`
`Biochemistry 2000, 39, 8888-8894
`
`Structural Determinants for Activity of Glucagon-like Peptide-2†
`
`Mark P. DaCambra,‡ Bernardo Yusta,§,| Martin Sumner-Smith,^ Anna Crivici,^ Daniel J. Drucker,§,|,# and
`Patricia L. Brubaker*,‡,§,#
`Departments of Physiology and Medicine, UniVersity of Toronto, Toronto, Canada, M5S 1A8, NPS Allelix Inc., Mississauga,
`Ontario, Canada, L4V 1V7, and The Banting and Best Diabetes Centre, The Toronto General Hospital, UniVersity of Toronto,
`Toronto, Canada, M5G 2C4
`ReceiVed March 3, 2000; ReVised Manuscript ReceiVed May 16, 2000
`
`ABSTRACT: Glucagon-like peptide-2 (GLP-2) is a 33 amino acid gastrointestinal hormone that regulates
`epithelial growth in the intestine. Dipeptidylpeptidase IV cleaves GLP-2 at the position 2 alanine, resulting
`in the inactivation of peptide activity. To understand the structural basis for GLP-2 action, we studied
`receptor binding and activation for 56 GLP-2 analogues with either position 2 substitutions or alanine
`replacements along the length of the peptide. The majority of position 2 substitutions exhibited normal to
`enhanced GLP-2 receptor (GLP-2R) binding; in contrast, position 2 substitutions were less well tolerated
`in studies of receptor activation as only Gly, Ile, Pro, R-aminobutyric acid, D-Ala, or nor-Val substitutions
`exhibited enhanced GLP-2R activation. In contrast, alanine replacement at positions 5,6,17, 20, 22, 23,
`25, 26, 30, and 31 led to diminished GLP-2R binding. Position 2 substitutions containing Asp, Leu, Lys,
`Met, Phe, Trp, and Tyr, and Ala substitutions at positions 12 and 21 exhibited normal to enhanced GLP-
`2R binding but greater than 75% reduction in receptor activation. D-Ala2, Pro2 and Gly2, Ala16 exhibited
`significantly lower EC50s for receptor activation than the parent peptide (p < 0.01-0.001). Circular
`dichroism analysis indicated that the enhanced activity of these GLP-2 analogues was independent of the
`R-helical content of the peptide. These results indicate that single amino acid substitutions within GLP-2
`can confer structural changes to the ligand-receptor interface, allowing the identification of residues
`important for GLP-2R binding and receptor activation.
`
`Glucagon-like peptide-2 (GLP-2) is a 33 amino acid
`proglucagon-derived peptide (PGDP) secreted by the L cell
`of the intestinal epithelium (1, 2). GLP-2 augments intestinal
`hexose transport and reduces gastric emptying in rats and
`pigs, respectively, within 30-60 min following peptide
`infusion (3, 4). GLP-2 administration to rodents for several
`days produces expansion of the small bowel epithelium via
`stimulation of crypt cell proliferation and inhibition of villus
`enterocyte apoptosis (5-7). GLP-2 also prevents parenteral
`nutrition-associated intestinal hypoplasia (8), and augments
`the endogenous adaptive response to intestinal resection in
`rats following major small bowel resection (9). The reparative
`actions of GLP-2 have also been observed in models of
`intestinal inflammation in that a GLP-2 analogue, h[Gly2]-
`GLP-2, ameliorated epithelial
`injury in the small and
`
`† This work was supported in part by grants from the Medical
`Research Council of Canada (to P.L.B. and D.J.D.) and NPS Allelix
`Inc. (to D.J.D.). D.J.D. is a Senior Scientist of the Medical Research
`Council of Canada, and a consultant to NPS Allelix Inc.
`* Correspondence should be addressed to this author at Room 3366,
`Medical Sciences Building, University of Toronto, Toronto, Ontario,
`M5S 1A8, Canada. Tel. and Fax: (416) 978-2593. E-mail: p.brubaker@
`utoronto.ca.
