Photoreceptor guanylate cyclases: a review

Bioscience Reports, Vol. 17, No, 5, 1997

REVIEW

Photoreceptor Guanylate Cyclases: A Review Edward N. Pugh, Jr.,1 Teresa Duda,2 Ari Sitaramayya,3 and Rameshwar K. Sharma2,4 Received June 18, 1997 Almost three decades of research in the field of photoreceptor guanylate cyclases are discussed in this review. Primarily, it focuses on the members of membrane-bound guanylate cyclases found in the outer segments of vertebrate rods. These cyclases represent a new guanylate cyclase subfamily, termed ROS-QC, which distinguishes itself from the peptide receptor guanylate cyclase family that it is not extracellularly regulated. It is regulated, instead, by the intracellularly-generated Ca2+ signals. A remarkable feature of this regulation is that ROS-GC is a transduction switch for both the low and high Ca2+ signals. The low Ca2+ signal transduction pathway is linked to phototransduction, but the physiological relevance of the high Ca2+ signal transduction pathway is not yet clear; it may be linked to neuronal synaptic activity. The review is divided into eight sections. In Section I, the field of guanylate cyclase is introduced and the scope of the review is briefly explained; Section II covers a brief history of the investigations and ideas surrounding the discovery of rod guanylate cyclase. The first five subsections of Section III review the experimental efforts to quantify the guanylate cyclase activity of rods, including in vitro and in situ biochemistry, and also the work done since 1988 in which guanylate cyclase activity has been determined. In the remaining three subsections an analytical evaluation of the Ca2+ modulation of the rod guanylate cyclase activity related to phototransduction is presented. Section IV deals with the issues of a biochemical nature: isolation and purification, subcellular localization and functional properties of rod guanylate cyclase. Section V summarizes work on the cloning of the guanylate cyclases, analysis of their primary structures, and determination of their location with in situ hybridization. Section VI summarizes studies on the regulation of guanylate cyclases, with a focus on guanylate cyclases activating proteins. In Section VII, the evidence about the localization and functional role of guanylate cyclases in other retinal cells, especially in "on-bipolar" cells, in which guanylate cyclase most likely plays a critical role in electrical signaling, is discussed. The review concludes with Section VIII, with remarks about the future directions of research on retinal guanylate cyclases. KEY WORDS: Calcium; guanylate cyclase; photoreceptor.

OUTLINE I. II.

Introduction: goals and scope of review Guanylate cyclase discovered during search for internal messenger of rod phototransduction

Departments of Psychology and Ophthalmology, University of Pennsylvania, 3815 Walnut Street, Philadelphia, PA 19104-6196. 2 Unit of Regulatory and Molecular Biology, Departments of Cell Biology, SOM, and Ophthalmology NJMS, University of Medicine and Dentistry of New Jersey, Stratford, NJ 08084. 3 Eye Research Institute, Oakland University, Rochester, MI 48309. 4 To whom correspondence should be addressed. 1

429 0144-8463/97/1000-0429$12.50/0 © 1W Plenum Publishing Corporation

430

III.

IV.

V.

VI.

