This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Qian, X. Articles by Goy, M. F. Search for Related Content PubMed PubMed Citation Articles by Qian, X. Articles by Goy, M. F.

Expression of GC-C, a Receptor-Guanylate Cyclase, and Its Endogenous Ligands Uroguanylin and Guanylin along the Rostrocaudal Axis of the Intestine¹

Department of Cell and Molecular Physiology, University of North Carolina (X.Q., S.P., M.F.G.), Chapel Hill, North Carolina 27599; and Department of Molecular Reproduction, Development, and Genetics, Indian Institute of Science (A.N., S.S.V.), Bangalore 560012, India

Address all correspondence and requests for reprints to: Dr. Michael F. Goy, Department of Cell and Molecular Physiology, Campus Box 7545, University of North Carolina, Chapel Hill, North Carolina USA 27599-7545. E-mail: mgoy{at}med.unc.edu

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Members of the receptor-guanylate cyclase (rGC) family possess an intracellular catalytic domain that is regulated by an extracellular receptor domain. GC-C, an intestinally expressed rGC, was initially cloned by homology as an orphan receptor. The search for its ligands has yielded three candidates: STa (a bacterial toxin that causes traveler’s diarrhea) and the endogenous peptides uroguanylin and guanylin. Here, by performing Northern and Western blots, and by measuring [¹²⁵I]STa binding and STa-dependent elevation of cGMP levels, we investigate whether the distribution of GC-C matches that of its endogenous ligands in the rat intestine. We establish that 1) uroguanylin is essentially restricted to small bowel; 2) guanylin is very low in proximal small bowel, increasing to prominent levels in distal small bowel and throughout colon; 3) GC-C messenger RNA and STa-binding sites are uniformly expressed throughout the intestine; and 4) GC-C-mediated cGMP synthesis peaks at the proximal and distal extremes of the intestine (duodenum and colon), but is nearly absent in the middle (ileum). These observations suggest that GC-C’s activity may be posttranslationally regulated, demonstrate that the distribution of GC-C is appropriate to mediate the actions of both uroguanylin and guanylin, and help to refine current hypotheses about the physiological role(s) of these peptides.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
AS A SECOND messenger, cGMP is thought to regulate a variety of metabolic and physiological processes, including phototransduction in the retina (1), relaxation of vascular smooth muscle (2), mucociliary clearance in the lung (3), electrolyte transport across epithelial monolayers (4), and modulation of synaptic strength in the nervous system (5). This signaling pathway is under the control of ligand-regulated enzymes that govern the relative rates of cGMP synthesis or degradation. Such enzymes include membrane-bound receptor-guanylate cyclases (rGCs) that are activated by peptide hormones (6), cytoplasmic guanylate cyclases that are activated by nitric oxide (7), and cGMP-selective phosphodiesterases that are activated by a receptor/G protein-coupled mechanism (8).

One member of the rGC family, a peptide-sensitive cyclase called GC-C (9), controls cGMP synthesis in intestinal epithelial cells. GC-C was first recognized because it is pathologically targeted by a bacterially produced peptide toxin, the heat-stable enterotoxin (STa) (10). Binding of STa to an extracellular (intraluminal) domain of GC-C enhances cGMP synthesis in epithelial target cells. The subsequent increase in cGMP levels stimulates electrogenic chloride secretion via the cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel (11) and also inhibits neutral sodium chloride uptake (12, 13). This leads to the luminal accumulation of sodium chloride and water (14) and thus to a severe form of diarrhea commonly referred to as traveler’s diarrhea. Gene knockout studies in rodents demonstrate that these effects on ion transport are mediated in small intestine primarily by the type II cGMP-dependent protein kinase (13, 15), whereas in colon they are mediated by an as yet unspecified pathway that is type II cGMP-dependent protein kinase independent (13).

Toxin-induced diarrhea promotes rapid transfer of pathogens from one host to another, thus providing toxigenic bacteria with a reproductive advantage. In contrast, the diarrhea has adverse health and reproductive consequences for the host; STa-induced diarrhea is a general health problem for all mammals, and for humans it is a leading world-wide cause of death among children under the age of 5 yr (16). Surprisingly, despite this strongly negative selective pressure, GC-C has been evolutionarily conserved in a wide variety of animal species, suggesting that it must play a critical role in some aspect of intestinal physiology. One suggestion consistent with this idea is that GC-C serves not only as a toxin receptor, but also, like other rGCs, as a receptor for one or more endogenous peptide hormones.

This hypothesis has been greatly strengthened by the identification of two candidate peptide ligands, uroguanylin and guanylin, both of which selectively activate GC-C, share significant structural homology with STa, and compete with STa for common binding sites (17, 18). Like GC-C, these putative ligands are expressed primarily in the intestine, although both the receptor and the peptides have been detected at very low levels in other epithelial tissues, such as kidney and lung (19, 20, 21, 22). In rodents, it appears from preliminary studies that uroguanylin levels are relatively high in proximal small intestine and much lower in colon, whereas guanylin shows the opposite pattern (19, 23, 24, 25, 26). Furthermore, intestinal uroguanylin and guanylin are made by distinct types of cells in the rat intestine; uroguanylin is made primarily by enterochromaffin cells (22, 24), whereas guanylin is made primarily by goblet cells (and perhaps also by a subset of columnar epithelial cells) (27, 28). These unique spatial and cellular distributions imply that the two peptides are likely to carry out distinct physiological functions.

Although initial studies have provided a partial assessment of guanylin and uroguanylin expression within the intestine, many regions of the bowel have not been evaluated. Similarly, expression of GC-C, the presumed guanylin/uroguanylin receptor, has not been well defined throughout the length of the intestine. This is particularly important in light of the already apparent lack of congruency in the distributions of its two putative ligands. Evaluation of the tissue expression patterns of these recently discovered peptides and their receptor is essential for understanding their physiological functions.

