other physiologically relevant cyclic nucleotides began. In 1963, cGMP was detected in the urine of rats [4]; several years later, guanylate cyclase was found in mammalian tissues [5^9]. In 1977, c NO 9 both as c NO gas and as c NO released from vasodilators like nitroglycerin and nitroprusside 9 was shown to activate guanylate cyclase [10^13]. At that point, though, c NO was unknown in biological systems, except when exogenously supplied by drugs like nitroglycerin, so it was assumed that c NO activation of sGC was a nonphysiological, if useful, phenomenon. The story soon became clear, however. In 1980, Furchgott et al. were able to demonstrate the existence of a substance produced by the endothelium that was required to mediate relaxation of blood

336

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

vessels [14]. They designated this substance endothelium-derived relaxing factor (EDRF). In 1986, Ignarro and Furchgott independently suggested that EDRF was c NO (reviewed in [15]); the next year, it was demonstrated that this was the case [16,17]. At the same time, our laboratory and others were un3 raveling the story of nitrite (NO3 2 ) and nitrate (NO3 ) production in mammals. After initial observations by Tannenbaum and coworkers that mammals produce 3 NO3 3 and that NO3 synthesis is increased following immune stimulation [18^20], it was shown that NO3 2 and NO3 3 were produced by immunostimulated macrophages from the amino acid arginine (reviewed in [21,22]). Shortly thereafter, our laboratory and others demonstrated that a soluble enzyme in murine macrophages, later puried and named nitric oxide synthase, was producing c NO from the amino acid arginine [23^25]. At this point, the major components of the c NO/cGMP signaling pathway had been identied: nitric oxide synthase produces c NO, which activates soluble guanylate cyclase, leading to a subsequent increase in the concentration of cGMP. 3. Signaling with the c NO/cGMP pathway Since the initial elucidation of the c NO/cGMP system, many good examples of signal transduction by this system have come to light, some better characterized than others. They include: smooth muscle relaxation and blood pressure regulation [26]; platelet aggregation and disaggregation [27]; and neurotransmission both peripherally, in non-adrenergic, non-cholinergic (NANC) nerves [26], and centrally, in the processes of long term potentiation and depression [28]. Understanding the c NO/cGMP pathway, however, requires an understanding of c NO's role as a signaling agent, the alternative targets for c NO, the alternative activators of sGC and, naturally, the targets for cGMP. It was initially a surprise that c NO was produced physiologically: prior to the demonstration of c NO synthesis by nitric oxide synthase, the biological relevance of c NO was as a pollutant; for example, as a component of automobile exhaust. Given the physical properties of c NO, however, it is now clear that c NO is uniquely well suited to its role as a signaling agent. Because it is small and carries no charge, c NO

diuses freely across cell membranes, obviating the need for a transport system. c NO is relatively unreactive for a radical species, so that it can diuse from a cell in which it is produced to neighboring cells without being destroyed before it reaches its target. Be3 cause c NO is oxidized to NO3 2 and NO3 under physiological conditions, no separate mechanism for its destruction is required [29^31]; however, at low concentrations of c NO and in the absence of proteins and metals which contribute to its destruction, the decomposition of c NO will be slow [29]. Most importantly, the electronic structure of c NO makes it an excellent ligand for heme, allowing it to bind to the sGC heme at a low, non-toxic concentration. The one well-dened target for c NO, of course, is sGC; however, several paths for c NO signal transduction which do not rely on the production of cGMP by sGC have been proposed, though they are less clearly dened than the c NO/cGMP pathway. First, c NO was shown to inhibit mitochondrial cytochrome oxidase at concentrations of c NO found physiologically [32,33]. Second, c NO was reported to activate calcium-dependent potassium channels in vascular smooth muscle directly, though it is possible that this only happens with levels of c NO higher than typically seen in vivo [34]. Third, c NO was shown to inhibit activation of NF-UB [35,36]. Fourth, c NO was found to activate G-proteins (such as protooncogene p21ras) by enhancing the rate of nucleotide exchange, leading to an increase in the GTP-bound form. In fact, a p21ras cysteine residue (C118) was shown to be S-nitrosylated in the presence of c NO [37^39]. Fifth, caspase, a cysteine protease involved in apoptosis, was shown to be inhibited by S-nitrosylation of its active-site cysteine [40]. Finally, both hemoglobin and the cardiac calcium release channel (i.e., the ryanodine receptor) were shown to be S-nitrosylated in vivo; moreover, this change resulted in a modication of their function [41^43]. As these alternative c NO targets become better dened, it will be important to ensure that they are physiologically relevant. c NO, especially at high concentration, can react with O2 through a series of reactions to form N2 O3 , which is a potent nitrosating agent reacting capable of reacting with nucleophiles such as the sulfur of cysteine residues. As a result, in vitro experiments with high c NO concentrations may lead to higher than normal levels of S-nitrosocysteine

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

337

formation. Therefore, while there is no doubt that Snitrosylation occurs in vivo, we believe its physiological signicance will not be established until two observations are made: rst, that S-nitrosylation 9 at in vivo levels of c NO 9 modies the function of target proteins and, second, that in vivo concentrations of S-nitrosylated proteins can elicit a physiological response. In addition to other targets for c NO, there are other potential physiological activators of sGC. These fall into two categories 9 non-c NO nitrogenous compounds and carbon monoxide (CO). Although most investigators are now convinced that c NO itself is the signaling agent which activates sGC, there are still some who believe that there are other nitrogenous products (not c NO) which activates sGC. Many alternatives have been proposed, such as nitrosothiols. It is not clear, however, that these agents are present in high enough concentration in vivo to perform any signicant signaling function. Likewise, it has been hypothesized that CO, which binds to the sGC heme and weakly activates the enzyme, may act as a signaling agent [44^46]. In contrast to c NO, which activates sGC approx. 400fold, CO only activates known isoforms of sGC approx. 5-fold [47]. Additionally, CO activation requires saturating concentrations of CO, a clearly non-physiological alternative. Although CO is produced by the enzyme heme oxygenase during the breakdown of heme, it is dicult to imagine that 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 a mechanism to increase the response of sGC to CO, we will have to reserve judgment on this potential pathway. Such a mechanism may have been found, however. It has been shown recently that CO and the synthetic compound YC-1 ([48]; see Fig. 4 for the structure) synergistically activate sGC to levels similar that of the c NO-stimulated enzyme [49,50]. If a physiological counterpart for YC-1 is identied, this could be a mechanism by which CO could act as a signaling agent. Regardless of the species that activates sGC, all signal transduction through sGC (and pGC as well) takes place through an increased concentration of cGMP. Cyclic GMP, like cAMP, is a second messenger; as such, it requires target proteins to have its

