J. Pcptide Re., 50, 1997. 222-230
`Printed in the Uiiitrd Srrirrs ofAinrriro
`
`1111 t r g l m reTr,r\ r d
`
`~
`
`Copwight 0 M~ink.\gaard, I9117
`JOURNAL OF PEPTIDE RESEARCH
`lSSN 13Y7-002X
`
`Synthesis, biological activity and isomerism of guanylate
`cyclase C-activating peptides guanylin and uroguanylin
`
`JOACHIM KLODT,' MICHAELA KUHN,' UTE C. MARX,' SILKE MARTIN,' PAUL ROSCH,'
`WOLF-GEORG FORSSMANN' and KNUT ADERMANN
`'Niedersachsisches Irisritutfur Peptid-Forschurig (IPF}, Hannovel; and
`Lehrstuhl fur Strukrur und Clieniie der Biopolymere, Uiiiversitat Bayreuth, Germany
`
`Received 12 February. revised 19 March. accepted for publication 10 May 1997
`
`Recently, the peptides guanylin and uroguanylin were identified as endogenous ligands of the membrane-bound
`guanylate cyclase C (GC-C) that is mainly expressed in the intestinal epithelium. In the present study, bioactive
`guanylin and uroguanylin have been prepared by solid-phase methodology using Fmoc/HBTU chemistry. The
`two disulfide bonds with relative 113 and 2/4 connectivity have been introduced selectively by air oxidation of
`thiol groups and iodine treatment of Cys(Acm) residues. Using this strategy, several sequential derivatives were
`prepared. Temperature-dependent HPLC characterization of the bioactive products revealed that guanylin-related
`peptides exist as a mixture of two compounds. The isoforms are interconverted within approximately 90 min,
`which prevents their separate characterization. This effect was not detected for uroguanylin-like peptides. Synthetic
`peptides were tested for their potential to activate GC-C in cultured human colon carcinoma cells (T84), known
`to express high levels of GC-C. The results obtained show that both disulfide bonds are necessary for GC-C
`activation. The presence of the amino-terminally neighboring residues of CyslO4 for guanylin and CyslO0 for
`uroguanylin has been found to be essential for GC-C stimulation. Unexpectedly, a hybrid peptide obtained
`from substitution of the central tripeptide AYA of guanylin by the tripeptide VNV of uroguanylin was not
`bioactive. 0 Munksgaard 1997.
`Ke! n.orcf.c: cGMP: ditulfide: guanylate cyclase C: guanylin: uroguanylin; peptide hormone: peptide synthesis;
`topological isomerism
`
`~
`
`Abbreviations: Acm, acetamidomethyl: Boc. trrt-butyloxycar-
`bonyl; cCMP, cyclic 3',5'-guanosine monophosphate: DCM.
`dichloromethane; DIEA. N,N-diisopropylerhylamine: EDT.
`ethanedithol; ESMS, electrospray mass spectrometry: Fnioc.
`fluorenylmethoxycarbonyl: GC-C, guanylate cyclase C: HBTU.
`2-( 1 H-benzotriazole- I -yl)- I . I ,3.3-tetraniethyluronium hexafluoro-
`phosphate; HOAc, acetic acid: HOBt. N-hydroxybenrotria7ole;
`Isc, short-circuit current; MeCN, acetonitrilc: NMP. N-methyl-
`pyrrolidinone; NMR, nuclear magnetic resonance: PHB. p-alkoxy-
`benzyl alcohol; TBME, tert-butylmethylether: TBTU. 2-( IH-benzo-
`triazole- 1 -yl)- I , 1,3.3-tetramethyluronium tetratluoroborate: TFA.
`trifluoroacetic acid.
`Residue numbers refer to the circulating proguanylin-(22-l IS)
`(7) and cDNA-deduced uroguanylin precursor containing I I 2
`amino acids (11) and are used throughout the text.
`Part of this work has been presented in abstract form at the 4th
`International Symposium. Solid Phase Synthesis & Comhinato-
`rial Chemical Libraries, Edinburgh. Scotland [Adennann. K.. Neitz.
`S., Marx. U., Rosch. P. & Forssmann. W.G. (1995) Abstract Pi 1.
`
`Guanylin and uroguanylin are novel mammalian pep-
`tide hormones that are ligands of intestinal receptor
`guanylate cyclase C (GC-C). GC-C activation increases
`the intracellular concentration of cyclic guanosine
`monophosphate (cGMP), which in turn modulates in-
`testinal water and electrolyte secretion (1-5). The en-
`docrine role of this new family of regulatory peptides
`with respect to epithelial ion transport and a functional
`link between intestine, kidney and other tissues are top-
`ics of current research in physiology. Guanylin, origi-
`nally isolated as a peptide of 15 residues from rat
`jejunum (6), also circulates in human plasma as a
`larger hormone containing 94 amino acids (7).