In this communication we survey uroguanylin, guanylin, and GC-C expression along the rostrocaudal axis of the intestine of the rat. Our measurements reveal that 1) uroguanylin is almost exclusively a small bowel peptide, with highest messenger RNA (mRNA) and propeptide expression in the jejunum, slightly lower levels in duodenum and ileum, and minimal levels in cecum and colon; 2) guanylin shows a nearly complementary pattern of expression, with mRNA and propeptide levels ascending to a peak in the cecum and declining slightly in the distal colon; 3) GC-C mRNA and STa-binding sites are both found at high and relatively constant levels in all regions of the small and large bowels, a distribution that is appropriate to mediate the effects of both uroguanylin and guanylin; and 4) STa-activated rGC activity is nonuniformly distributed, with relatively high levels of activity at the proximal and distal extremes of the intestine (duodenum and distal colon) and relatively low levels in the middle (ileum), suggesting that some form of posttranslational regulation of GC-C uncouples ligand binding from cyclase activity in specific parts of the intestine.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Tissue preparation
Age-matched male Sprague Dawley rats (150–300 g) were purchased from local suppliers and maintained for several days on a 12-h light, 12-h dark cycle with continuous access to water and standard rat chow. All animals were killed in the midmorning hours. Although this does not provide the highest levels of peptide expression, which have recently been shown to peak at night (29), the relatively constant time of sacrifice serves to minimize circadian variability. Animals were decapitated after urethane or CO₂ anesthesia, and intestinal tissues were rapidly removed. Some studies employed the entire length of the small and large bowels, subdivided into segments (17 contiguous, equal-sized pieces for small intestine, 1 piece for cecum, and 3 contiguous, equal-sized pieces for large intestine). In the figures, these pieces are numbered consecutively as segments 1–21. Other studies focused on a set of 5 standardized, reproducibly identifiable regions: duodenum (a slightly enlarged region 1–2 cm in length located just distal to the stomach, corresponding to segment 1), proximal jejunum (a 2- to 4-cm length of small bowel 2 cm distal to the duodenum, corresponding to segment 3), distal ileum (a 2- to 4-cm length of small bowel just proximal to the cecum, corresponding to segment 17), proximal colon [the proximal third (5 cm) of the large bowel just distal to the cecum, corresponding to segment 19], and distal colon [the distal third (5 cm) of the large bowel just proximal to the rectum, corresponding to segment 21]. In all cases, the luminal surface was exposed by cutting along the long axis of the bowel, and the tissue was rinsed with cold physiological saline solution (140 mM Na⁺, 120 mM Cl^-, 25 mM HCO₃, 1.2 mM Mg²⁺, 1.2 mM Ca²⁺, 2.4 mM K₂HPO₄, 0.4 mM KH₂PO₄, and 10 mM glucose; bubbled with 95% O₂-5% CO₂).

Subsequent processing depended on the type of analysis being performed. 1) For prouroguanylin and proguanylin Western blots, mucosal tissue was scraped away from the underlying smooth muscle with a microscope slide. We observed comparable results using extracts prepared from whole intestine, but the increase in specific activity obtained with mucosal extracts greatly improves the signal to noise ratio. Scrapings were immediately frozen on a metal plate chilled with dry ice and then homogenized in buffer B [50 mM HEPES, pH 7.4, containing the following protease inhibitor cocktail: 1 mM EDTA, 0.01% bacitracin, 2.5 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, 38 µM pepstatin A, 35 µM trans-epoxysuccinyl-L-leucylamido(4-guanidino)butane, 100 µM bestatin, 55 µM leupeptin, and 2 µM aprotinin]. Homogenates were then cleared by centrifugation at 50,000 x g for 60 min at 4 C, and the pellet was discarded. 2) For STa binding assays, mucosal scrapings were prepared as described above and homogenized in buffer C (50 mM HEPES, pH 7.5, containing 1 µg/ml leupeptin, 1 µg/ml aprotinin, 2 mM phenylmethylsulfonylfluoride, 2 mM dithiothreitol, and 100 mM NaCl). The suspension was first centrifuged for 10 min at 1,000 x g at 4 C, and the resulting supernatant fraction was then centrifuged for 75 min at 30,000 x g. The final pellet was resuspended in buffer C without dithiothreitol at a protein concentration of 5 mg/ml. Binding assays could not be performed reliably on membranes derived from whole intestine due to excessive levels of nonspecific binding. 3) For Northern blots, intact isolated tissues were frozen as quickly as possible in liquid nitrogen (to prevent ribonuclease activity) and stored at -80 C for subsequent RNA extraction. Although the specific activities of uroguanylin, guanylin, and GCC mRNAs would undoubtedly be higher in RNA isolated from mucosal scrapings, we found that scraping results in an unacceptable level of RNA degradation. 4) For GC-C bioassay experiments, isolated tissues were subdivided into smaller pieces (3 x 3 mm) and stored briefly (<30 min) at 4 C in physiological saline until assayed. These assays were performed with intact tissues to minimize changes in activity that occur during membrane purification; in our laboratories, GC-C-dependent cyclase activity is always much higher in intact tissues than in isolated mucosal membranes.

Northern blotting
Total RNA was purified from frozen tissues, blotted to positively charged nylon membranes (Roche Molecular Biochemicals, Indianapolis, IN), and hybridized with ³²P-labeled uroguanylin, guanylin, or GC-C probes, as previously described (30). Templates for probe synthesis were amplified by PCR from appropriate plasmids, then gel purified and recovered from low melt agarose using the Wizard PCR Preps DNA Purification System (Promega Corp., Madison, WI). Labeling was performed with either a DECAprime II DNA labeling kit (Ambion, Inc., Austin, TX) or a Strip-EZ DNA labeling kit (Ambion, Inc.) using ³²P-labeled deoxy-ATP or deoxy-CTP (New England Nuclear Corp., Boston, MA). After hybridization, an initial autoradiographic image was obtained on Kodak X-OMAT Blue XB-1 film (Eastman Kodak Co., Rochester, NY), and probe binding was then quantitated on a STORM 480 PhosphorImaging System (Molecular Dynamics, Inc., Sunnyvale, CA).

We tested several potential normalization standards (including ßactin and ubiquitin), hoping to use them to correct for sample to sample differences in RNA recovery. However, the relative expression levels of all standard mRNAs tested were themselves found to vary considerably from one region of the intestine to another (data not shown). Therefore, we normalized our results to the amount of RNA loaded in each lane (40 µg), as determined spectrophotometrically. We also monitored the intensity of the ethidium bromide-stained 18S and 28S ribosomal RNA bands (observed before transferring the samples to the blotting membrane) to confirm that the spectrophotometric measurements had allowed for uniform sample loading.

For each probe (uroguanylin, guanylin, and GC-C) we tested RNA isolated from a set of contiguous tissue samples that had been obtained from a single animal (white symbols). In addition, to evaluate animal to animal variability, we measured uroguanylin, guanylin, and GC-C mRNA levels in several independent animals, focusing on the five standardized regions of the intestine described above. Because these measurements were performed independently over a period of time, with corresponding differences in exposure times and probe specific activities, we normalized the set of data points obtained for each animal so that they were distributed over a range of values that was comparable for all animals. Specifically, for each animal that was analyzed, we quantitated expression in the five standardized regions, summed these five values to give a weighting factor, and then normalized the data for each animal so that all of the weighting factors would be brought to a common value. The normalized datasets were then combined, and the mean values and SEs (black symbols) were plotted at the appropriate locations in Fig. 1 for uroguanylin, Fig. 2 for guanylin, and Fig. 8 for GC-C. The results of our limited survey with multiple animals are fully consistent with the more extensive study performed on the single animal and confirm that the characteristic distributions of each mRNA species are observed reproducibly.