eect. Moreover, the diversity of targets allows cGMP to have wide-ranging eects that can dier by cell and tissue type. There are three known targets for cGMP which mediate the transmission of the c NO/cGMP pathway signal downstream from guanylate cyclase: cGMP-dependent protein kinase [51], cGMP-regulated phosphodiesterase [52,53] and cGMP-gated ion channels [54]. The rst of these, cGMP-dependent protein kinase, phosphorylates target proteins in response to an increase in cGMP concentration. For example, in smooth muscle cells, cGMP-dependent protein kinase phosphorylates the inositol 1,4,5-triphosphate receptor, leading to a decrease in Ca2 concentration and, ultimately, smooth muscle relaxation. The second, cyclic nucleotide phosphodiesterase, catalyzes the hydrolysis of the 3P-phosphodiester bond of cAMP and cGMP to yield AMP and GMP, respectively. Of the seven known families, three (PDE2, PDE5 and PDE6) are allosterically regulated by cGMP and one (PDE3) is inhibited by the binding of cGMP in its substrate site. Phosphodiesterases allow crosstalk between cGMP and cAMP signaling pathways because they cause the concentration of one cyclic nucleotide to inuence the degradation of the other. The third, cyclic nucleotide-gated ion channels, are non-specic cation channels found in a variety of tissues. Most notably, these channels are found in the retina and in olfactory epithelium where they are involved, respectively, in visual phototransduction and in olfaction. 4. The nucleotide cyclase family: sGC, pGC and AC Just as sGC is a component of the larger c NO/ cGMP signaling pathway, so is it a member of a larger group of related enzymes: the nucleotide cyclases. This family of enzymes converts nucleotide triphosphates (GTP and ATP) to cyclic nucleotide monophosphates (cGMP and cAMP) and includes adenylate cyclase (AC), particulate guanylate cyclase (pGC) and soluble guanylate cyclase. As a family, these enzymes are involved in a broad array of signal transduction pathways mediated by the cyclic nucleotides cAMP and cGMP-vision, blood pressure regulation, response to hormones and many others. An evolutionarily conserved catalytic domain denes these three groups of enzymes as a single family; in

338

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

Fig. 1. The guanylate cyclase reaction and c NO signal transduction.

other words, the genes that code for the catalytic portions of these enzymes are derived from a common ancestor. This is unsurprising, given that they catalyze virtually identical reactions. Moreover, because of their relatedness and similar function, insights into the structure, catalytic mechanism and possibly even regulation of one member of this family may be generalizable to other members. Adenylate cyclase, for example, has been studied for nearly four decades. As a result, there is a rich literature of ndings with AC that may, because of the similarity of AC to sGC, prove useful in understanding sGC. Unlike the guanylate cyclases, which contain both receptor and catalytic functions within the same protein, ACs are stimulated by hormone receptors acting through heterotrimeric G-proteins [55]. AC is an integral membrane protein and is assumed to be monomeric. All of the known mammalian isoforms are activated by Gs K ; other activators are isoform specic [56]. AC is also activated by the small molecule forskolin, a diterpene derivative isolated from the Indian plant Coleus forskohli. Forskolin was originally shown to have a variety of cardiovascular eects [57] and was later shown to activate AC [58]; it has served as a useful tool in the study of AC, but no mammalian counterpart has been iden-

tied. As shown in Fig. 2, AC contains two groups of six transmembrane helices and two intracellular loops. The intracellular loops each contain a region 9 C1 and C2 , respectively 9 now known to come together in a head-to-tail fashion to form a functional catalytic domain. (Because the domain is made up of two regions from a single polypeptide chain, it is sometimes referred to as `pseudo-heterodimeric'.) Two crystal structures of catalytic domains have, in fact, been solved recently. The rst was a structure of a C2 C2 homodimer [59], the second of a C1 C2 heterodimer [60]; both were engineered constructs which lacked the transmembrane domains and were, thus, soluble. The crystal structure of C1 C2 supports the conclusion that this pseudo-heterodimeric catalytic domain has two pseudo-symmetric binding sites: one binds substrate, the other forskolin. In addition, Gs K binds to a site in the plane of the catalytic domain opposite the substrate binding site. Particulate guanylate cyclase has also provided insight into the structure and function of sGC. After the purication of cGMP and guanylate cyclase, it was observed that guanylate cyclase activity could be separated into a membrane-bound (particulate) and cytosolic (soluble) fractions. As its name implies, particulate guanylate cyclase is the membrane-bound

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 cell-surface receptor enzyme: it contains an extracellular peptide receptor domain and an intracellular catalytic domain separated by a single transmembrane domain (Fig. 2). pGC is now generally thought to be a homodimer of subunits of approx. 120 kDa [61]. Seven subclasses of mammalian pGC (GC-A through GC-G) have been identied. Three of these have known peptide ligands (GC-A, atrial natriuretic peptide and brain natriuretic peptide; GC-B, C-type natriuretic peptide; GC-C, the heat stable enterotoxins of Escherichia coli and guanylin); the rest (GC-D, olfactory epithelium; GC-E and GC-F, retina; GC-G lung, intestine, skeletal muscle) are so-called `orphan' receptors because their activating ligands have not been determined [55,61,62]. For the purposes of understanding sGC, pGC is useful because it helps to dene the characteristics that are specic to guanylate cyclases, for example, the identity of the amino acids which allow the dierent members of the nucleotide cyclase family to distinguish GTP from ATP [63]. Also, given the similarity of their catalytic domains, it is possible that the general

mechanism of activation of pGC will be similar to that of sGC. 5. Soluble guanylate cyclase genetics and isoforms Since the original identication of sGC, a great deal of information about sGC genetics has become available. Thus far, a total of 12 sGC cDNAs have been cloned from ve dierent species (for identity and similarity comparisons see Fig. 3). The rst were cloned on the basis of protein sequence from puried bovine and rat K1L1 heterodimer. The cDNA sequences for bovine and rat L1 subunits, respectively, were reported in 1988 [64,65], followed by the bovine and rat K1 subunits in 1990 [66,67]. Subsequently, K1L1 pairs were cloned from Homo sapiens (originally called K3 and L3) ([68,69], localized to chromosome 4q32, [70]), Drosophila melanogaster [71^73] and the sh Oryzias latipes [74]. Putative sGC sequences exist in the Caenorhabditis elegans genome, but these have yet to be conrmed with any empirical studies beyond the sequencing itself.

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 denitions 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 cDNAs for two additional sGC subunits have been cloned: L2 from rat kidney [75] and K2 from human fetal brain ([76], localized to chromosome 11q21-q22 [77]. In transient transfection experiments, the K2 subunit has been shown to form an active heterodimer with the L1 subunit [76]. Although the L2 subunit has been recently reported to form an active heterodimer with the K1 subunit in transient expression experiments [78], similar experiments in our laboratory (unpublished results) and others (unpublished results, see [79,80]) were unable to generate an active heterodimer containing the L2 subunit. Thus, the issue of the L2 subunit partner remains to be conclusively proven. A unique feature of the L2 subunit is that its predicted amino acid sequence contains a COOH-terminal consensus sequence

(-CVVL) for isoprenylation/carboxyl-O-methylation [75], but it is not known whether such a modication exists on the mature polypeptide. Because full-length sequences of the K2 and L2 subunits from species other than the ones from which they were originally isolated have yet to be identied, one might question their signicance. However, a partial sequence for the bovine K2 has been determined [76] and a putative L2 subunit partial sequence exists in a mouse placenta EST database (GenBank accession No. AA023957, from the Soares mouse placenta library). These subunits, then, may well be common to all mammals and perhaps to other sGC-containing species. The potential existence of novel sGC isoforms (i.e., in addition to the K1L1) raises an important issue about the distinction between sGC subunits and

