`f. Klodt
`M. Meyer
`H. Gerlach
`P Rosch
`W -G. Forssmann
`K. Adermann
`
`One peptide, two topologies:
`
`
`structure and interconversion
`
`dynamics of human
`
`uroguanylin isomers
`
`
`
`guanylin; heat-stable enterotoxin; isomerization;
`
`
`
`
`
`solution structure; topological stereoisomer; uroguanylin
`
`
`
`
`
`The peptide hormone uroguanylin stimulates chloride
`
`Key words:
`
`
`
`Authors' affiliations:
`U. C. Marx, Niedersiichsisches Institut fur
`
`
`
`
`
`
`Peptid-Forschung IIPF), Hannover, Germany,
`
`
`and Lehrstuhl fur Struktur und Chemie der
`Abstract:
`
`
`
`
`Biopolymere, Universitiit Bayreuth, Bayreuth,
`Germany.
`
`
`
`
`
`
`
`
`
`
`secretion via activation of intestinal guanylyl cyclase C (GC-C). It is
`
`
`
`[ Klodt, M Meyer, W-G. Forssmann, and K.
`
`
`
`Adermann,
`
`
`Niedersiichsischcs Institut fur
`
`
`
`
`Peptid-Forschung IIPFI, Hannover, Germany.
`
`H. Gerlach,
`
`
`Lehrstuhl fur Organische Chemie 11,
`
`
`
`
`Universitiit Bayreuth, Bayreuth, Germany.
`
`P. Rosch, Lehrstuhl fur Struktur und Chemie
`
`
`
`
`
`der Biopolymere, Universitiit Bayreuth,
`
`
`Bayreuth, Germany.
`
`
`
`
`
`
`
`unambiguous structure-function relationship of the isomers, we
`
`
`
`
`
`determined the solution structure of the separated uroguanylin
`
`
`
`
`
`isoforms using NMR spectroscopy. Both isomers adopt well-defined
`
`
`
`structures that correspond to those of the isomers of the related
`
`
`
`
`
`characterized by two disulfide bonds in a 1-3/2 4 pattern that
`
`
`
`
`
`causes the existence of two topological stereoisomers of which
`
`
`
`
`
`only one induces intracellular cGMP elevation. To obtain an
`
`
`
`
`
`
`
`
`
`peptide guanylin. Furthermore, the structure of the GC-C
`
`the agonistic
`
`Escherichia
`Niedersiichsisches Institut fur Peptid-Forschung
`
`
`
`
`
`
`with guanylin isomers, the conformational interconversion of
`
`
`
`
`
`coli heat-stable enterotoxin. Compared
`
`
`
`uroguanylin isomers is retarded significantly. As judged from
`
`
`
`
`
`
`
`
`
`activating uroguanylin isomer A closely resembles the structure of
`
`
`
`
`
`
`
`
`
`
`Correspondence to:
`Dr. Knut Adermann
`
`
`
`(!PF)
`Feodor-Lynen-Strasse 31
`
`
`
`
`D-30625 Hannover, Germany
`
`Phone: int +49-511-5466262
`Fax: int +49-511-5466102
`at 37°C with an equilibrium
`
`E-mail: knutadermann@compuserve.com
`
`pH-dependent mutual isomerization
`
`
`
`
`
`isoforms are stable at low temperatures, but are subject to a slow
`
`
`
`
`
`
`
`chromatography and NMR spectroscopy, both uroguanylin
`
`
`
`Ta cite this article:
`
`Marx, U.C., Klodt, J., Meyer, M., Gerlach, H., Rosch,
`
`
`
`
`
`
`
`P., Forssmann, W.-G., Aderrnann, K. (1998) One peptide,
`
`
`
`two topologies: structure and interconversion
`
`
`
`isomers. dynamics of human uroguanylin
`
`f. Peptide
`Res. P, 229-240.
`
`
`
`
`
`magnetic resonance; NOE, nuclear Overhauser effect, also used for
`
`
`
`
`
`
`
`ISSN 1397-002X
`
`
`
`NOESY cross peak; NOESY, NOE spectroscopy; RMSD, root-mean-
`
`229
`
`
`
`isomer ratio of approximately 1: 1 . The conformational exchange is
`
`
`
`
`
`most likely under the sterical control of the carboxy-terminal
`
`
`
`
`
`leucine. These results imply that GC-C is activated by ligands
`
`
`
`
`
`
`
`
`
`
`
`exhibiting the molecular framework corresponding to the structure
`
`
`
`
`
`of uroguanylin isomer A.
