Analysis of Drosophila cGMP-dependent Protein Kinases and Assessment of Their in Vivo Roles by Targeted Expression in a Renal Transporting Epithelium*

cGMP-dependent protein kinase (cGK) forms encoded by the dg2 ( for ) gene are implicated in behavior and epithelial transport in Drosophila melanogaster . Here, we provide the first biochemical characterization and cellular localization of cGKs encoded by the major transcripts of dg2 : dg2P1 and dg2P2 . cGMP stimulates kinase activity of DG2P1 (EC 50 : 0.13 (cid:1) 0.039 (cid:2) M ) and DG2P2 (EC 50 : 0.32 (cid:1) 0.14 (cid:2) M ) in Malpighian tubule and S2 cell extracts. DG2P1 and DG2P2 are magnesium-requiring enzymes and were inhibited by 10 and 100 (cid:2) M of a cGK inhibitor, 8-(4-chlorophenylthio)guanosine-3 (cid:1) ,5 (cid:1) -cyclic monophosphorothioate, Rp isomer; whereas DG1, the cGK encoded by the D. melanogaster dg1 gene, was un-affected. DG2P1 and DG2P2 were localized in the plasma membrane in S2 cells, whereas DG1 was localized in the cytosol. The D. melanogaster fluid-transport-ing for

Signaling via guanosine cGMP is an important regulatory process for diverse physiological processes. Among the known cellular targets of cGMP are cyclic nucleotide-gated channels, cGMP-regulated phosphodiesterases, and cGMP-dependent protein kinases (cGKs) 1 (1)(2)(3). Much of the knowledge of cGK function has come from both biochemical and cell culture analyses in vitro, as well as in vivo work using mouse gene knockouts (reviewed in Refs. [2][3][4]. Mammalian cGKI is involved in smooth muscle and platelet function (5,6) and axonal guidance (7); cGKII is important in the function of transporting epithelia, including lung and intestine (8 -12) and in regulation of the circadian clock (13). Although major roles of cGKs in physiological function have been established in vertebrates, considerably less is known about cGK in other organisms. cGMP signaling has been implicated in behavioral and neural phenotypes in invertebrates, including Caenorhabditis elegans (14,15). In the insects Drosophila melanogaster (16) and Apis mellifera (17,18), cGK has been shown to influence feeding and social behavior. However, the studies in insects have been based on genetic analysis of a gene encoding cGK, dg2 (foraging, for), whereas little is known about the actual enzymes encoded by dg2.
Given the relevance of cGK to behavior, it is thus of interest to investigate the biochemistry of the dg2-encoded cGK family in Drosophila. Furthermore, the elegant transgenic tools afforded to Drosophila workers (21) allows study of cGKs in an organotypic context, using targeted gene technology. Finally, well defined physiological phenotypes in Drosophila provide a basis for evaluating targeted genes in this organism.
The insect Malpighian (renal) tubule is a fluid-transporting organ analogous to vertebrate kidney/liver and is essential to insect survival. D. melanogaster tubules, which constitute an important genetic model for transporting epithelia (22), display increased fluid transport rates when stimulated by either exogenous cGMP, which enters tubule cells via a cyclic nucleotide transporter (23), nitric oxide (NO), or neuropeptide (capa)generated NO/cGMP (24 -27). NO-and capa-stimulated cGMP signaling is compartmentalized to principal cells in the main, fluid-secreting segment of tubules (27,28), containing the electrogenic vacuolar H ϩ -ATPase (V-ATPase) (29), which energizes fluid transport. Electrophysiological studies show that cGMP signaling modulates V-ATPase activity (25), suggesting that cGMP signaling may regulate ion transport in tubules. Previous work using PCR has also shown that tubules express both dg1 and dg2 genes (24), suggesting the importance of cGK in this tissue; which could be valuable approach for studying Drosophila cGKs in vivo.
In this report, we provide the first biochemical characterization of transcript-specific dg2-encoded cGKs. The roles of Drosophila cGKs were also investigated in the context of epithelial transport via targeted overexpression of both dg1 and dg2 transcripts in transgenic lines using a cell-specific GAL4 driver. Analysis of tubule function from such lines has allowed delineation of the roles of individual Drosophila cGKs in an organotypic context for the first time.

