ELSEVIER
`
`REGULATORY
`
`PEPTIDES
`
`2 75 99,
`
`

`

`46
`
`Z Li et al
`
`/ Regulatory Peptides
`
`–
`68 (1997) 45 56
`
`tract, by analogy to the way that the other membrane-
`associated guanylate cyclases (GCA and GCB) serve as
`receptors for peptide ligands produced by tissues like the
`heart and brain [7]. Recently, a candidate intestinal pep-
`tide, called guanylin, was purified from jejunal extracts [8].
`Guanylin is believed to be a natural
`ligand for GCC
`because it (a) elevates intracellular cyclic GMP levels in
`GCC-expressing cells [8], (b) competes with STa for
`binding to GCC [8–10], and (c) stimulates the secretion of
`chloride by intestinal epithelial cells [9,11–15]. One key
`feature shared by guanylin and STa is a set of four
`conserved cysteines connected by specific disulfide bonds;
`this provides the secondary structure required for bio-
`logical activity [8,16].
`The mRNA encoding GCC is strongly expressed in the
`intestine [6,17], exclusively in epithelial cells [18,19]. No
`other tissue tested displays significant
`levels. However,
`binding studies indicate that receptors for guanylin are
`present
`in the kidney [20,21]. Furthermore, STa and
`guanylin can stimulate sodium and potassium excretion by
`the isolated, perfused kidney [22,23] and can elevate cyclic
`GMP in organ cultured kidney slices [21,24]. These
`observations suggest
`that guanylin, or a guanylin-like
`peptide, may play a role in regulating kidney function. One
`possibility is that guanylin is produced locally in the
`kidney. Indeed, low levels of guanylin probe hybridization
`have been reported on Northern blots of mRNA isolated
`from rat kidney [25]. However, this observation was not
`confirmed in comparable studies with mouse [26] or
`human [10] kidney, and, furthermore, only small amounts
`of guanylin-like bioactivity are present
`in rat kidney
`extracts [8,27].
`A second possibility is that guanylin or a related peptide
`is delivered to the kidney from some other source. In an
`effort to identify such a peptide, Hamra et al. [28] and Kita
`et al. [29] analyzed opossum and human urine for the
`presence of guanylin-like peptides. They found small
`amounts of guanylin, and larger amounts of a second,
`structurally-similar
`peptide,
`which
`they
`named
`uroguanylin. Uroguanylin can bind to the STa/guanylin
`receptor: its EC for activating cyclic GMP synthesis in
`50
`GCC-expressing cells is intermediate between that of
`guanylin and STa [28,29], and it can competitively dis-
`125
`place
`I-STa binding [28,29]. Uroguanylin is similar in
`amino acid sequence to guanylin and STa, and it retains
`their characteristic disulfide bond structure. Thus, guanylin
`and uroguanylin define a family of naturally occurring
`peptides that are structurally and functionally related.
`These studies leave open the question of what tissue
`serves as the biological source of uroguanylin found in the
`urine. In our current study, we show that small intestine is
`potentially one such source: (a) Rat duodenal extracts
`contain a peptide that displays guanylin-like bioactivity,
`but is chromatographically distinct from guanylin. (b) This
`second peptide
`is much more
`closely
`related to
`uroguanylin (80% amino acid identity with human and
`
`opossum uroguanylin) than to guanylin (50% amino acid
`identity with rat guanylin). (c) Cloning of a rat uroguanylin
`cDNA confirms that this second peptide represents the rat
`isoform of uroguanylin. (d) Northern blots show relatively
`selective expression of uroguanylin mRNA in proximal
`small intestine.
`
`2. Experimental procedures
`
`2.1.
`
`Tissue and Extract Preparation
`
`"
`
`!
`
`Tissues were removed from Sprague–Dawley rats (250–
`275 g) under urethane anesthesia (1.6 g urethane/kg
`administered via i.p.
`injection). For RNA isolation,
`the
`whole intestinal tract was removed and put in ice cold
`Ringer’s-glucose (130 mM Na , 120 mM Cl
`, 25 mM
`2!
`2!
`HCO , 1.2 mM Mg , 1.2 mM Ca
`, 2.4 mM K HPO ,
`"
`3
`2
`4
`0.4 mM KH PO , 10 mM glucose). Tissues were then
`2
`4
`isolated as rapidly as possible, frozen on dry ice, and
`stored at " 80&C until used for RNA purification.
`For the preparation of peptide-containing extracts for
`Western blots,
`the mucosal
`layer of each tissue was
`stripped free of the muscle layers, and homogenized in a
`buffer containing protease inhibitors (25 mM HEPES
`pH $ 7.4, 1 mM phenylmethylsulfonylfluoride, 10 )M
`trans-epoxysuccinyl-L-leucylamido(4-guanidino)butane, 10
`)g/ml aprotinin, 10 )g/ml
`leupeptin, 1 mM ben-
`zamidine). After homogenization, the extracts were cen-
`trifuged at 10 000 # g for 20 min and insoluble material
`was discarded. The protein concentration was determined
`by the bicinchoninic acid method (BCA kit obtained from
`Pierce).
`For the preparation of peptide-containing extracts for
`HPLC fractionation, rat intestines were frozen on dry ice,
`and kept at " 80&C until used. After thawing, the intes-
`tines were divided into regions corresponding to duodenum
`and colon. Duodenal tissue was taken as the 4 to 5 cm
`segment of intestine immediately distal to the stomach.
`Colon was taken to include both proximal and distal
`segments of the large bowel (posterior to the caecum and
`anterior to the sigmoid colon). Tissue pieces were split
`lengthwise and rinsed with normal saline. The tissue was
`minced in 10 volumes 1 M acetic acid, placed in a boiling
`water bath for 5 min, and then homogenized. Boiled
`extracts were centrifuged at 4&C for 20 min at 230 000 # g.
`The resulting supernatant fractions were filtered through
`Whatman No. 2 paper and applied to Waters C Sep-Pak
`18
`cartridges. Unbound and weakly bound material was
`washed through with a solution of 10% acetonitrile–0.1%
`trifluoroacetic acid (TFA) in water. Tightly bound material
`was eluted with 60% acetonitrile–0.1% TFA. The eluted
`material was dried under vacuum and reconstituted either
`in bioassay medium (Hanks’ buffered salt solution con-
`taining 1 mM IBMX) for bioassay on T84 cells, or in 10%
`acetonitrile/0.1% TFA for HPLC analysis (see below).
`
`