Pugh, Duda, Sitaramayya and Sharma

Phototransduction 1970-1985: cyclic GMP and Ca2+ were candidate internal messengers Discovery of rod guanylate cyclase activity Cyclic GMP identified as the internal messenger of phototransduction Quantification of rod guanylate cyclase activity in vitro and in situ Rod GC activity in vitro exhibits cooperative dependence on Ca2+ Rod GC activity measured in intact retinas Kinetic characterization of rod GC activity in situ by Hodgkin & Nunn Rod GC activity has been measured in single truncated rod outer segments Summary evaluation of measurements of guanylate cyclase activity in vitro and in situ GC activation by lowered Ca2+ serves to increase the range of light intensities to which the rod can respond by maintaining cyclic GMP-activated channels open GC activation increases the signal amplification of the light-adapted response Role of GC activation in speeding dim-flash photoresponse in light adaptation Isolation, purification and characterization of guanylate cyclase Earliest studies recognized that rod outer segment guanylate cyclase is membrane bound Solubilization of ROS-GC was achieved with non-ionic detergent and high salt concentration Atrial natriuretic factor receptor guanylate cyclase was the first purified and functionally characterized vertebrate membrane guanylate cyclase ROS-GC is a 110-112 kDa protein Cloning of retinal guanylate cyclases Natriuretic factor receptors were the first cloned guanylate cyclases from retina First molecular cloning of a membrane guanylate cyclase that was not a surface receptor was from retina Molecular cloning of the wild-type ROS-GC established a guanylate cyclase subfamily that was distinct from the surface receptor subfamily Recent studies indicate that retGC is the human analogue of ROS-GC Molecular cloning of retGC2, the second member of ROS-GC subfamily. One genetic feature peculiar to ROS-GC subfamily is its exceptionally large mRNA size Regulation of retinal guanylate cyclases activity ATP regulates the activities of membrane guanylate cyclases, a feature that distinguishes the membrane guanylate cyclase transduction system from the adenylate cyclase system ATP inhibits basal ROS-GC activity Modulation of the cloned ROS-GC GCAP/-modulated domain resides in the kinase-like region of the ROS-GC intracellular segment

Photoreceptor Guanylate Cyclases

vn. vra. IX.

x.

431

ROS-GC activity is also stimulated by high (micromolar) concentrations of Ca2+ suggesting its linkage to retinal processes unrelated to phototransduction GCAP2-modulated ROS-GC transduction mechanism is different from that of GCAP7 S100 protein-modulated ROS-GC domain is distinct from the low Ca2+ GCAP7-modulated domain Guanylate cyclases in "on-bipolar" cells and other retinal neurons Concluding remarks and future directions Acknowledgments References

I. INTRODUCTION TO GUANYLATE CYCLASES, AND GOALS AND SCOPE OF THE REVIEW Guanylate cyclases (GCs) are enzymes that catalyze the conversion of GTP to cyclic GMP. Because cyclic GMP serves as an intracellular messenger in a large variety of cells (Mayer, 1994; Warner et al., 1994; MacFarland, 1995; Biel et al., 1995; Imai, 1995; Kaupp, 1995; Lincoln et al., 1995; Menini, 1995; Finn et al., 1996), GCs and the intracellular and extracellular factors that modulate their activity are now of considerable interest. Known GCs belong to two families: soluble cyclases, and membrane-bound or particular cyclases. The soluble cyclases form a homologous family of proteins comprising two subunits, one of 70 kDa, and a second of 82 kDa; a heme group is bound to each subunit (Stone and Maretta, 1995), and the cyclase is activated by the binding of nitric oxide to the heme (Katsuki et al., 1977; Bredt and Snyder, 1994). As we shall review in detail, the membrane-bound cyclases that have been characterized to date comprise a homologous family quite distinct from the soluble cyclases, with two subfamilies: the natriuretic peptide-activated subfamily, whose members are activated by the binding of atrial naturetic peptide or a related peptide to an extracellular domain, and the calcium-activated subfamily, whose members are activated by one or more calcium-binding proteins that interact with intracellular domains of the cyclase. This review will focus on the membrane-bound family of guanylate cyclases. Moreover, most of the review will be devoted to the specific members of this family that are found in the outer segments of vertebrate rod photoreceptors. These cyclases will be called "ROS-GC's" because of their expression in Rod Outer Segments. The principal reason for this restriction is, as the review will show, that the role that ROS-GCs play in the rod phototransduction signalling cascade is now thoroughly understood. Thus, it is possible to compare the relevant quantitative properties of the phototransducing rod with the kinetic properties determined in vitro of the isolated, and the cloned/expressed enzyme.