View larger version (33K): [in this window] [in a new window]   Figure 1. Distribution of uroguanylin mRNA along the rostrocaudal axis of the intestine. a, Top, Autoradiogram from a Northern blot of total RNA (40 µg/lane) hybridized with radiolabeled uroguanylin probe. RNA was isolated from 21 contiguous pieces of tissue comprising the full length of the small and large bowels, as indicated at the bottom of the figure. All samples were electrophoresed at the same time (on two gels), transferred, and hybridized as a group, but some individual lanes have been isolated photographically for clarity. RNA obtained from segment 16 was degraded, so this sample has been deleted. Middle, 28S and 18S ribosomal RNA present in each RNA sample, visualized with ethidium bromide, and photographed just before the RNA was transferred to the membrane; this confirms that comparable amounts of intact RNA were loaded in each lane. Bottom, Phosphorimager quantitation of probe hybridization. White symbols give relative magnitudes of the hybridization signals, each aligned with the appropriate lane of the autoradiogram (n = 1). Black symbols (mean ± SEM; n = 3) include data from the blot shown in the autoradiogram averaged together with data obtained from additional blots performed on a smaller subset of tissue segments isolated from two other animals (one of which is shown in b). b, Top, Autoradiogram obtained from a Northern blot performed on a subset of tissue segments, as indicated. The exposure time for this autoradiogram was effectively about 20 times longer than the exposure time for the autoradiogram in a to demonstrate the existence of a minor 1.1-kb transcript (marked by the asterisk) that hybridizes with the uroguanylin probe. The overexposed uroguanylin transcript (0.6–0.8 kb) is indicated by the arrow. Bottom, Ethidium bromide staining demonstrates that comparable amounts of intact RNA have been loaded in each lane.

View larger version (37K): [in this window] [in a new window]   Figure 2. Distribution of guanylin mRNA along the rostrocaudal axis of the intestine. a, The autoradiogram (top), ethidium bromide stain (middle), and phosphorimager quantitation (bottom) were all obtained as described in Fig. 1. The RNA samples were the same as those used in Fig. 1a, but independent aliquots were electrophoresed on a pair of new gels, transferred, and hybridized with a radiolabeled guanylin probe. As before, the RNA obtained from segment 16 was degraded, so this sample has been deleted. White symbols give relative magnitudes of the hybridization signals from each tissue segment, aligned with the appropriate lanes of the autoradiogram (n = 1). As in Fig. 1, the black symbols (mean ± SEM; n = 6) include additional data obtained from blots performed on a smaller subset of tissue segments isolated from five other animals (one of which is shown in b). b, Top, Autoradiogram obtained from a Northern blot performed on a subset of tissue segments, as indicated. This is the same blot illustrated in Fig. 1b after stripping and rehybridizing with a guanylin probe. The exposure time for this autoradiogram was effectively about 20 times longer than the exposure time used for the autoradiogram in a. The overexposed guanylin transcript (0.6–0.8 kb) is indicated by the arrow. No additional hybridizing species were observed with this probe. Bottom, Ethidium bromide staining demonstrates that comparable amounts of intact RNA have been loaded in each lane.

View larger version (42K): [in this window] [in a new window]   Figure 8. Distribution of GC-C mRNA along the rostrocaudal axis of the intestine. The autoradiogram (top), ethidium bromide stain (middle), and phosphorimager quantitation (bottom) were all obtained as described in Fig. 1. The RNA samples were the same as those used in Fig. 1a, but independent aliquots were electrophoresed on a new gel, transferred, and hybridized with a radiolabeled GC-C probe. As before, the RNA obtained from segment 16 was degraded, so this sample has been deleted. White symbols give relative magnitudes of the hybridization signals from each tissue segment, aligned with the appropriate lanes of the autoradiogram (n = 1). As in Fig. 1, the black symbols (mean ± SEM; n = 7) include additional data obtained from blots performed on a smaller subset of tissue segments isolated from six other animals.

GC-C bioassay
Small pieces of tissue isolated from each intestinal region were placed in groups of three to six in glass shell vials, each vial containing 1 ml oxygenated physiological saline solution with or without STa at 1–4 µg/ml (equivalent to 100–400 U/ml), supplemented with 0.5 mM isobutylmethylxanthine (IBMX) and in some experiments additionally with 0.1 mM zaprinast. Tissues were incubated for 30 min at 37 C and intermittently bubbled with 95% O₂-5% CO₂. Reactions were stopped by freezing tissue pieces on a metal plate chilled with dry ice. Each individual piece was then homogenized in 6% trichloroacetic acid (TCA) and centrifuged to separate TCA-insoluble protein from TCA-soluble cGMP. The pellet was dissolved in 1 N NaOH and assayed for protein content (Bio-Rad Laboratories, Inc., Hercules, CA; Bradford assay). The cGMP level was quantitated by RIA using standard procedures and was normalized to protein to correct for small differences in the sizes of the original pieces of tissue. In some experiments, a portion of the incubation medium was also analyzed by RIA to measure cGMP that was secreted from the tissue. Secreted cGMP was normalized to total protein, taken as the sum of the protein contributed by all of the pieces of tissue that had been coincubated in the vial.

STa integrity assay
T84 cell membranes were prepared as described previously (31). Membranes were diluted to a protein concentration of 1 mg/ml in buffer B (see above), and stored at -80 C until used. To initiate cyclase activity, a portion of the diluted membranes was thawed and incubated with an equal volume of buffer A [50 mM HEPES, 8 mM MgCl₂, 2 mM IBMX, 4 mM GTP, 60 mM phosphocreatine, 800 µg/ml creatine phosphokinase (250 U/mg), and 1 mg/ml BSA] at 37 C, with or without a sample of STa (or control medium) to be tested. At various times thereafter, 80-µl samples were pipetted from the incubation mixture into 200 µl 6% (wt/vol) TCA. Each sample was then extracted four times with 1 ml diethyl ether to remove the acid, and the cGMP content was determined by RIA. The rate of cGMP synthesis is given by the slope of a line fit to 0, 5, and 10 min points by linear regression analysis. Under subsaturating conditions the fold stimulation of cGMP synthesis (stimulated rate/basal rate) is proportional to the concentration of intact STa in the test solution.

STa binding assay
Binding was performed with freshly isolated membranes at pH 7.5 as described previously (32), using radiolabeled STY72F peptide as the ligand and 100–200 µg protein for each assay point. Assays contained an input radioactivity of approximately 100,000 dpm in either the presence or absence of unlabeled STa peptide. Incubations were continued for 1 h, after which the samples were filtered through GF/C filters (Whatman, Clifton, NJ) and washed with chilled PBS containing 0.2% BSA. Filters were dried and monitored for radioactivity. Specific binding was always more than 80% of the total binding and is taken as the difference between total and nonspecific binding. Scatchard analysis was performed with the LIGAND program (33), which allows direct comparison between single and multiple binding site models; our data were best fit by assuming a single binding site.