`
`Abbreviations:
`
`
`
`cGMP, cyclic 3 ',5 '-guanosine monophosphate;
`
`
`
`Clean-TOCSY, TOCSY with suppression of NOESY-type cross peaks;
`
`
`
`
`
`
`
`
`
`DG, distance geometry; DQF-COSY, double-quantum filtered COSY;
`
`
`
`
`
`
`
`DSS, 2,2-dimethyl-silapentane-5-sulfonic acid; GC-C, guanylyl
`
`
`
`
`
`cyclase C; JR-NOESY, 2D NOESY spectrum acquired with a jump
`
`
`
`
`
`return observe pulse; MD, molecular dynamics; NMR, nuclear
`
`
`
`
`
`Dates:
`Received 23 January 1998
`
`
`
`Revised 2 March 1998
`
`Accepted 21 March 1998
`
`
`
`
`
`Copyright© Munksgaard 1998
`
`Bausch Health Ireland Exhibit 2010, Page 1 of 12
`Mylan v. Bausch Health Ireland - IPR2022-00722
`
`
`
`U.C. Marx et al . Structure and dynamics of uroguanylin isomers
`
`
`
`square deviation; SA, simulated annea l ing; ST, Escherichia
`coli
`
`
`
`heat-stable enterotoxin; STaR, E. coli heat-stable enterotoxin
`
`amino acids of guanylin and a 22-residue guanylin fragment
`produced by tryptic digestion of the recombinant pro
`hormone consist of two species each (molar ratio, 1:1) which
`can not be separated (16). One of those isoforms (A) adopts
`a well-defined structure with a right handed spiral confor
`mation similar to that of ST obtained from X-ray crystal
`lography [22). The second isoform (B) adopts a less
`well-defined left-handed spiral conformation, and therefore,
`is suggested to have a lower affinity to the receptor GC C
`(16). This isomerism was confirmed by NMR spectroscopy
`for guanylin containing 17 amino acids (23). Two compo
`nents were observed clearly during high-performance liq
`uid chromatography (HPLC) analysis for this peptide only
`and the smaller guanylin of 15 residues. Rapid intercon
`version between the two compounds, however, prevented
`the characterization of the isolated molecular species (15,
`23). Recently, the synthesis of two different topological iso
`mers of human uroguanylin was reported (24). Polyclonal
`antisera against these isoforms were used to detect
`uroguanylin and its corresponding 10-kDa prohormone in
`urine and plasma (11). Although not shown on the structural
`level of the peptides, this finding suggests that both stere
`oisomers of uroguanylin are present in the body. In addi
`tion to the possible existence of other specific receptors for
`the GC-C-active isomer of uroguanylin, nothing is known
`about the structure and function of the peptide's B form
`that does not cause an increase of intracellular cGMP.
`The purpose of this study was to establish an unambigu
`ous structure-activity relationship of uroguanylin stereoi
`somers by combining the NMR-spectroscopical structure
`determination with the GC-C-activating potential and the
`dynamical characteristics of uroguanylin isomers. The re
`sults obtained are a prerequisite to model the interaction
`of ligands with the target protein GC-C with respect to the
`inhibition of GC-C activity and enterotoxic infections.
`
`
`
`Experimental Procedures
`
`
`
`Peptide synthesis, analytical c h romatography and polarimetry
`
`Uroguanylin-16 and uroguanylin-24 were synthesized on a
`preloaded TentaGel-S-PHB-LeuFmoc resin (Rapp Polymere).
`Disulfides of uroguanylin-16 were introduced selectively by
`air oxidation of Cys4 and Cys12 followed by iodine treat
`ment of acetamidomethyl-protected cysteine residues 7 and
`15. After formation of the second disulfide, two stereoiso
`mers were obtained and separated by reversed-phase HPLC
`at a temperature of 15 °C. Uroguanylin-24 was prepared cor
`
`
`
`
`
`receptor; TOCSY, total correlation spectroscopy; TPPI, time
`
`
`
`
`
`
`
`proportional phase incrementation.
`
`Uroguanylin is a small mammalian peptide hormone that
`is known to be involved in the regulation of epithelial wa
`ter and electrolyte transport. It is related structurally and
`functionally to the agonistic peptide guanylin (1-3). A tar
`get protein for uroguanylin is guanylyl cyclase C (GC C),
`also known as an Escherichia coli heat-stable enterotoxin
`(ST) receptor (STaR) (4, 5 ). Activation of GC-C increases in
`tracellular production of cyclic guanosine monophosphate
`(cGMP) (3, 6), thereby stimulating cGMP dependent protein
`kinase type II (cGKII) [7, 8). cGKII in turn is responsible for
`chloride secretion by the activation of cystic fibrosis trans
`membrane conductance regulator (CFTR), resulting in in
`testinal fluid and chloride accumulation (9 ).