Cloning and Isolation of dg1 and dg2 Transcripts: Expression in Drosophila S2 Cells
Intact D. melanogaster Malpighian tubules were dissected and used to make template cDNA for PCR reactions. Primers specific for the DG1 gene transcript and each of the two major DG2 gene transcripts were used to create PCR products encompassing the appropriate open reading frames. These were sequenced to confirm their identity and then cloned into the pMT/V5-HIS-TOPO vector (Invitrogen) suitable for expression in Drosophila S2 cells.
To create plasmids for expression of DG1 and DG2 proteins containing a C-terminal V5 epitope tag (GKPIPNPLLGLDST, derived from a small epitope (Pk) present on the P and V proteins of the paramyxovirus of simian virus 5), PCR mutagenesis was carried out that allowed removal of the stop codon from each of the open reading frames. To ensure fidelity of the coding regions, proofreading Taq polymerase (Promega) was used in the reactions. The mutated PCR products were sequenced and then cloned into the pMT/V5-HIS-TOPO vector (Invitrogen). Expression constructs for vertebrate cGKI␣ (31) and cGKII (32) were also sub-cloned into the pMT/V5-HIS-TOPO vector. These plasmids were then used for the transient transfection of S2 cells under conditions of Cu 2ϩ -inducible expression. Cells were cultured according to standard protocols (30). Approximately 3 ϫ 10 6 cells were used in each transfection, which was performed using calcium phosphate according to standard techniques (Invitrogen); transfection efficiencies were routinely 5-15%.
Malpighian Tubules-Immunostaining protocol for tubules was performed as previously described (34) using rabbit polyclonal anti-cGKI␣ antibody (11) (1/100), followed by a Texas Red-labeled anti-rabbit secondary antibody (1/5000) (Amersham Biosciences). Higher concentrations of primary antibody were used, but no significant improvement in labeling intensity was observed. Tubules were viewed under fluorescence using an Axiocam imaging system (Zeiss) or a confocal system (Bio-Rad Radiance).

Western Blotting
Samples of 3 ϫ 10 6 dg2P1-, dg2P2-, and dg1-transfected S2 cells were collected by centrifugation at 5,000 ϫ g for 3 min. Cells were resuspended in fresh Schneider's medium (S2 cell culture medium, Invitrogen), spun down once more, and resuspended in homogenization buffer (20 mM Tris (pH 7.5), 250 mM sucrose, 2 mM EDTA, 100 mM NaCl, 50 mM ␤-mercaptoethanol, 2 g/ml leupeptin, 5 g/ml aprotinin, and 1 g/ml phenylmethylsulfonyl fluoride-protease inhibitors obtained as protease inhibitor mixture, Sigma). Cells were disrupted by sonication and centrifuged at 21,000 ϫ g for 30 min, and the supernatants were collected. The protein content of each sample was estimated using the Bradford assay. Samples (10 g of protein each) were loaded onto each lane for Western analysis, performed according to standard protocols (33) using the ECL system (Amersham Biosciences).
Preliminary experiments showed that optimal measurements of cGK activity were achieved using GLASStide as the substrate and a reaction time of 30 min. This prolonged time was necessary to obtain reliable radioactivity counts from tubule samples. For S2 cell extracts, similar V max readings were obtained for reaction times from 3 to 60 min; therefore, for consistency, a 30-min reaction time was used for both cell and tubule preparations.
S2 Cells-Cells were collected and washed once in Schneider's medium before collection and resuspension in homogenization buffer (25 mM Tris, pH 7.4, 150 mM sucrose, 2 mM EDTA, 100 mM NaCl, 50 mM ␤-mercaptoethanol, 2 g/ml leupeptin, 5 g/ml aprotinin, and 1 g/ml phenylmethylsulfonyl fluoride). Fifty microliters of homogenization buffer was used per 10 6 cells, i.e. 150 l per transfection. Each individual transfection was assayed for cGMP-dependent kinase activity, which was routinely observed to be 20-fold that of untransfected cells, after adjustment for transfection efficiency.
Fly Tissue-Approximately 200 heads, 200 bodies, or 800 tubules from 1-week-old Oregon R (wild-type) adults were dissected and placed in 200 l of homogenization buffer as above. Tissue was homogenized and centrifuged for 5 min at 13,000 ϫ g to remove debris. Protein concentration of homogenates was determined by the Bradford assay, and homogenates were adjusted to equivalent protein concentrations for use in cGK assays.