`‡ Department of Physiology, University of Toronto.
`§ Department of Medicine, University of Toronto.
`| The Banting and Best Diabetes Centre, The Toronto General
`Hospital.
`^ NPS Allelix Inc.
`# These authors were equal contributors to this study.
`
`large intestine of mice following induction of experimental
`enteritis in vivo (10, 11).
`The diverse number of GLP-2 actions in the gastrointes-
`tinal tract remains poorly understood. A GLP-2 receptor
`(GLP-2R) has recently been cloned from hypothalamic and
`intestinal cDNA libraries and appears to be a new member
`of the glucagon/secretin 7-transmembrane, G-protein-coupled
`receptor (GPCR) superfamily (12). The GLP-2R is expressed
`in a highly tissue-specific manner, suggesting that transcrip-
`tional regulation of GLP-2R expression represents an im-
`portant control mechanism for regulating the specificity of
`GLP-2 action. Consistent with the structure of the related
`glucagon, GLP-1, and secretin receptors,
`the GLP-2R
`contains a large extracellular amino terminus thought to be
`involved in ligand binding. Analysis of GLP-2R signaling
`in fibroblasts demonstrates coupling of the GLP-2R to
`activation of adenylyl cyclase and production of cAMP (12,
`13). Although GLP-2 also activates AP-1-dependent signal-
`ing pathways, no stimulation of intracellular calcium influx
`was observed following activation of the rat GLP-2R in
`transfected BHK cells in vitro (13).
`The amino acid sequence of GLP-1 is identical in mouse,
`rat, and human species (14-16). Although not as highly
`conserved as GLP-1, the GLP-2 amino acid sequence is also
`conserved throughout vertebrate evolution [(17) and Figure
`1], with rat and human GLP-2 differing by only one amino
`acid (Thr19 in the rat versus Ala19 in human). Despite the
`emerging biological importance of GLP-2, little information
`
`10.1021/bi000497p CCC: $19.00 © 2000 American Chemical Society
`Published on Web 07/06/2000
`
`CFAD Exhibit 1020
`
`1
`
`
`
`Structure-Function Studies of GLP-2
`
`Biochemistry, Vol. 39, No. 30, 2000 8889
`
`FIGURE 1: Sequences of hGLP-21-33, GLP-17-36NH2, glucagon, hGIP, and Heloderma suspectum exendin-41-39. Amino acids conserved
`between GLP-2 and the other glucagon-related peptides are shaded.
`
`is available correlating GLP-2 structure with determinants
`of GLP-2 action. In studies analyzing the metabolism of
`circulating GLP-2, the first two amino acids (His1-Ala2) were
`found to be cleaved by the enzyme dipeptidylpeptidase IV
`(DP-IV), rendering the peptide biologically inactive (18, 19).
`Consistent with these findings, an Ala2Gly substitution in
`GLP-2 renders the peptide DP IV-resistant, and enhances
`the biological effectiveness of GLP-2 in vivo (18).
`Structure-function analyses of glucagon-like peptide
`action have identified specific residues critical for receptor
`binding and signal transduction. For example, His1, Asp9,
`and Ser16 have been defined as a catalytic triad in glucagon
`(20-23), while His1, Phe6, and Phe22-Ile23 appear to be
`important for the biological action of GLP-1 (24-26) (Figure
`1). To understand the specific structural determinants im-
`portant for GLP-2 binding and receptor activation, we have
`now carried out an analysis of position 2 analogues and an
`alanine substitution scan of the GLP-2 molecule.