`Uroguanylin of which different molecular forms were
`identified in urine (3, 6), blood (10, 11), and the intestine (12,
`13) is related closely to guanylin on the level of primary
`structures (Fig. 1). Both peptides are characterized by two
`disulfide bonds in relative positions 1 3 and 2 4, which are
`crucial for biological activity (2, 14, 15 ). The disulfides are
`located within a short sequence of 12 amino acid residues.
`This structural element leads to the existence of two dis
`tinct topological isoforms of each peptide (15, 16). The en
`terotoxin ST exhibits an equivalent pattern of disulfides,
`but has an additional disulfide loop which may enhance its
`conformational rigidity (17-19 ). In binding assays, both pep
`tides compete with ST for their common receptor GC C.
`However, the endogenous peptides are less potent activa
`tors of GC-C than ST is (10, 20, 21).
`The topological stereoisomers of guanylin and uroguany
`lin exhibit the same cysteine connectivity, but are confor
`mationally distinct molecules. A synthetic derivative of 13
`
`Uroguan y li n ( u r i n e )
`
`N D DCELCVNVACTGCL
`
`1 �--�
`
`1 6
`
`
`
`U r o gu an y li n ( b l o o d )
`FKTLRTIANDDCELCVNVACTGCL
`
`Guany l i n
`
`PGTCEICAYAACTGC
`
`
`
`E . coli
`�
`
`ent e r o t o x i n ST
`
`NTFYCCELCCNPACTGCY
`
`1. Amino acid sequences and disulfide pattern of GC-C
`activating peptides. The primary structures of human uro
`guanylin and guanylin are shown. ST is from a human strain of
`(41) with the GC C binding and toxic domain located
`between the outer cysteine residues.
`
`Figure
`
`E. coli
`
`2 3 0 I r Peptide
`
`R e s . 5 1 , 1998 I 2 2 9-240
`
`Bausch Health Ireland Exhibit 2010, Page 2 of 12
`Mylan v. Bausch Health Ireland - IPR2022-00722
`
`
`
`respondingly. A detailed synthetic procedure was reported
`recently by Klodt et al. (15 ). The identity of synthetic pep
`tides was confirmed by electrospray mass spectrometry and
`sequence analysis by Edman degradation. For the cGMP
`T84 cell assay, synthetic peptides were used according to
`the net peptide content as determined by amino acid analy
`sis. The HPLC study of the mutual conversion of purified
`uroguanylin isoforms was carried out using a C18 column
`(Nucleosil C18 PPN, Macherey & Nagel, 2 x250 mm, 5 µm,
`100 A, buffer A: 0.06% trifluoroacetic acid; buffer B: 0.06%
`trifluoroacetic acid in 80% acetonitrile; flow rate, 0.2 mL/
`min; UV detection at 215 nm; peptide concentration, 1 µg/
`10 µL). Polarimetry was carried out on a Perkin-Elmer 241
`polarimeter. Sample concentration was 1 mg/mL (0.60 mM)
`using 10-cm cells. The optical rotation at 18 °C was detected
`at 5 89 nm (Na), 5 78 nm (Hg), 5 4 6 nm (Hg), 436 nm (Hg)
`and 365 nm (Hg). Starting from pure isoforms the changes
`of the optical rotation were examined for 120 h. For the first
`24 h, the samples were stored at room temperature; from
`24 to 120 h, the samples were incubated at 3 7 °C between
`the measurements.
`
`NMR spectroscopy
`
`Two-dimensional NMR spectra were obtained on commer
`cial Bruker AMX6oo and AMX400 spectrometers at n°C by
`standard methods (25, 26). Peptide concentrations were: 1.5
`mM, pH 3.9 (uroguanylin-24); 4.9 mM, pH 3.3 (uroguanylin-
`16, isomer A); and 4.1 mM, pH 3.3 (uroguanylin-161 isomer B),
`in H2O/D 2O (9:1, v/v, 500 µL). The H2O resonance was
`presaturated by continuous coherent irradiation at the H2O
`resonance frequency before the reading pulse, except the JR
`NOESY. The sweep widths in w1 and CO2 were 9 ppm (5434.8
`Hz on AMX6oo and 3 623.2 Hz on AMX400 spectrometer).