Reverse Transcriptase-PCR
Expression of dg1 and dg2 transcripts was determined by performing reverse transcriptase (RT)-PCR on tubule cDNA preparations. Twenty tubules were dissected, poly(A) ϩ RNA-extracted (Dynal mRNA direct kit), and reverse-transcribed with Superscript Plus (Invitrogen) as described previously (24). One microliter of the reverse transcription reaction, containing cDNAs derived from one tubule (ϳ160 cells), was used as a template for the PCR performed with appropriate genespecific primer pairs based on published sequences. The sequences of primers used in this study are listed in Table I. PCR cycle conditions were as follows: 94°C (1 min) and then 36 cGK in Drosophila Renal Epithelia cycles of (94°C (30 s), 58°C (30 s), and 72°C (1 min)) and 72°C (1 min). PCR products obtained from such RT-PCR experiments were cloned using the Invitrogen topoisomerase (TOPO TA Cloning) system. Cloned plasmids were purified using Qiagen kits and sequenced to confirm their identity. The cloned PCR products shared 100% sequence identity with appropriate dg1 and dg2 transcripts (data not shown).

Generation of Transgenic Lines
Complete coding sequences for DG2-P1, DG2-P2, and DG1 were sub-cloned into pP{UAST} vectors and used to transform Drosophila embryos according to standard techniques as previously described (33). The chromosomal location of each insertion was determined by standard crossing schemes.

Fluid Transport Assays
Intact Malpighian tubules were isolated from parental lines (c42) and c42/UAS-transgenic lines, as specified in the legend of Fig. 5 into 10-l drops of Schneider's medium under liquid paraffin. Fluid secretion rates were measured in tubules as detailed elsewhere (38).
Basal rates of fluid transport were measured for 30 min, and tubules were either left untreated (controls) or treated with known stimulants of fluid transport: either cGMP (10 Ϫ4 M) (33) or capa-1 (10 Ϫ7 M) (27). The concentrations of cGMP and capa-1 used were those that elicited maximum fluid transport responses compared with controls (24,27). After stimulation, fluid secretion rates were measured for a further 40 min. Data are expressed as rates of fluid transport (nanoliters/min) Ϯ S.E.

Statistics
Where appropriate, statistical significance was assessed using Student's t test for unpaired samples, taking p Ͻ 0.05 as the critical value.

