THE JOURNAL OF BIOLOGICAL CHEMISTRY
`
`Vol. 267, No. 11, Issue of April 15, pp. 7402-7406,1992
`Printed in (I. S. A.
`Isolation and Characterization of Exendin-4, an Exendin-3 Analogue,
`from Heloderma suspecturn Venom
`FURTHER EVIDENCE FOR AN EXENDIN RECEPTOR ON DISPERSED ACINI FROM GUINEA PIG
`PANCREAS*
`
`John EngSBlI, Wayne A. KleinmanS, Latika Singhll, Gurcharn Singhi(, and Jean-Pierre Raufmanll
`From the $Solomon A. Berson Research Luboratov, Veterans Affairs Medical Center, Bronx, New York 10468, the $Department
`of Medicine, Mount Sinai School of Medicine, New York, New York 10029, and the I( Department of Medicine, Division of
`Digestive Diseases, State University of New York-Health Science Center, Brooklyn, New York 11203-2098
`
`(Received for publication, August 12, 1991)
`
`The recent identification in Heloderma horridurn
`gressively inhibit exendin-3-induced increases in cAMP (2).
`Because venom from a closely related lizard, Heloderma sus-
`venom of exendin-3, a new member of the glucagon
`superfamily that acts as a pancreatic secretagogue,
`pecturn, has been shown to contain helodermin (3), a peptide
`prompted a search for a similar peptide in Heloderma
`closely related in structure to helospectin (4), a search was
`suspecturn venom. An amino acid sequencing assay for
`undertaken for a His’ peptide in H. suspecturn venom that
`peptides containing an amino-terminal histidine resi-
`might be analogous to exendin-3. We report the isolation from
`due (His’) was used to isolate a 39-amino acid peptide,
`H. suspecturn venom of such an analogue that has been named
`exendin-4, from H. suspecturn venom. Exendin-4 dif-
`exendin-4. Exendin-4, unlike exendin-3, is not a pancreatic
`fers from exendin-3 by two amino acid substitutions,
`secretagogue. Instead, it interacts exclusively with the newly
`
`Gly2-GluS in place of Ser2-Asp’, but is otherwise iden-
`described exendin receptor (2) to increase pancreatic acinar
`tical. The structural differences make exendin-4 dis-
`CAMP.
`tinct from exendin-3 in its bioactivity. In dispersed
`acini from guinea pig pancreas, natural and synthetic
`MATERIALS AND METHODS
`exendin-4 stimulate a monophasic increase in CAMP
`H. suspectom venom (lots HS19SZ and HS2OSZ) was purchased
`beginning at 100 p~ that plateaus at 10 nM. The ex-
`from Miami Serpentarium Laboratories (Salt Lake City, UT). Di-
`endin-4-induced increase in CAMP is inhibited pro-
`phenylcarbamyl chloride-treated trypsin was purchased from Sigma.
`gressively by increasing concentrations of the exendin
`Endoproteinase Asp-N was purchased from Boehringer Mannheim.
`receptor antagonist, exendin-(9-39) amide. Unlike ex-
`His’ Assay and Amino Acid Sequencing-Amino-terminal amino
`endin-3, exendin-4 does not stimulate a second rise in
`acid analysis was performed by a single cycle of Edman degradation
`acinar cAMP at concentrations >lo0 nM, does not stim-
`using an automated gas phase protein sequenator in combination
`ulate amylase release, and does not inhibit the binding
`with an on-line PTH-amino acid analyzer (Applied Biosystems, Fos-
`of radiolabeled vasoactive intestinal peptide to acini.
`ter City, CA). PTH-His was positioned to elute between PTH-Ala
`This indicates that in dispersed pancreatic acini, ex-
`and PTH-dehydro-Ser. Purified peptides and peptide fragments were
`endin-4 interacts only with the recently described ex-
`sequenced with the gas phase sequenator.
`endin receptor.
`(25 mg)
`Isolation of His’ Peptides from Heloderma Venom-Venom
`was dissolved in distilled water (10 mg/ml) and passed through a CIS
`Sep-Pak cartridge (Waters Associates, Milford, MA). The Cls car-
`tridge was washed with 5 ml of water and eluted with 2 ml of 0.1%
`trifluoroacetic acid, 60% acetonitrile. Peptides in the eluate were
`separated by HPLC on an 8 mm X 10-cm pBondapak C18 Radial-Pak
`column (Waters Associates). The column was eluted with a linear
`gradient (20-60%) of acetonitrile in 0.13% heptafluorobutyric acid at
`a flow rate of 1 ml/min. One-minute fractions were collected, and
`aliquots were assayed for His’ content.