Chemicals
Zaprinast (Calbiochem, La Jolla, CA) was made as a 100-fold concentrated stock solution (10 mM) in 1 M NaOH. IBMX (Sigma, St. Louis, MO) was dissolved directly in saline solution at a final working concentration of 0.5 mM. Protease inhibitor cocktail (for mammalian cell extracts) and STa were purchased from Sigma. The anti-cGMP antibody and protein A used for the RIA were obtained from Woods Assays (Portland, OR), chemiluminescence reagent was obtained from NEN Life Science Products (Boston, MA), and protein assay reagent was purchased from Bio-Rad Laboratories, Inc.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Regional expression of uroguanylin and guanylin mRNAs
Previous studies have examined uroguanylin and guanylin expression at a limited number of discrete points along the rostrocaudal length of the rat intestine. The autoradiograms in Figs. 1a and 2a present results from a Northern blot analysis of a single intestine, subdivided into 21 contiguous pieces that span the full length from duodenum to distal colon. The RNA samples obtained from each region have been hybridized with probes for uroguanylin (Fig. 1a) or guanylin (Fig. 2a), revealing the RNA expression patterns for both peptides simultaneously within a single animal. Each probe is specific, hybridizing exclusively (or, in the case of uroguanylin, predominantly) to a single transcript (marked with the arrows in Figs. 1a and 2a).

Note that an additional, low abundance 1.1-kb mRNA species was detected with the uroguanylin probe after long exposure times (marked with the asterisk in Fig. 1b). No such species was detected after long exposure times with the guanylin probe (Fig. 2b). The significance of this minor transcript is not clear. It is much larger than any of the uroguanylin complementary DNA clones isolated to date, and its tissue distribution (with highest expression in proximal large bowel) does not match the distribution of the uroguanylin propeptide as detected by Western blotting (see below). Further analysis is required to determine whether it represents an alternately spliced product of the uroguanylin gene or is the product of an independent gene with sufficient homology to hybridize with the uroguanylin probe. One possibility is that it encodes the rat form of lymphoguanylin (34), a recently discovered peptide, to date identified only in opossums, that has strong homology to uroguanylin.

We used a phosphorimager to quantify relative uroguanylin and guanylin probe hybridization from each region of the intestine. These results are presented in the lower portions of Figs. 1a and 2a, aligned beneath the relevant lanes of the autoradiograms. The white symbols indicate individual values obtained from the single animal shown in the upper autoradiogram, and the black symbols indicate average values for selected regions that were analyzed from multiple animals. Comparison of uroguanylin and guanylin reveals significant differences in abundance along the length of the intestine. Uroguanylin mRNA is low in duodenum, increases rapidly to a high level that remains relatively constant throughout the jejunum and most of the ileum, declines somewhat in terminal ileum, and then declines rapidly to very low levels in cecum and colon (detectable only after long exposure times; see Fig. 1b). In contrast, guanylin mRNA increases relatively linearly from very low levels in duodenum to a peak in terminal ileum, remaining high in cecum, and declining to about half the maximal level along the length of the colon. These results are fully consistent with the previous, more limited Northern blot surveys of intestinal guanylin and uroguanylin expression cited above.

Regional expression of uroguanylin and guanylin propeptides
We used Western blots to measure prouroguanylin and proguanylin levels along the length of the intestine. Each propeptide was evaluated with two different antipeptide antibodies. The two antiprouroguanylin antibodies (6910 and 6912) were independently raised against short sequences of amino acids taken from two regions of the prouroguanylin molecule that have no homology with proguanylin. Similar considerations apply to the two antiproguanylin antibodies (2538 and 6240). The specificities of these antibodies have been extensively investigated and described previously (24, 27).

Properties of the antibodies
Figure 3a shows representative immunoblots obtained with the two antiprouroguanylin antibodies. Two identical samples from a single protein extract were run side by side on a 15% acrylamide gel; one lane was blotted with antibody 6910, and the other with antibody 6912. Each antibody labels what appears to be a single polypeptide in the 6- to 10-kDa range (marked by the arrow). With either antibody, on occasional blots this single band could be resolved into a closely migrating doublet (not shown here, but see examples below). The two closely migrating polypeptides in the doublet had comparable electrophoretic mobilities when detected with either 6910 or 6912; as the two antibodies were raised against synthetic peptides representing different portions of the prouroguanylin sequence, this argues strongly that both of the immunoreactive molecules must be derivatives of prouroguanylin (perhaps representing intact prouroguanylin and a lower mol wt metabolite).

View larger version (38K): [in this window] [in a new window]   Figure 3. Representative Western blots performed with antiprouroguanylin antibodies 6910 and 6912 (left) and antiproguanylin antibodies 2538 and 6240 (right). Two identical 40-µg protein samples from small intestine (left) were electrophoresed on the same gel and electroblotted at the same time, and the blotting membrane was then cut into two separate pieces for incubation with the two antiprouroguanylin antibodies, as indicated. Similarly, two identical 40-µg protein samples from colon (right) were analyzed in parallel using the two antiproguanylin antibodies, as indicated. Apparent molecular mass scales are provided, in kilodaltons. Antibodies 6910 and 6912 both recognize an approximately 10-kDa polypeptide (marked with the arrow), which sometimes resolves into a doublet. We believe that this closely migrating doublet is comprised of two related forms of prouroguanylin. Antibodies 2538 and 6240 recognize a pair of well resolved immunoreactive polypeptides (marked by the arrows) that we believe are related forms of proguanylin. Antibody 6240 recognizes an additional, smaller molecule (marked by the arrowhead) that may also be related, and two 29- to 30-kDa proteins that are too large to be related. All four antibodies also display minor, nonspecific cross-reactivity with very high molecular mass material (not shown). See text and Refs. 24 and 27 for details.

Additional arguments support the identification of the 6- to 10-kDa bands as authentic products of the uroguanylin and guanylin genes. 1) Polypeptides migrating in the 6–10 kDa range have appropriate molecular masses for prouroguanylin (predicted to be 9.4 kDa) and proguanylin (predicted to be 10.2 kDa), along with smaller processing fragments. 2) No immunoreactive 6- to 10-kDa polypeptides are detectable in tissues that lack uroguanylin or guanylin mRNA (Qian, X., D. T. Devries, and M. F. Goy, manuscript in preparation), whereas, as will be described below, the relative levels of the polypeptides detected by 6910/6912 and 2538/6240 along the rostrocaudal axis of the intestine agree reasonably well with the relative rostrocaudal abundance of the uroguanylin and guanylin mRNA transcripts, respectively. 3) Biological activity (the ability to activate GC-C) copurifies with the immunoreactive prouroguanylin- and proguanylin-like polypeptides on a reverse phase HPLC column (27, 35).