`

`Z Li et al
`
`/ Regulatory Peptides
`
`–
`68 (1997) 45 56
`
`47
`
`For measuring region-specific responses to STa, seg-
`ments of tissue approximately 4 cm in length were placed
`in a dissecting pan filled with 37&C Ringer’s-glucose
`bubbled continuously with 95% O –5% CO Tubular
`2
`2.
`sections were cut longitudinally, exposing a flat luminal
`surface, and luminal contents were discarded. Individual
`pieces of tissue (0.5 cm # 0.5 cm) were excised, placed in
`shell vials containing standard Ringer’s-glucose solution at
`37&C, and exposed to test solutions (see below).
`
`2.2.
`
`Northern blots
`
`Selected regions of the uroguanylin, guanylin, and GCC
`genes were amplified by PCR and subcloned into plasmid
`vectors (pBS, Stratagene), as described below and in a
`previous publication [18]. The cDNA inserts were isolated
`and used as templates for the synthesis of randomly-
`32
`primed [ P]-labeled cDNA probes (DECAprime II kit,
`Ambion).
`Total RNA was isolated by standard techniques [30],
`fractionated on a 1% agarose formaldehyde gel (2.2 M
`formaldehyde), and transferred to a nylon membrane
`(ICN). Membranes were treated with 10 ml prehybridiza-
`tion solution (50% formamide, 5X SSPE, 5X Denhardt’s,
`0.25mg/ml sperm DNA, 0.5% SDS) for 3 h and then
`hybridized in 10 ml Northern blot hybridization solution
`(50% formamide, 5X SSPE, 1X Denhardt’s, 0.1 mg/ml
`6
`sperm DNA, 0.1% SDS, 10% dextran sulfate with 10
`32
`cpm/ml of each [ P]-labeled probe) at 42&C for 24 h
`(SSPE $ 150 mM NaCl, 11.5 mM NaH PO , 1 mM
`2
`4
`EDTA, pH 7.4). Membranes were then washed twice (15
`min each) at room temperature in 2X SSC with 0.1% SDS
`(SSC $ 150 mM NaCl, 15 mM Na citrate, pH 7), followed
`by two 30 min washes at 55&C in 0.1X SSC with 0.1%
`SDS.
`
`2.3.
`
`Western blots
`
`Samples of tissue extracts or HPLC fractions were dried
`under vacuum, boiled in electrophoresis sample buffer, and
`fractionated on 15% SDS–polyacrylamide gels made
`according to standard procedures [31] except
`for
`the
`composition of the electrode buffers (upper $ 100 mM
`Tris-OH, 100mM tricine, 0.1% SDS, pH 8.3;
`lower $
`200mM Tris-Cl, pH 8.9). After electrophoresis, samples
`were transferred to 0.1 )m nitrocellulose membranes
`(Schleicher and Schuell) using a TE 22 transphor ap-
`paratus (Hoefer Scientific). The membranes were blocked
`with 3% BSA, washed, and incubated for 1 h at room
`temperature with a 1:500 dilution of antiserum 2538, a
`polyclonal antiserum that was raised against a 14 amino
`acid synthetic peptide whose sequence appears near the
`amino terminus of the rat guanylin prohormone [32]. The
`membranes were then washed and incubated with a
`secondary antibody (horseradish peroxidase-conjugated
`sheep-antirabbit IgG diluted 1:10 000, Boehringer-Mann-
`
`heim) for 60 min at room temperature. After an additional
`wash,
`the membranes were treated with chemilumines-
`cence reagent as specified by the manufacturer (Boeh-
`ringer-Mannheim), and exposed to Kodak XAR-5 film.
`
`2.4.
`
`Bioassay
`
`The studies described below have made use of two
`different bioassay procedures. The first procedure was used
`to evaluate the relative ability of colon and duodenum to