`
`EXPERIMENTAL PROCEDURES
`Peptides. Wild-type rat GLP-21-33 (rGLP-2) and human
`(h) [Gly2]GLP-2 (American Peptide Co. Inc., Sunnyvale, CA)
`were used as control peptides. Exendin-4, glucagon, GLP-
`1, and GIP were from Bachem California Inc. (Torrance,
`CA); 26 hGLP-2 analogues, each with a different amino acid
`or amino acid derivative substituted for Ala2, and 31 h[Gly2]-
`GLP-2 analogues, each with an XxxnfAlan mutation, were
`synthesized by Chiron Technologies Inc. (Raleigh, NC). The
`peptide synthesis was performed on Chiron Mimitopes
`proprietary Multiple Peptide Synthesis grafted HEMA
`crowns. Free Fmoc-protected R-amino acids were in situ
`activated with DIC/HOBt using DMF as the solvent. Fmoc
`deprotection was completed after each coupling cycle using
`20% v/v piperidine in DMF. Side chain deprotection and
`cleavage were carried out using a TFA/scavenger mixture.
`Side chain deprotection byproducts were removed using an
`ether/petroleum ether/methanol solvent mixture.
`Synthetic peptides from Chiron were purified using RP-
`HPLC and a C-18 preparative HPLC column with an
`acetonitrile gradient in 1% TFA/distilled water. Quality
`control analysis was carried out for each peptide using both
`analytical HPLC and mass spectrometric analysis. Peptide
`purity ranged from 97 to 100%, with a mean purity of 99%.
`A repeat quantitative peptide amino acid analysis was also
`carried out using an aliquot from the actual peptide stock
`used in each experiment, and predicted peptide concentrations
`were corrected for actual peptide content for all full dose-
`response assays. The stability and demonstrated DP-IV
`resistance of [Gly2]GLP-2 led to its selection as the base
`peptide for the alanine scanning studies on the structural
`determinants of GLP-2 action.
`
`125I-Labeled GLP-2. 125I (2(cid:2) 108 Bq in NaOH; Amersham
`Life Science, Oakville, ON, Canada) was used to iodinate
`rGLP-2 on His1 (27). Tracer was purified by passage through
`a SepPak cartridge of C18 silica (Waters Associates, Milford,
`MA) to a specific activity of approximately 21 TBq/mmol.
`GLP-2R Binding Assay. Analysis of receptor binding was
`carried out using a membrane fraction prepared from BHK
`cells stably transfected with the rat GLP-2 receptor (rGLP-
`2R-BHK) (13). The cells were washed with PBS, and
`harvested in the same buffer used for membrane preparation.
`Ten milliliters of incubation buffer (25 mM Hepes, 140 mM
`NaCl, 0.9 mM MgCl2, 5 mM KCl, 1.8 mM CaCl2, pH 7.4,
`17 mg/L Diprotin A, and 100 (cid:237)M phenanthroline) was added
`to approximately 10 (cid:2) 107 cells and homogenized with a
`Polytron tissue homogenizer for 15 s. The homogenate was
`centrifuged at 1000g for 10 min at 4 (cid:176) C, and the supernatant
`(membrane-enriched fraction) was aliquoted and stored at
`-70 (cid:176)C until required. The protein concentration was
`determined by the method of Lowry (28).
`Polypropylene tubes (1.5 mL) were used for the receptor
`binding assay, each containing peptides dissolved in 50 (cid:237)L
`of incubation buffer. Tracer (101 kBq/50 (cid:237)L of buffer) and
`25 (cid:237)g of freshly thawed membrane protein were gently mixed
`in 400 (cid:237)L of incubation buffer, and the tubes were incubated
`with shaking for 2 h at 4 (cid:176)C. The reaction was terminated
`by centrifugation at 13000g at 4 (cid:176)C for 15 min. The
`membrane pellets were washed 3 times with 1.5 mL of cold
`50 mM Tris buffer, pH 7.4, and the amount of bound
`radioactivity was ascertained in the pellet fraction. Total
`binding was defined as the amount of 125I-GLP-2 bound to
`the membranes in the absence of nonradiolabeled GLP-2 and
`was approximately 1.3% of total 125I-GLP-2 added per tube.