`Quadrature detection was used in both dimensions with the
`time proportional phase incrementation (TPPI) technique in
`co,. In CO2 4000 data points were collected, and 512 data points
`in co1 • Zero filling to 1000 data points was used in w1. Spec
`tra were multiplied with a squared sinebell function phase
`shifted by rr,/41 rc/3, and rc/2 for the NOESY spectra, by rc/4
`for the Clean-TOCSY and the DQF-COSY spectra. Baseline
`and phase correction of the sixth order was used. Data were
`evaluated on X-Window workstations with the NDee pro
`gram package (Software Symbiose, Bayreuth). For the se
`quence-specific assignment of spin systems and the
`evaluation of the NOESY distance constraints for the pep
`tides, data from the following 600 MHz spectra were used:
`DQF-COSY, Clean-TOCSY with a mixing time of So msec,
`NOESY with a mixing time of 200 msec and JR-NOESY with
`
`U.C. Marx et al . Structure and dynamics of uroguanylin isomers
`
`a mixing time of 200 msec. For estimation of the coupling
`constants, DQF-COSY spectra were recorded with 8000 data
`points in w2 and 512 data points in co, and were processed
`without window function. Coupling constants were mea
`sured between the antiphase peaks of the resonances using
`a Lorentzian function for peak fitting. The time dependent
`10 NMR spectra were obtained from the AMX400 spectrom
`eter (400 MHz) at 2 5 °C. For the first 24 h, samples were
`stored at room temperature between the measurements;
`from 24 to 92 h, samples were stored at 3 7 °C.
`
`Structure calculations and analysis
`
`The total number of nontrivial unambiguous NOESY cross
`peaks used for structure calculation was 84 for the A form
`of uroguanylin-16 and 69 for the B form (Table 1). These
`cross peaks were divided into three groups according to
`their relative intensities: strong, 0.2 to 0.3 nm; medium,
`0.2 to 0-4 nm; weak, 0.2 to 0.5 nm. In a pseudoatom ap
`proach, 0.05 nm was added to the upper distance limit for
`distances involving unresolved methyl or methylene pro
`ton resonances. Also included by an iterative strategy were
`s <\> and 3 x' dihedral angle restraints for isomer A and 4 (\l
`and 2 x' dihedral angle restraints for the B form. Deviations
`of 30° from the calculated angles were allowed in the cal-
`
`
`t..,%�,� '$ • Jf�l�f9)t � �ntrH.>�(h:.<�S.: ���� $:#t�Ni ft<.'>�'N :S�«��t���
`
`
`
`
`
`
`t��t��:�i!!�t�:�$���,�!.!:!,;t����tt�}���!l�\t��,�t�!����ttt��tt�,,��"�����-
`
`� f·f ·:�: {� ,
` : �j{:��··:;�
`H··] t· :-.�-':f
`�t·n>, . :�
`,f
`:: g .
`
`�:-,:_,�$
`'-�-=�
`�:}.�;�
`� . t:$
`
`�4:ii
`:f{��
`·-t�%
`�t���
`�t�ii
`
`t:�t?. �:..:. B)>�����
`�... ��� f: tfr>���
`)�;.�}�'
`:· �?i:�>-.l�:·S�}�
`1Ji /·��···����1
`
`
`�}.:£f)�f'
`
`-:\;< :) �;:
`�f���tr
`
`~5/<� ���""" """ �"""�,� �� A*<"""* ��- �_ .,�
`
`����i�i�+i��+�fii+������~.�-
`
`. ,.,. ••• i�t
`
`::�:t{������
`Jh� ��}l:��$ ��h��t�}��t���J�f
`
`
`··;f:::=�t�::�\����::t ::_.:_����:t�*�����
`
`-��-��=-��,����������:j:::::::::::4.� ..:::.,.,-----,--------,_,,.-
`
`f. Peptide
`Res. ;,, 1998 / 229-240 I 2 3 l
`
`Bausch Health Ireland Exhibit 2010, Page 3 of 12
`Mylan v. Bausch Health Ireland - IPR2022-00722
`
`
`
`A
`
`41
`Glyl
`
`Vall O
`
`I
`
`Thr1 3
`Gius
`�la!
`
`I
`
`I
`
`Leu6
`
`3.8
`ijva
`18
`4.0
`I I
`
`4.2 [
`
`4.4 8
`
`eul 6
`C)'
`
`Asp3
`
`�
`
`C s l 2
`
`4.6
`
`Cysl 5
`
`9
`Asn
`
`4.8
`
`'
`
`'
`
`As
`
`I p2
`-
`
`
`
`
`
`isomers U.C. Marx et al . Structure and dynamics of uroguanylin
`
`culation without penalty. Structure calculations were per
`formed using a modified ab mitio SA protocol with an ex
`tended version of X-PLOR 3.1 program package (27). The
`structure calculation included floating assignments of
`prochiral groups (28) and a reduced presentation for
`nonbonded interactions for part of the calculation (29) to
`increase efficiency. The calculation strategy is similar to
`those described previously (30, 31). Structure parameters
`were extracted from the standard files parallhdg.pro and
`topallhdg.pro of X-PLOR V3.840 (27). The disulfide bonds
`were included explicitly. For each fragment, 30 structures
`were calculated and 10 structures for each isoform were se
`lected with the criterion of the lowest overall energy for
`further characterization. SYBYL 6.o (TRIPOS), RASMOL V.