`respond to STa (Fig. 1b). Shell vials containing 1 ml
`standard Ringer’s-glucose solution with 0.5 mM IBMX
`were placed in a 37&C water bath. Vials were continuously
`bubbled with 95% O –5% CO throughout the experi-
`2
`2
`ment. During the period of temperature and gas equilibra-
`tion, 100 units/ml STa (Sigma) was added to the appro-
`priate vials. Pieces of tissue from each region of the gut
`were then placed in the vials for 30 min. Reactions were
`stopped by removing the tissues from the vials and quick
`freezing them on a metal plate resting on a bed of dry ice.
`The frozen tissue was then homogenized in 6% trichloro-
`acetic acid (TCA) and centrifuged to separate TCA-insolu-
`ble protein from TCA-soluble cyclic GMP. The protein
`pellet was dissolved by heating (37&C) in 1 M NaOH.
`Cyclic GMP was quantitated by RIA [33,34] and protein
`was determined using a dye-binding assay (Bio-Rad
`Bradford Assay) with BSA as standard. Results are
`reported as pmol cyclic GMP/mg protein.
`The second procedure was used to evaluate HPLC
`fractions for guanylin-like activity (Figs. 2–5). This bioas-
`say, based on the method of Currie et al. [8], employs a
`reporter cell line (T84 cells) to detect the presence of GCC
`ligands, as manifested by an increase in intracellular cyclic
`GMP levels. T84 cells are derived from a human colon
`carcinoma, and retain many properties of crypt epithelium,
`including expression of high levels of GCC [35].
`In
`addition, because T84 cells express very low levels of the
`other known membrane cyclases [8,36], they provide a
`relatively specific bioassay system for ligands that target
`GCC. The cells were grown to confluency in 12- or
`24-well plastic trays. Cells were then washed two times
`with HBSS and incubated for 10 min at 37&C in HBSS
`containing 1 mM IBMX. HPLC fractions were reconsti-
`tuted in bioassay medium, then applied to the cells for an
`additional 30 min. The reaction was stopped by removing
`the incubation solution and replacing it with 0.5 ml ice
`cold 6% TCA. Cells were scraped and transferred to plastic
`microcentrifuge tubes and the wells were rinsed with an
`additional 0.5 ml ice cold 6% TCA. Protein and cyclic
`GMP content were determined as above. Because the
`protein content per well
`is quite uniform for T84 cell
`cultures within a single plating, experimental results are
`reported as pmol cyclic GMP/well rather than pmol cyclic
`GMP/mg protein.
`Activation of HPLC fractions: In our initial studies, the
`biological activity of our HPLC fractions was quite low.
`
`

`

`68 (1997) 45 56
`
`a
`
`100
`-8-- 75
`N2 2 50
`
`mo
`
`25
`
`0
`
`b
`
`_J
`
`15
`10
`20
`fraction number
`
`25
`
`0I
`
`1.0
`
`0.8
`
`0.6 g
`
`0.4 <
`
`0.2
`
`0.0
`
`4 iVe blot Tr
`30
`
`1 . 0
`0.8
`0.6 00
`
`0.4 a
`
`0.2
`
`0.0
`
`0
`
`100
`
`co
`
`0
`
`C
`
`a
`
`100
`
` 75
`z
`E
`.5 50
`
`E
`8u
`0
`
`25
`0
`
`
`6-
`
` 6
`
`1'5'
`10
`20
`fraction number
`
`-11 1Vebtr
`30
`
`25
`
`a S D J
`
`I C
`
`
`IP
`
`kb
`9.5
`7.5
`—
`_ 4.4
`
`-
`
`2.4
`
`eed- 1.35
`
`—
`
`0.78
`
`15
`
`C
`
`0
`STa:
`
`-
`
`kDa
`
`- 29
`
`.4 - 18
`
`- 14.3
`
`35-
`
`30-
`25-
`20-
`15-
`10-
`
`5-
`
`ignal density (arbitrary units)
`
`- 6.2
`- 2.3 s
`
`100
`