432

Pugh, Duda, Sitaramayya and Sharma

By making such comparisons, it can be concluded definitively that calcium is the exclusive intracellular signal that governs ROS-GC activity under normal intracellular conditions, to establish with certainty what is the mechanism by which calcium regulates cyclase, and to say what role this calcium regulation plays in the light-signalling function of the rod. In short, in the case of the rod photoreceptor, a complete story can be told connecting molecular biological and biochemical details of ROS-GC with the normal function of the cell. It is hoped that by laying out this complete story, a useful paradigm for the understanding of the functional and quantitative roles of other guanylate cyclases will be established. To put the functional role of ROS-GC into perspective, then, consider Fig. 1, which illustrates a typical vertebrate rod. Vertebrate rods are highly differentiated retinal neurons that generate electrical signals in response to photons captured by the integral membrane protein rhodopsin, a member of the great superfamily of seven-helix, G-protein-coupled receptors (cf. http://receptor.mgh.harvard.edu/ GCRDBHOME.html). The biochemical process by which rods generate these signals is known as phototransduction. Rod photoreceptors are possessed by nearly all vertebrates and are universally specialized for generating reliable electrical signals to single photons of light (Baylor et al., 1979; Baylor, et al., 1984). Phototransduction takes place in the outer segment region of the cell, which contains the disc or lamellar membranes, and in which all the molecular components of the phototransduction cascade reside. Our review of photoreceptor guanylate cyclase will proceed as follows. In Section II we will give a brief history of the investigations and ideas surrounding the discovery of rod GC. This section can be omitted by the reader eager to get Fig. 1. Left panel. An illustration of a typical vertebrate rod. In the dark a circulating current (arrows) is present, which is outward in the inner segment and carried primarily by K+; in the outer segment the net charge flow is inward, with about 90% of the inward flow carried by Na+ and 10% by CE?+ ions. Na + /K + exchange pumps in the inner segment membrane and Na^/K^-Ca 2 * exchangers in the outer segment membrane (see also right panels) maintain the overall ionic gradients against the dark flows. The capture of a photon (hi>) by a rhodopsin molecule in one of the disc membranes of the outer segment initiates the phototransduction G-Protein cascade. Right upper panel. The components of the phototransduction cascade are shown in the dark/resting steady-state. A steady-state cGMP concentration, cG, of 3-4/iM is maintained by the resting activity of guanylyl cyclase (rate, a dark ) in the presence of a relatively small basal phosphodiesterase activity (rate constant, /3darlc) (see text eqns 1-2); the cG binds rapidly and reversibly to the cGMP-activated channels, keeping a few percent of the total channels open. The influx of Ca2+ through the open channels is balanced by the outward extrusion through the Na + /Ca 2+ -K + exchanger. In the resting state most of the GCAPs have calcium bound to them, and are unable to activate cyclase. Right lower panel. The capture of a photon by a rhodopsin molecule activates the trimeric G-Protein, transducin, by catalyzing a GDP/GTP exchange; the activated moiety Ga-GTP in turn activates the phosphodiesterase (PDE) by relieving inhibition imposed by the y-subunits. Thus, the rate constant /3 of PDE-catalyzed hydrolysis of cGMP is strongly elevated, causing cG to decline from its resting concentration, and net loss of binding to the cGMP-gated channels, which then close. The closure of the channels results in a decline in Ca2+ as the exchanger continues to pump out Ca2*. The GCAPs now lose the Ca2+ that is bound to them, and have increased affinity for GC, to which they bind, activating it. This creates a higher rate of cGMP synthesis a, which results in the presence steady illumination in a new steady-state level of cG (see equations 1-4 of text). For the right-hand panels the following is the symbol key: R = rhodopsin; G = G-protein (transducin); PDE = phosphodiesterase; GC = guanylate cyclase; GCAP = guanylate-cyclase activating protein.