`Uroguanylin was isolated from urine and represents a
`second activator of GC-C (8, 9). Recently, an
`amino-terminally elongated form of 24 amino acid resi-
`dues was shown to represent a plasma-circulating form
`of uroguanylin (10, 11). It was also reported that
`prouroguanylin is present in plasma and urine in pa-
`tients with chronic renal failure (12).
`
`222
`
`MYLAN - EXHIBIT 1034
`
`

`

`Both guanylin and uroguanylin share a significant
`sequence homology with a specific family of bacterial
`enterotoxins (Fig. 1). In particular, Escherichia coli
`heat-stable enterotoxin STa causing secretory diarrhea
`in mammals was shown to activate GC-C with a sig-
`nificantly higher efficacy than guanylin and uroguany-
`lin (13, 14). It was predicted that the four cysteine
`residues of guanylin and uroguanylin form two disul-
`fide bonds in a connectivity that is also contained in
`E. coli STa ( I 5, 16). Synthesis of the three possible dis-
`ulfide isomers of guanylin confirmed that only one of
`the several possible disulfide isomers acts as a GC-C
`activator and therefore is able to increase the intracellu-
`lar cGMP level (2, 17). Similarly, uroguanylin with
`equivalent disulfide connectivity was shown to stimu-
`late GC-C in a concentration-dependent manner (9,lO).
`It is generally accepted that the other two disulfide iso-
`mers of uroguanylin are biologically inactive.
`Considering the molecular interaction of guanylin
`and uroguanylin with the extracellular domain of their
`receptor, the disulfide loops drastically reduce their con-
`formational flexibility, resulting in a highly con-
`strained three-dimensional structure. In this context, a
`unique feature of guanylin and uroguanylin is that both
`peptides exist in two distinct conformations. Although
`not separable by chromatographic methods, it was re-
`ported that, for synthetic guanylin-( 103-1 15) and the
`tryptic fragment (94-1 15) obtained from recombinant
`proguanylin, two isomers are present in aqueous solu-
`tion, as detected by NMR spectroscopy ( 1 8). In that
`study, it was concluded that one of the two forms adopts
`a conformation similar to that observed for E. coli
`enterotoxin STa (19). We have shown that guanylin-(99-
`1 15) also appears as a mixture of two compounds which
`are detectable by means of low-temperature HPLC (20).
`Very recently, it was reported that solid-phase synthe-
`sis of human uroguanylin-(97-l12) yields two topo-
`logical isomers which could be separated by HPLC
`(21). In another study, it was reported that only one
`of these isomers is bioactive and different native forms
`of uroguanylin have been identified using polyclonal
`antisera raised against the two conformational isomers
`of synthetic uroguanylin (1 2).
`In this communication, we describe a general strat-
`egy that allows solid-phase synthesis of guanylin/
`uroguanylin-type peptides with regic-selective introduc-
`tion of disulfide bonds. Using this tool, several deriva-
`tives of guanylin and uroguanylin have been
`
`Guanylin (99-1 15)
`Uroguanylin (97- 1 12)
`Uroguanylin (89-1 12)
`E co/i S la
`
`GC-C-activating peptides guanylin and uroguanylin
`
`synthesized to gain insight into sequential and confor-
`mational requirements for the interaction between the
`peptide regulators and their receptor.
`
`EXPERIMENTAL PROCEDURES
`Reagents for peptide synthesis. Fmoc amino acids
`(L-configuration) were purchased from Orpegen
`(Heidelberg, Germany) and PerSeptive Biosystems
`(Wiesbaden, Germany). Protective groups were:
`Arg(Pbf), Asn(Trt), Asp(OtBu), Cys(Acm), Cys(Trt),
`Glu( 0 t Bu) , Ly ~(Boc), Ser( tBu), Thr( tBu) , Tyr( tB u) .
`Syntheses were carried out using TentaGel resins (Rapp
`Polymere, Tubingen, Germany). HBTU was obtained
`from Perkin-Elmer/ABI (Weiterstadt, Germany),
`TBTU was from Peboc (Llangefni, Wales). NMP was
`purchased from Merck (Darmstadt, Germany). Other
`reagents and solvents were of analytical or higher grade
`and obtained from Merck or Fluka (Neu-Ulm, Ger-
`many). E. coEi heat-stable enterotoxin STa was from
`Sigma (Deisenhofen, Germany).
`
`Solid-phase peptide synthesis. Peptide assemblies
`were carried out in NMP at a scale of 0.1 mmol using a
`preloaded Fmoc-Cys(Trt)-TentaGel S Trt or S PHB resin
`(0.23 mmol/g) for guanylin-related peptides and
`Fmoc-Leu-TentaGel S Trt (0.19 mmol/g) for uro-
`guanylin-related peptides. For the synthesis of guanylin
`peptides, CyslO4 and Cysl12 were Acm-protected,
`whereas CyslO7 and C y s l l 5 were Trt-protected.