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

341

Fig. 4. Activators and inhibitors of sGC. Shown in panel A are compounds known to interact directly with sGC. As described in the text, YC-1 is an activator of sGC that acts synergistically with CO to yield activity comparable to that observed with c NO. ODQ is a relatively potent inhibitor of the enzyme, although the mechanism of action is not clearly established. It has been reported to oxidize the ferrous heme moiety [124]. Shown in panel B are compounds that have been reported to aect catalytic activity. Methylene blue and LY83583 are reported to inhibit the enzyme and isoliquiritigenin has been reported to activate the enzyme. Studies with these compounds have typically been carried out with crude preparations of the enzyme and so the exact mechanism of action is not clear.

existence) of sGC isoforms other than the K1L1 is 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 coexpression 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 the same cell) does basal and c NO-stimulated activity result [67,73,81]. The sGC requirement for co-expression is similar to that of bacterial luciferase, which has been investigated extensively in the Baldwin laboratory ([82], for a review of co-translational protein folding, see [83]). It is possible that sGC, like bacterial luciferase, has one subunit that can, in addition to normal heterodimerization, form a kinetically trapped homodimer or fold to a dimerization-incompetent monomer. 6. Purication from mammalian tissue and expression systems Although the existence of other isoforms of sGC besides the K1L1 has been suggested by cloning and expression experiments, only the K1L1 has been puried from mammalian tissue. In the past, a number of dierent sources for sGC, including lung, liver, and brain, were used for purications. Early purications (see, for example, [84,85]) produced enzyme with little or no bound heme [86]. Because of the high sGC content and ready availability, bovine lung is now used as a source for purication from mammalian tissue. The groups currently purifying sGC from bovine lung use a variety of purication procedures [47,87^92]. Most of the more recent purications have been modied to increase the yield of heme-containing sGC. In addition, yields have been increased: purication from tissue can yield up to 3^ 5 mg of pure protein (S. Kim, M.A. Marletta, unpublished data; [92]). Because isolation of sGC from bovine lung is typically arduous, there has been strong motivation for

sGC isoforms. Our denition of an sGC isoform is a catalytically competent heterodimer (in principle, however, it could be a homodimer). For AC (which exists as a monomer) and pGC (which exists as a homodimer), the identication of dierent isoforms 9 with dierent tissue or developmental distribution, dierent regulation and/or dierent kinetic parameters 9 has been relatively simple: each distinct gene product is a distinct isoform. For sGC, the picture is somewhat more complicated. Given that four distinct sGC subunits have been cloned, there are 16 possible homodimeric and heterodimeric combinations of sGC subunits. Each of these combinations, if shown to exist in vivo, could be rightly called an isoform. The only isoform, therefore, that has been shown to exist at the protein level in animal tissues is the K1L1. Again, as mentioned above, putative K2L1 and K1L2 isoforms have been demonstrated in transient mammalian expression experiments [76,78], but until these subunits are shown to exist in vivo and, ideally, puried from tissue sources, they cannot be considered bona de sGC isoforms. Because the role (even the

342

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

the development of alternative sources of sGC, namely, the development of expression systems. Both transient and stable expression of sGC in mammalian expression systems using COS cells (an African green monkey kidney cell line) and HEK293 cells (a human embryonic kidney cell line) has successfully produced active cyclase in several laboratories [67,73,76,78,81]. Unfortunately, while these systems have been useful for exploring the catalytic activity of sGC isoforms and mutants, the low level of protein produced has made them unsuitable as sources of puried protein. In order to take advantage of an expression system that can generate amounts of sGC suitable for purication, several labs (including ours) have used expression and purication of sGC in the baculovirus/Sf9 system [93^96]. In general, purication of sGC from baculovirus/Sf9 expression systems is more straightforward because it requires a lower purication factor. Obviously, the ability to express heterodimeric sGC at high levels in E. coli would be a tremendous boon to the eld. Many laboratories have doubtless tried to do just that and have been, to this point, unsuccessful 9 and here we can judge only by our own experience, given the lack of published reports 9 because sGC tends to form insoluble inclusion bodies when expressed in E. coli. To our knowledge, the only sGC construct that has been successfully overexpressed in E. coli is the L1 subunit heme binding region that has been expressed and puried in our laboratory [97,98]. 7. Heme environment Although there has been some debate in the past, it is clear now that sGC binds one protoporphyrin IX type heme per heterodimer. The early history of the determination that sGC is a hemoprotein is reviewed in [99]. After initial observations that heme was needed to allow stimulation of sGC by c NO in crude preparations [100,101], sGC was shown to contain a prosthetic heme moiety that was important in activation by c NO and c NO donors [85,102^104]. Because the early purications yielded sGC with heme stoichiometries less than one (often far less than one) [86], it was assumed that the heme stoichiometry was, in fact, one. After the cloning of both K and L sGC subunits and the cloning of some of the pGC iso-

forms, it became clear that the COOH-terminal portion of both sGC subunits was most likely catalytic in nature. In addition, as our laboratory noted [47,87], the NH2 -terminal portions 9 upstream from the conserved catalytic regions 9 of both the K and L subunits were homologous to each other (i.e., 30% identical and 39% similar), though less homologous at the extreme NH2 terminus. This observation led us to hypothesize that both subunits might bind heme. Initially, our heme stoichiometry determination for sGC supported that hypothesis [87]; however, recent work in our laboratory has demonstrated that there had been an error in the quantitative amino acid analysis used to determine sGC concentration. As a result, we have denitively determined the heme stoichiometry to be one heme per heterodimer ([93]; S. Kim, P.E. Brandish, M.A. Marletta, unpublished data). The hypothesis that the sGC heme is ligated by a histidine residue, much like the well-studied heme proteins hemoglobin and myoglobin, is not new; however, it has been only within the last several years that that suspicion has been conrmed and the ligand identied as histidine 105 of the L1 subunit (the numbering applies to bovine, human and rat L1). Early observations of the sGC heme Soret (i.e., the maximal heme absorbance peak) as isolated [102] hinted that sGC was likely to bind heme in a way similar to deoxyhemoglobin and deoxymyoglobin, 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 mutant of sGC in which histidine 105 of the bovine L1 subunit had been mutated to phenylalanine (H105F) and transiently expressed in COS cells was found to have lost stimulation by c NO and, in addition, to lack any heme absorbance [94]. Second, electronic absorption spectroscopy on pure sGC revealed a spectrum that was entirely consistent with a histidine-ligated heme, namely, a Soret peak at 431 nm and a broad K/L band at 562 nm [47,105]. Additional observations with electronic absorption spectroscopy [106], magnetic circular dichroism spectroscopy [106], resonance Raman spectroscopy (RR) [107,108] and Fourier transform infrared spectroscopy (FTIR) [109] conrmed that the heme ligand was histidine. Subsequent observations in our laboratory denitively identied L1-His105 as the heme ligand: the