`2.6 (32) and MOLSCRIPT (33) were used for visualization
`of the structure data. To elucidate the stability of the struc
`tures, we calculated local RMSD using SYBYL 6.o as well
`as XPLOR 3.1. The geometry of the structures was analyzed
`using PROCHECK (34)1 PROMOTIF (35) and XPLOR 3.1.
`
`cG MP T84 cel l assay
`
`Synthetic peptides were tested to assess the specific CC
`C-stimulating potency as described previously (e.g. 10, 21).
`T84 cells were preincubated with 1 mM isobutylmethyl
`xanthine for 5 min. Then, peptides were added to the me
`dium in a concentration range of 10-9 to 10-6 M and cells
`were incubated for 60 min. Incubation was stopped by re
`moval of medium and addition of ice cold ethanol. The
`amount of intracellular cGMP induced by the peptides was
`determined using a specific radioimmunoassay (36).
`
`
`
`Results and Discussion
`
`The isomers of uroguanylin-16 and the amino-terminally
`extended uroguanylin-24 were synthesized and separated
`using standard chromatography techniques as described
`under "Experimental Procedures". The GC-C-activating
`and earlier eluting compound during HPLC is referred to
`as isomer A.
`
`Chemical shift ana lysis
`
`With very little resonance overlap in the 2D NMR spectra
`of the separated isoforms of uroguanylin-16 (Fig. 2), the se
`quence-specific assignment of each species could be per
`formed by standard methods (26, Table 2). The amide
`proton chemical shifts of Cys7, Val8, Cys12 and Thn3 dif-
`2 3 2 I /. Peptide
`
`Res. 52, ,998 / 2 29-240
`
`9.0 8.8 8.6 8 .4 8.2 8.0 7.8 7.6
`ro2 (ppm)
`
`B
`
`Gly l 4
`
`Vall o
`Valf
`
`' ,
`
`0
`
`" ll
`
`unp
`
`A.la '1
`
`1 Lcµ6
`j I eul 6
`
`0
`
`Thrl3
`
`3 . 8
`
`4.0
`
`4.2 [
`
`4.4 8
`
`•
`
`Cys t
`
`1Cysl 5
`
`As p2
`
`1L ,.
`
`Cys4
`
`C s
`7
`Asn9 y
`
`4.6
`
`4.8
`
`�
`
`9.0 8 . 8 8.6 8.4 8.2 8.0 7.8 7.6
`co 2 (ppm)
`
`Figure 2 . NOESY spectra ( 200 msec) of uroguanylin 16 isomers.
`
`
`
`
`
`
`Fingerprint region of (A) isomer A and (BJ of isomer B. The
`
`intraresidual cross peaks between the amideand Ca-protons of
`
`
`each amino acid are labeled and the chain tracing along the
`
`sequence is indicated.
`
`fer by up to 1 ppm between the two isoforms, indicating
`that the backbone folds in these regions are different (Fig.
`3A). The differences in the chemical shifts of the Ca-pro
`ton resonances with a maximum of 0.3 ppm are less sig
`nificant, but obvious for the residues Cys7 and Cys12 (Fig.
`3B). Amide proton chemical shifts of the first six and Ca.
`proton chemical shifts of the first three residues are virtu
`ally identical for both isoforms, indicating that the
`backbone folds of the amino-terminus of both isoforms are
`similar. In the 2D NMR spectra of a 1:1 mixture of both iso
`mers of the NH 1-terminally extended uroguanylin-24 (data
`
`Bausch Health Ireland Exhibit 2010, Page 4 of 12
`Mylan v. Bausch Health Ireland - IPR2022-00722
`
`
`
`U.C. Marx et al . Structure and dynamics of uroguanylin isomers
`
`,twJ�
`�>*·\�
`
`not shown) only one set of spin systems was found for the
`residues Phe1 to Aspn, indicating that the u amino-termi
`nal amino acids have the same conformation, most prob
`ably unstructured or extended, as judged from their
`Ca-proton chemical shifts. For the 13 carboxy terminal
`amino acids, an unambiguous and independent assignment
`of resonances of the A and B forms was possible because
`of clear differences in the chemical shifts of both isoforms.
`The backbone chemical shift values of the last 13 amino
`acids comprising the Cys-rich region of uroguanylin-16 and
`-24 differ by less than 0 . 0 8 ppm for the corresponding
`isoforms with one exception, indicating that the isoforms
`of uroguanylin 24 most probably have a tertiary fold which
`closely resembles the corresponding isomers of uro
`guanylin-16. Thus, the additional 8 amino acids at the
`amino terminus of uroguanylin-24 apparently do not influ
`ence the global fold of the backbones of the two isomers.