dg2 Encodes Bona Fide cGKs
To characterize dg2-encoded proteins, V5 epitope-tagged, cloned full-length transcripts were expressed in Drosophila S2 cells. Fig. 1A shows the localization of dg2-encoded proteins DG2P1 and DG2P2 in this cell line, achieved by immunocytochemistry using either anti-V5 antibody or a vertebrate cGK antibody, anti-cGKI (11). Because dg2-encoded cGK has been proposed to be closely related to vertebrate cGKI (32), we investigated the possibility that an anti-cGKI antibody would recognize DG2. Both DG2P1 and DG2P2 show a clear plasma membrane localization using both anti-V5 and anti-cGKI antibody; furthermore, co-localization between staining patterns using anti-V5 and -cGKI antibodies is demonstrated. We also show that both antibodies recognize specific DG2 bands of the correct predicted sizes by Western blotting (Fig. 1C), suggesting that anti-cGKI antibody does indeed recognize dg2-encoded cGK.
In contrast to the clear membrane localization of DG2 proteins, DG1 is localized rather more diffusely in S2 cells, showing staining adjacent to the plasma membrane. Furthermore, whereas there is some possible co-localization observed be-  (11) antibody (middle panels, red) using confocal microscopy; co-localizations with both antibodies are also shown (right panels, yellow). Control (untransfected cells) were also subjected to immunocytochemistry with either anti-V5 or anti-cGKI antibody. Single cells are shown, except for DG2P2 and control cells. In all panels, cell nuclei were stained with DAPI (blue). Scale bars are shown for dg2P1-transfected cells stained with anti-V5 antibody; except for controls, all other images are shown at the same magnification. B, for comparison, vertebrate cGKs were also expressed in S2 cells and localized with appropriate antibodies: i, cGKI␣-transfected S2 cells labeled with anti-cGKI antibody; ii, cGKII-transfected cells labeled with anti-cGKII antibody (11). The scale bar shown in A for dg2P1-transfected cells also applies to the images shown in B. C, S2 cells transfected with V5-tagged dg2P1 and dg2P2 transcripts and dg1 were subjected to Western blotting, using either anti-V5 antibody (left panel) or anti-cGKI antibody (right panel). Molecular mass markers were used to estimate the sizes of expressed protein bands. The expected sizes of expressed tagged proteins were observed, with the exception of DG1 (DG2P1, 123 kDa; DG2P2, 86 kDa; DG1, 90 kDa expected, 75 kDa observed). The absence of DG1 labeling by anti-cGKI antibody is discussed in the text.

cGK in Drosophila Renal Epithelia
tween DG1 staining with anti-V5 and -cGKI antibodies, Western blotting reveals that the vertebrate antibody may not recognize full-length transfected DG1 in S2 cells (Fig. 1C). Although the predicted size of V5-tagged DG1 is ϳ90 kDa, the band observed in the anti-V5 Western is ϳ75 kDa, suggesting proteolytic cleavage of DG1. Indeed, a 65-kDa subunit that retains catalytic activity has been shown to occur upon induction of DG1 expression in Sf9 cells (20): the 65-kDa fragment was not recognized by an N-terminal antibody to DG1, suggesting that it represents the C-terminal end of DG1. Our data support this, in that the anti-V5 antibody recognizes DG1 tagged at the C terminus as a smaller band than the predicted size. Autophosphorylation of DG1 at 84 kDa has been demonstrated (20), 2 suggesting that under some conditions DG1 does exist at the correct molecular mass, although it seems to be proteolytically cleaved in our studies when exogenously expressed and isolated for Western analysis. Finally, staining of DG1 in transfected intact S2 cells (Fig. 1A) and dg1 transgenic tubules (see Fig. 4, Aiii-Ciii) may be explained by lack of in situ proteolysis and a close conservation between the amino acid sequences of vertebrate and Drosophila cGK at the C terminus.
To compare the localizations of Drosophila cGKs to those of well defined cGKs, we transfected S2 cells with constructs encoding vertebrate cGKI␣ and cGKII. Fig. 1B shows the results of such experiments: cGKI␣, a cytosolic protein, shows staining patterns similar to that of DG1, suggesting that DG1 may be a cytosolic enzyme. By contrast, cGKII shows a clear plasma membrane localization, as predicted by previous work (11,39,40), a localization that is similar to that observed for DG2P1 and DG2P2.
All overexpressed Drosophila transcripts resulted in cGK activity. The V5 epitope tag did not affect the cGK activity associated with each transcript compared with its respective untagged constructs (data not shown). Kinetic analysis of overexpressed dg2P1 and dg2P2 (Fig. 2Ai) show that these encode bona fide cGKs, with EC 50 for cGMP of 0.13 and 0.32 M, respectively (Table II). Furthermore, DG2P1 and DG2P2 are sensitive to cAMP; but at ϳ20 M (Table II). Finally, we confirm that, in our hands, dg1 encodes a cGK, with an EC 50 for cGMP that is directly comparable with results from a previous study: 0.13 Ϯ 0.04 (Table II) versus 0.19 Ϯ 0.06 (20).
Drosophila cGKs, like their mammalian counterparts, require high Mg 2ϩ . Fig. 2Aii shows that the maximum activity occurs at 8 -10 mM Mg 2ϩ for DG2P2 and at 20 mM for DG2P1.