`

`control
`(n = 3)
`
`figure 2a
`fraction 21
`(n = 2)
`
`figure 2b
`fraction 23/24
`(n = 6)
`
`guanylin
`(n = 3) =IN
`
`uroguanylin
`(n = 2)
`
`STa
`(n = 2)
`
`figure 4
`fraction 16
`(n = 1)
`
`0
`
`0.5
`
`1
`
`20 40 60 80
`
`cGMP pmol/well
`
`2.5.
`
`2.5.1.
`
`1
`
`68 (1997) 45 56
`
`100
`
`1.0
`
`as
`>,O 0.5-
`‘i 2
`
`8 k 0.0
`
`0.04
`
`0.03
`
`0.02 co
`
`0.01
`0.00
`
`20
`15
`10
`fraction number
`
`25
`
`30
`
`0.02
`
`0.01
`
`2--
`
`0.00
`
`100
`
`120
`
`•
`
`.
`
`.
`
`.
`
`.
`
`.
`
`80
`60
`40
`fraction number
`
`100
`
`ao
`
`0
`
`2.0
`
`3
`g d 1.5-
`05
`22 1.0-
`
`E 0.5
`
`0
`
`r
`
`.
`
`.
`
`.
`
`.
`
`20
`
`0
`
`2.5.2.
`
`2
`
`

`

`50
`
`Z Li et al
`
`/ Regulatory Peptides
`
`–
`68 (1997) 45 56
`
`Step
`3
`2.5.3.
`Active fractions from step 2 were pooled and applied to
`a C
`analytical column (Vydac), pre-equilibrated with
`18
`10% buffer B (buffer A, 10 mM ammonium acetate pH
`6.2; buffer B, 99% acetonitrile ! 10 mM ammonium
`acetate pH 6.2). The column was eluted isocratically for 10
`min with 10% buffer B, followed by a linear gradient to
`40% buffer B over 160 min, and a final 10 min elution at
`100% buffer B. Fractions (1.5 ml) were collected at a flow
`rate of 1 ml/min, and analyzed by bioassay, as shown in
`Fig. 5.
`Synthetic 15 amino acid rat guanylin and opossum
`uroguanylin (used as HPLC standards) were generously
`provided by Drs. Ding Chang (Peninsula Labs) and
`Leonard Forte (University of Missouri), respectively.
`
`2.6.
`
`Peptide characterization
`
`Sequencing was performed on a sample of peptide
`adsorbed to a 0.22 micron PVDF membrane, using a
`Perkin Elmer/ABI model 491 sequencer with on-line PTH
`amino acid detection.
`
`2.7.
`
`cDNA analysis
`
`Rat duodenal RNA was prepared as described previous-
`ly [18], and reverse transcribed with SuperScript II using
`an oligo dT primer (Gibco BRL). The resulting cDNA was
`then subjected to PCR, using degenerate primers designed
`from areas of high homology in published sequences of
`human and opossum uroguanylin [37,38]. The sense
`primer, TACATCCAGTA(CT)(GC)A(AG)GCCTTCC, and
`antisense primer, GCAGCC(GT)GTACA(GC)GC(AC)-
`ACGTT, correspond, respectively, to base pairs 104–124
`and 334–354 of the human transcript [38]. PCR was
`performed for 40 cycles (denaturation for 1 min at 94&C,
`annealing for 1 min at 55&C, and extension at 72&C for 2
`min) followed by a final extension for 10 min at 72& C
`using Taq polymerase (Boehringer Mannheim). A 250 bp
`product was amplified, as expected. This PCR product was
`subcloned into the pBluescript II SK vector using the
`T/A cloning method [39]. DNA was sequenced at
`the
`UNC-CH Automated DNA Sequencing Facility on a
`Model 373A DNA Sequencer (Applied Biosystems) using
`TM
`the Taq DyeDeoxy
`Terminator Cycle Sequencing Kit
`(Applied Biosystems). Sequence analysis revealed a high
`degree of homology at the nucleotide level to the human
`and opossum forms of uroguanylin (72 and 60%, respec-
`tively), confirming that we had amplified the appropriate
`target sequence.
`To determine the full-length uroguanylin cDNA se-
`quence, we used the PCR product to produce a random-
`primed probe for screening a rat duodenal cDNA library
`constructed in (GT-11 [40] (a gift from Dr. Andrew
`Leiter). Phage plaques were adsorbed onto nitrocellulose
`filters, hybridized with the probe in 50% formamide, 0.8 M
`
`"
`
`NaCl, 20 mM PIPES pH 6.5, 0.5% SDS and 100 )g/ml
`salmon sperm DNA at 42&C, and washed in 0.1X SSC/
`0.1% SDS at 55&C. Positive plaques were purified by
`sequential low-density plating and bacteriophage DNA was
`isolated with a Qiagen Lambda kit (Qiagen). The cDNA
`insert was excised with EcoRI and subcloned into the
`pBluescript II SK vector for sequencing, as described
`above.
`
`"
`
`3. Results
`
`Guanylin and its receptor are not distributed in
`3.1.
`parallel in the GI tract
`
`Previous studies have shown that guanylin mRNA is
`expressed in a rostrocaudal gradient, ranging from quite
`low in duodenum to quite high in colon [9,32,41,42]. In
`contrast, as shown in Fig. 1a, the mRNA encoding the
`guanylin receptor
`(GCC)
`is expressed at high levels
`throughout the GI tract. These data reveal a nonparallel
`distribution of guanylin and its receptor, with the mismatch
`particularly evident in duodenum.
`in
`To verify that
`the high level of GCC transcript
`duodenum corresponds to a high level of
`functional
`receptor, we applied a sub-saturating dose of a GCC-
`specific ligand (STa) to excised pieces of tissue in organ
`culture, and measured intracellular levels of cyclic GMP
`after 30 min of exposure (a time at which the response is
`still proceeding linearly). This experiment was performed
`in the presence of a phosphodiesterase inhibitor, isobutyl
`methylxanthine (IBMX), to minimize the effects of phos-
`phodiesterase enzymes on cyclic GMP metabolism. The
`agonist-dependent
`increase in cyclic GMP levels (the
`amount of cyclic GMP in STa-stimulated tissue minus the
`amount in unstimulated control tissue) is slightly greater in
`duodenum than it is in colon (Fig. 1b), indicating that
`duodenum does express substantial
`levels of functional
`receptor.
`To verify that the low level of guanylin transcript in
`duodenum corresponds to a low level of guanylin prop-
`eptide, we performed Western blots on extracts of
`duodenum and colon, using an antibody that recognizes the
`N-terminus of the propeptide [32]. Extracts were prepared
`from four separate animals;
`in each case the level of
`proguanylin in duodenum, as measured densitometrically,
`was less than 11% of the level in colon (mean $ 8.7%
`%1.0% (SEM)) (Fig. 1c).
`
`HPLC analysis of guanylin-like peptides in
`3.2.
`intestinal extracts
`
`The ligand/receptor mismatch described above suggests
`the possibility that duodenum might produce a ligand that
`resembles guanylin in its receptor specificity, but
`is
`biochemically distinct. In order to compare guanylin-like
`
`