`Enzyme Cleavages-Purified exendin-4 (5-20 nmol) was digested
`with 0.2 pg of trypsin or with 0.2 pg of endoproteinase Asp-N. The
`peptide fragment exendin-(9-39) amide was prepared as described
`previously (2). Although this fragment was previously referred to as
`exendin-3-(9-39) amide (2), the name has been shortened to exendin-
`(9-39) amide to indicate that the carboxyl-terminal 31 amino acids
`of exendin-3 and exendin-4 are identical. Peptide fragments were
`purified by HPLC on a Nova Cls Radial-Pak column (Waters Asso-
`ciates).
`Amino Acid Analysis-Peptides were dried and hydrolyzed with
`gas phase 6 M HC1 at 150 “C for 60 min. Amino acids were analyzed
`with an automated amino acid derivatizer (Applied Biosystems) con-
`nected to an on-line phenylthiocarbamyl-derivative amino acid ana-
`lyzer.
`Mass Spectrometry-The mass of the COOH-terminal fragment
`generated by trypsin digestion of exendin-4 was determined by fast
`atom bombardment-mass spectrometry. Mass accuracy of greater
`than kO.1 unit was achieved by peak matching to appropriate cesium
`chloride cluster ions. Fast atom bombardment-mass spectrometry
`was performed by the Laboratory for Macromolecular Analysis at the
`
`An assay for His’ peptides was recently used to identify the
`presence of helospectin and a new, previously unrecognized
`His1-Phe‘ peptide in Heloderma horridurn venom (1). This
`new peptide, designated exendin-3, is a pancreatic secreta-
`gogue. At concentrations greater than 100 nM, exendin-3
`interacts with VIP’ receptors on guinea pig pancreatic acini
`to stimulate an increase in cellular cAMP and amylase release
`(2). At lower concentrations (0.1-10 nM), however, exendin-3
`interacts with a putative exendin receptor that causes an
`increase in acinar cAMP but not amylase release. This con-
`clusion is based on the observation that increasing concentra-
`tions of a specific antagonist, exendin-3-(9-39) amide, pro-
`* This paper was supported in part by the Department of Veterans
`Affairs. The costs of publication of this article were defrayed in part
`by the payment of page charges. This article must therefore be hereby
`marked “aduertisement” in accordance with 18 U.S.C. Section 1734
`solely to indicate this fact.
`ll To whom reprint requests and correspondence should be ad-
`dressed Solomon A. Berson Research Laboratory, Veterans Affairs
`Medical Center, Bronx, NY 10468. Tel.: 212-584-9000 (ext. 6069).
`’ The abbreviations used are: VIP, vasoactive intestinal peptide;
`PTH, phenylthiohydantoin; HPLC, high performance liquid chro-
`matography.
`
`7402
`
`SANOFI-AVENTIS Exhibit 1025 - Page 7402
`
`IPR for Patent No. 8,951,962
`
`

`
`7403
`
`Purification, Structure, and Bioactivity of Exendin-4
`RESULTS
`Albert Einstein College of Medicine using a Finnigan MAT-90 mass
`spectrometer.
`Similar to the findings in H. horridurn venom, two His'
`Dispersed Pancreatic Acini-Dispersed acini from guinea pig pan-
`peptides, one 5-10-fold more abundant than the other, were
`creas were prepared by digestion with collagenase and incubated with
`detected in H. suspecturn venom. The smaller, earlier eluting
`VIP, natural exendin-4, natural exendin-3, or synthetic exendin-4.
`His' peptide peak (Fig. 1) was determined to be helodermin
`The incubations were performed in the absence or presence of in-
`by sequence analysis (data not shown). The major His' pep-
`creasing concentrations of exendin-(9-39) amide. Amylase release,
`tide peak in Fig. 1 proved to be a His'-Phe' peptide analogue
`cellular CAMP, and binding of 'zsI-VIP were measured as described
`previously (2).
`of exendin-3.