Photoreceptor Guanylate Cyclases

433

434

Pugh, Duda, Sitaramayya and Sharma

to the more substantive sections. In the first five subsections of Section III we review experimental efforts to quantify the GC activity of rods, including in vitro and in situ biochemistry, and work done since 1988 in which GC activity has been determined from analysis of cGMP-activated currents (see Fig, 1). In the remaining three subsections of Section III we present an analytical evaluation of the activity of GC in rods, and derive quantitative conclusions about the role of GC activity and its regulation by internal calcium in phototransduction. Section IV returns to the issues of the biochemical nature, isolation and purification, subcellular localization and functional properties of rod GC. Section V summarizes work on the cloning of GCs, analysis of their primary structure, and determination of their location with in situ hybridization. Section VI summarizes work on the regulation of GCs, with a focus on "GCAPs" or guanylate cyclase activation proteins. In Section VI we discuss evidence about the localization and functional role of GCs in other retinal cells, especially in "on-bipolar" cells, in which GC likely plays a critical role in the electrical signalling. We concluded in Section VIII with remarks about future directions that research on retinal GCs will likely take.

II. GUANYLATE CYCLASE WAS DISCOVERED IN PHOTORECEPTORS DURING THE SEARCH FOR THE INTERNAL CYTOPLASMIC MESSENGER OF ROD PHOTOTRANSDUCTION Phototransduction 1970-1985: Both Cyclic GMP and Ca2+ Were Candidate Internal Messengers In the late 1960s and early 1970s electrophysiological and structural research led investigators to the conclusion that an "internal messenger" was a necessary intermediary of phototransduction in vertebrate photoreceptors (Baylor and Fuortes, 1970; reviewed in Lamb, 1986; Stryer, 1986a; Pugh, 1987). Key support for this conclusion came from ultrastructural research which showed that the rod disc membranes containing the light-capturing rhodopsin molecules were physically separated from the plasma membrane, which in turn contains the cation conductance reduced during the light response (Baylor & Fuortes, 1970; Hagins et al, 1970). Thus, a cytoplasmic messenger was required, and a principal thrust of photoreceptor research from 1970 through 1985 was toward identifying this messenger. Throughout the period, most research focussed on two candidate messengers, Ca2+ and cyclic GMP. The hypothesis that Ca2+ serves as the internal messenger of rod phototransduction was proposed by Yoshikami and Hagins (1970), whose experiments demonstrated that elevated Caf + leads to rapid closure of the light-controlled conductance. Bitensky et al. (1971) first proposed that a cyclic nucleotide (cyclic AMP) might serve as the internal messenger; by 1973 it was clear that if the messenger was a cyclic nucleotide, it would be cyclic GMP. Miller and Nicol

Photoreceptor Guanylate Cyclases

435

subsequently reported a number of electrophysiological findings that supported the cyclic GMP hypothesis (Nicol and Miller, 1978, 1979; Miller, 1982), while Bitensky and his collaborators went on to characterize the biochemistry of the rhodopsin/G-protein/phosphodiesterase cascade (Bitensky et al., 1973; Bitensky el al., 1978). Many other scientists made critical contributions during this period to the understanding of the cascade (reviewed in Miller, 1981; Stieve, 1986). Perhaps the most important lesson for current science to retain about the period of phototransduction research from 1970-1985 is that substantial evidence was presented supporting both Ca2+ and cyclic GMP hypotheses (reviewed in Miller et al., 1981). Thus, whatever the identity of the internal excitational messenger would turn out to be, research in this period produced a rich body of evidence demonstrating powerful regulatory effects of both Ca2+ and cyclic GMP in the light-controlled biochemistry and electrophysiology of vertebrate photoreceptors (for reviews of research at the end of this period, see Stryer, 1986; Lamb, 1986; Pugh & Cobbs, 1986).