cGK in Drosophila Renal Epithelia
DG1 has been shown to be maximally active (histone phosphorylation assay) at 20 mM Mg 2ϩ (20).

cGK Activity in Vivo
Assay of cGK in adult Drosophila reveals highest activity in heads and tubules, with very much lower activity in bodies (Fig. 3A). Although this assay does not distinguish the cGK isoforms, the results are consistent with previous work on DG1, which showed high levels of activity in heads but not bodies (20). Tubules contain almost as much cGK activity as heads (tubules: 10. 8 Ϯ 1.3 pmol of ATP/min/mg; versus heads: 14.9 Ϯ 0.9 pmol of ATP/min/mg), suggesting that cGK action is of major significance in tubules. This is supported by the data shown in Fig. 3B, in which all major transcripts of dg2, i.e. P1, P2, and P3, as well as dg1, are expressed in tubules.

Cell-specific Targeting of DG2 and DG1 in Vivo
To test the in vivo effects of individual cGKs, DG2P1, DG2P2, and DG1, we utilized the Malpighian tubule model, which express both cGK genes (42) and dg2-and dg1-associated transcripts (Fig. 3). Transgenic lines for dg2P1, dg2P2, and dg1 under UAS control were generated. This allows tissueand/or cell-specific overexpression of these transgenes using an appropriate GAL driver (21), thus ensuring specificity, and avoiding potential problems associated with global gene expression. We used the principal cell-specific c42 GAL4 driver to generate animals in which DG2P1, DG2P2, and DG1 were expressed in only the principal cells of the main, fluid-transporting segment of the tubule (37) and in which cGMP signaling is of known significance to tubule function (43).
Specificity of the c42 GAL4 driver has been previously established, both by use of an aequorin transgene (36) and by counter-staining cell nuclei (33): principal cells and the smaller, stellate cells have nuclei of different sizes (33,37). In this work, we also used the stellate cell-specific GAL4 driver, c710 (37), which specifies expression to only stellate cells (36,37) in the main segment, to demonstrate and verify specificity of c42driven expression. In Fig. 4, expression of DG2P1 (Ai-Ci), DG2P2 (Aii-Cii), and DG1 (Aiii-Ciii) was detected by immunocytochemistry using anti-cGKI␣ antibody (11).
DG2P1-This cGK isoform is expressed in the apical membrane of tubule principal cells (28,33), indicated by white arrows; stellate cells containing small nuclei remain unstained (Fig. 4, yellow arrow, panel Ai); close inspection of the confocal image (panel Bi and inset) shows staining in the apical microvilli of tubule principal cells, confirming the apical localization shown in panel Ai. Panel Ci shows a control, the specific c710-driven expression of DG2P1, in only stellate cells (yellow arrow; note that at this magnification, punctate staining can be observed in these cells).
DG2P2-DG2P2 is expressed in both the apical, and basolateral membrane of tubule principal cells (Fig. 4