`The new peptide has been named exendin-4. Its amino acid
`Peptide Synthesis-Exendin-4 and exendin-(9-39) amide were syn-
`thesized on solid phase support (5-(4-(9-fluorenylmethyloxycarbnyl)-
`sequence was determined by direct sequencing of the intact
`aminomethyl-3,5-dimethoxyphenoxy)valeric acid resin) using acti-
`peptide and by sequence analysis of overlapping peptide frag-
`vated Fmoc (N-(9-fluorenyl)methoxycarbonyl) amino acids on a Mil-
`ments generated by digestion with trypsin. The COOH-ter-
`minal trypsin fragment was further analyzed by mass spec-
`ligen 9050 peptide synthesizer (Milligen, Burlington, MA). Cleavage
`trometry. Sequence analysis of exendin-4 showed that it con-
`and deprotection of the peptides were performed in trifluoroacetic
`acid containing anisole, thioanisole, and ethanedithiol as scavengers.
`tains a Gly at position 2 and a Glu at position 3, but is
`The crude synthetic peptide mixtures were purified by preparative
`otherwise identical with exendin-3. The structure of the car-
`HPLC.
`boxyl-terminal trypsin fragment of exendin-4 (T4), as deter-
`mined by a combination of sequencing and mass spectrometry,
`was shown to be identical with that of exendin-3. The exper-
`imental monoisotopic mass of 1022.48 confirmed that the
`COOH terminus is amidated. Thus, exendin-4, like exendin-
`3, is a 39-amino acid peptide that has an amidated carboxyl
`terminus. Exendin-4 has a calculated mass of 4184 units. The
`full amino acid sequence of exendin-4 is shown in Fig. 2 with
`the sequences of related peptides for comparison.
`The bioactivity of the new peptide was examined using
`dispersed acini from guinea pig pancreas. In Fig, 3, the effects
`of natural and synthetic exendin-4 on acinar cAMP are com-
`pared with those of exendin-3. As described previously (2),
`exendin-3 causes a biphasic increase in CAMP. In contrast,
`the increase in cAMP observed with increasing concentra-
`tions of natural or synthetic exendin-4 is monophasic (KD 60
`= 0.2-0.4 nM). In terms of potency and efficacy, this increase
`in cAMP corresponds to the first phase of exendin-3-induced
`increases in cAMP (exendin concentrations between 0.1 and
`10 nM). As noted previously with exendin-3 (2), the first phase
`increase in cAMP is not associated with amylase release or
`inhibition of binding of radiolabeled VIP. It is the second
`phase increase in cellular cAMP that correlates with stimu-
`lation of VIP receptors. At concentrations up to 1 PM, exen-
`din-4 does not alter binding of 1261-VIP to dispersed pancreatic
`acini (Table I) and thus does not interact with VIP receptors
`to stimulate further increases in cAMP (Fig. 3) or alter basal
`amylase release (Fig. 4).
`Because the exendin fragment, exendin-(9-39) amide, is a
`specific exendin receptor antagonist (2), this fragment was
`used to determine whether exendin-4 interacts with the ex-
`endin receptor. As shown in Fig. 5 (top), increasing concen-
`
`MB Clg 0.1 % TFA/20-60% ACN 1""
`
`FIG. 1. His' peptides in 2%. suspecturn venom. Venom (25 mg)
`was concentrated on a C18 Sep-Pak cartridge that was eluted with
`0.1% trifluoroacetic acid (TFA) containing 60% acetonitrile (ACN).
`Peptides in the eluate were separated by reverse phase HPLC. Elution
`fractions were assayed for His' by amino-terminal amino acid se-
`quencing. The smaller peak of His' peptide is helodermin. The larger
`peak is exendin-4. ME, NBondapak.
`
`EXENDIN-4
`
`45
`+
`
`
`
`EXENDIN-3
`
`
`
`HELOSPECTIN as DADA L A K L @ L a K Y LBS I L G s s TOP R as s
`
`%
`Homology
`100
`
`FIG. 2. Amino acid sequence of
`exendin-4 and its comparison with
`peptide sequences of other members
`of the glucagon superfamily. Exen-
`din-4 differs from exendin-3 only by
`amino acid substitutions at positions 2
`and 3 from the amino terminus. GLP,
`glucagon-like peptide; PHI, peptide his-
`tidine isoleucine; GZP, gastric inhibitory
`peptide; GRF, growth hormone-releasing
`factor; PACAP, pituitary adenylate cy-
`clase-activating polypeptide.