`Nonspecific binding (NSB) was defined as the amount of
`125I-GLP-2 binding to the membranes in the presence of 1
`(cid:237)M nonradiolabeled GLP-2 and was consistently 0.1% of
`total binding. The percent 125I-GLP-2 specific binding for
`any given dose of unlabeled GLP-2 or GLP-2 analogue was
`therefore determined as: 100 (cid:2) [cpm bound in the presence
`of unlabeled peptide - NSB cpm]/[total cpm - NSB cpm].
`cAMP Determination. Levels of cAMP were determined
`using rGLP-2R-BHK cells in 24 well plates (13). Peptides
`were diluted in 300 (cid:237)L of DMEM containing 100 (cid:237)M
`3-isobutyl-1-methylxanthine (IBMX; Sigma Chemical Co.,
`St. Louis, MO), and incubated with the cells for 10 min at
`37 (cid:176)C. Control incubations (DMEM with IBMX in the
`absence of peptides) were carried out for each experiment.
`The reaction was stopped by the addition of 1 mL of absolute
`ethanol at -20 (cid:176) C. Cells were then homogenized, and cAMP
`levels determined in dried aliquots using a cAMP radioim-
`munoassay kit (Biomedical Technologies, Stoughton, MA).
`Protein levels were determined as above.
`
`2
`
`
`
`8890 Biochemistry, Vol. 39, No. 30, 2000
`
`DaCambra et al.
`
`FIGURE 2: GLP-2R binding (A: as a percent of total specific binding) and activation (B: as the fold increase over basal cAMP production)
`by position 2 analogues of GLP-2. Each analogue was tested in triplicate at 100, 500, and 1000 pM, and the data for the 1000 pM dose are
`shown. The rGLP-2 sequence is shown at the top of the figure.
`
`Circular Dichroism Spectroscopy. Peptides were dissolved
`in water to a final concentration of 35 (cid:237)M, and the UV
`spectra between 180 and 260 nm were determined in a
`Molecular Dichroism Spectrometer (Aviv, Lakewood, NJ)
`at room temperature using a 0.1 cm cell path length. Each
`peptide was scanned 5 times, and the absorption spectra were
`averaged. The molar ellipticity was determined as [the
`absorption at 222 nm (cid:2) 100]/[nM peptide (cid:2) the number of
`amino acids in the peptide]. The percent helicity was then
`calculated as 100 (cid:2) (the molar ellipticity/34 000), where
`34 000 represents the molar ellipticity of a peptide with 100%
`R-helicity.
`Data Analysis. Screening assays were performed in
`triplicate, while full dose-response curves were carried out
`in duplicate or triplicate in each of 3-4 different experiments.
`Data are expressed as mean ( SEM. Half-maximal inhibition
`of GLP-2 binding (inhibitory concentration or IC50), and the
`half-maximal effective concentration (EC50) and the maxi-
`mum effect (Emax) for stimulation of cAMP were calculated
`using GraphPad Prism 2.00 Software (GraphPad Software
`Inc., San Diego, CA). All peptides tested exhibited two-site
`competition binding, and it was assumed for the purposes
`of the IC50 calculations that both sites express equal
`probability of ligand binding. Because of difficulties in
`accurately assessing the high-affinity IC50 for individual
`curves, only the mean for the averaged data is reported.
`Statistical differences between groups were assessed by
`ANOVA using n-1 post-hoc custom hypotheses tests on SAS
`software (Statistical Analysis System, Cary, NC).