Quantitative analysis of prouroguanylin expression
Using antibodies 6910 and 6912, we analyzed uroguanylin propeptide expression along the complete length of the intestines of four independent animals. The upper portion of Fig. 4 shows the relevant portion of a representative Western blot from one of these animals, performed with antibody 6910 (comparable results are obtained with antibody 6912). A barely resolved pair of uroguanylin-related polypeptides can be seen on this blot in the 6–10 kDa range (marked by the arrow). For quantitation, both of these bands were summed together to provide a measure of total prouroguanylin-related polypeptide expression (note that the ratio of the two bands appears to be fairly constant in all segments). Similar densitometric analysis was performed on blots obtained for each of the other three animals, all of which look similar to the blot illustrated in the figure. Results for each animal were then normalized so that the values would fall within a range that was comparable for all animals (see Materials and Methods), and the mean relative expression levels were calculated at each position along the intestine and aligned beneath the appropriate lanes of the autoradiogram in Fig. 4. Comparison of these quantitative polypeptide measurements with the quantitative mRNA measurements in Fig. 1 reveals very good correspondence overall, indicating that steady state uroguanylin polypeptide levels are, in general, proportional to steady state uroguanylin mRNA levels.

View larger version (22K): [in this window] [in a new window]   Figure 4. Distribution of prouroguanylin along the rostrocaudal axis of the intestine. Top, Chemiluminescent signals after Western blot analysis with antiprouroguanylin antibody 6910 generated from a set of tissue extracts (40 µg protein/lane) obtained from a single, representative animal. Extracts were prepared from 21 contiguous pieces of tissue comprising the whole length of small and large bowels. All samples were concurrently electrophoresed, electroblotted, and incubated with primary and secondary antibodies, but some individual lanes have been isolated photographically for clarity. Only the relevant portion of the blot is shown, and an apparent molecular mass scale is provided, in kilodaltons. Bottom, Densitometric quantitation of chemiluminescent signals. Each point is aligned with the corresponding lane of the Western blot shown at the top, and gives the relative prouroguanylin expression level (mean ± SEM), calculated from normalized data obtained from four independent animals, each analyzed as described above.

View larger version (50K): [in this window] [in a new window]   Figure 5. Distribution of proguanylin along the rostrocaudal axis of the intestine. Chemiluminescent signals were generated by tissue extracts (40 µg protein/lane) after Western blot analysis with antiproguanylin antibody 2538. Four individual animals are illustrated. In each case, extracts were prepared from 21 contiguous pieces of tissue comprising the whole length of small and large bowels. For each animal, all samples were concurrently electrophoresed, electroblotted, and incubated with primary and secondary antibodies. Some individual lanes have been isolated photographically for clarity. Only the relevant portion of each blot is shown, and apparent molecular mass scales are provided, in kilodaltons.

View larger version (45K): [in this window] [in a new window]   Figure 6. Variable proguanylin expression in small intestine is consistently detected with two independent antibodies. Duplicate samples from representative tissue extracts were analyzed in parallel with antibodies 2538 (a) and 6240 (b). Test samples were obtained from consecutive small bowel segments of a single animal, chosen to span a region of low expression (segments 11–12) bounded by regions of high expression (segments 10 and 13). The arrows and the arrowhead mark proguanylin-related peptides identified previously (see Fig. 3b). A shift toward lower molecular mass is evident in those samples in which total proguanylin expression is low.

Random proteolysis during sample processing would provide a nonbiological mechanism for producing unexpectedly low guanylin polypeptide expression in some regions but not others. We think this is unlikely, however, for several reasons: 1) we routinely add a broad spectrum cocktail of protease inhibitors to our homogenization and sample processing buffers; 2) we always confirm bulk protein integrity by Coomassie or Ponceau staining a portion of each sample after electrophoresis (Fig. 7, left panel); and 3) when we test samples for expression of both guanylin and uroguanylin, we fail to see significant uroguanylin degradation, even in samples in which guanylin levels are very low (Fig. 7, center and right panels).

View larger version (40K): [in this window] [in a new window]   Figure 7. Variable proguanylin expression in small intestine is not a consequence of general proteolysis during sample processing. Samples from two representative tissue extracts were analyzed in parallel with Coomassie stain (left), antiproguanylin antibody 2538 (center), and antiprouroguanylin antibody 6910 (right). As can be seen with antibody 2538, the test samples were chosen to compare a region of high proguanylin expression (segment 13) to an adjacent region of low proguanylin expression (segment 14). The Coomassie stain indicates general preservation of intact protein in both samples. Antibody 6910 reveals that prouroguanylin remains intact in both samples. Molecular mass scales are given in kilodaltons.

As described above, we investigated GC-C expression along the full length of the intestine of a single animal (Fig. 8, autoradiogram), using the same RNA samples that had been assayed for uroguanylin and guanylin expression. We quantitated GC-C mRNA levels at each point (Fig. 8, white symbols), including five standardized regions where the results were confirmed by obtaining measurements from multiple animals (Fig. 8, black symbols). Interestingly, the Northern blot results demonstrate that, in marked contrast to its peptide ligands, GC-C is expressed with comparable abundance in most regions of the small and large bowels. This observation is in good agreement with previous, less extensive regional Northern blot analyses of intestinal GC-C mRNA expression in the mouse (36) and rat (26, 29).

Note that GC-C levels in duodenum (segment 1) are somewhat lower than the levels observed in other regions. Low GC-C expression in duodenum appears to correlate with the generally lower levels of ligand found at this location (almost no guanylin and somewhat reduced uroguanylin levels). This could reflect a restricted need for GC-C-mediated signaling at this location, or it may be an artifact resulting from the high levels of proteases and nucleases typically found in duodenum.

Regional distribution of STa-binding sites
The relatively uniform presence of GC-C mRNA throughout the intestine does not necessarily imply that functional GC-C protein will be equally well represented in all segments. As an indirect index of GC-C protein expression, we used [¹²⁵I]STa to measure the numbers and affinities of ligand-binding sites present in four standardized regions of the intestine (segments 1, 3, 17, and 20). Representative Scatchard plots for each segment are presented in Fig. 9, and the average maximal binding capacity (B_max) and K_m values calculated from multiple assays are provided in Table 1. The Scatchard analyses demonstrate that each segment expresses primarily, and perhaps exclusively, a single class of binding site, and that the binding affinity of that site does not differ significantly from one segment to another. The relative B_max for each segment (reflecting total numbers of STa-binding sites per mg protein) is generally in good agreement with the relative GC-C mRNA level measured in that segment (compare to Fig. 8), although the slightly reduced levels of GC-C mRNA observed in duodenum are not matched by a comparable decrease in STa-binding sites. Overall, these data are consistent with the idea that GC-C protein expression levels are approximately proportional to GC-C mRNA levels. They also correlate well with previous regional STa binding assays performed on mouse intestinal tissue at a single, subsaturating dose of STa (36).