Discovery of Rod Guanylate Cyclase Activity Rod guanylate cyclase (GC) entered the stage of phototransduction research in a report by Pannbacker (1973), who had previously determined that the rod phosphodiesterase activity discovered by Bitensky and Miller was in fact specific for cyclic GMP (Pannbacker et al., 1972). Early contributions to the characterization of the rod GC were made by a number of other investigators (Goridis et al., 1973; Chader et al., 1974; see Table I) culminating in investigations of substrate and divalent cation specificity by Krishnan et al. (1978) and Fleishman & Denisevich (1979). The latter investigations established the critical requirement of rod GC activity for Mg2+ or Mn2+ coordination of the enzyme-GTP complex, and also discovered inhibition of GC by Ca2+. The inhibition by Ca2+ was proved not to occur through a competition with the GTP-coordinating divalent cation, but was also found to have a non-physiological K\ of about 1 mM. In the late 1970s and early 1980s a number of studies, beginning with that of Cohen et al. (1978), and including those of Bownds and co-workers (Woodruff & Bownds, 1979; Cote et al. 1984) and Lolley & Racz (1982) demonstrated that the retinal pool of cyclic GMP that could be readily reduced by light was also calcium-sensitive. Specifically, manipulations that were expected to lower Ca2+ were found to greatly increase the light-sensitive pool of cyclic GMP, and those expected to increase Ca2+ decreased the pool. Thus, the scene was set for a resolution of the conflict between the Ca2+ and cyclic GMP hypotheses, by means of a scheme of reciprocal interactions. Nonetheless, as late as 1985 two critical issues remained unresolved: (1) the actual identity of the excitational messenger of phototransduction; and (2) a mechanism by which Ca2+, at the submicromolar concentrations maintained by cells might regulate GC.

436

Pugh, Duda, Sitaramayya and Sharma Table I. Experimental characterization of rod guanylate cyclase activity in vitro

Species

bovine

mouse bovine bovine bovine frog

T CO

37 37 37 37 37

Km (GTP) (mM)

Vmnjc (nmol/ min/mg of protein)

0.27 0.04 0.80 0.79 (Mg) 0.14 (Mn) 0.23

5.7

*c.t

(s )

24 (i)

bovine bovine lizard

30 30 30 37 37 30 37 30 30 30

human human human bovine bovine

bovine

1.1

30 30 37 37

0.76 0.16 0.08 0.27 1.07

20 (g) 23 (g)

-30 90

3.9

240 200

2 1.7

2.7 100-700 (a) 3262 (a)

Reference

Krishnan et al. (1978) Fleischman et al. (1979)

10.0

0.2

bovine bovine

bovine bovine

m (Ca)

Goridis et at. (1973) Troyer et al. (1978)

rat

toad

Km (Ca) (nM)

2.7 (a) 1.1 (a) 0.2- 1.3 (a) 6.3 (a) 220 (c) 50-100(c,d) 280 (c) 50 (e) 261 (f)

100 100 (h,f) 150 (i)

1.7 1.7-1.9

2 2.5

Hakkiera/. (1990) Coccia et al. (1994) Lolley et al. (1982) Koch et al. (1988) Koch et al. (1990) Gorczyca et al. (1994) Dizhoor el al. (1995) Denton et al. (1992) Hayashi et al. (1991) Koch (1991) Aparicio et al. (1995) Dizhoor et al. (1994) Lowe et al. (1995) Laura et al. (1996) Dizhoor et al. (1996) Frins et al. (1996)

Duda et al. (1996) Goraczniak et al. (1997)

Table notes. Columns labelled Km, Vmax and Arcal give the standard kinetic parameters; K1/2 is the concentration of Ca2+ that yielded half-maximal enzyme activity, and /«(Ca) is the estimated Hill coefficient of the dependence of activity on Ca2*, as expressed in eqn (2) of the text, (a) purified enzyme; (b) our estimate based on the data; (c) recombinant cyclase + bovine GCAP-2; (d) RetOC-2; (e) recombinant GCAP-2; (f) recombinant GCAP-1; (g) fully activated at low calcium concentration; (h) recombinant cyclase + recombinant GCAP-1; (i) recombinant ROS-GC2 + recombinant GCAP2.