, panel Aii), with clear absence of staining in stellate cells (yellow arrow).
Orientation of the tubule basolateral membrane has been previously assigned by localization of the Na ϩ /K ϩ ATPase (28). Confocal microscopy of the c42/UAS-dg2P2 tubules confirms both apical and basolateral expression of DG2P2 (Bii). Finally, we show clear cGK staining of only stellate cells (yellow arrow) as a result of targeted expression via the c710 GAL driver (Cii).
DG1-Interestingly, DG1 is expressed in the cytosol of principal cells of the main segment, with some association at the basolateral membrane (Fig. 4, panels Aiii and Biii). Note the absence of staining in stellate cells (Biii) using c42. However, specific targeting of DG1 to the stellate cells via c710 reveals expression throughout the cytoplasm (Ciii).
Thus, we show that transcripts encoded by the same gene (DG2P1 and DG2P2) have differential localizations in vivo. Furthermore, we show that DG1 may be targeted to both the membrane and cytosol in tubules.
Interestingly, the localizations of DG2P1, DG2P2, and DG1 are fairly consistent for both S2 cells and tubules. However, whereas DG1 is primarily cytosolic in S2 cells, it is additionally   (11) was used to examine localization of overexpressed cGK in intact tubules from c42/UAS-dg2P1 (panels i), c42/UAS-dg2P2 (panels ii), and c42/UAS-dg1 (panels iii), as described under "Materials and Methods." Tubules were counterstained with the nuclear stain DAPI, which allows principal and stellate cells to be distinguished based on nuclear size (33). Tubules were viewed under UV light using an Axiocam imaging system (Zeiss) (panels A and C) or a Bio-Rad Radiance confocal system (panels B). The scale bar is shown for panel Ai and is identical for Aii and Aiii. Similarly, apart from the inset in panel Bi, the scale bar shown for Bi is identical for Bii and Biii. Finally, scale bars are shown for Ci, Cii, and Ciii. In all images, the tubule diameter can be taken as 35 m. No staining was observed in antibody-treated parental c42 tubules, nor in UAS-dg2P1, UAS-dg2P2, or UAS-dg1 lines in the absence of either primary or secondary antibody (data not shown). Panel Ai, DG2P1 is expressed at the apical membrane of principal cells (white arrows). Aii, DG2P2 is expressed at the apical and basolateral membrane of principal cells (white arrows). Aiii, DG1 is expressed at the apical membrane and in the cytosol of principal cells. Note unstained stellate cells containing small nuclei in Ai and Aii (yellow arrows), compared with those principal cells with large nuclei (33). Panels B, confocal images. Bi, expression of DGP2 in apical microvilli of principal cells. Inset, high magnification image showing apical microvilli from one principal cell (scale bar represents 20 m). Bii, expression of DG2P2 in both the apical and basolateral membranes of principal cells. Biii, DG1 is expressed throughout the principal cells. Note unstained stellate cells with small nuclei (yellow arrows). Panels C, to confirm specificity of expression using GAL4 drivers, UAS-dg2P1, UAS-dg2P2, and UAS-dg1 adults were crossed to the fly line bearing the stellate cell-specific GAL4 driver, c710 (36,37). Panels Ci (c710/UAS-dg2P1), Cii (c710/UAS-dg2P2), and Ciii (c710/UAS-dg1) show anti-cGKI antibody immunostaining of stellate cells (yellow arrows) in tubules from progeny of these crosses. D, cGK activities were assayed in tubule preparations from wild-type and transgenic c42/UAS-dg2P1, c42/UAS-dg2P2, and c42/UAS-dg1 flies. Data are expressed as -fold increases over wild-type cGK activity, Ϯ S.D., n ϭ 3.

cGK in Drosophila Renal Epithelia
present in the basolateral membrane of tubules, suggesting that polarity of tubule cells may contribute to the latter localization.
Targeted overexpression of each cGK isoform in only tubule principal cells results in significantly enhanced cGK activity (Fig. 4D). Interestingly, overexpression of DG2P1 and DG2P2 elicited a much greater increase in cGK activity than did DG1 overexpression, in comparison to wild-type flies: 32.2 Ϯ 2.8-fold (DG2P2), 45 Ϯ 0.16-fold (DG2P1), and 6.16 Ϯ 0.7-fold (DG1). In all experiments, cGK activity in parental UAS and c42 lines showed similar activity to wild-type Oregon R animals (data not shown). Finally, although successful targeted expression of cGKs in stellate cells was achieved by use of the c710 GAL4 driver, these lines were not characterized further due to the small numbers of stellate cells (ϳ30) in each tubule.