`Uroguanylin was synthesized with Trt protection for cys-
`teines in positions 103 and 111, and Acm protection
`at CyslOO and Cysl08. As a standard procedure, all
`peptides were synthesized on a 433A peptide synthe-
`sizer (Perkin-Elmer/ABI) with conditional conductiv-
`ity monitoring of Fmoc deprotections. In addition,
`guanylin-(99-115) and uroguanylin-(89-112) were
`synthesized manually on preloaded TentaGel S PHB
`resins (2 mmol) using double coupling cycles for each
`residue (10 mmol Fmoc amino acid, 9.5 mmol TBTU,
`4 mmol HOBt and 20 mmol DIEA). Acylations were
`performed for 2 h and monitored by ninhydrin assay
`(22). Fmoc groups were cleaved with 20% piperidine
`in NMP for 20 min. After completion of syntheses,
`resins were washed with DCM and dried under vacuum.
`For resin cleavage and deprotection, the dry peptidyl
`resins were treated at room temperature with a fresh
`mixture of TFA/EDT/H20 (94:3:3, 20 mL/g) for 1.5-
`2 h. Subsequently, the peptide-containing cleavage cock-
`tail was filtered and slowly dropped into chilled TBME
`(100 mW20 mL TFA). Precipitated crude peptides were
`separated by centrifugation, washed twice with cold
`TBME and dried under vacuum without lyophilization.
`
`FIGURE 1
`Primary structures of GC-C-activating peptides. Conserved rebi-
`dues are shaded.
`
`Disulfide formation and purification. To introduce the
`first disulfide bond by air oxidation, the crude peptide
`was dissolved in water (0.5-1 mg/mL). pH was ad-
`
`223
`
`

`

`J. Klodt et al.
`justed between 7.8 and 8.5 with diluted NH?. The so-
`lutions were stirred until no free thiol was detectable
`by Ellman's reagent (23) and C 18-HPLC monitoring.
`Mixtures were acidified with TFA (pH 2.5), loaded di-
`rectly onto a preparative C18-HPLC column and puci-
`fied (Vydac C18, Hesperia, CA, 47 x 300 mm, 300 A,
`15-20 pm, buffer A; 0.1% TFA, buffer B: 0.1% TFA
`in MeCN/water, 80:20, gradient 0-1 00% B in 60 min,
`flow 50 mL/min, UV detection at 230 nm). Fractions
`containing the pure monocyclic intermediate as de-
`tected by ESMS and C 18-HPLC were collected and
`lyophilized. Cleavage of Acm groups and introduction
`of the second disulfide bond were carried out by dis-
`solving the purified monocyclic Acm2-peptides in
`deairized HOAdO.1 M HC1 (4:1, pH 2, 1 mg/mL). A
`solution of iodine (10-20 eq, 0.05 % I ) in HOAc was
`added and after 30-60 min the excess iodine was re-
`duced by the addition of fresh 0.1 M sodium thiosul-
`fate, resulting in a colorless solution. After dilution
`with water (2 volumes), the bicyclic products were
`obtained from preparative C18-HPLC (Vydac C18, 20
`x 250 mm, 300 A, 10 pm, buffer A: 0.1% TFA; buffer
`B: 0.1 % TFA in MeCNIwater 80:20, gradient 0-100%
`B in 60 min, flow 8 mL/min, 230 nm). Purification
`was performed at 50 "C (guanylin peptides) or room
`temperature (uroguanylin peptides). In the case of in-
`sufficient purity, HPLC purification of bicyclic prod-
`ucts was repeated.
`
`Analytical methods. Analytical HPL$ was performed
`on a Vydac C I8 column (5 pm, 300 A, 4.6 x 250 mm,
`buffer A: 0.06% TFA; buffer B: 0.055% TFA in 80%
`MeCN, flow rate 0.8 mL/min, 215 nm). Tempera-
`ture-dependent HPLC characterization of purified pep-
`tides was carried out using an adjustable column oven
`with a Nucleosil C18 PPN column (2 x 250 mm, 5
`pm, 100 A, flow rate 0.2 mL/min, 215 nm, Macherey
`& Nagel, Diiren, Germany). Mass spectrometry was
`performed using a triple-stage quadrupole mass spec-
`trometer (Sciex API 111, Perkin-Elmer). Peptide
`samples were dissolved in MeCN/water (1 : 1) contain-
`ing 0.2% AcOH to a concentration of 0.1 mg/mL and
`infused into the ion spray using a syringe pump at a
`flow rate of 5 pL/min. The ion spray interface was
`set to a positive potential of 5 kV. Amino acid analy-
`sis was carried out on a Aminoquant 1090L analyzer
`(Hewlett-Packard, Waldbronn, Germany). Sequence
`analysis was performed on a 494A protein sequencer
`(Perkin-Elmer). Two-dimensional NMR spectroscopy
`(NOESY, Clean-TOCSY, DQF-COSY) was performed
`on a Bruker AMX600 spectrometer at 284 K using stan-
`dard methods (24,25). Synthetic peptides were dissolved
`in H,O/D,O (9:l) at a concentration of 5-10 mg/mL.