`
`RESULTS
`
`For analysis of the properties of the position 2 and alanine
`scanning mutants, we utilized a previously described stable
`BHK fibroblast cell line transfected with the rat GLP-2
`receptor (13). Characterization of the specificity of GLP-2R
`binding and receptor activation demonstrated that the cloned
`rat GLP-2R responds specifically to nanomolar concentra-
`tions of GLP-2, but not to equivalent concentrations of related
`members of the glucagon peptide superfamily (12, 13),
`including GLP-17-36NH2, glucagon, GIP, and exendin (data
`not shown). Following observations that the amino acid at
`position 2 was a key determinant of GLP-2 degradation and
`bioactivity in vivo (18, 19), we initially analyzed the receptor
`binding and signaling properties of a series of position
`2-substituted GLP-2 molecules. The 26 analogues with Ala2
`substitutions included all physiologic amino acids as well
`as L-pencillamine, (cid:226)-alanine, D-alanine, phosphotyrosine,
`R-aminobutyric acid, norvaline, and phenylglycine. The
`majority of position 2 substitutions exhibited normal to
`enhanced GLP-2 receptor binding, as illustrated by their
`ability to displace specific binding of 125I-GLP-2 (Figure 2A).
`In contrast, position 2 substitutions were less well tolerated
`in studies of receptor activation. Only the natural amino acids
`Gly, Ile, or Pro or the synthetic amino acids R-aminobutyric
`acid, D-Ala, or nor-Val at position 2 exhibited enhanced GLP-
`2R activation in BHK-GLP-2R cells (Figure 2B). Hence,
`although the amino acid at position 2 is not a critical
`determinant of receptor binding, coupling of the GLP-2R to
`
`3
`
`
`
`Structure-Function Studies of GLP-2
`
`rGLP-2R binding (A: as a percent of total specific
`FIGURE 3:
`binding) and activation (B: as the fold increase over basal cAMP
`production) by alanine-substituted analogues of GLP-2. Each
`analogue was tested in triplicate at 100, 500, and 1000 pM, and
`the data for the 1000 pM dose are shown. An asterisk indicates
`that Gly2 and the native alanine at positions 18 and 19 were neither
`substituted nor tested.
`
`cAMP generation is clearly sensitive to the specific residue
`present at this position of the GLP-2 molecule.
`For the alanine scan, Ala was substituted for the native
`amino acid at all positions in h[Gly2]GLP-2 except for Gly2
`and the Ala residues already at positions 18 and 19. In
`contrast to the modest effects of varying position 2 amino
`acids on GLP-2 binding, alanine replacement at positions
`5,6,17, 20, 22, 23, 25, 26, 30, and 31 led to diminished GLP-
`2R binding (Figure 3A). Several alanine-substituted mol-
`ecules, notably Thr12Ala and Asp21Ala, exhibited normal
`GLP-2R binding yet markedly reduced generation of cAMP
`(Figure 3B). Furthermore, His1Ala and Asp3Ala exhibited
`markedly reduced receptor activation despite only modest
`changes in receptor binding, further establishing the impor-
`tance of the GLP-2 amino terminus for coupling of the GLP-
`2R to adenylyl cyclase-dependent signal transduction.
`Several analogues were observed to exhibit normal to
`enhanced receptor binding and/or cAMP generating ability,
`including h[Gly2]GLP-2, h[D-Ala2], and h[Pro2]-GLP-2 (Fig-
`ure 2). A full analysis of GLP-2R binding and cAMP
`generation for these peptides over a broad range of concen-
`trations is shown in Figure 4 and Table 1. rGLP-2 displayed
`a two-site binding curve, with 50% inhibition of binding at
`2.2 pM (high affinity; IC50-1) and 49.2 ( 10.5 nM (low
`affinity; IC50-2). Receptor activation by rGLP-2, as assessed
`by cAMP generation, exhibited an EC50 of 14.0 ( 2.9 nM
`and an Emax of 28.4 ( 2.2 pmol of cAMP/mg of protein.