`

`Z Li et al
`
`/ Regulatory Peptides
`
`–
`68 (1997) 45 56
`
`51
`
`peptides expressed in duodenum to those expressed in
`colon, we prepared aqueous extracts of each tissue, frac-
`tionated the extracts by HPLC, and assayed an aliquot of
`each fraction for
`its ability to stimulate cyclic GMP
`synthesis in cultured T84 cells, as described in the Section
`2. In parallel with these bioassays, we performed Western
`blots on a sample of each HPLC fraction, using the
`antibody described in Fig. 1c. For convenience, only the
`region of the blot containing immunoreactive proguanylin
`is shown in the figures below.
`Colonic extracts contain both guanylin and proguanylin.
`Fig. 2a illustrates the HPLC analysis of a colonic extract.
`The Western blot inset at the bottom of the figure shows
`that proguanylin elutes in fraction 21. When the fractions
`are bioassayed, a peak of cyclic GMP-promoting activity is
`also seen in fraction 21. In the course of characterizing this
`peak, we noticed that its biological activity increases with
`time if it is incubated at 37&C (Fig. 3). Such an increase is
`not
`observed when
`synthetic
`guanylin,
`synthetic
`uroguanylin, or commercially-purified STa are incubated
`under similar conditions (Fig. 3). We do not yet know the
`mechanism by which this time-dependent enhancement of
`activity occurs; however, as discussed below, we consider
`it likely that proguanylin, which is biologically relatively
`inactive [10,25,42,43], has co-eluted with a protease
`capable of converting it to a smaller, more active peptide.
`Hamra et al. have previously demonstrated that specific
`proteases enhance the activity of proguanylin [44].
`Duodenal extracts contain biologically-active material
`that is distinct from both guanylin and proguanylin. When
`duodenal extracts are analyzed by HPLC, proguanylin-like
`immunoreactivity is again observed in fraction 21 (West-
`ern blot inset, Fig. 2b), though, as expected, the amount is
`much less than can be seen in comparable colonic extracts
`(Western blot inset, Fig. 2a). The T84 cell assay confirms
`that this duodenal proguanylin is associated with a small
`peak of biological activity, which, as above, can be
`enhanced by preincubation at 37&C. However, the most
`conspicuous aspect of the chromatogram is the presence of
`a second, much larger peak of activity in fractions 23 and
`24. This material differs from proguanylin in two ways: it
`is retained more tightly by the column, and it fails to react
`with our proguanylin-specific antibody. In addition,
`the
`retention time of the material is significantly different from
`that of synthetic rat guanylin (fraction 13) or synthetic
`opossum uroguanylin (fraction 15). The novel duodenal
`peak does, however, share one property with proguanylin:
`its activity is also enhanced by preincubation at 37&C (Fig.
`3). If an intestinal protease is indeed responsible for this
`activation phenomenon, then the data suggest that such a
`protease is likely to come off the HPLC column in a broad
`peak that overlaps both proguanylin (fraction 21) and the
`material in fractions 23–24.
`is a
`In order to test whether the duodenal material
`peptide, we incubated it with a mixture of exogenously-
`added proteases (trypsin, chymotrypsin, elastase, amino-
`
`peptidase, and S aureus V8 protease). After a 2 h
`incubation, its biological activity was completely destroyed
`(data not shown). This, in conjunction with the purification
`and sequencing studies described below, confirms the
`peptide nature of the active material in fractions 23–24.
`
`Preincubation alters the chromatographic properties
`3.3.
`of the duodenal peptide
`
`If a proteolytic mechanism is responsible for the activa-
`tion of the duodenal peptide, then the size of the molecule
`should be altered once it has been activated, and this
`should be reflected by a change in its chromatographic
`properties. To test this prediction, we partially purified the
`duodenal peptide on our standard reverse phase HPLC
`column (as in Fig. 2b), allowed it to preincubate at 37&C,
`rechromatographed it on the same HPLC column, and
`bioassayed the resulting fractions. Although biological
`activity is still quite evident (Fig. 4), it now elutes from the
`column much earlier (fraction 16), consistent with cleav-
`age to a smaller (less hydrophobic) peptide. Furthermore,
`the material in fraction 16 has become fully active: no
`enhancement of the T84 cell response can be induced by
`additional preincubation at 37&C (Fig. 3).
`
`3.4.
`
`Purification of the duodenal peptide
`
`In order to achieve further purification, the material in
`fraction 16 (Fig. 4) was again applied to a C-18 column,
`but now subjected to a new set of elution parameters using
`a different ion pairing reagent and a different pH (see
`Section 2). Activity was recovered as a single, sharp peak
`that aligns precisely with a major peak of UV absorbance,
`and appears well separated from other contaminants (Fig.
`5).
`This material was submitted for amino acid sequencing,
`revealing the presence of approximately 15–18 pmol of the
`peptide shown at the top of Fig. 6a. During the sequencing
`reaction, no amino acid could be identified in cycles 7, 10,
`15, and 18, indicating that the amino acid at each of those
`positions is most likely cysteine. This is consistent with the
`observation that the blank cycles align exactly with a set of
`four cysteines that are absolutely required for biological
`activity in all known guanylin-like peptides (guanylin,
`uroguanylin, and STa-see Fig. 6a). Also,
`in the first
`sequencing cycle, both threonine and glutamate were
`found in approximately equal abundance and thus the
`residue at this position is ambiguous. The sequence of the
`purified peptide suggests that it may be the rat homolog of
`uroguanylin.
`
`Sequence and tissue distribution of the rat
`3.5.
`uroguanylin transcript
`
`In order to confirm and extend our peptide sequence
`analysis, we used PCR to generate a 250 bp nucleotide
`
`