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

343

NH2 -terminal portion of the L1 subunit (L11385 ) expressed and puried from E. coli was capable of binding heme with spectral characteristics virtually identical to the heterodimeric enzyme [97]; mutants of His105 did not bind heme [98]; and heme binding to the H105G mutant could be rescued by including imidazole during growth and purication [98]. The majority of groups studying sGC purify the enzyme in a ve-coordinate, high-spin histidyl complex that resembles deoxyhemoglobin and deoxymyoglobin in terms of ligation, oxidation state and spectral characteristics. Unlike hemoglobin and myoglobin, however, sGC has an extremely low anity for oxygen, which fails to bind to the sGC heme even under 100% O2 [47]. The molecular basis for this property is still under investigation. The failure of sGC to bind oxygen enables sGC to act as a receptor for c NO by allowing c NO to bind directly to the ferrous-deoxy heme; in contrast, c NO will react with ferrous-oxy complexes (like oxyhemoglobin) to give nitrate and ferric heme [110]. The Soret position for sGC heme (431 nm) coupled with a broad K/L peak at 562 nm [47,105] is consistent with ve-coordinate high-spin histidyl-ligated proteins and model compounds [111]. Despite this seeming consensus, however, there have been reports that sGC, as isolated, is a mixture of ve-coordinate high-spin histidyl heme and six-coordinate low-spin bis-histidyl heme [95,106,108,112]. It is not clear whether heme reconstitution, or some other dierence in procedure, is the cause of these multiple populations of enzyme. Extensive study of the sGC heme in its c NO-bound (NO-sGC) and CO-bound (CO-sGC) forms has given us a clear picture of the characteristics of these species. NO-sGC is a ve-coordinate nitrosyl heme, in which the Fe-His bond has been broken [47,92,107,108,112,113]. A combination of electronic absorption spectroscopy, electron paramagnetic resonance spectroscopy (EPR) and RR has been used to show this. When CO is bound to the sGC heme, the Fe-His bond apparently remains intact, forming a six-coordinate heme [47,92,106,108,112]. Observations by electronic absorption spectroscopy and RR have demonstrated this to be the case. Table 1 shows a summary of electronic absorption spectroscopy maxima for sGC in its as isolated, CO-bound and c NO-bound forms; Table 2 shows RR and FTIR vibrations for the heme ligands of these forms.

8. Catalysis Several lines of evidence have converged to unambiguously dene the catalytic domain of sGC: sequence comparisons to AC and pGC to dene an evolutionarily conserved region, site-directed mutagenesis to produce inactive enzyme, the construction of deletion mutants that retain catalytic function, and the creation of mutants with altered substrate specicity. The rst method that was used to dene the catalytic regions of AC, pGC and sGC was alignment of their predicted amino acid sequences (once the cDNA sequences became available). These alignments revealed a highly homologous region approx. 200 amino acids long that was common to all three enzymes. Two regions were found in the AC sequence (C1 and C2 ), one region in pGC monomers and one region in each half of the sGC heterodimer (K1 and L1). Several additional pieces of evidence showed that these regions in the K1 and L1 subunits of sGC came together to form the catalytic domain. First, the mutation of two aspartic acid residues in the sGC K1 subunit each abolished catalytic activity in co-expressed heterodimers [114]. Second, Wedel and coworkers demonstrated that co-expressed NH2 -terminal deletion mutants of K1 (deletion of the rst 366 amino acids) and L1 (deletion of the rst 305 amino acids) retained catalytic activity [115]. Finally, any remaining doubt that these regions formed the catalytic domain was removed by the recent nding that sGC can be made into an c NO-responsive adenylate cyclase with changes to these regions alone: by mutating one residue in the K1 subunit and two in the L1 subunit in the putative catalytic regions that were predicted 9 on the basis of the AC
Table 1 Peak positions and extinction coecients from electronic absorption spectroscopy of sGC sGC form Ferrous CO complex NO complex Ferric Soret 431 423 399 391 (111) (145) (85) (110) L 562 541 537 511 (12) (12) (11) (14) K 567 (12) 572 (11) Por-Fea 770 (0.2)

642 (4)

Values for this table were drawn from [47,105]. Extinction coefcients (mM31 cm31 ) are given in parentheses. a Por-Fe refers to the putative porphyrin to iron charge transfer band.

344

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

catalytic domain crystal structures 9 to be important in nucleotide base binding [59,60], Sunahara and coworkers were able to change the nucleotide specificity of sGC [116]. Many aspects of the catalytic mechanism of sGC are poorly understood. The type of reaction catalyzed by sGC 9 the attack of an activated hydroxyl on the K-phosphate of a nucleotide triphosphate 9 is a common one. This is, for example, the same reaction catalyzed in the polymerization of deoxynucleotide triphosphates in the synthesis of DNA; however, in the case of DNA it is an intermolecular attack of the 3P-hydroxyl on an K-phosphate. In fact, after publication of the AC C2 C2 structure, it was suggested that AC catalytic domain contains a fold similar to the `palm' domain in prokaryotic DNA polymerase I [117]. It is possible that some of the features of the DNA polymerase mechanism will hold true for the nucleotide cyclases. Thus, the outline of the reaction is clear: the 3P-hydroxyl is activated in some way to become more nucleophilic and attacks the Kphosphate, causing the release of the L- and Q-phosphates as inorganic pyrophosphate. What is not clear is how the 3P-hydroxyl is activated, whether pyrophosphate release is concerted or proceeds through a pentavalent phosphorous transition state and what role divalent metal cations play in the mechanism. Despite these questions, however, some possibilities for the sGC mechanism have been eliminated. For example, the displacement of PPi by the 3P-hydroxyl of GTP has been shown to occur with inversion of

stereochemistry, which indicates that the reaction proceeds by a single displacement reaction [118]. 9. Model of sGC structure and function In order to synthesize some of the latest sGC developments into a model 9 albeit a working model 9 we have divided sGC enzymology into four subareas that represent what are, in our view, the most important open questions: structure, activation, catalytic mechanism and deactivation. No X-ray crystal structure of sGC or of any sGC domain has been solved; as a result, any model of sGC structure must draw heavily on related proteins: The high degree of homology (approx. 30% identity and 40% similarity) between the catalytic domain of AC and the catalytic domain of sGC, in addition to their almost identical catalytic function, suggests that these catalytic domains are descended from a common ancestor and, as such, are likely to share a common three-dimensional fold. In fact, the Hurley group has used the AC catalytic domain structure determined in his laboratory to create a homology model of the catalytic domain of sGC [63]. Such models have been used to correctly predict the amino acid residues that control the substrate specicity of both ACs and GCs. For example, the substrate specicity of retinal guanylate cyclase (a pGC) was changed from GTP to ATP [119], sGC was converted into an c NO-activated AC and AC was converted into a non-specic (but still Gs K-activated) purine nucleotide cyclase [116]. As a result, the X-ray crystal structures of the AC C1 C2 heterodimer [60] and C2 C2 homodimer [59] appear to be excellent models for the sGC catalytic domain. We would concur, then, with the prediction made by others [63], that the COOH-terminal portions of the K and L sGC subunits form a head-to-tail catalytic domain. If the pseudo-heterodimeric AC catalytic domain is any guide, sGC may well have only one substratebinding site. As for the rest of the protein, the lack of direct evidence (even by analogy) makes it more dicult to choose among the possibilities; still, there are ndings upon which we can continue to build the model. The homology between the K and L subunits of sGC (33% identity and 42% similarity overall, although

Table 2 Resonance Raman and FTIR vibrations for sGC heme ligands sGC form Ferrous X(Fe-His) CO complex X(Fe-CO) N(Fe-C-O) X(C-O) NO complex X(Fe-NO) X(N-O) cm31 204 472 562 1987 525 1677 Description Fe-His stretch Fe-CO stretch Fe-C-O bending C-O stretch Fe-NO stretch N-O stretch

The values in this table were drawn from [107,109]. Similar values have been reported by other laboratories [91,92,108,112] ; some dierences have been reported for reconstituted sGC [91,108,112].