Cyclic GMP Identified as the Internal Messenger of Phototransduction; Mechanism of Rod GC Dependence on Ca2+ Discovered The resolution of the first issue occurred in 1985: the identity of the internal messenger of rod phototransduction was settled by the discovery of cyclic GMP-gated channels in frog rod outer segments by Fesenko et al., (1985), and the simultaneous publication of compelling evidence against the calcium hypothesis (Matthews et al., 1985; Lamb et al., 1986). Subsequent studies, culminating in the cloning and re-expression of the cyclic GMP-gated channels (Kaupp et al, 1989), left no doubt that cyclic GMP was the cytoplasmic link between the discassociated G-Proteins activated by rhodopsin, and the cyclic GMP channels. A significant step toward resolution of the second issue was made when Pepe et al. (1986) published the first report of calcium sensitivity of rod GC activity in the submicromolar concentration range. Two years later Koch and Stryer (1988) reported a definitive submicromolar calcium regulation of rod GC activity, and

Photoreceptor Guanylate Cyclases

437

provided evidence that the exquisite sensitivity of GC to Ca2+ (K1/2 ~ 50 nM) was conferred by a soluble protein. Since control of Ca2+ in the rod by the reciprocal balance of inward flux through the cyclic GMP-activated channels and the outward pumping by the Na + /Ca 2+ -K + exchanger of the rod plasma membrane was by then firmly established (McNaughton et al, 1989; see Fig. 1), a resolution of the principal critical issues involving Ca2+ and cyclic GMP was at hand: these two pivotal agents in the drama of phototransduction were clearly interlocked in a reciprocal feedback loop in which GC was a key intermediary (reviewed in Pugh & Lamb, 1990). The focus of research on rod GC by the late 1980s thus shifted toward isolating and characterizing the protein(s) that conferred upon GC its exquisite sensitivity to Ca2+, and on quantifying the contribution of GC to the photoresponses of rods.

III. QUANTIFICATION OF ROD GUANYLATE CYCLASE ACTIVITY IN VITRO AND IN SITU To understand the role that GC plays in the physiological function of the photoreceptor, its kinetic parameters must be measured, and estimates made of GC activity under physiologically relevant conditions. In order to put all the measurements we will review into proper physiological perspective, we emphasize at the outset of this section the fundamental fact implicit in Fig. 1: in rod and cone photoreceptors, the primary function of GC is to maintain sufficient free cyclic GMP in the cytoplasm so that an adequate number of cyclic GMP-activated cation channels are open and available for signalling newly captured photons of light. To elucidate this primary function, we can formulate the scheme in Fig. 1 in the following simplified manner. At any given time in the photoreceptor the quantity of the cytoplasmic messenger, cyclic GMP, is determined by the balance of the activities of the synthesizing enzyme, GC, and the hydrolytic enzyme, phosphododiesterase (PDE):

In equation (1) and hereafter in the text a refers to the rate of cyclic GMP synthesis (units: /iM s"1) in the outer segment, and /3 refers to the rate constant of cyclic GMP hydrolysis (unit: s"1) by PDE. In the intact photoreceptor both a and /3 are "variables" that depend on a number of factors, including especially the recent illumination history. Since GC activity depends on factors that can vary in situ, determination of a requires consideration of the effects of such factors, as well as estimation of their concentrations. These factors include the concentrations of its normal substrate GTP, of calcium, of guanylate cyclase activating protein(s) (GCAPs), and possibly of other cofactors such as Mg2+ and ATP. We will summarize first measurements made in vitro.

438

Pugh, Duda, Sitaramayya and Sharma Table II.