`

`68 (1997) 45 56
`
`a
`
`C
`T/E I ATCI F
`C
`O E L,'
`C
`N
`D
`PN T
`C
`SH T
`C
`P GT
`C
`P N T
`C
`NTFYC
`C
`
`E
`E
`E
`E
`E
`E
`E
`
`L C
`L C
`L C
`I C
`I C
`C
`C
`L C
`
`V
`I NVACT
`I NVAC T
`v N V A C T
`AYAACT
`AFAAC t.
`AYAACT
`A Y A A C T
`C
`P litie A
`
`GC
`GC
`G C•L
`G C
`G C,
`GC
`GC
`,GCY
`
`rat uroguanyl in
`opossum uroguanyl in
`human uroguanyl in
`rat guanyl in
`opossum guanyl in
`human guanyl in
`mouse guanyl in
`E Col i STa
`
`b
`
`I
`
`cgttgtcgactgtcoggcagoaacccataggt2tgagctgggaagccggg -1
`ATGTCAGGAAGCCAACTGIGGGCTGCTGTACTCCIGCTGCTGGIGCTGCAGAGTGCCCAG
`60
`M
`S
`G
`S
`O
`L
`W
`A
`A
`V
`L
`L
`L
`L
`V
`L
`O S
`A
`O
`GGTGTCTACATCAAGTACCATGGCTTCCAAGTCCAGCTAGAATCGGTGAAGAAGCTGAAT 120
`21 G
`V
`Y
`I
`K
`Y
`H
`G
`F
`O
`L
`N
`V
`O
`L
`E S
`V
`K
`K
`GAGTTGGAAGAGAAGCAGAIGTCCGATCCCCAGCAGCAGAAAAGTGGCCICCICCCCGAT 180
`41 E
`L E E K 0 M S D P G
`O O
`K S G L L
`P D
`GIGTGCTACAACCCCGCCITGCCCCIGGACCTCCAGCCTGTTIGTGCATCCCAGGAAGCT
`V
`C
`Y
`N
`P
`A
`L
`P
`L
`D
`L
`O
`P
`V
`C
`A
`S O
`E
`A
`GCCAGCACCTICAAGGCCTIGAGGACCATTGCCACTGATGAATGTGAGCTGTGTATAAAT
`A
`S
`T
`F
`K
`A
`L
`R
`T
`I
`A
`T
`D
`E
`C
`E
`L
`C
`I
`NI
`GITGCCIGTACGGGCTgCtgatgaciatgactccagacaccttacccccacagcctaccct
`V
`A
`C
`T
`G
`CI stop
`gcccatacttaggtaccattgacataattaccaccctcccagcacaaatggatccatagc 420
`ciagacaatatggatgcagagccgccatatttggtccccaggcagctgcaccggaataaaa 480
`atctgacagtcgacao
`496
`
`ei
`
`RI
`
`loI
`
`240
`
`300
`
`360
`
`C
`
`d D J I pCdC H K U Lu S T
`
`L ESVKKL
`
`L ESVKOL
`LESVKKL
`LESVKKL
`
`rat uroguanyl in
`opossum uroguanyl in
`human uroguanyl in
`rat guanyl in
`human guanyl in
`mouse guanyl in
`
`