`h[Gly2]GLP-2 was found to bind and activate the rGLP-2R
`in a fashion similar to that of rGLP-2, with IC50-1 and -2
`
`Biochemistry, Vol. 39, No. 30, 2000 8891
`values of 4.8 pM and 42.9 ( 5.6 nM, respectively, and an
`EC50 and Emax of 9.2 ( 0.6 nM and 27.2 ( 2.8 pmol of
`cAMP/mg of protein, respectively. While the high- and low-
`affinity binding of h[D-Ala2]GLP-2 was not different from
`that of rGLP-2, the IC50-2 for h[Pro2]GLP-2 was 5-fold
`higher than that of both rGLP-2. In contrast, h[Pro2]GLP-2
`values to activate the rGLP-2R were 2-3-fold-enhanced
`compared to rGLP-2 (p < 0.01-0.001; Figure 4 and Table
`1).
`In comparison with h[Gly2]GLP-2, the base peptide used
`for the Ala scanning studies, both h[Gly2,Ala5]GLP-2 and
`h[Gly2,Ala16]GLP-2 were found to exhibit similar binding
`characteristics (Figure 4 and Table 1). Although the Emax
`was slightly reduced for h[Gly2,Ala5]GLP-2 in comparison
`with the base peptide (P < 0.01), the EC50 for h[Gly2,Ala16]-
`GLP-2 was more than 5-fold lower than that of h[Gly2]GLP-
`2 (P < 0.001; Figure 4 and Table 1).
`To establish whether the biological activities of the various
`GLP-2 analogues were affected by the R-helicity of the
`peptides, circular dichroism (CD) scanning was performed
`on rGLP-2 as well as the five analogues studied in detail.
`Two additional peptides which exhibited markedly reduced
`abilities to activate the rGLP-2R in the scanning assays,
`h[Thr2]GLP-2 and h[Gly2,Ala6]GLP-2, were also analyzed
`for comparative purposes (Figure 5). rGLP-2 exhibited a far-
`UV CD spectrum consistent with a helical content of
`approximately 11%. Although similar degrees of R-helicity
`were observed for h[D-Ala2]GLP-2, h[Pro2]GLP-2, h[Gly2,-
`Ala5]GLP-2, h[Gly2,Ala6]GLP-2, and h[Gly2,Ala16]GLP-2
`(13-15%), h[Gly2]GLP-2 and h[Thr2]GLP-2 exhibited marked
`increases in R-helicity (18 and 24%, respectively). Thus, the
`R-helical content of GLP-2 and its analogues did not appear
`to correlate with their ability to bind and/or activate the
`rGLP-2R.
`
`DISCUSSION
`GLP-2 has been shown to increase intestinal mass and
`enhance intestinal adaptation in normal rodents as well as
`in a number of pathophysiological models of intestinal
`resection and/or
`inflammation (29). Understanding the
`structural determinants of GLP-2 function may provide useful
`information for the design of more potent GLP-2 analogues
`or GLP-2 antagonists. The recent cloning of the GLP-2R
`(12) has therefore provided an opportunity to determine the
`structure/function relationships of GLP-2 molecules in vitro.
`Initial analyses of the rGLP-2R in stably expressed BHK
`cells demonstrated a high degree of specificity for GLP-2
`compared to other structurally related peptides, consistent
`with the findings of GLP-2R specificity using transfected
`COS or EBNA cells (12). The related peptides glucagon and
`GIP bound the rGLP-2R very poorly and failed to activate
`cAMP biosynthesis, while no physiologically significant
`binding of GLP-1 or the lizard GLP-1 analogue exendin-4
`was detected [(12) and our study]. Furthermore, carboxy-
`terminal extensions to the GLP-2 molecule generally retained
`the ability to bind and activate the GLP-2R, whereas amino-
`terminal truncations exhibited significant reductions in bind-
`ing and GLP-2R activation (12).