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

345

higher in the COOH-terminal catalytic region than in the NH2 -terminal regulatory region) suggests that the fold of the two subunits may be pseudo-symmetrical. The K and L subunits of sGC have a predicted amphipathic helical region just NH2 -terminal to the predicted catalytic domain. These helices may be evolutionarily related to the predicted amphipathic helices NH2 -terminal to the catalytic region of pGC, which have been demonstrated to mediate dimerization [120]. This conclusion is supported by the homodimerization of the L11385 [97] and by the dimerization of K1 and L1 subunit NH2 -terminal deletion mutants [115], which retain the putative dimerization helices. The conformational change which shifts the AC catalytic domain to its active state upon binding of Gs K may be similar to the conformational change which shifts the sGC catalytic domain to its active state when c NO binds to the sGC heme. It is possible, then, that the sGC heme domain (the NH2 -terminal portion of the L subunit) abuts the catalytic domain in the same way that Gs K abuts the AC catalytic domain [60]. This would put both the K and L subunit NH2 -terminal domains in the plane of the catalytic domain (which is roughly disk-shaped) 9 the heme domain opposite the substrate site (in the same way Gs K is opposite the AC substrate site) and the K subunit NH2 -terminal portion next to the substrate site (opposite the site analogous to where forskolin binds in the AC structure). Our model of sGC activation by c NO does not dier substantially from the following simplied picture. Binding of c NO to the sGC heme severs the bond between the heme Fe and the ligating histidine, presumably puckers the heme toward the distal nitrosyl ligand and, through some combination of changes in heme geometry, causes a conformational change that activates the enzyme. Recent work, however, has rened our view of the heme environment of sGC in a way that allows us to rene our working model of activation. As mentioned above, YC-1 is an activator of sGC that has the interesting property of synergistically activating sGC in the presence of CO (see Fig. 4 for the structure of YC-1 and other sGC activators and inhibitors). Electronic absorption spectroscopy studies have shown that sGC activated by YC-1 and CO has a six-coordinate heme, even though its level of activity is similar to the c NO-activated enzyme [49,121]. Therefore, the breaking of the

Fe-His bond which occurs with c NO binding may be correlated with the molecular event that brings about activation, but may not directly cause it. In other words, it is possible that the breaking of the Fe-His bond can be eliminated as a step in the activation of sGC. There may be, then, some other eect of c NO binding to the heme that leads to activation. We are currently pursuing this possibility in our laboratory. Our model of the sGC catalytic mechanism is based largely on proteins that catalyze similar reactions. In addition to this, we have incorporated what is known about the nucleotide cyclase catalytic mechanisms through crystal structures and mutational studies. The active site functions/characteristics that need to be elucidated are as follows: the number of substrate binding sites; the number of Mg2 ions participating in substrate binding and/or catalysis; the amino acid residues important for substrate binding; the amino acid residues involved in Mg2 binding; and the amino acid residues involved in the catalytic mechanism (e.g., activation of the 3P-OH and transition-state stabilization). There are almost certainly no more than two substrate-binding sites in sGC; given that the catalytic domain of sGC, like that of AC, is made up of homologous but non-identical components, we hypothesize that there is only one substrate-binding site, as in AC. This hypothesis also ts in well with the obvious asymmetry of the sGC regulatory region: the L1 subunit binds heme, the K1 does not. As for important residues, the mutation of rat K1D513 and K1D529 each abolished catalytic activity when transiently expressed with wild-type L1 subunit [114], but it is dicult to pinpoint the role that these residues may be playing without further study. The nal component of our model of sGC structure and function is the question of deactivation, which remains an open question because of the following observation: the deactivation of sGC in organ bath studies has suggested that the half-life of sGC activation is on the order of minutes [17], whereas the half-lives of most nitrosyl-heme complexes are longer. Two hypotheses have been advanced to explain this discrepancy. First, the sGC heme may simply have dierent characteristics that allow c NO to dissociate more rapidly. Second, there may be some extrinsic factor 9 a small molecule or protein 9 which assists in the deactivation either by

346

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350

directly interacting with the ferrous-nitrosyl heme or via an allosteric eect that increases the rate of c NO dissociation from the heme. Studies of the sGC heme, for example, showed that its properties were dierent from most histidine-ligated heme proteins and model compounds: the half-life (extrapolated to 37C) for c NO dissociation was on the order of 2 min [122]. Recent work in our laboratory has shown that thiol compounds like glutathione, at concentrations similar to that found in vivo, increase the rate of c NO release from the heme [93]. Compounds that trap c NO, like oxyhemoglobin, also increase the rate of c NO 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 under these conditions complicates the determination of such rates [93]. We believe, therefore, that the relatively fast c NO o-rate for the sGC heme [122] and contribution of thiol compounds like glutathione account for most, if not all, of the in vivo rate of deactivation. 10. Conclusion We have attempted in this review to provide some insight into the structure and function of soluble guanylate cyclase. In doing so, we have placed sGC in its larger context by discussing the history of the sGC eld, the physiological importance of the c NO/ cGMP pathway and its components, and the nucleotide cyclase family to which sGC belongs. We have delved into the current understanding of sGC by discussing what is known about sGC isoforms, expression and purication, heme environment and catalysis. And nally, we have presented a working model of sGC structure, activation, catalytic mechanism and deactivation. Naturally, a great deal of work remains to be done to complete our understanding of sGC. Two important milestones in achieving that progress will be the identication and characterization of sGC subunits and isoforms from several species and the development of an sGC expression system capable of producing large quantities of the heterodimeric enzyme. With or without these breakthroughs, the most pressing problems to address for sGC are: (1) determining the structure,

(2) understanding the catalytic mechanism and (3) piecing together the molecular details of activation and deactivation. As these questions are answered, a complete view of sGC structure and function will emerge. Acknowledgements We would like to thank all of the Marletta lab for many thought-provoking discussions and for their insightful comments on this manuscript.