`
`
`
`
`
`tocol (27) was used with explicit inclusion of disulfide
`bonds. None of the ten selected structures shows NOE vio
`lations more than 0.03 nm, and no structure has angle vio
`lations more than 5 °. Although only a few nonsequential
`NOEs were assigned, and no typical NOEs and no consecu
`tive 3 f NHcxH coupling constants for any regular secondary
`structure element were found, the NMR data were suffi
`cient to define well structured global folds for both
`isoforms. The well-defined global fold of the region from
`Cys4 to Cys15 is reflected by the low backbone root-mean
`square deviation (RMSD) of 0.07 5 nm for the A form and
`0.063 nm for the B form (Fig. 4A). The structures of the iso
`mers A and B of uroguanylin are rather different: the back
`bone RMSD between the average structures of the A and B
`forms from residue Cys4 to Cys15 is 0.46 nm. For each frag
`ment, the ten final structures have been deposited in the
`Brookhaven Protein Databank (accession numbers 1UYA
`and 1UYB).
`The structure of the A form of uroguanylin-16 contains
`Structure calculation and analysis of uroguanylin-16 isomers
`three loops, Cys4 to Cys7, Cys7 to Cys12 and Cys12 to
`Cys15 . These loops are arranged in a right-handed spiral
`which is stabilized by disulfide bonds Cys4-Cys12 and Cys7
`Cys15 (Fig. 4B, C). A view along the spiral axis shows that
`the connected cysteines are in line after one complete spi
`ral turn. The three amino-terminal amino acids are unstruc
`tured. No elements of regular secondary structure were
`
`
`
`Eighty-four and sixty-nine distance restraints from the 200
`msec-NOESY spectra at 11 °C were collected for uro
`guanylin-16 isomers A and B and were used together with
`8 and 6 dihedral restraints, respectively, in restrained MD
`calculations (Table 1). For structure calculations of both
`isoforms, a modified ab initio simulated annealing ISA) pro-
`
`
`
`/. Peptide Res. s>, 1998 / 229-240
`
`I 2 3 3
`
`Bausch Health Ireland Exhibit 2010, Page 5 of 12
`Mylan v. Bausch Health Ireland - IPR2022-00722
`
`
`
`U.C. Marx et al . Structure and dynamics of uroguanylin isomers
`
`A
`
`"' 0.5
`•
`z
`
`0.0
`
`· -0.5
`z
`
`B
`
`0.2
`
`0.1
`
`"
`
`-0.1
`
`-0.2
`
`::r:
`tl
`IQ -0.3
`
`Sequence position
`
`2 4
`
`10 1 2 14
`8
`6
`Sequence position
`
`1 6
`
`Figure 3. Chemical shift differences of the backbone protons of
`uroguanylin-16 isomers. (A) Differences of the chemical shifts of
`the backbone amide protons of isomers A and B, (B) differences of
`the chemical shifts of Ca-protons of isomers A and B . The values
`were taken from Table 2.
`
`found for the ten calculated conformations except �- and
`-y-turns. For 80% of the A form structures, the three loops
`are characterized by type IV � turns having low RMSD val
`ues within the structure family. For five structures, an in
`verse -y-turn from Asn9 to Ala11 also was found. The
`structure of the 13 carboxy-terminal amino acids of the B
`form can be depicted as a distorted left-handed spiral (Fig.
`4B, C). Where Cys4 and Cys7 are superimposed with their
`corresponding half-cystines Cys12 and Cys1 5 , the loop be
`tween Cys4 and Cys7 is parallel to the backbone segment
`from Cys12 to Cys7. As for the A form, the three amino
`terminal amino acids are unstructured and no regular sec
`ondary structure elements except reverse turns were
`identified for the 10 final structures of the B form. All 1 0
`structures show type IV �-turns between Cys4 and Cys7
`and between Cys12 and Cys1 5 with comparable backbone
`angles within the structure family.
`With a backbone RMSD of 0-46 nm for residues Cys4
`to Cys15 between the average structures of A and B forms,
`
`
`
`234 I T- Peptide Res. S l, 1998 / 229-240
`
`the backbone folds of the two isomers differ significantly.
`To identify the segments with the highest degree of differ
`ences in backbone fold between the two isomers, we cal
`culated the local RMSD values between the two average
`structures using a five-amino acid window ( 3 7, Fig. 5 ) . The
`RMSD values around Cys4 and the loop around Asn9 and
`Vaho are reduced, indicating that the local structures of the
`A and B forms of uroguanylin-16 resemble each other in
`these sequence regions. From Leu6 to Val8 and from Alan
`to Thn3 the local RMSD is increased, indicating that the
`backbone folds of the A and B forms differ most signifi
`cantly within this region. The two segments of higher
`RMSD comprise the residues Cys7 and Cys12, for which sig
`nificant differences of the (j> and 'l' angles between the two
`isoforms also were found. Thus, the local backbone folds
`within the three loops are less different for the two
`isoforms, but their arrangement determines the overall
`backbone fold, mostly restricted by the backbone direction
`around Cys7 and Cys12. This feature also is reflected by the
`high differences in the chemical shifts of these residues be
`tween the A and B forms (Fig. 3 ) .