Species

Method

Measurements of rod guanylate cyclase activity in situ T (°C)

rabbit toad salamander bullfrog

isolated rod truncated rod

37 22 22-24 22-24

salamander

truncated rod

22-24

180 180

Km (GTP) (mM)

0.25

(MM

"max (/iM

»~')

s'1)

6.7 0.5 3

30.2 5.4 20-25 -20

"dark

5-30

Km

(Ca) (nM)

m (Ca) (#)

—

1.7-2.1

Reference

Amesefa/. (1986) Dawis et al. (1988) Hodgkin & Nunn (1988) Kawamura & Murakami (1989) Koutalos et al. (1995)

Table notes. The 18O method is described in the text. The isolated (intact) rod method and truncated rod methods are based on measurements of the cGMP-activated currents, as described in the text. Values derived from the intact rod require an estimate of the concentration of free cGMP in the outer segment, which has been estimated in various ways to be 3-4 pM (cf. Pugh and Lamb, 1990). a dark is the rate of synthesis of cGMP in the dark/resting state (referred to the cytoplasm of the outer segment), while a max is the rate of synthesis under strong illumination that drives Ca^+ to a very low level (see Table III), maximally activating cyclase; m is the Hill coefficient of calcium-dependent activity (see eqn 2 of text).

Rod GC Activity in vitro Exhibits a cooperative, Inhibitory Dependence on Ca?+ Tables I and II summarize from many studies estimates of the kinetic parameters of rod outer segment guanylate cyclase (ROS-GC), and their dependence on the principal factors that govern its activity. From the perspective of the physiological function of GC, none of the factors that vary in the rod is more important than Ca2+. The dependence of rod GC in the presence of GCAP on calcium concentration, Ca2+, can be characterized by a form of the Hill equation:

Here AT1/2 is the value of Ca2+ at which the activity is half-maximal, and m the Hill coefficient. The two additional factors that are critical for cyclase activity are the concentration of its substrate, GTP, and of the coordinating divalent cation, Mg2+. Measurements of the GTP content of frog rod outer segments by Robinson and Hagins (1979), and by Biernbaum and Bownds (1979) show that the normal concentration is about 2mM. However, Biernbaum and Bownds (1985) showed that even with exposure to very high light intensities the concentration of GTP does not decline more than about 30%. This seems to be because the ATP concentration is also 1-2 mM, and highly stabilized, and because GTP is readily regenerated from ATP and GDP and GMP. Thus, the concentration of substrate for GC in normal rods is always well above the enzyme Km, and so the enzyme runs near the Vmax determined by the number of activated copies of the enzyme. There are two reports of the Mg2+ concentration in the rod outer segment. Somlyo and Wals (1985), using electron-probe X-ray analysis of cryosections estimated Mg2+ to be 11 mM in the frog rod outer segment at rest, and 9mM

Photoreceptor Guanylate Cyclases

439

after exposure to bright light. Sitaramayya et al. (1991) reported previously unpublished data of Hagins and colleagues who estimated Mg2+ in bullfrog rod outer segments in two experiments at 8 and 15 mM respectively. Moreover, since Mg2+ readily permeates the cyclic GMP-activated channels, Mg2+ can be expected to be roughly in equilibrium with the dark, resting membrane potential, ca. -30 mV. Since Mg2+ is of the order of 3-5 mM, it follows from the Nernst equation that Mg2+ should be about 3-fold higher than Mg2+, or about 10-15 mM, in general agreement with the experimental estimates. In sum, then, the only known physiological regulation of the activity of ROS-GC in situ is that by the binding of GCAP, and the binding affinity of GCAP for ROS-GC is in turn determined by the concentration of free calcium, Ca2+ (Fig. 1). We now turn to a summary and analysis of in situ measurements of GC activity. Rod GC Activity in situ was First Measured in Intact Retinas As noted above, a number of studies in the late 1970s and 1980s that utilized isolated retina preparations established that the light-sensitive pool of cyclic GMP rises very rapidly when the retina is exposed to low external calcium (Ca2+). Because of the powerful Na + /Ca 2+ -K + exchanger of the rod outer segment (cf Fig. 1), low Ca2"1" in the presence of normal Na