Phenotypic Consequences of Targeted Expression of cGKIsoforms
Targeted overexpression of individual cGKs in tubule has revealed differential localization in vivo. To determine the physiological effects of targeted overexpression of cGKs, tu-bules from c42/UAS-dg2P1, c42/UAS-dg2P2, c42/UAS-dg1, and the parental c42 line were assayed for transport phenotype. Because cGMP signaling in principal cells has been shown to modulate fluid transport rates (43), this was conducted in the absence and presence of two stimulants of tubule fluid secretion: exogenous cGMP (24,33), which enters the tubule via cyclic nucleotide transporters (23), and the endogenous nitridergic neuropeptide, capa-1 (27).
pression of DG2P2 and DG1, but not DG2P1, in principal cells (Fig. 5, C and D). Although enhancement of capa-1-stimulated transport observed in tubules is statistically significant in c42/ UAS-dg1 and c42/UAS-dg2P2 compared with c42 tubules, the enhancement is most marked in c42/UAS-dg2P2 tubules. Thus, the roles of dg2-encoded transcripts with respect to fluid transport in tubule principal cells may be dependent on the source of cGMP, whereas DG1 is of primary significance in fluid transport stimulated by both cGMP and capa-1 in the tubule. DISCUSSION Although the effects of the dg2 gene have been well documented in Drosophila, the proteins encoded by dg2 transcripts have not been studied. We provide the first biochemical characterization of dg2-encoded cGKs and show that the major dg2 transcripts, dg2P1 and dg2P2, encode bona fide cGKs. We have also demonstrated expression of dg2P3 in adult tubules (Fig. 3) and have unpublished evidence that dg2P4 expression also occurs in tubules (data not shown). However, in our hands, dg2P3 and dg2P4 do not display cGK activity when expressed in S2 cells nor when overexpressed in tubules. This is surprising, given that these transcripts encode proteins highly similar to DG2P1 and DG2P2. They contain the canonical domains for cGMP binding and kinase activation, however, they lack the autoinhibitory region and possibly also the appropriate residues for autophosphorylation contained in the other cGKs. Thus, DG2P3 and DG2P4 may be inactive in vivo. Given the lack of cGK activity associated with dg2P3 and dg2P4, this study focused on dg2P1 and dg2P2.
We show that DG2P1 and DGP2 are expressed at the plasma membrane in S2 cells and in vivo, in tubules (below); by contrast, DG1 is cytoplasmic. Analysis of vertebrate cGK expression in S2 cells supports previous work that shows that cGKI is a soluble enzyme (44,45), whereas cGKII is localized to the plasma membrane (32), confirming that cGKs are not mislocalized in S2 cells. In general, post-translational modifications resulting in protein attachment to the membrane occur via the N-terminal (myristoylation) or C-terminal (glycosylphosphatidylinositol (GPI) lipid anchors, farnesylation, or geranylgeranylation). cGKII has been shown to be anchored to the membrane via myristoylation of an N-terminal Gly residue (46). However, close inspection of the DG2 sequences does not reveal any such conserved residue. Use of an N-terminal myristoylation residue predictor (NMT predictor; available at mendel. imp.univie.ac.at/myristate/SUPLpredictor.htm (47)) shows no evidence for such residues in DG2P1, DG2P2, or DG1. Furthermore, a search of possible GPI anchor sites using big-PI (available at mendel.imp.univie.ac.at/sat/gpi/gpi_server.html (47)) shows that no probable GPI anchor sites exist in the Drosophila cGKs. In the absence of either myristoylation or GPI anchoring, the possibility remains then, that DG2 proteins may interact with specific anchoring proteins to allow membrane tethering in the tubule. As such, it is entirely possible that, in Drosophila tubules, anchoring of cGKs occurs via proteins similar to vertebrate G-kinase anchor proteins (48).