`
`Structure comparison with guanyli n a n d E . coli enterotoxin ST
`
`Uroguanylin shows a high sequence homology with the
`related peptide guanylin (Fig. 1) and has a similar ability to
`stimulate GC-C (3, 1 5 ) . The chemical shifts of the backbone
`protons of the cysteines and the following residues of
`uroguanylin resemble those of the guanylin containing 13
`
`amino acids found by Skelton et al. (16 ) and guanylin con
`
`taining 17 amino acids (23). This phenomenon is most strik
`mg for Cys7 and Cys12 and their consecutive residues. The
`backbone RMSD values for the cysteine rich region be
`tween the average structure of uroguanylin-16 isoforms and
`the structures of guanylin-13 isoforms (Brookhaven Protein
`Databank, 1GNA and 1GNB) is 0.14 nm for the A forms and
`0.15 nm for the corresponding B forms. In contrast, the
`RMSD values between the average s tructures of the
`noncorresponding isoforms are higher than 0.4 nm. Al
`though the guanylin isomers could be characterized only
`in a mixture, this strong conformity of the structures now
`renders it all but certain that guanylin isomer A is actu
`ally the GC-C-stimulating isomer.
`A comparison of the average structures of uroguanylin-
`16 isomers with the crystal structure of the toxic domain
`of the heat-stable enterotoxin ST ( Fig. 11 (221 PDB acces
`sion number 1ETN) shows that the A form of uroguanylin-
`16 closely resembles the structure of ST; the backbone
`RMSD for C y s 4 to C y s 1 5 ( numbering according t o
`
`Bausch Health Ireland Exhibit 2010, Page 6 of 12
`Mylan v. Bausch Health Ireland - IPR2022-00722
`
`
`
`
`
`U.C. Marx et al . Structure and dynamics of uroguanylin isomers
`
`
`
`
`
`Isomer A
`
`lsomer B
`
`Cys4
`
`A
`
`4. (A) Best-fit superposition from Cys4 to
`Figure
`Cys15 of the backbone atoms of the 10 final
`structures of each uroguanylin-16 isomer. (left)
`isomer A; (right) isomer B. (BJ Lowest energy
`solution structures of uroguanylin 16 isomers.
`(left) isomer A; (right) isomer B. The lower
`views were obtained from a 90° rotation about a
`horizontal axis. (C) Schematic view of the
`backbone folds and the disulfide connectivities
`of uroguanylin-16 isoforms. (left) isomer A;
`(right) isomer B. Cysteine residues are repre
`sented by encircled numbers.
`
`8
`
`COOH
`
`�
`
`C 11 ---t-t
`
`uroguanylin-16 ) between the A form of uroguanylin and
`ST is 0.11 nm, whereas the RMSD value for the B form is
`0.45 nm. The known higher activation potency of ST may
`be related to the additional disulfide bond which causes
`a higher rigidity of its three-dimensional structure and,
`thus, a possibly more efficient interaction with the recep
`tor. Structure calculations of uroguanylin-16 with an ad
`ditional distance restraint between protons that occupy
`the positions of fictitious sulfur atoms of a third disulfide
`bridge between residues 3 and 8 show that a third disul
`fide bridge is possible for the A form structure without
`distortion of the peptide backbone. The same calculation
`for the B form resulted in a higher overall energy of these
`structures and a slight violation of the additional fictitious
`
`distance restraint, indicating that a third disulfide bridge
`for the B form structure could be possible but needs a
`higher distortion of the peptide backbone ( data not shown).
`A third disulfide bond apparently would lead to a prefer
`ence of a structure similar to the A form isomer that was
`found for ST ( 2 2 ) .
`Because GC C i s not activated by uroguanylin isomer B,
`the rigid framework provided by the structures of the A
`form isomers of uroguanylin and guanylin as well as of ST
`is necessary for the binding of these ligands to the extra
`cellular domain of GC-C and for the activation of the in
`tracellular catalytic domain. The well-defined topology of
`A-type isomers, however, does not seem to be sufficient to
`activate the receptor; this was demonstrated by the chi-
`
`
`
`/. Peptide Res 52, 1998 / 229-140 I 2 3 5
`
`Bausch Health Ireland Exhibit 2010, Page 7 of 12
`Mylan v. Bausch Health Ireland - IPR2022-00722
`
`
`
`U.C. Marx et al . Structure and dynamics of uroguanylin isomers
`
`0.15
`
`0.10
`
`0.05
`
`2
`
`4
`
`6
`
`8 1 0
`Sequence position
`
`12 14
`
`Figure 5 . Local RMSD values between the A and B forms of
`uroguanylin 16 calculated with a five amino acid window (37).