`
` VIP
`
`95
`x
`20
`33
`45
`
`53
`
`26
`
`33
`
`21
`
`28
`
`10
`
`30
`5
`25
`10
`
`15
`20
`35
`40
`. +
` . .
` . . . + .
` . + .
`.
` .
` .
`. + .
` . + .
`.
`.
` . + .
`.
`. . . + .
`
`.
`.
`f
`.
`
`( H G E G T F T S D L S K Q M E E E A V R L F I E W L K N G G P S S G A P P P S I I
`H S D I G T F T S D L S K Q M E E E A V R L F I E W L K N G G P S S G A P P P S I I
`E Y
`~
`L
`
`HELODERMIN @ S D A I H E E Y m L L A K L m L Q K Y L A S I
`SECRETIN
`@ S D t ~ l E ~ R L ~ D S A R L O R L L a G ~ V I
`H S I 3 l - ] Y a Y
`L D S R R A O D ~ V Q ~ M ~ T
` Y L H G Q W A K E m A m V K m R I
`m A m l
`D E M N T I L D N L O A ~ C ~ ~ N F ~
`I a T K I T D R
`O A D W V I F T S D ] F D R L L G Q L S A K K Y L ~ S ~ I
`
`I
`
`I
` L ~
`Y T R L R K ~
`~ K K Y L N S I
`~ S D A V ~ D N
` IT^
`Y A ~
`~
`I
`~ Y D J Y O I
`
` I R ~ O D ~ V N ~ L A ~ K G K K S D W K H N
`A ~ D K
`
` L G S R T O P P P I
`
`
`
`
`
`V @ S
`
`OA DBSBS
`
`GLUCAGON
`
`GLP-1
`
`GLP-2
`
`PHI
`
`GIP
`
`GRF
`
`
`
`
`
` Y A D A I ( F N S Y R ~ V L G O L S A ~ ~ K L L ~ D IM S R Q ~ G E S N ~ E R G A R A R L I
`
`
`
`18
`
`P A C A P ~
`
`~ S D ~ I
`
`( F D S Y ~ R Y R K ~ ~ K K Y L A A V L G K R V K Q R V K N K I
`
`
`
`SANOFI-AVENTIS Exhibit 1025 - Page 7403
`
`IPR for Patent No. 8,951,962
`
`

`
`7404
`
`:: j 16
`
`Purification, Structure, and Bioactivity of Exendin-4
`
`Natural
`Exendin-3
`
`Synthetic
`Exendin-4
`
`I
`-5
`
`I
`-11
`
`I
`-10
`
`I
`-8
`
`I
`-7
`
`1
`
`,
`-6
`-9
`CONCENTRATION (log M)
`FIG. 4. Effect of natural exendin-3 and synthetic exendin-
`4 on amylase release from dispersed pancreatic acini. Acini
`were incubated with indicated concentrations of agents for 30 min at
`37 "C. In each experiment, each value was determined in duplicate,
`and results given are means f S.E. from four separate experiments.
`In these experiments, amylase release with 30 p~ carbachol, 1 pM
`VIP, and 1 p~ secretin was 14.5 f 1.5, 11.5 f 0.6, and 11.2 k 0.1%,
`respectively.
`
`F
`
`-12
`
`-11
`
`-9
`-10
`EXENDIN-4
`
`-7
`
`-8
`(log M)
`
`-6
`
`-5
`
`-11
`
`-10
`
`-9
`
`-8
`
`Exendin-4
`-7
`-6
`CONCENTRATION (log M)
`FIG. 3. Effect of natural and synthetic exendin-4 and nat-
`ural exendin-3 on cellular cAMP in dispersed pancreatic
`acini. Acini were incubated with indicated concentrations of agents
`for 30 min at 37 "C. In each experiment, each value was determined
`in duplicate, and results given are means 2 S.E. from four separate
`experiments.
`
`-5
`
`TABLE I
`Effect of VZP, exendin-4, and exendin-(9-39) amide on binding of
`'*'Z- VIP to dispersed pancreatic acini
`Acini were incubated with 50 pM '2sI-VIP, alone, or in combination
`with the indicated concentrations of peptides for 30 min at 37 "C.
`Values for binding were calculated as the percentage of radioactivity
`that was saturably bound in the absence of peptide (% control).