`
`T84 cell cGMP bioassay. Determination of the
`GC-C-stimulating activity of synthetic peptides was evalu-
`ated as described elsewhere in detail (1, 7, 10). Cultured
`
`224
`
`human colon carcinoma (T84) cells were incubated with
`synthetic peptides for 60 min in the presence of the phos-
`phodiesterase inhibitor isobutylmethylxanthine (IBMX, 1
`8
`mv, Si ma). The peptides were tested in concentrations
`of 10- to 10f' M . The given concentrations refer to
`the net peptide content specified by amino acid analy-
`sis. Effects on intracellular cGMP level were compared
`with those of E. coli enterotoxin STa. cGMP was mea-
`sured using a specific radioimmunoassay (26).
`
`lor1 transport (Ussing chamber experiments). Female
`Wistar rats maintained under standard temperature,
`light conditions and with free access to water and food
`were used. For the assessment of active electrogenic
`ion transport, specimens of proximal colon were dis-
`sected and serosal and muscular layers separated. Sheets
`of isolated mucosa were mounted in Ussint; chambers
`(27) with an exposed surface area of 1 cm' and auto-
`matically voltage-clamped as described previously (3,
`4). Short-circuit current (Isc) was recorded continu-
`ously. After an equilibrium period of 30 min, cumu-
`lative concentrations of synthetic peptides were added
`to the luminal side of the colonic mucosa.
`
`RESULTS AND DISCUSSION
`
`Synthesis
`In agreement with earlier observations (2, 21) and in
`contrast to other peptides like endothelin (28) and low
`molecular weight toxins (29), oxidative folding of
`tetrathiol-containing precursors of guanylin and
`uroguanylin did not result in bioactive products that
`increase intracellular cGMP concentration. The re-
`duced peptides had a poor solubility in aqueous buffers.
`Further attempts to synthesize guanylin with different sets
`of orthogonal side-chain protective groups for the cys-
`teine residues like acetamidomethyl (Acm), tert-butyl
`(tBu), tert-butylthio (StBu), p-methoxybenzyl (Mob)
`or allyloxycarbonylaminomethyl (Allocam) resulted in
`products that were only slightly soluble in buffers re-
`quired for disulfide formation. To improve solubility,
`guanylin was initially synthesized with additional
`amino-terminal Glu-Asp residues which do not influ-
`ence the bioactivity. Guanylin and uroguanylin were
`synthesized using Acm and Trt protection for cysteine
`residues with Cys(Trt) at the C-terminal position. De-
`tectable by conductivity monitoring, qualitative ninhy-
`drin assay or HPLCESMS of intermediates, guanylin and
`uroguanylin represent difficult peptides. Starting from
`Alal 10 (guanylin) and Val106 (uroguanylin), acylations
`and cleavage of Fmoc groups tended to proceed
`slowly. For the manual synthesis of larger amounts of
`peptides, acylations were carried out for 2 h whereas
`Fmoc peptidyl resins were treated for 20 min with pi-
`peridine. Exchanging the pairs of cysteine protective
`groups resulted in a product of slightly less purity.
`The first loop between the unprotected cysteines of
`
`

`

`all synthetic compounds was introduced by air oxida-
`tion at pH 8. The proceeding oxidation process could
`be monitored easily by Ellman's test (23), HPLC com-
`parison of thiol peptides and monocyclic products or
`ESMS analysis after iodoacetamide treatment (30) (Fig.
`2). Depending on the sequence, a period of 12-72 h
`was necessary for this step. For the subsequent work-up
`of monocyclic Acm2-peptides, it was essential to
`acidify the mixture to pH 2.5 (HCl or TFA). As de-
`tected by MS and HPLC, direct lyophilization or pu-
`
`2+
`[M+2H] 1372
`
`1
`
`[ M + H j 1848
`
`[ M + 2 H r 1300
`
`[M+H]+ 1703
`
`------
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`5
`
`10
`
`IS
`
`20
`
`25
`
`30
`
`Time (min)
`
`FIGURE 2
`HPLC characterization of linear, monocyclic and purified
`peptides. ( a ) human uroguanylin-(89-112) FKTLRTIAN-
`DDCELCVNVACTGCL; ( b ) human guanylin-(99-115),
`EDPGTCEICAYAACTGC. Top traces, linear peptides after
`TFA cleavage; middle traces, monocyclic Acml-peptides;
`bottom traces, bicyclic, purified products. HPLC gradient (Vydac
`C 18): linear and monocyclic peptides, 20-70% buffer B in 30
`min at room temperature; bicyclic peptide, 10-50% buffer B in
`40 min at 40 "C. Masses given were determined by ESMS.