View larger version (26K): [in this window] [in a new window]   Figure 9. Scatchard analysis of STa binding to mucosal membranes isolated at defined positions along the rostrocaudal axis of the intestine. Four regions were tested, as indicated, and LIGAND software was used to establish that in each case the experimental data are best fit by a one-binding site model. Average binding constants were calculated for each region (on the basis of three independent assays) and are presented in Table 1.

Figure 10 shows intracellular and extracellular cGMP levels measured under control and STa-stimulated conditions at standardized points along the rat intestine. Initial studies (Fig. 10a) focused on intracellular cGMP levels in tissue treated with IBMX, a general phosphodiesterase inhibitor that blocks most of the known cGMP-selective phosphodiesterase isoforms (8). Under these conditions, STa responsiveness was found to be greatest in duodenum and colon and to drop off dramatically in other regions of the bowel. The most striking feature of these results is the barely detectable response to STa in distal portions of the small intestine. In terminal ileum, for example, the increment in cGMP synthesis induced by STa (stimulated level minus basal level) is approximately 6% of the increment induced in duodenum. The STa-dependent increment in cGMP synthesis has been calculated for selected intestinal segments and is plotted in Fig. 11 (gray symbols) to provide a direct comparison with GC-C mRNA levels (replotted from Fig. 8 using black symbols) and STa-binding site densities (B_max values obtained from Table 1, plotted as white symbols). The mismatch is evident; STa-dependent cyclase activity varies markedly from region to region, whereas GC-C mRNA and STa-binding sites remain relatively constant.

View larger version (27K): [in this window] [in a new window]   Figure 10. GC-C-mediated bioassay responses of isolated tissue explants obtained from defined regions along the rostrocaudal axis of the small and large bowel. a, Intracellular cGMP levels measured in the presence of IBMX under basal conditions (black symbols) and after STa application (white symbols). Most regions were tested several times, using tissue explants obtained from three to seven independent animals, and every test employed multiple determinations (three to nine individual pieces of tissue) to measure control and STa-stimulated cGMP levels at each rostrocaudal location. The asterisks denote intestinal segments where the difference between the stimulated level and the basal level achieves statistical significance by Student’s unpaired two-tailed t test (**, P < 0.001; *, P < 0.01). b, After exposure to zaprinast, distal ileum (right side) remains insensitive to STa, whereas duodenum (left side) remains sensitive. Intracellular cGMP (white bars) was measured in three animals (six individual pieces of tissue per animal); extracellular cGMP (hatched bars) was measured in the same three animals (a single determination per animal). **, P < 0.0001 relative to unstimulated controls; *, P < 0.05. c, Extracellular cGMP measured under basal conditions (black symbols) and after an STa challenge (white symbols) at various positions along the length of the bowel. At each position we determined the mean ± SEM using data obtained from three to seven independent animals. Results were pooled from measurements made in the presence of IBMX and in the presence of IBMX plus zaprinast. We observed no difference between these two treatments, except that the amount of secreted cGMP was generally larger in the presence of zaprinast than in its absence. In these experiments relatively small amounts of cGMP were secreted into large volumes of incubation medium, so that every measurement has a substantial error associated with it; thus, the only segment where the effect of STa has achieved a strict criterion for statistical significance (P < 0.02, by Student’s paired, two-tailed t test) is duodenum. However, it is apparent that secreted cGMP tends to parallel intracellular cGMP, a trend that we observed in every animal tested.

View larger version (18K): [in this window] [in a new window]   Figure 11. GC-C mRNA levels and STa-binding site densities match each other, but do not match STa-activated cGMP responses at specific points along the length of the intestine. Selected data have been replotted from Table 1 and Figs. 8 and 10a for direct comparison.

The apparent lack of STa sensitivity in terminal ileum could also potentially be explained by a local mechanism that inactivates STa, for example through a degradative enzyme or a clearance receptor. To evaluate this possibility we collected STa-containing test medium after exposing it for 30 min to various intestinal tissues and measured how much biological activity remained in the medium. We also collected medium that had been exposed to tissues in the absence of STa to evaluate whether endogenous GC-C ligands, such as uroguanylin or guanylin, might have been secreted by the tissue during the incubation. In each case, activity was assayed by applying the conditioned medium to GC-Ccontaining membranes isolated from T84 cells. No significant difference was observed between STa-containing test media that had been exposed to duodenum or to terminal ileum (Fig. 12). Thus, tissue-specific depletion of STa cannot account for the region to region differences observed in Fig. 10. Figure 12 also shows that tissue-specific secretion of endogenous agonists or antagonists (if any such exist) into the incubation medium is unlikely to account for the differences.

View larger version (23K): [in this window] [in a new window]   Figure 12. Differential responses to STa are not due to tissue-specific degradation of the toxin. We collected medium with (+) or without (-) 2 µg/ml STa after 30 min of preincubation with either duodenum or distal ileum (n = 6), and measured the effects of the recovered medium on the rate of cGMP synthesis by T84 membranes. Control experiments (no tissue preincubation) establish the basal rate of cGMP synthesis by T84 cell membranes and the response to standard untreated STa at 2 µg/ml. No statistically significant difference is observed between STa exposed to duodenum and STa exposed to distal ileum, nor are detectable amounts of biologically active material secreted into the medium by either tissue under these conditions.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have begun to investigate this latter possibility by analyzing the intestinal expression of GC-C, the only rGC identified to date that responds to either uroguanylin or guanylin. Although GC-C responds to both ligands, it is actually about 10 times more sensitive to uroguanylin than to guanylin (18), implying that it might serve preferentially as a uroguanylin receptor. Our Northern blot measurements, however, demonstrate that GC-C mRNA is expressed relatively uniformly in all regions of the intestine, rather than matching the distribution of one or the other peptide. In addition, binding assays reveal relatively constant numbers of STa-binding sites along the length of the bowel, suggesting that GC-C protein expression parallels GC-C mRNA expression.