`In studies utilizing the rat GLP-2R in transfected 293-
`EBNA cells, Munroe et al. found IC50-1 (60 pM) and -2 (259
`nM) values for displacement of their tracer (125I[Tyr34]GLP-
`2) that are 5-10-fold higher than reported here with BHK
`
`4
`
`
`
`8892 Biochemistry, Vol. 39, No. 30, 2000
`
`DaCambra et al.
`
`rGLP-2R binding (A) and activation (B) by rGLP-2, h[Gly2]GLP-2, h[D-Ala2]GLP-2, h[Pro2]GLP-2, h[Gly2,Ala5]GLP-2, and
`FIGURE 4:
`h[Gly2,Ala16]GLP-2 (n ) 4-5). Membranes for binding studies were prepared from BHK cells stably transfected with the rGLP-2R, while
`cAMP were levels were determined in rGLP-2-BHK cells in culture.
`cells (12). Similarly, rGLP-2 was (cid:24)40-fold less potent in
`stimulating cAMP in COS cells. In contrast, the EC50 for
`rGLP-2 was (cid:24)0.6 nM in stably transfected 3G2R cells (12)
`and (cid:24)0.06 nM for both rGLP-2R and h[Gly2]-GLP-2 in
`studies with BHK-GLP-2R cells (13). Consistent with the
`results of our previous studies using BHK-GLP-2R cells (13),
`we did not observe significant differences between h[Gly2]-
`GLP-2 and rGLP-2 in their abilities to bind and activate the
`rGLP-2R. Hence, the minor differences in receptor binding
`and activation that have been reported to date may be
`attributable to potential differences in the different cell lines
`utilized for studies of GLP-2R signaling.
`Initial studies of GLP-2 bioactivity identified DP-IV-
`mediated cleavage at
`the position 2 alanine as a key
`
`determinant of GLP-2 biological activity in vivo. Substitution
`of the position 2 alanine with a glycine conferred resistance
`to degradation and enhanced biological potency in vivo (18,
`19, 30). The related peptides GIP and GLP-1 are also
`substrates for DP-IV-mediated cleavage, and modifications
`to the amino terminus of these peptides confer enhanced
`resistance to inactivation and substantially greater bioactivity
`in vivo (31-33). Analysis of GLP-2 molecules with position
`2 substitutions identified several amino acid substitutions that
`resulted in enhanced biological activity compared to the
`native peptide, including proline and D-alanine. Given the
`short incubation time period involved in assessment of
`peptide-stimulated cAMP formation in BHK cells and the
`observation that a position 2 proline renders a peptide highly
`
`5
`
`
`
`Structure-Function Studies of GLP-2
`
`Table 1: Summary of rGLP-2R Binding and Activation
`Characteristics for rGLP-2, h[Gly2]GLP-2, h[D-Ala2]GLP-2,
`h[Pro2]GLP-2, h[Gly2,Ala5]GLP-2, and h[Gly2,Ala16]GLP-2 (n )
`3-4)a
`
`activation
`binding
`Emax (pmol of
`EC50
`IC50-2
`IC50-1
`GLP-2
`(nM)
`cAMP/mg of protein)
`(nM)
`(pM)
`analogues
`28.4 ( 2.2
`49.2 ( 10.5 14.0 ( 2.9
`2.2
`rGLP-2
`27.2 ( 2.8
`42.9 ( 15.6 9.2 ( 0.6
`h[Gly2]
`4.8
`18.9 ( 1.3c
`46.3 ( 10.9 7.2 ( 0.3c
`h[D-Ala2]
`3.1
`21.0 ( 1.2b
`228.0 ( 6b
`5.1 ( 0.3d
`h[Pro2]
`6.6
`18.0 ( 0.9e
`569.0 ( 371
`5.8 ( 0.4
`h[Gly2,Ala5]
`5.7
`27.2 ( 1.2
`89.0 ( 32
`1.7 ( 0.5f
`h[Gly2,Ala16]
`0.2
`a Data were calculated from the curves shown in Figure 4. No SEM
`was assessed for the IC50-1 due to difficulties in calculating individual
`values at the low end of each curve. bP < 0.05 vs rGLP-2. c P < 0.01
`vs rGLP-2. d P < 0.001 vs rGLP-2. e P < 0.01 vs h[Gly2]GLP-2. f P
`< 0.001 vs h[Gly2]GLP-2.