`Nonsaturable binding, determined using 10 p~ unlabeled VIP, was
`(10% of total binding in all experiments. In each experiment, each
`value was determined in duplicate, and results given are means f
`S.E. from five separate experiments.
`'261-VIP bound
`
`Concentration
`
`100 pM
`1 nM
`10 nM
`100 nM
`1 pM
`
`VIP
`
`Exendin-l
`
`99.6 f 3.3 105.5
`82.3 f 2.1
`22.0 f 1.2
`5.6 f 1.2
`3.2 f 1.2
`
`% control
`f 2.6 97.9
`104.5 k 2.9
`104.5 f 0.6
`103.5 f 2.8
`106.6 f 2.0 106.7
`
`Erendin49-39)
`amide
`f 2.4
`100.5 f 1.8
`100.0 f 0.8
`101.7 f 0.8
`f 3.6
`
`trations of the antagonist cause a progressive rightward shift
`in the exendin-4 dose-response curve. Analysis of these data
`by the method of Schild (5) (Fig. 5, bottom) yields a single
`regression line having a slope of 1.3. By
`this method, the
`apparent affinity of the antagonist for the exendin receptor
`is 3.5 nM. These results indicate that, like exendin-3 at
`concentrations 4 0 0 nM (Z), exendin-4 interacts with exendin
`receptors on dispersed pancreatic acini to stimulate an in-
`crease in cellular cAMP but not amylase release.
`
`DISCUSSION
`We report the discovery in H. suspectum venom of a new
`bioactive, 39-amino acid peptide that is designated exendin-4
`to indicate its structural similarity to three peptides (helos-
`pectin, helodermin, and exendin-3) that were previously found
`in Helodermatidae venoms. The nearly superimposable dose-
`response curves for cAMP production with synthetic, as com-
`pared with natural exendin-4 provides further validation for
`the amino acid sequence of the new peptide.
`Exendin-4 differs from exendin-3 by two amino acid sub-
`stitutions near the amino terminus. This has a unique effect
`on the biological activity of the new peptide. Although exen-
`din4 retains the ability to interact with exendin receptors,
`its ability to interact with VIP receptors is abolished. Previous
`
`4.0 -
`3.5
`
`3,0
`
`h
`
`- I
`- - a 0, 2.0 1.5
`
`~ :
`
`2.5
`
`1.0 -
`
`y = 11 .o + 1.3x
`r = 0.98
`p < 0.002
`
`/
`
`;
`
`
`
`0.5 -
`
`a
`
`-5
`
`0.0
`-9
`
`-6
`-7
`-8
`(log M)
`EXENDIN(9-39)NHz
`FIG. 5. Effect of increasing concentrations of exendin-@-
`39) amide on exendin-4-induced increases in pancreatic aci-
`nar CAMP. Top, results for cellular cAMP measured after acini were
`incubated with the indicated concentrations of agents for 30 min at
`37 "C. In each experiment, each value was determined in duplicate,
`and results given are means f S.E. from three separate experiments.
`Bottom, data (solid symbols) are plotted by the method of Schild (5).
`Linear regression was calculated by least squares analysis (Minitab,
`Minitab Inc., State College, PA). Dose ratio (DR) is the ratio of the
`concentration of exendin-4 required to cause a half-maximal increase
`in cAMP in the presence of the indicated concentration of esendin-
`(9-39) amide to that in the absence of the antagonist.
`
`SANOFI-AVENTIS Exhibit 1025 - Page 7404
`
`IPR for Patent No. 8,951,962
`
`

`
`Purification, Structure, and Bioactivity of Exendin-4
`7405
`amide binds to exendin receptors, as evidenced by inhibition
`studies using the same cell model have shown that carboxyl-
`of the actions of exendin-3 (2) and exendin-4 (Fig. 5)). ( d )
`terminal fragments of VIP and secretin, such as VIP-(10-28)
`Differences at amino acid positions 2 and 3 between exendin-
`and secretin-(5-27), lose bioactivity but maintain their ability
`to interact with VIP receptors on pancreatic acini (6-10).
`3 and -4 do not interfere with the interaction of the peptides
`with exendin receptors (Fig. 3).
`Hence, these fragments function as VIP receptor antagonists.