`

`Z Li et al
`
`/ Regulatory Peptides
`
`–
`68 (1997) 45 56
`
`53
`
`opossum
`and
`human
`for
`established
`previously
`prouroguanylin (70% and 73% identity at the nucleotide
`level, and 67% and 66% identity at the amino acid level,
`respectively). In addition, as has been noted in these other
`species [37,38], there are two discrete regions of homology
`between rat prouroguanylin and rat proguanylin. The first
`region (boxed in Fig. 6b, and compared across all species
`for which sequences are available in Fig. 6a) comprises the
`biologically active C-terminus of each propeptide. The
`second region (underlined in Fig. 6b, and compared across
`all species for which sequences are available in Fig. 6c) is
`found closer to the N-terminus, separated from the con-
`served C-terminal domain by about 50 residues. The
`significance of this second conserved motif is not clear, but
`its retention across multiple species suggests that it has
`some as-yet unknown biological function.
`The tissue distribution of the uroguanylin transcript is
`shown in Fig. 6d. The size of the uroguanylin transcript is
`approximately 600 bp, just slightly smaller than the size of
`the guanylin transcript. It
`is clear that,
`like guanylin,
`uroguanylin mRNA is expressed primarily in the intestine.
`However, within the intestine,
`its distribution is quite
`distinct from that of the guanylin transcript (compare Fig.
`6d to Fig. 1a): uroguanylin is most prominent in proximal
`small intestine, while guanylin is most prominent in distal
`large intestine. This confirms the distribution of guanylin
`and uroguanylin peptides determined in our biochemical
`studies (Fig. 2).
`In addition, there is a limited amount of uroguanylin
`expression outside the intestine,
`in kidney,
`testis, and
`possibly spleen (the exposure time shown in Fig. 6d for
`non-intestinal tissues is 20 times longer than the exposure
`time for intestinal tissues).
`
`4. Discussion
`
`Guanylin was initially discovered in a search for endog-
`enous agonists for
`the STa receptor. Guanylin fulfils
`several of the criteria expected of such an agonist: it is
`present in the intestine, it competes with STa for binding to
`the receptor, and it activates cyclic GMP synthesis when
`applied to cells that express the receptor. Therefore, in a
`simplest-case scenario, guanylin expression might be ex-
`pected to be high in tissues where receptor expression is
`high, and low in tissues where receptor expression is low.
`However, within the intestine the tissue distributions of
`guanylin and its receptor are surprisingly non-parallel. We
`were particularly struck by the fact that duodenum expres-
`ses high levels of receptor but low levels of peptide. In the
`experiments described above, we have confirmed this
`mismatch at both the mRNA and polypeptide levels.
`These observations led us to consider the possibility that
`duodenum might produce its own endogenous ligand that
`resembles guanylin in its target specificity. Our HPLC
`
`analysis of duodenal extracts has confirmed this idea, and
`allowed us to purify a second peptide with guanylin-like
`biological activity. Interestingly, guanylin and the new
`peptide have complementary distributions: duodenum has
`low levels of guanylin and high levels of the new peptide,
`while colon has high levels of guanylin and little or none
`of the new peptide. Alignment of the sequence of the
`duodenal peptide with the appropriate regions of
`rat
`guanylin and uroguanylin (Fig. 6a) reveals that the new
`peptide is more closely related to uroguanylin (80%
`identity when compared across species, with nearly all
`differences representing conservative amino acid substitu-
`tions) than it is to guanylin (47–53% identity, with few of
`the differences representing conservative amino acid sub-
`stitutions). These observations suggest that the duodenal