Direct measurement of STa-activated cyclase activity, however, yields the surprising finding that ligand sensitivity is not constant along the length of the rat intestine. It is most pronounced at the proximal and distal extremes (corresponding to regions with high uroguanylin and guanylin levels, respectively) and is low or undetectable throughout the terminal regions of the small intestine. A strikingly similar rostrocaudal distribution of STa-dependent cGMP synthesis has previously been noted after luminal introduction of STa into ligated segments of rat intestine (26). This in vivo result not only provides support for our current in vitro observations, but also indicates that the observed segment to segment variability is not an artifact that occurs as a consequence of removing the intestine from the animal. In addition, comparable regional patterns of STa sensitivity have been reported along the length of the raccoon (39), opossum (39), duck (40), turkey (40), skink (41), and alligator (41) intestine. Although parallel mRNA measurements were not performed in these studies, it seems possible that a similar mismatch between RNA levels and enzyme activity may be a general property of GC-C expression in many animals.

Although we do not understand the basis for the RNA/cyclase activity mismatch in rats, we believe that it does not simply reflect region to region differences in GC-C protein levels, because this would not be consistent with the simplest interpretation of our STa binding studies or with available GC-C Western blot data (29, 42). Several alternative explanations should therefore be considered. 1) One hypothesis, which we favor, is that some form of posttranslational modification differentially regulates GC-C’s catalytic activity in specific regions of the intestine without otherwise affecting its RNA or protein expression levels or STa binding capacity. Possible mechanisms include dephosphorylation, which has been shown for other members of the rGC family to produce a desensitized receptor that still retains an active binding site but no longer synthesizes cGMP in response to ligand binding (43, 44), and limited proteolysis, which has been shown to release the intracellular (catalytic) portion of GC-C while leaving the membrane-anchored ligand-binding domain intact and able to interact with agonists (45). Indeed, a recently published regional Western blot analysis of GC-C expression within the rat intestine lends support to this latter possibility (29). In this study the total amount of material that was detected by the GC-C antibody appears fairly constant from segment to segment (consistent with the constant RNA expression levels observed in our current work), but the molecular masses of the immunoreactive species varied considerably. In fact, the largest species (150 kDa) was seen almost exclusively in those segments where we find biological activity (duodenum, cecum, and colon), whereas the smallest fragment (85 kDa) was seen most abundantly in those segments where activity is low (jejunum and ileum). A third species (140 kDa) was relatively uniformly distributed in all segments. It is tempting to speculate that the 150-kDa species may represent intact, fully functional GC-C, whereas the 140-kDa species may represent an inactive (perhaps incompletely glycosylated) form, and the 85-kDa species may represent a completely inactivated C-terminal proteolytic fragment. 2) A second hypothesis to explain the RNA/activity mismatch could be that one or more additional STa-sensitive rGCs (distinct from GC-C) is expressed in duodenum and distal colon, leading to elevated rates of STa-activated cGMP synthesis in these tissues. Several observations, however, argue against this idea. First, STa-dependent cGMP synthesis is essentially eliminated in the GC-C knockout mouse (with the caveat that not every region of the intestine has been carefully evaluated in these mice) (37). Second, multiple components were not observed in Scatchard analysis of our STa binding data. 3) A third mechanism that could generate differential toxin sensitivity would be a region-specific diffusion barrier (for example, an excessively thick mucus coating) that restricts access to GC-C in some segments but not in others. If so, this barrier must be consistently present in the terminal small bowel of every animal tested and must be difficult to detect by visual inspection, as we failed to observe any structural correlate in live or histologically prepared tissues. 4) A final possibility is that some regions of the intestine may degrade or eliminate cGMP more effectively than other regions. The conventional routes for elimination of intracellular cGMP are degradation by phosphodiesterase enzymes (8) and active export into the extracellular fluid compartment (46). In our studies we have employed a pair of phosphodiesterase inhibitors (IBMX and zaprinast) that, in combination, are effective at inhibiting all known cGMP-hydrolyzing enzymes (8). This suggests that differential phosphodiesterase activity is not a likely explanation for the observed region to region variability, although we cannot rule out the unexplored possibility that a novel IBMX- and zaprinast-resistant phosphodiesterase is preferentially expressed in ileum and terminal jejunum. With regard to cGMP export, direct measurement of extracellular cGMP levels after STa stimulation fails to reveal differential rates of cGMP export at any point in the intestine.

Interestingly, a mismatch between protein and RNA expression is also observed for guanylin, occurring most markedly in the small intestine. One possible contributor to such a mismatch might be recently documented circadian variations in guanylin expression (29), especially if RNA and protein levels were to vary out of phase with one another. Arguing against this idea, however, is the observation that circadian variability is actually much more striking for uroguanylin than for guanylin (29), and yet uroguanylin RNA and protein expression follow each other rather closely. Furthermore, we killed all of our test animals within a relatively narrow time window, thereby limiting circadian fluctuations. An alternate possibility is that the proguanylin molecule is particularly susceptible to proteolysis by some as yet unidentified protease that is nonuniformly distributed within the small bowel. If so, then this proteolytic activity must be highly selective for proguanylin, arguing that it is likely to be a processing event that occurs as a natural step in the proguanylin life cycle. A final possibility is that variability could reflect local differences in guanylin secretion. For example, low steady state levels could be characteristic of regions with high secretory activity followed by rapid degradation or loss of the secreted materials, or high steady state levels could be characteristic of regions with high secretory activity followed by trapping and accumulation of the secreted products within the mucous layer that coats the apical surface of the epithelium. Whatever the explanation, the segment to segment variability suggests that some aspect of guanylin metabolism is highly dynamic, leading to conspicuous fluctuations in the protein to RNA ratio along the length of the small bowel. This contrasts markedly with the relatively constant protein to RNA ratio observed for uroguanylin in the same tissue segments.

What can we conclude about the function(s) of uroguanylin and guanylin from an analysis of their distributions at the cell and tissue levels? In the case of guanylin, our current observations reinforce and extend the proposal that guanylin-mediated fluid and electrolyte transport plays a role in the hydration of mucin (required for the proper formation of mucus) (24). This hypothesis is consistent with several observations. First, mucin and guanylin are both produced by goblet cells (27, 28, 47) and are both vectorially secreted into the lumen (47, 48), a compartment that is monitored by apically targeted receptors such as GC-C. Second, successful hydration of secreted mucin requires luminal sodium ions and water (49), both of which are provided when GC-C is activated (50). Third, guanylin expression is prominent throughout the colon, a tissue that is otherwise designed for efficient withdrawal of water and electrolytes from the lumen. Interestingly, despite the paucity of these essential components, the production of well hydrated mucus actually becomes increasingly important for lubrication and protection as the luminal contents of the colon solidify. One mechanism that could allow successful production of mucus and, at the same time, dehydration of the lumen would be cosecretion of guanylin and mucin. Spatial and temporal linkage between these molecules would provide a restricted local mechanism capable of supplying calibrated amounts of fluid only on demand. Note that GC-C activity is also high in all regions of colon, providing a route by which guanylin can actively control ion transport in this tissue compartment.