`In addition to the discovery of a new bioactive peptide, this
`In contrast, structural modifications in the amino-terminal
`report provides further evidence for the existence on guinea
`portion of exendin-3 (as in exendin-(9-39) and exendin-4)
`pig pancreatic acini of an exendin receptor that mediates an
`result in the loss of both secretagogue activity and the ability
`increase in cellular CAMP. This increase in CAMP does not
`of these peptides to bind to VIP receptors. Several conclusions
`stimulate amylase release, and its function, if any, is at present
`regarding structure-function relationships of these peptides
`unknown. Currently, the only peptides that are known to
`can be drawn from these observations.
`interact with this newly discovered receptor are exendin-3
`The following are true in terms of interactions with VIP
`and -4. Since the new exendin receptor predicts the existence
`receptors. (u) In contrast to VIP and secretin, removal of
`of an endogenous mammalian ligand, we suggest that exendin-
`amino-terminal amino acids from exendin-3 (e.g. exendin-(9-
`(9-39) amide will be useful in the search for mammalian
`39) amide) abolishes the ability of the peptide to interact with
`peptides that bind to exendin receptors.
`high affinity VIP-preferring receptors that mediate the major
`increase in amylase release from guinea pig pancreatic acini
`REFERENCES
`(2, 7). (6) Likewise, amino acid substitutions at positions 2
`1. Eng, J., Andrews, P. C., Kleinman, W. A., Singh, L., and Rauf-
`and 3 in the exendin-3 sequence (i.e. exendin-4) is sufficient
`man, J.-P. (1990) J. Bwl. Chem. 265, 20259-20262
`to abolish interaction of the peptide with this same class of
`2. Raufman, J.-P., Singh, L., and Eng, J. (1991) J. Biol. Chem. 266,
`VIP receptors. ( c ) Thus, in contrast to VIP and secretin, the
`2897-2902
`peptide regions of exendin-3 and -4 that determine their
`3. Hoshino, M., Yanaihara, C., Hong, Y.-M., Kishida, S., Katsu-
`maru, Y., Vandermeers, A., Vandermeers-Piret, M.-C., Rohber-
`binding affinity for VIP receptors coincide with the peptide
`echt, P., Christophe, J., and Yanaihara, N. (1984) FEBS Lett.
`regions that activate VIP-dependent adenyl cyclase. Both
`178,233-239
`regions localize to the amino terminus.
`4. Parker, D. S., Raufman, J.-P., O’Donohue, T. L., Bledsoe, M.,
`In terms of interactions with exendin receptors the follow-
`Yoshida, H., and Pisano, J. J. (1984) J. Biol. Chem. 259,
`ing are true. (a) The amino-terminal regions of exendin-3 and
`11751-11755
`5. Schild, H. 0. (1949) Br. J. Pharmacol. 4, 277-280
`-4 that activate exendin-dependent adenyl cyclase are distinct
`6. Bissonnette, B. M., Collen, M. J., Adachi, H., Jensen, R. T., and
`from the remaining middle and carboxyl-terminal peptide
`Gardner, J. D. (1984) Am. J. Physiol. 246, G710-G717
`sequences that influence the peptides’ binding affinity for
`7. McArthur, K. E., Wood, C. L., O’Dorision, M. S., Zhou, Z . 4 . Z.,
`exendin receptors. (6) The intrinsic biological activity of
`Gardner, J. D., and Jensen, R. T. (1987) Am. J. Physiol. 252,
`exendin-3 and -4, in terms of increasing acinar CAMP, resides
`G404-G412
`8. Christophe, J. P., Conlon, T. P., and Gardner, J. D. (1976) J.
`in the amino-terminal portion of the molecule ( i e . exendin-
`Bwl. Chem. 251,4629-4634
`(9-39) amide does not stimulate an increase in CAMP). ( c )
`9. Gardner, J. D., Rottman, A. J., Natarajan, S., and Bodanszky, M.
`The amino acid sequence of exendin-3 and -4 required for
`(1979) Biochim. Biophys. Acta. 583,491-503
`binding to exendin receptors resides in the middle and car-
`10. Robberecht, P., Conlon, T. P., and Gardner, J. D. (1976) J. Biol.
`boxyl-terminal portions of the molecule (i.e. exendin-(9-39)
`Chem. 251,4635-4639
`
`SANOFI-AVENTIS Exhibit 1025 - Page 7405
`
`IPR for Patent No. 8,951,962