`
`GC-C-activating peptides guanylin and uroguanylin
`
`rification of a pH 8 solution of monocyclic peptides re-
`sulted in products with a poor solubility, possibly
`caused by further oxidation of cysteines. To obtain fi-
`nal products of a high purity, additional HPLC puri-
`fications of the monocyclic Acm2-peptides were
`performed. Cleavage of Acm groups with subsequent
`formation of the second disulfide with excess iodine
`at pH 2 (HCl) was a suitable method with suppression
`of disulfide rearrangement (3 1). The absence of HCl
`led to unsatisfactory products. Different solvents like
`mixtures of water, acetic acid and methanol did not
`affect iodine treatment. Alternative attempts at disul-
`fide formation using on-resin procedures and the
`solvent-dependent selectivity of iodine treatment of
`free thiols and Acm-protected cysteine residues failed
`(32); dimerization was a significant side effect. Over-
`all yields were between 2 and 8% because of the re-
`peated HPLC purifications of mono- and bicyclic
`peptides. With guanylin-(99-115) and uroguanylin-
`(89-1 12) as examples, Fig. 2 shows analytical HPLC
`of intermediate compounds and final products which
`were unambiguously characterized by ESMS. Using
`this general synthetic procedure, guanylin and
`uroguanylin as well as several derivatives have been
`prepared and characterized (Table 1). As recently com-
`prehensively reviewed (31,33), our results of the syn-
`thesis of guanylin and uroguanylin are another
`indication that suitable reaction conditions for each
`single peptide containing multiple disulfide bonds must
`be worked out accurately.
`Unlike other polycyclic peptides with multiple cys-
`tines within a short sequence segment, low-tempera-
`ture HPLC analysis of purified guanylin-(99-115) (4)
`showed two components in a ratio of about 1 : 1. ESMS
`analysis resulted in only one signal with the expected
`molecular weight of 1703 Da. Reverse-phase HPLC
`with a stepwise increase of column temperature led to
`one sharp absorption at high temperature. Cislrruns
`isomerism of the Asp100-Pro101 amide bond was
`excluded as the reason for this observation by
`temperature-dependent HPLC analysis of guanylin-
`(101-1 15) (3) with an amino-terminal Pro residue.
`Shown in Fig. 3, a and b, peptides 3 and 4 exhibit the
`identical HPLC characteristic at different temperatures.
`A corresponding temperature effect was observed with
`guanylin-( 104-1 15) (l), [desThrl03]guanylin-(99-
`115) (7) and compounds 12 and 13. In contrast,
`guanylin-( 103-1 15) (2) appears as a single HPLC peak.
`This is consistent with an earlier report (1 8). Our pre-
`liminary NMR experiments with guanylin-(99-115) (4)
`confirmed the presence of two distinct compounds in
`a ratio of about 1 : 1 (20). The guanylin/uroguanylin hy-
`brid 9 had a major HPLC peak, but two-dimensional
`NMR spectroscopy revealed the presence of two com-
`ponents in equivalent amounts (data not shown). As
`expected, monocyclic Cys(Acm)2 intermediates 10, 11
`and 18 were homogeneous during HPLC. A reversed in-
`
`225
`
`

`

`2938.5 I 1470.5'
`251 1.9 I 1257.0'
`2614.1 I 1372.0'
`2469.9 I 1235.0'
`2484.9 I 1242.0'
`2599.0 I I300.0'
`1667.9 I 1667.Sh
`1515.6 I 1516.0"
`1458.7 I 14S9.0h
`1845.9 I I 846.0h
`1845.9 I I 846.0h
`1709.9 I 17 I O.Oh
`1582.7 I 1582.Sh
`I6O1.8 I 1601.0h
`1573.8 I 1573.Sh
`1702.9 I 1703.Sh
`1702.9 I I 703.Sh
`1458.7 I 1459.0h
`1304.5 I 1304.Sh
`1203.4 I 1203.Sh
`
`Calc./ESMS
`Mass (Da)
`
`n.d., not determined.
`'[M+2H]".
`h[M+H]+.
`gBiotinAhx, N-(Biotinyl-6-aminocaproyl).
`'Sequence from guinea pig uroguanylin obtained by translation of the corresponding mRNA (EMBL database, Accession no. 274738).
`eMonocyclic Acmz-intermediates.
`dSequences from porcine and guinea pig guanylin obtained by translation of the corresponding mRNA (EMBL database, Accession nos. 273607 and 274736.
`'Monocyclic Acmz-intermediates.
`'Monocyclic Acm2-intermediates.