`• heavy atoms; ■
`RMSD, backbone.
`
`meric peptide EDPGTCEICVNVACTGC investigated in
`our laboratory (15 ). Here, the central residues AYA of
`guanylin were replaced by VNV of uroguanylin. Showing a
`guanylin-like isomerism, this derivative generated two iso
`mers, as detected by HPLC and 2D NMR spectroscopy, but
`had an unexpectedly poor potency to stimulate GC-C in a
`minimum concentration of 10-
`M. To understand the indi
`6
`vidual contribution of single amino acids for receptor bind
`ing and activation, further experiments using systematic
`substitution of amino acids contained in uroguanylin and
`guanylin are necessary.
`
`
`
`
`
`
`
`Stability and conversion of uroguanylin isomers
`
`Because of the known interconversion equilibrium of
`guanylin isomers, the stability of the separated uro
`guanylin isomers was investigated regarding the influence
`of temperature, pH and solvent. The effect of temperature
`on the stability of uroguanylin 16 isomers A and B in aque
`ous solution at pH 4.5 was determined by HPLC after dif
`ferent periods of incubation up to 7 days . The results
`obtained show that both isomers pass through a tempera
`ture-dependent conversion generating the corresponding
`stereoisomeric peptide without detectable decomposition
`or disulfide exchange at this pH (Fig. 6A, B). Integration
`of HPLC peaks obtained for isomers A and B showed that
`both isomers of uroguanylin-16 are completely stable at
`0°C, whereas a temperature of 60 °C caused an accelerated
`formation of the complementary isomer within about 4
`h. At 3 7 °C, 2 5 % of both uroguanylin-16 isomers are inter
`converted within 24 h. This and conversion experiments
`carried out at higher temperatures resulted in a mixture
`containing uroguanylin isomers A and B in an approxi-
`
`r Peptide Res. 52, 1998 / 229-240
`2 3 6 I
`
`mately 1:1 ratio with a slight preference toward isomer A.
`It was not possible to shift this equilibrium ratio. Thus,
`isomer A is the thermodynamically slightly preferred mo
`lecular species of uroguanylin. From the chromatographi
`cally detected conversion, it is clear that once one of the
`uroguanylin isomers has turned into its stereoisomer, it
`is subjected to a dynamic equilibrium reconstituting the
`original isomer and vice versa. No conversion of the lyo
`philized uroguanylin isomers after long term storage of up
`to 3 months at -20 °C was observed. Further experiments
`at 3 7 °C clearly show a significant effect of the medium's
`acidity on the transition kinetics between the stereoiso
`meric forms of uroguanylin containing 16 and 24 amino
`acids (Fig. 6C). Although a decrease of pH below 4.5 did
`not change the kinetics of interconversion, a slight alka
`line pH adjusted either by ammonium hydrogen carbon
`ate or by the medium used for cGMP T84 cell assay
`drastically reduced the rate of interconversion. Therefore,
`in principle, it is possible that, under conditions reflect
`ing the pH of blood and within the wide range of mucosal
`acidity (21), both uroguanylin isomers may exist without
`generating a significant amount of their corresponding ste
`reoisomer. This feature clearly distinguishes uroguanylin
`from the related peptide guanylin whose isoforms have
`been shown to interconvert much more rapidly (15, 16).
`The slow development of the equilibrium b etween
`uroguanylin isomers at alkaline pH indicates that the ion
`ization state of the isomeric molecules strongly influences
`the kinetics of transition between the isome1s of
`uroguanylin 16 and uroguanylin 24. Thus, the terminal
`carboxyl, ionizable side-chains of Asp2, Asp3 and Glu5,
`or those groups able to form intrachain hydrogen bonds,
`may be involved in the control of stabilization of the two
`isomers. After 3 days at alkaline pH, both isoforms decom
`posed because of disulfide exchange. Comparison of the
`conversion of uroguanylin 16 isomers with the isomers of
`the N-terminally extended uroguanylin-24 resulted in
`identical kinetics for isoforms A and B at a pH of 4.5 and
`7. 7 for both peptides (Fig. 6C). This result clearly demon
`strates that the isomerization is not affected by the amino
`terminal region of uroguanylin. Corresponding to the
`equilibrium isomer ratio with a slightly preferred isomer
`A, B-type isomers of both peptides are converted some
`what faster than A-type isomers. The interconversion of
`uroguanylin 16 isomers also was studied by recording the
`time-dependent change of the o