References
[1] T.W. Rall, E.W. Sutherland, Formation of a cyclic adenine ribonucleotide by tissue particles, J. Biol. Chem. 232 (1958) 1065^1076. [2] J.G. Hardman, G.A. Robison, E.W. Sutherland, Cyclic nucleotides, Annu. Rev. Physiol. 33 (1971) 311^336. [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, Isolation of adenosine 3P,5P-monophosphate and guanosine 3P,5P-monophosphate from rat urine, Biochem. Biophys. Res. Commun. 11 (1963) 330^334. [5] A.A. White, G.D. Aurbach, Detection of guanyl cyclase in mammalian tissues, Biochim. Biophys. Acta 191 (1969) 686^ 697. [6] G. Schultz, E. Bo hme, K. Munske, Guanyl cyclase. Determination of enzyme activity, Life Sci. 8 (1969) 1323^1332. [7] N.D. Goldberg, S.B. Dietz, A.G. O'Toole, Cyclic guanosine 3P,5P-monophosphate in mammalian tissues and urine, J. Biol. Chem. 244 (1969) 4458^4466. [8] E. Ishikawa, S. Ishikawa, J.W. Davis, E.W. Sutherland, Determination of guanosine 3P,5P-monophosphate in tissues and of guanyl cyclase in rat intestine, J. Biol. Chem. 244 (1969) 6371^6376. [9] J.G. Hardman, E.W. Sutherland, Guanyl cyclase, an enzyme catalyzing the formation of guanosine 3P,5P-monophosphate from guanosine triphosphate, J. Biol. Chem. 244 (1969) 6363^6370. [10] S. Katsuki, W. Arnold, C. Mittal, F. Murad, Stimulation of guanylate cyclase by sodium nitroprusside, nitroglycerin and nitric oxide in various tissue preparations and comparison to the eects of sodium azide and hydroxylamine, J. Cyclic Nucleotide Res. 3 (1977) 23^35. [11] N. Miki, Y. Kawabe, K. Kuriyama, Activation of cerebral guanylate cyclase by nitric oxide, Biochem. Biophys. Res. Commun. 75 (1977) 851^856. [12] W.P. Arnold, C.K. Mittal, S. Katsuki, F. Murad, Nitric oxide activates guanylate cyclase and increases guanosine

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 3P :5P-cyclic monophosphate levels in various tissue preparations, Proc. Natl. Acad. Sci. USA 74 (1977) 3203^3207. K. Schultz, G. Schultz, Sodium nitroprusside and other smooth muscle-relaxants increase cyclic GMP levels in rat ductus deferens, Nature 265 (1977) 750^751. R.F. Furchgott, J.V. Zawadzki, The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine, Nature 288 (1980) 373^376. S. Moncada, R.M. Palmer, E.A. Higgs, The discovery of nitric oxide as the endogenous nitrovasodilator, Hypertension 12 (1988) 365^372. L.J. Ignarro, G.M. Buga, K.S. Wood, R.E. Byrns, G. Chaudhuri, Endothelium-derived relaxing factor produced and released from artery and vein is nitric oxide, Proc. Natl. Acad. Sci. USA 84 (1987) 9265^9269. R.M. Palmer, A.G. Ferrige, S. Moncada, Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor, Nature 327 (1987) 524^526. D.A. Wagner, V.R. Young, S.R. Tannenbaum, Mammalian nitrate biosynthesis: incorporation of 15 NH3 into nitrate is enhanced by endotoxin treatment, Proc. Natl. Acad. Sci. USA 80 (1983) 4518^4521. L.C. Green, K. Ruiz de Luzuriaga, D.A. Wagner, W. Rand, N. Istfan, V.R. Young, S.R. Tannenbaum, Nitrate biosynthesis in man, Proc. Natl. Acad. Sci. USA 78 (1981) 7764^ 7768. L.C. Green, S.R. Tannenbaum, P. Goldman, Nitrate synthesis in the germfree and conventional rat, Science 212 (1981) 56^58. M.A. Marletta, M.A. Tayeh, J.M. Hevel, Unraveling the biological signicance of nitric oxide, Biofactors 2 (1990) 219^225. M.A. Marletta, Nitric oxide: biosynthesis and biological signicance, Trends Biochem. Sci. 14 (1989) 488^492. J.B. Hibbs Jr., R.R. Taintor, Z. Vavrin, E.M. Rachlin, Nitric oxide: a cytotoxic activated macrophage eector molecule, Biochem. Biophys. Res. Commun. 157 (1988) 87^94. M.A. Marletta, Mammalian synthesis of nitrite, nitrate, nitric oxide, and N-nitrosating agents, Chem. Res. Toxicol. 1 (1988) 249^257. J.M. Hevel, K.A. White, M.A. Marletta, Purication of the inducible murine macrophage nitric oxide synthase. Identication as a avoprotein, J. Biol. Chem. 266 (1991) 22789^ 22791. T.D. Warner, J.A. Mitchell, H. Sheng, F. Murad, Eects of cyclic GMP on smooth muscle relaxation, Adv. Pharmacol. 26 (1994) 171^194. W.A. Buechler, K. Ivanova, G. Wolfram, C. Drummer, J.M. Heim, R. Gerzer, Soluble guanylyl cyclase and platelet function, Ann. NY Acad. Sci. 714 (1994) 151^157. S.R. Jarey, S.H. Snyder, Nitric oxide: a neural messenger, Annu. Rev. Cell Dev. Biol. 11 (1995) 417^440. R.S. Lewis, W.M. Deen, Kinetics of the reaction of nitric oxide with oxygen in aqueous solutions, Chem. Res. Toxicol. 7 (1994) 568^574.

347

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22] [23]

[24]

[25]

[26]

[27]

[28] [29]