Our observations about uroguanylin are also relevant to previously formulated hypotheses concerning its physiological function(s) (51, 52, 53). Uroguanylin is a small bowel peptide, made by enterochromaffin (EC) cells (22, 24). Intestinal EC cells form an epithelial sensory system that responds to noxious luminal stimuli (54). For example, when gastric acid enters the intestine, it triggers the secretion of alkali by a reflex mediated in part by EC cells (54, 55). Unlike many other types of enteroendocrine cell, whose secretory products are released exclusively basolaterally, EC cells discharge a portion of their secretory products into the lumen of the intestine (55, 56). Thus, when activated by acid-related stimuli, small bowel EC cells are very likely capable of delivering uroguanylin to luminally oriented GC-C, thereby activating CFTR via the pathway outlined above. Interestingly, this would contribute directly to the alkaline reflex, because CFTR, along with its widely recognized permeability to chloride ions, also has high permeability to bicarbonate ions. Indeed, recent experiments with isolated duodenal explants have shown that apical application of GC-C ligands produces marked CFTR-dependent alkalinization of the apical fluid compartment (52, 53). In this context, restriction of uroguanylin primarily to small bowel makes sense, because the intestinal phase of digestion (with its requirement for a slightly alkaline pH) occurs within this tissue compartment. Further, it is intriguing to note that GC-C activity appears to be down-regulated in distal regions of the small intestine; this could provide a reserve population of receptors for rapid up-regulation if gastric acid delivery becomes excessive.

It is also noteworthy that small amounts of uroguanylin are produced by enterochromaffin-like (ECL) cells in the stomach (57). ECL cells play an important role in gastric acid production by releasing histamine, a locally acting hormone that triggers acid secretion by parietal cells. If ECL cells also release uroguanylin into the gastric lumen, then the uroguanylin would flow into the intestine along with the acidic contents of the stomach and provide an additional stimulus for alkali secretion. In addition, as GC-C is expressed at low levels within the stomach itself (and is particularly abundant in a subpopulation of nonparietal cells within the gastric glands) (26), it also appears likely that uroguanylin plays an as yet undefined local role in gastric physiology.

In addition to the foregoing analysis of uroguanylin’s potential functions within the digestive system, there are convincing reasons to consider targets of uroguanylin action that might occur outside of the intestine. 1) Although EC and ECL cells can deliver secretory products to the lumen, it is generally acknowledged that they secrete preferentially into the basolateral fluid compartment, making their secretory products available to the blood. 2) Consistent with this, uroguanylin has been detected in both plasma (58, 59, 60) and urine (18, 22). Plasma levels of uroguanylin increase further in patients with renal disease or in normal subjects after prolonged periods of elevated salt intake (61). 3) High affinity STa-binding sites are found outside of the intestine, most prominently in kidney (62). 4) Natriuresis, kaliuresis, and diuresis are observed when uroguanylin is introduced into the renal artery of an isolated, perfused kidney (63). Taken together, these observations have led to the hypothesis that uroguanylin is the hormonal mediator of a gut/kidney endocrine axis that serves to couple increased salt intake (detected by the intestine) to increased salt excretion (accomplished by the kidney) (51). Because dietary salt is completely absorbed within the small intestine, this hypothesis is consistent with our observation that high levels of uroguanylin expression are observed only in this part of the bowel.

Whether uroguanylin actually participates in salt and/or pH homeostasis, as outlined above, remains to be established by rigorous experimental analysis. In particular, confirmation that uroguanylin secretion is triggered by salt or acid entry into the small intestine remains a critical experimental objective. Similarly, the postulated role for guanylin in mucus hydration has not yet been validated by direct evidence that guanylin and mucin are cosecreted. Thus, at present, our physiological speculations are not overly constrained by an excess of experimental data. However, even at this stage, it is apparent that whatever physiological roles are ultimately established for this new peptide family and its guanylate cyclase-coupled receptor, they will need to be consistent with the elaborate and well defined tissue distribution that is described by our current studies.

	Acknowledgments

 
We thank Josh Knowles, Zhiping Li, Ashley Perkins, and Dominique Goyeau for generously providing some of the Northern blot data presented here; Karen Jones for help with Western blots; Sharon Milgram for the gift of antibody 6240; Julia Vorobiov and the University of North Carolina Center for Gastrointestinal Biology and Disease (CGIBD) Immunotechnologies Core for iodinating cGMP and for producing antibodies 6910, 6912, and 2538; Yiwei Rong and the CGIBD Advanced Culture Technologies Core for T84 cell culture; and Kathleen Dunlap and Peter Mohler for providing helpful comments on the manuscript.

	Footnotes

 
¹ This work was supported by grants from the American Heart Association (AHA 5–40152/0–110-4224) and the NSF (NSF 5–37411/0–110-4224). The University of North Carolina Center for Gastrointestinal Biology and Disease was supported by NIH Center Grant DK-34987.

² Current address: Southwestern Medical Center, Dallas, Texas 75235.

Received January 21, 2000.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Fesenko EE, Kolesnikov SS, Lyubarsky SL 1985 Induction by cyclic GMP of cationic conductance in plasma membrane of retinal rod outer segment. Nature 313:310–313[CrossRef][Medline]
2. Lincoln TM 1989 Cyclic GMP and mechanisms of vasodilation. Pharmacol Ther 41:479–502[CrossRef][Medline]
3. Geary CA, Davis CW, Paradiso AM, Boucher RC 1995 Role of CNP in human airways: cGMP-mediated stimulation of ciliary beat frequency. Am J Physiol 268:L1021–L1028
4. Forte LR, Hamra FK 1996 Guanylin and uroguanylin: intestinal peptide hormones that regulate epithelial transport. News Physiol Sci 11:17–24[Abstract/Free Full Text]
5. Boulton CL, Southam E, Garthwaite J 1995 Nitric oxide-dependent long-term potentiation is blocked by a specific inhibitor of soluble guanylyl cyclase. Neuroscience 69:699–703[CrossRef][Medline]
6. Foster DC, Wedel BJ, Robinson SW, Garbers DL 1999 Mechanisms of regulation and functions of guanylyl cyclases. Rev Physiol Biochem Pharmacol 135:1–39[Medline]
7. Murad F 1998 Nitric oxide signaling: would you believe that a simple free radical could be a second messenger, autacoid, paracrine substance, neurotransmitter, and hormone? Recent Prog Horm Res 53:43–60
8. Dousa TP 1999 Cyclic-3',5'-nucleotide phosphodiesterase isozymes in cell biology and pathophysiology of the kidney. Kidney Int 55:29–62[CrossRef][Medline]
9. Schulz S, Green CK, Yuen PST, Garbers DL 1990 Guanylyl cyclase is a heat-stable enterotoxin receptor. Cell 63:941–948[CrossRef]