Interestingly, analysis of the primary structures of Drosophila and vertebrate cGK places cGKII closer to DG1 than to cGKI and relates cGKI␣ and cGKI␤ to DG2 (32). However, the cellular locations of DG1 and DG2P1/DG2P2 do not map neatly to those of cGKII/I, either in S2 cells or in the tubule, although only DG1, like cGKI, has some cytosolic localization. DG2 has been associated with insect behavior, and cGKII is widespread in mammalian brain, in contrast to cGKI, which is more restricted to cerebellum and a few neuronal cell types (49). cGKII, like DG2, is highly expressed in brain and kidney, and even more in intestinal mucosa (32,50). cGKI is most highly expressed in smooth muscle, platelets, cerebellar Purkinje cells, and to a lesser extent in fibroblasts, cardiac myocytes, and certain endothelial cells (2,3,10).
We show that both DG2P1 and DG2P2 are activated by micromolar concentrations of cGMP and require high Mg 2ϩ concentrations for activity, and we also confirm previous work showing that Drosophila dg1 encodes a cGK (20). Interestingly, both DG2 isoforms, although not DG1, are sensitive to the cGK inhibitor, (R p )-8-CPT-cGMPS, which inhibits vertebrate cGKII at 100-fold lower concentrations than cGKI␣ (41). Thus, the gene products of both Drosophila and vertebrate cGK genes have markedly different preferences for this cGMP analog inhibitor.
Previous studies on DG1 have shown that this enzyme acts as a dimer (20). Preliminary work utilizing gel filtration shows that DG2P1 and DG2P2 may also be dimers 3 and are ancestral to the dimeric vertebrate cGKs.
Investigation of organotypic roles for these cGKs required a suitable genetic model; as such, we utilized the fluid-transporting Drosophila Malpighian tubule, in which cGMP signaling is an important modulatory event. Furthermore, previous work has shown that cGMP signaling in the polarized tubule principal cells can modulate epithelial function, which can be assessed by measurement of fluid transport rates.
By utilizing the UAS/GAL4 system for targeted gene expression, we show for the first time that the cGKs encoded by different dg2 transcripts are localized to distinct membranes in a polarized cell in vivo. DG2P1 is localized to the apical membrane, whereas DG2P2 is localized to both the apical and basolateral membrane, suggesting different roles for individual cGK isoforms in vivo. This also suggests that both apical and basolateral membranes play a role in cGMP signaling. Interestingly, cGKII has also been shown to localize to the apical surface of intestinal mucosa cells (32). By contrast, however, DG1 expression in tubule principal cells occurs at the basolateral membrane and in the cytosol, although it is cytosolic in S2 cells. Immunostaining for DG1 in native animals also reveals expression in cell bodies of optic lamina in adult Drosophila heads (20), supporting a cytosolic role for endogenous DG1. These and other results suggest that DG1 may have additional roles in cGMP signaling. Recent microarray analysis shows that dg1 mRNA expression is ϳ16-fold higher in the tubule compared with the rest of the fly (53). Thus, the epithelial context of DG1 is particularly significant in Drosophila and is confirmed by the phenotypic consequences of DG1 overexpression in tubule.
Assessment of the impact of targeted cGKs on epithelial transport reveals differential roles for dg2 transcripts and distinct roles of dg2 versus dg1 in stimulated fluid transport. DG1 enhances cGMP-stimulated transport, whereas DG2P2 increases capa-1 stimulated transport.
cGMP-stimulated transport is also enhanced to lesser extents by DG2P1 and DG2P2, which suggests that differentially localized cGKs are able to transduce the cGMP signal. This is in contrast to capa-1 stimulation, which is enhanced by DG2P2 and to a lesser extent by DG1, both localized to the basolateral membrane. Importantly, the capa response, which is due to both stimulation of calcium signaling and cGMP production (27), has been shown to be associated with L-type calcium channels located in the basolateral membrane. Thus, it is possible that a basolateral pool of cGMP is critical to the capa response, explaining the role of cGKs localized to the basolateral membrane. Interestingly, mammalian cGKI has been shown to inhibit the L-type calcium channel in cardiac myocytes (51). Thus, the combination of in vitro and organotypic studies for Drosophila cGKs has revealed new insights into the function of these important enzymes in tubules.