[30] J.M. Fukuto, Chemistry of nitric oxide: biologically relevant aspects, Adv. Pharmacol. 34 (1995) 1^15. [31] J.F. Kerwin Jr., J.R. Lancaster Jr., P.L. Feldman, Nitric oxide: a new paradigm for second messengers, J. Med. Chem. 38 (1995) 4343^4362. [32] C. Giulivi, Functional implications of nitric oxide produced by mitochondria in mitochondrial metabolism, Biochem. J. 332 (1998) 673^679. [33] G.C. Brown, Nitric oxide regulates mitochondrial respiration and cell functions by inhibiting cytochrome oxidase, FEBS Lett. 369 (1995) 136^139. [34] V.M. Bolotina, S. Najibi, J.J. Palacino, P.J. Pagano, R.A. Cohen, Nitric oxide directly activates calcium-dependent potassium channels in vascular smooth muscle, Nature 368 (1994) 850^853. [35] D. Sekkai, F. Aillet, N. Israel, M. Lepoivre, Inhibition of NF-kappaB and HIV-1 long terminal repeat transcriptional activation by inducible nitric oxide synthase 2 activity, J. Biol. Chem. 273 (1998) 3895^3900. [36] M. Spiecker, H.B. Peng, J.K. Liao, Inhibition of endothelial vascular cell adhesion molecule-1 expression by nitric oxide involves the induction and nuclear translocation of IkappaBalpha, J. Biol. Chem. 272 (1997) 30969^30974. [37] H.M. Lander, P.K. Sehajpal, A. Novogrodsky, Nitric oxide signaling: a possible role for G proteins, J. Immunol. 151 (1993) 7182^7187. [38] H.M. Lander, J.S. Ogiste, S.F. Pearce, R. Levi, A. Novogrodsky, Nitric oxide-stimulated guanine nucleotide exchange on p21ras, J. Biol. Chem. 270 (1995) 7017^7020. [39] H.M. Lander, D.P. Hajjar, B.L. Hempstead, U.A. Mirza, B.T. Chait, S. Campbell, L.A. Quilliam, A molecular redox switch on p21(ras). Structural basis for the nitric oxidep21(ras) interaction, J. Biol. Chem. 272 (1997) 4323^4326. [40] J. Li, T.R. Billiar, R.V. Talanian, Y.M. Kim, Nitric oxide reversibly inhibits seven members of the caspase family via S-nitrosylation, Biochem. Biophys. Res. Commun. 240 (1997) 419^424. [41] J.S. Stamler, L. Jia, J.P. Eu, T.J. McMahon, I.T. Demchenko, J. Bonaventura, K. Gernert, C.A. Piantadosi, Blood ow regulation by S-nitrosohemoglobin in the physiological oxygen gradient, Science 276 (1997) 2034^2037. [42] L. Xu, J.P. Eu, G. Meissner, J.S. Stamler, Activation of the cardiac calcium release channel (ryanodine receptor) by poly-S-nitrosylation, Science 279 (1998) 234^237. [43] A.J. Gow, J.S. Stamler, Reactions between nitric oxide and haemoglobin under physiological conditions, Nature 391 (1998) 169^173. [44] T. Ingi, J. Cheng, G.V. Ronnett, Carbon monoxide: an endogenous modulator of the nitric oxide-cyclic GMP signaling system, Neuron 16 (1996) 835^842. [45] A. Verma, D.J. Hirsch, C.E. Glatt, G.V. Ronnett, S.H. Snyder, Carbon monoxide: a putative neural messenger, Science 259 (1993) 381^384. [46] R. Zakhary, K.D. Poss, S.R. Jarey, C.D. Ferris, S. Tonegawa, S.H. Snyder, Targeted gene deletion of heme oxygen-

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

[47]

[48] [49]

[50]

[51]

[52]

[53]

[54]

[55] [56]

[57]

[58]

[59]

[60]

[61]

[62]

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 [77] F. Yu, D. Warburton, S. Wellington, R.S. Danziger, Assignment of GUCIA2, the gene coding for the K2 subunit of soluble guanylyl cyclase, to position 11q21-q22 on human chromosome 11, Genomics 33 (1996) 334^336. [78] G. Gupta, M. Azam, L. Yang, R.S. Danziger, The L2 subunit inhibits stimulation of the K1/L1 form of soluble guanylyl cyclase by nitric oxide. Potential relevance to regulation of blood pressure, J. Clin. Invest. 100 (1997) 1488^1492. [79] B. Mayer, D. Koesling, E. Bo hme, Characterization of nitric oxide synthase, soluble guanylyl cyclase, and Ca2 /calmodulin-stimulated cGMP phosphodiesterase as components of neuronal signal transduction, Adv. Second Messenger Phosphoprotein Res. 28 (1993) 111^119. [80] D. Koesling, P. Humbert, G. Schultz, in: S.R. Vincent (Ed.), Nitric Oxide in the Nervous System, Academic Press, San Diego, CA, 1995, pp. 43^50. [81] C. Harteneck, D. Koesling, A. So ling, G. Schultz, E. Bo hme, Expression of soluble guanylyl cyclase. Catalytic activity requires two enzyme subunits, FEBS Lett. 272 (1990) 221^223. [82] A.C. Clark, S.W. Raso, J.F. Sinclair, M.M. Ziegler, A.F. Chaotte, T.O. Baldwin, Kinetic mechanism of luciferase subunit folding and assembly, Biochemistry 36 (1997) 1891^1899. [83] A.N. Fedorov, T.O. Baldwin, Cotranslational protein folding, J. Biol. Chem. 272 (1997) 32715^32718. [84] D.L. Garbers, Purication of soluble guanylate cyclase from rat lung, J. Biol. Chem. 254 (1979) 240^243. [85] M.S. Wolin, K.S. Wood, L.J. Ignarro, Guanylate cyclase from bovine lung. A kinetic analysis of the regulation of the puried soluble enzyme by protoporphyrin IX, heme, and nitrosyl-heme, J. Biol. Chem. 257 (1982) 13312^13320. [86] S.A. Waldman, F. Murad, Cyclic GMP synthesis and function, Pharmacol. Rev. 39 (1987) 163^196. [87] J.R. Stone, M.A. Marletta, Heme stoichiometry of heterodimeric soluble guanylate cyclase, Biochemistry 34 (1995) 14668^14674. [88] A. Mu lsch, R. Gerzer, Purication of heme-containing soluble guanylyl cyclase, Methods Enzymol. 195 (1991) 377^ 383. [89] P. Humbert, F. Niroomand, G. Fischer, B. Mayer, D. Koesling, K.D. Hinsch, H. Gausepohl, R. Frank, G. Schultz, E. Bo hme, Purication of soluble guanylyl cyclase from bovine lung by a new immunoanity chromatographic method, Eur. J. Biochem. 190 (1990) 273^278. [90] E.A. Dierks, J.N. Burstyn, The deactivation of soluble guanylyl cyclase by redox-active agents, Arch. Biochem. Biophys. 351 (1998) 1^7. [91] E.A. Dierks, S.Z. Hu, K.M. Vogel, A.E. Yu, T.G. Spiro, J.N. Burstyn, Demonstration of the role of scission of the proximal histidine-iron bond in the activation of soluble guanylyl cyclase through metalloporphyrin substitution studies, J. Am. Chem. Soc. 119 (1997) 7316^7323. [92] T. Tomita, T. Ogura, S. Tsuyama, Y. Imai, T. Kitagawa, Eects of GTP on bound nitric oxide of soluble guanylate cyclase probed by resonance Raman spectroscopy, Biochemistry 36 (1997) 10155^10160.