`"Reversed order of disulfide bond formation using Cys(Acm)l07/115 and Cys(Trt)l04/112.
`
`n.d.
`Ala 2.13 Arg 0.99 Val 0.96 Met 0.95 Ile 1.00 Leu 3.08 Asx 3.99 Glx 2.35 Gly 1.20 Thr 1.82
`n.d.
`Ah 2.07 Arg 0.99 Val 2.02 Phe 0.97 Ile 0.98 Lci 3.05 Asx 4.06 Cly 1.17 Thr 2.21 Lys 0.85
`Ala 2.08 Arg I .OI Val I .94 Phe 1.00 IIc 0.93 Leu 3.10 Asx 3.09 Glx 0.71 Gly 1.23 Thr 2.57 Lys 1.02
`Ala 2.12 Arg 1.04 Val 1.94 Phe 0.98 Ile 0.98 Lcu 3.05 Asx 3.95 Glx 0.71 Gly 1.1 I Thr 2.31 Lys 0.97
`Ala 1.04 Val 1.89 Leu 2.04 Asx 3.84 Glx 0.72 Gly 1.19 Thr 0.93
`Ala 3.08 Ilc 0.98 Asx 0.78 Glx 0.65 Gly I .03 Thr I .70 Tyr 0.82 Pro I .02
`Ala 4. I2 Ilc 0.93 Glx 0.63 Gly I. 15 Thr 0.78 Tyr 0.88 Pro I .04 Ser 0.45
`n.d.
`n.d.
`Ala I .OI Ile 0.97 ASX I .83 Glx I .42 Gly 2.02 Thr I .77 Pro I .02 Val I .93
`Ala 2.02 11s 0.97 Asx 0.81 Glx 1.31 Gly 4.06 Thr I .72 Pro 1.06
`Ala 3.06 Ile 0.94 Asx 0.66 Glx I. I2 Gly 2.02 Thr 0.90 Tyr 0.8 I Pro I .01
`Ala 3.22 111: 0.93 Asx 0.90 Glx 0.72 Gly I .Y8 Thr I .63 Tyr 0.79 Pro 0.98
`Ala 3.1 I Ilc 0.95 Asx 0.89 Glx 1.37 Gly 2.23 Thr 1.73 Tyr 0.86 Pro 1.12
`Ala 3.02 Ilc 0.94 Asx 0.97 Glx I .56 Gly 2.0 I Thr I .6 I Tyr 0.84 Pro I .OX
`Ala 3.31 Ile 1.00 Glx 0.63 Gly 0.09 Thr 1.56 Tyr 0.94 Pro 0.92
`Ala 3.09 Ile 0.95 Glx 0.59 Gly 1.02 Thr 1.60 Tyr 0.88
`Ala 3.06 Ilc 0.98 Glx 0.68 Gly I. 13 Thr 0.84 Tyr 0.75
`
`(Biotin Ahx)FKTLRTIANDDCELCVNVACTGCL
`LQALRTMDNDECELCVNI ACTGC
`FKTLRTIANDDC( Acm)ELCVNVAC(Acm)TGCL
`FKTLRT[ANDDCI,CVNVACTGCl,
`FKTLRTIANDCELCVNVACTGCL
`FKTLRTI ANDDCELCVNVACTGCL
`NDDCELCVNVACTGCL
`PNTCEICAYAACTGC
`PSTCEICAYAACAGC
`EDPGTCEIC(Acm)AYAACTGC( Acrn)
`EDPGTC( Acm)EICAYAAC( Acm)TGC
`EDPGTCEICVNVACTGC
`EDPGTCEIC AGGACTGC
`EDPGCEICAYA ACTGC
`EDPGTCICAYAACTGC
`EDPGTCEICAYAACTCC
`EDPGTCEICAYAACTCC
`PGTCEICAYAACTGC
`TCEICAYAACTGC
`CEICAYA ACTGC
`
`2Qg
`19'
`I 8'
`17
`I6
`15
`14
`I3
`I2d
`11'
`I Oh
`9
`8
`7
`6
`5
`4
`3
`2
`I
`
`Amino acid composition
`
`No. Sequence
`
`

`

`troduction of disulfides did not alter the low-temperature
`HPLC pattern of guanylin-(99-115) (5).
`Rechromatography of the collected peaks I and I1
`from an analytical HPLC run of guanylin 4 at 11 "C
`(Fig. 3b) unexpectedly did not result in single peaks.