`peptide represents the rat isoform of uroguanylin, whose
`sequence has not yet been determined.
`In a previous study, Hamra et al. [28] analyzed extracts
`of opossum intestinal mucosa by preparative isoelectric
`focusing, and found two distinct peaks of guanylin-like
`activity. Purification and sequencing of one of the peaks
`(pI $ 5.2) showed that
`it was the opossum form of
`guanylin. The other peak was not purified, but its isoelec-
`tric point (pI $ 3.0) was consistent with the idea that it
`could be uroguanylin. In a subsequent publication, Hamra
`et al. [44] showed that opossum intestine contains both
`proguanylin and a second inactive propeptide whose
`biological activity could be enhanced by proteolysis with
`chymotrypsin. Purification and C-terminal sequencing of
`this second peptide revealed that
`it was distinct from
`proguanylin, and that it had properties more closely related
`to those of uroguanylin than to those of guanylin. Our
`present study provides a direct biochemical demonstration
`that the rat intestine produces a peptide with an N-terminal
`sequence corresponding to uroguanylin. These results,
`together with the analysis of
`rat uroguanylin mRNA
`expression presented in Fig. 6d, provide convincing evi-
`dence that uroguanylin is produced by the rat intestine. In
`addition, recent cloning studies have identified human [38]
`and opossum [37] cDNAs encoding a uroguanylin prop-
`eptide, and have demonstrated expression of a transcript
`encoding this propeptide in the human and opossum
`intestine.
`In the course of identifying and purifying uroguanylin,
`we learned that
`the biological activity of the partially
`purified material could be enhanced by incubating it at
`37&C. This enhancement was accompanied by a shift in its
`HPLC retention time. We have observed a similar phenom-
`enon with partially purified proguanylin extracted from the
`colon. As it is unlikely that this is a spontaneous process,
`we believe it is most likely that a co-eluting protease is
`converting inactive or weakly active precursors into active
`products. As of yet, the peptide processing sites respon-
`sible for generating biologically active guanylin or
`uroguanylin in vivo have not been identified. Thus, the
`N-terminally extended form of uroguanylin that we have
`
`

`

`54
`
`Z Li et al
`
`/ Regulatory Peptides
`
`–
`68 (1997) 45 56
`
`identified may actually be the form that is active in the
`intestine.
`Interestingly, an N-terminally extended form of guanylin
`(containing an extra aspartate residue) has recently been
`detected in extracts of rat ileum and colon, using an RIA
`directed against
`the proguanylin C-terminus [45]. This
`N-terminally extended peptide accounts for about 50% of
`the total guanylin-like immunoreactivity present
`in the
`colon. We have not detected this peptide in our present
`studies, nor should we have for two reasons: (a) it
`is
`inactive when assayed on T84 cells, and thus would fail to
`show up in our bioassays, and (b) it lacks the proguanylin
`N-terminus, and thus would fail to show up on our Western
`blots.
`The affinity of GCC for uroguanylin (opossum or
`human)
`is about 10-fold higher
`than its affinity for
`guanylin (rat or human) [28,29]. Thus, features that are
`found in uroguanylin, but not
`in guanylin, offer infor-
`mation about structural elements that specify the strength
`of the ligand/receptor interaction. Of particular interest are
`two residues that are basic or uncharged in guanylin but
`acidic in uroguanylin (stippled arrowheads), and one
`residue that contains an aromatic ring in guanylin but an
`acid amide in uroguanylin (solid arrowhead). At all three
`positions, our duodenal peptide follows the consensus
`sequence of uroguanylin rather than that of guanylin, and
`thus we would expect its affinity to be comparable to that
`of opossum or human uroguanylin. Dos