349

[93] P.E. Brandish, W. Buechler, M.A. Marletta, Regeneration of the ferrous heme of soluble guanylate cyclase from the nitric oxide complex: acceleration by thiols and oxyhemoglobin, Biochemistry (1998) in press. [94] B. Wedel, P. Humbert, C. Harteneck, J. Foerster, J. Malkewitz, E. Bo hme, G. Schultz, D. Koesling, Mutation of His-105 in the L1 subunit yields a nitric oxide-insensitive form of soluble guanylyl cyclase, Proc. Natl. Acad. Sci. USA 91 (1994) 2592^2596. [95] G. Gupta, J. Kim, L. Yang, S.L. Sturley, R.S. Danziger, Expression and purication of soluble, active heterodimeric guanylyl cyclase from baculovirus, Protein Expr. Purif. 10 (1997) 325^330. [96] W.A. Buechler, S. Singh, J. Aktas, S. Muller, F. Murad, R. Gerzer, High-level expression of biologically active soluble guanylate cyclase using the baculovirus system is strongly heme-dependent, Adv. Pharmacol. 34 (1995) 293^303. [97] Y. Zhao, M.A. Marletta, Localization of the heme binding region in soluble guanylate cyclase, Biochemistry 36 (1997) 15959^15964. [98] Y. Zhao, J.P. Schelvis, G.T. Babcock, M.A. Marletta, Identication of histidine 105 in the L1 subunit of soluble guanylate cyclase as the heme proximal ligand, Biochemistry 37 (1998) 4502^4509. [99] L.J. Ignarro, Haem-dependent activation of guanylate cyclase and cyclic GMP formation by endogenous nitric oxide: a unique transduction mechanism for transcellular signaling, Pharmacol. Toxicol. 67 (1990) 1^7. [100] P.A. Craven, F.R. DeRubertis, D.W. Pratt, Electron spin resonance study of the role of NOc catalase in the activation of guanylate cyclase by NaN3 and NH2 OH. Modulation of enzyme responses by heme proteins and their nitrosyl derivatives, J. Biol. Chem. 254 (1979) 8213^8222. [101] P.A. Craven, F.R. DeRubertis, Restoration of the responsiveness of puried guanylate cyclase to nitrosoguanidine, nitric oxide, and related activators by heme and hemeproteins. Evidence for involvement of the paramagnetic nitrosyl-heme complex in enzyme activation, J. Biol. Chem. 253 (1978) 8433^8443. [102] R. Gerzer, E. Bo hme, F. Hofmann, G. Schultz, Soluble guanylate cyclase puried from bovine lung contains heme and copper, FEBS Lett. 132 (1981) 71^74. [103] L.J. Ignarro, J.N. Degnan, W.H. Baricos, P.J. Kadowitz, M.S. Wolin, Activation of puried guanylate cyclase by nitric oxide requires heme. Comparison of heme-decient, heme-reconstituted and heme-containing forms of soluble enzyme from bovine lung, Biochim. Biophys. Acta 718 (1982) 49^59. [104] R. Gerzer, F. Hofmann, G. Schultz, Purication of a soluble, sodium-nitroprusside-stimulated guanylate cyclase from bovine lung, Eur. J. Biochem. 116 (1981) 479^486. [105] J.R. Stone, M.A. Marletta, Spectral and kinetic studies on the activation of soluble guanylate cyclase by nitric oxide, Biochemistry 35 (1996) 1093^1099. [106] J.N. Burstyn, A.E. Yu, E.A. Dierks, B.K. Hawkins, J.H. Dawson, Studies of the heme coordination and ligand bind-

350

J.W. Denninger, M.A. Marletta / Biochimica et Biophysica Acta 1411 (1999) 334^350 ing properties of soluble guanylyl cyclase (sGC): characterization of Fe(II)sGC and Fe(II)sGC(CO) by electronic absorption and magnetic circular dichroism spectroscopies and failure of CO to activate the enzyme, Biochemistry 34 (1995) 5896^5903. G. Deinum, J.R. Stone, G.T. Babcock, M.A. Marletta, Binding of nitric oxide and carbon monoxide to soluble guanylate cyclase as observed with resonance Raman spectroscopy, Biochemistry 35 (1996) 1540^1547. A.E. Yu, S.Z. Hu, T.G. Spiro, J.N. Burstyn, Resonance Raman-spectroscopy of soluble guanylyl cyclase reveals displacement of distal and proximal heme ligands by NO, J. Am. Chem. Soc. 116 (1994) 4117^4118. S.Y. Kim, G. Deinum, M.T. Gardner, M.A. Marletta, G.T. Babcock, Distal pocket polarity in the unusual ligand binding site of soluble guanylate cyclase: implications for the control of c NO binding, J. Am. Chem. Soc. 118 (1996) 8769^8770. M.P. Doyle, J.W. Hoekstra, Oxidation of nitrogen oxides by bound dioxygen in hemoproteins, J. Inorg. Biochem. 14 (1981) 351^358. E. Antonini, M. Brunori, Hemoglobin and Myoglobin in their Reactions with Ligands, vol. 21, North-Holland Publ. Co., Amsterdam, 1971. B. Fan, G. Gupta, R.S. Danziger, J.M. Friedman, D.L. Rousseau, Resonance Raman characterization of soluble guanylate cyclase expressed from baculovirus, Biochemistry 37 (1998) 1178^1184. J.R. Stone, R.H. Sands, W.R. Dunham, M.A. Marletta, Electron paramagnetic resonance spectral evidence for the formation of a pentacoordinate nitrosyl-heme complex on soluble guanylate cyclase, Biochem. Biophys. Res. Commun. 207 (1995) 572^577. P.S. Yuen, L.K. Doolittle, D.L. Garbers, Dominant negative mutants of nitric oxide-sensitive guanylyl cyclase, J. Biol. Chem. 269 (1994) 791^793. B. Wedel, C. Harteneck, J. Foerster, A. Friebe, G. Schultz, D. Koesling, Functional domains of soluble guanylyl cyclase, J. Biol. Chem. 270 (1995) 24871^24875. R.K. Sunahara, A. Beuve, J.J.G. Tesmer, S.R. Sprang, D.L. Garbers, A.G. Gilman, Exchange of substrate and inhibitor specicities between adenylyl and guanylyl cyclases, J. Biol. Chem. 273 (1998) 16332^16338. P.J. Artymiuk, A.R. Poirrette, D.W. Rice, P. Willett, A polymerase I palm in adenylyl cyclase?, Nature 388 (1997) 33^34. P.D. Senter, F. Eckstein, A. Mu hme, The stereo lsch, E. Bo chemical course of the reaction catalyzed by soluble bovine lung guanylate cyclase, J. Biol. Chem. 258 (1983) 6741^ 6745. C.L. Tucker, J.H. Hurley, T.R. Miller, J.B. Hurley, Two amino acid substitutions convert a guanylyl cyclase, RetGC-1, into an adenylyl cyclase, Proc. Natl. Acad. Sci. USA 95 (1998) 5993^5997. E.M. Wilson, M. Chinkers, Identication of sequences mediating guanylyl cyclase dimerization, Biochemistry 34 (1995) 4696^4701. A. Friebe, D. Koesling, Mechanism of YC-1-induced activation of soluble guanylyl cyclase, Mol. Pharmacol. 53 (1998) 123^127. V.G. Kharitonov, V.S. Sharma, D. Magde, D. Koesling, Kinetics of nitric oxide dissociation from ve- and six-coordinate nitrosyl hemes and heme proteins, including soluble guanylate cyclase, Biochemistry 36 (1997) 6814^ 6818. V.G. Kharitonov, M. Russwurm, D. Magde, V.S. Sharma, D. Koesling, Dissociation of nitric oxide from soluble guanylate cyclase, Biochem. Biophys. Res. Commun. 239 (1997) 284^286. J. Garthwaite, E. Southam, C.L. Boulton, E.B. Nielsen, K. Schmidt, B. Mayer, Potent and selective inhibition of nitric oxide-sensitive guanylyl cyclase by 1 H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one, Mol. Pharmacol. 48 (1995) 184^ 188.

[116]

[107]

[117]

[108]

[118]

[109]

[119]

[120]

[110]

[111]

[121]

[112]

[122]

[113]

[123]

[114]

[124]

[115]