`Proving the interconvertibility between the two iso-
`mers, the mixture of both compounds was recovered
`after HPLC of isolated peaks I and I1 (Fig. 3d). There-
`fore, racemization of cysteine residues during peptide
`assembly can be excluded as the reason for two HPLC
`peaks with identical molecular weights. Although our
`attempts to separate both isoforms failed, repeated
`cycles of analytical HPLC, peak collection and
`rechromatography led to concentrated peaks I and 11,
`respectively (Fig. 3d). Variation of the conditions of
`HPLC analysis of peptides 3 and 4, the HPLC (phos-
`phate buffer, pH 3) and storage of the collected
`isoforms corresponding to peaks I and I1 (Fig. 3, a
`and b) for different periods (up to 30 days) at differ-
`ent temperatures, even at +50 "C and liquid nitrogen
`temperature, did not change the original HPLC char-
`acteristics. From the results of NMR spectroscopy, the
`half-life of guanylin isomers was estimated to be at
`least in the range of seconds (18). Regarding the
`rechromatography of the collected isoforms, which was
`performed about 90 min after collection of both iso-
`mers, a maximum half-life of 90 min of the isoforms
`can be concluded. However, the mechanism of
`interconversion remains unclear. Because of a high ac-
`
`GC-C-activating peptides guanylin and uroguanylin
`
`tivation barrier for a mutual exchange of isoforms,
`which has been described to be in excess of 20 kcaV
`mol(1 8), an intermediate opening of one disulfide loop
`followed by a cystine rearrangement may be consid-
`ered as an alternative mechanism for the formation of
`topoisomers. This pathway seems to be unlikely be-
`cause the formation of positional disulfide isomers by
`the attack of a free thiol on the second disulfide would
`be a likely side effect. There are no experimental indi-
`cations for a cystine rearrangement. Recently, Chino et
`al. (21) reported the isolation of two isoforms of syn-
`thetic uroguanylin-(97-112) (14). Depending on the
`order of introduction of the disulfide bonds, the isomers
`were formed in a different ratio. Although we can con-
`firm this result, only one compound was isolated after
`synthesis of uroguanylin-(89-112) (15) and other
`uroguanylins (Fig. 3c). Our preliminary NMR studies
`of synthetic peptide 15 demonstrated that only one
`compound is present in solution (data not shown).
`However, it is possible that the second isomer was not
`noticed during chromatographic product separation.
`Our results show that guanylins 3 and 4 exist as a
`mixture of two isoforms which are interconverted. As
`suggested for uroguanylin (21), the order of introduc-
`tion of the two disulfide loops had no influence on the
`product composition. This was revealed by synthetic
`peptides 4 and 5 which showed an identical HPLC pat-
`tern. An amino-terminal truncation of the first two resi-
`dues of guanylin in peptide 2 only influences the
`
`I
`
`11°C Peak I \ , Peak I1
`
`11 "C
`
`I
`
`I
`l l 0 C l i J (
`11°C
`
`11 "C
`
`1
`
`20
`
`30
`
`40 20
`
`30
`
`40
`Time (min)
`
`40
`
`so
`
`25
`
`30
`
`35
`
`FIGURE 3
`(a) HPLC analysis of guanylin-(lOl-
`115) 3 at 11,25 and 35 "C; (b) HPLC
`analysis of guanylin-(99-115) 4 at 1 I ,
`25 and 35 "C; (c) HPLC analysis of
`uroguanylin-(89-112) 15 at 11 and
`35 "C; (d) HPLC of peak I (bottom)
`and I1 (top) at 11 "C obtained after
`collection from run (b), peaks were
`separated two times and reanalyzed
`under the same conditions. Gradient:
`1 5 4 5 % buffer B in 60 min.
`
`221
`
`

`

`J. Klodt et al.
`
`detectability of the two isomers by HPLC. Mutations
`or deletions at the N-terminal region of peptides 1, 7,
`9, 12 and 13, and the C-terminal mutation in 12 (Thr
`to Ala) also resulted in two HPLC peaks. In contrast,
`the hybrid mutant 9 appears as a single peak, but con-
`sists of two compounds as detected by two distinct sets
`of NMR resonances (data not shown). Chino et al. (21)
`suspected that the carboxy-terminal leucine residue
`present in uroguanylin-(97-112) (14) may stabilize the
`isomers because the absence of this leucine residue re-
`sulted in a peptide showing a HPLC characteristic simi-
`lar to guanylin. On the other hand, uroguanylin 19 with
`a carboxy-terminal cysteine was homogeneous during
`reverse-phase HPLC. Summarizing the present data, it
`is possible that a C-terminally attached elongation is
`crucial for the stability of topoisomers, but other resi-
`dues, in particular between the inner cysteine residues,
`also influence this phenomenon by their sterical bulk.
`Deduced theoretically by Mao (34), guanylin and
`uroguanylin represent an actual example of topologi-
`cal stereoisomerism. Further studies dealing with the
`sequence dependency of the formation of guanylin
`isoforms and of the influence of variables like pH, sol-
`vent and temperature are currently under investigation
`in our laboratory. Stability and NMR studies of syn-
`thetic uroguanylins 14 and 15 are also in progress.
`
`Biological activity
`Guanylin and uroguanylin peptides had a concentra-
`tion-dependent effect on the intracellular cGMP level
`at concentrations between lo-* and lo4 M, whereas E.
`coli enterotoxin STa was already active at lo-‘ M (Fig.
`4a). Different monocyclic derivatives, isolated as in-
`termediates during synthesis, were biologically inac-
`tive indicating that the two disulfide bonds and,
`therefore, the rigid three-dimensional structure is nec-
`essary for receptor activation. Guanylin lacking
`Thr103, either as a truncated (1) or deletion peptide
`(71, was only effective at a concentration of
`M.
`Thus, Thrl03 is a crucial residue for receptor activa-
`tion. This finding corresponds with earlier reports about
`the ability of rat guanylin to bind to intestinal mem-
`branes (35). Interestingly, amino acid exchange of resi-
`due 102 of guanylin did not influence its activity on
`cCMP generation (Fig. 4a, compounds 12 and 13). Pep-
`tide 6 lacking the endocyclic Glu105 was inactive. Re-
`versing the order of disulfide introduction in human
`guanylin-(99-115) had no effect on HPLC pattern and
`did not influence its bioactivity (4 and 5). From this re-
`sult, it can be concluded that peptides 4 and 5 have a
`similar composition. Compound 8 with the non-
`voluminous residues AGGA between the inner cys-
`teines was inactive, indicating an important role for this
`region in receptor activation. For uroguanylin peptides,
`changes in the N-terminal region, like a specific
`biotinylaminocaproylation (20) or amino acid exchange
`(19) did not influence the increase of cGMP concen-
`
`22%
`
`Pep1,ds
`
`i R l
`
`/‘ --
`/
`/
`
`~
`
`lt1-7M
`
`‘ 1
`
`15
`
`16
`
`17
`PIptldc
`
`18
`
`I(,
`
`21)
`
`yya
`
`c
`
`FIGURE 1
`Effects of synthetic peptides on intracellular cGMP concentration
`of cultured T84 cells. (a) Guanylin-related peptides and (b)
`uroguanylin derivatives (C = control; approxiinately 106 cells/
`well: 50 WLlwell).
`
`tration in T84 cells (Fig. 4b). It is noticeable that dele-
`tion of GlulOl (17) did not influence the bioactivity of
`uroguanylin, whereas deletion of Asp99 resulted in a
`peptide (16) with only minor activity. Superficially, the
`latter effect seems to be similar to guanylin 7 lacking
`residue Thrl03, but the second Asp residue of
`uroguanylin complicates interpretation of a functional
`connection between reduction in bioactivity and the
`amino acid residue 103 in guanylin and residue 99 in
`uroguanylin. An unexpected result was obtained from
`the cGMP bioassay of compound 9. Being a hybrid of
`guanylin and uroguanylin, this compound had only mi-
`nor activity at lo-’ M.
`In the intestinal epithelium, a final effect of GC-C-
`activating peptides is the cGMP-mediated activation of
`the CFTR (cystic fibrosis transmembrane conductance
`regulator) chloride channels, thereby enhancing chlo-
`ride secretion (1,2,5). Selected synthetic peptides were
`tested in their ability to modulate ion transport using
`rat intestinal mucosa mounted in Ussing chambers. As
`shown earlier, the guanylin-induced changes of
`short-circuit current (Isc) can be used as an index of
`
`

`

`chloride secretion (7). The results obtained from
`concentration-dependent effects of synthetic peptides
`on Isc corresponded to those from T84 cell bioassay.
`Figure 5 shows the changes of Isc observed after the
`addition of increasing doses of synthetic peptides.
`Guanylin 3 and uroguanylin 14 were bioactive in con-
`centrations between lo-* and 10“ M. Amino-terminally
`longer guanylin 4 and uroguanylin 15 were slightly
`less bioactive starting at lo-’ M. As in the cGMP as-
`say, [desThrl03Jguanylin-(99-115) (7) had only a mi-
`nor effect at 10 M. These results indicate a connection
`between the peptide structure and GC-C activation and
`chloride secretion.
`
`CONCLUSIONS
`The presented data of the synthesis of guanyliduro-
`guanylin peptides demonstrate a particular necessity to
`use T84 cell bioassay as an additional tool to charac-
`terize the obtained products. Besides disulfide isomers
`with different cysteine connectivity and well-known ra-
`cemization of protected cysteine derivatives during
`solid-phase synthesis (36), the novel form of topologi-
`cal stereoisomerism described may lead to a large and
`confusing number of products having the same mo-
`lecular weight. The formation of topological stere-
`oisomers should be considered in the synthesis of any
`multiple-cysteine peptide. Our results demonstrate that,
`up to now, all published functional studies with bioac-
`tive guanylin reported in the literature have not been
`carried out with a single defined molecule but with a