`
`FIGURE 5: Circular dichroism spectroscopy of rGLP-2, h[Gly2]GLP-
`2, h[D-Ala2]GLP-2, h[Pro2]GLP-2, and h[Thr2]GLP-2 (A) and of
`h[Gly2]GLP-2, h[Gly2,Ala5]GLP-2, h[Gly2,Ala16]GLP-2, and h[Gly2,-
`Ala6]GLP-2 (B). Each peptide was scanned 5 times between 180
`and 260 nm, and the absorption spectra were averaged.
`
`susceptible to DP-IV-mediated cleavage, it seems unlikely
`that these substitutions enhance the activity of the GLP-2
`molecule solely due to reduced degradation in vitro. Fur-
`thermore, given the complex determinants of peptide deg-
`radation and clearance and factors modulating activation of
`
`Biochemistry, Vol. 39, No. 30, 2000 8893
`
`the endogenous intestinal GLP-2 receptor in vivo, analysis
`of peptide binding and bioactivity using cell lines in vitro
`may not always correlate with studies of peptide bioactivity
`in whole animal systems in vivo.
`Analysis of alanine-substituted GLP-1 molecules did not
`reveal any GLP-1 derivatives that exhibited increased binding
`affinity or adenylate cyclase-stimulating activity in vitro (24).
`Although several alanine-substituted GLP-2 derivatives
`exhibited increased receptor binding, none of the alanine
`substitutions, with the exception of h[Gly2,Ala16]GLP-2,
`exhibited greater receptor activation in the cAMP assay
`compared to the base peptide h[Gly2]GLP-2. Despite normal
`binding characteristics, h[Gly2,Ala16]GLP-2 exhibited a
`significantly lower EC50 than h[Gly2]GLP-2. The native
`amino acid at position 16 in GLP-2 (Asn) has a large
`uncharged polar R chain, in contrast to the smaller R groups
`found in Gly16 and Ser16 of GLP-1 and glucagon, respectively
`(Figure 1). Amino acid 16 serves as part of the catalytic triad
`in glucagon, and is essential for signal transduction; replace-
`ment of Asp16 completely abrogates the glucagon signal,
`without affecting binding (34). In contrast, Gly16 does not
`appear to be crucial for the biological function of GLP-1
`(24, 25). As the Asn16Ala substitution did not appear to
`significantly alter the structure of GLP-2 (as assessed by the
`R-helical content of the peptide), these findings suggest that
`replacement of the large uncharged polar Asn with the
`similar, but smaller R chain of Ala may enhance the
`interaction of GLP-2 with amino acids in the rGLP-2R
`essential for signal transduction.
`Replacement of His1 in GLP-1 with Ala renders a 500-
`fold decrease in GLP-1 receptor binding affinity (24-26)
`while changes to the His1 in glucagon affect a modest
`decrease in binding affinity with complete abolition of
`receptor activation (21). Consistent with these findings, the
`His1Ala substitution in GLP-2 did not markedly affect
`binding affinity in the screening assay, but did decrease
`production of cAMP. Similarly, hGLP-2 analogues with
`amino-terminal extensions have a decreased ability to activate
`the rGLP-2R (12). These findings reinforce the importance
`of histidine at position 1 for signal transduction in multiple
`members of the glucagon peptide superfamily.
`Phe6 is another highly conserved amino acid in the
`glucagon family of peptides (Figure 1). This amino acid is
`important for interaction of GLP-1 with its receptor (24).
`Consistent with this finding, a Phe6Ala substitution in GLP-2
`reduced both binding and activation (although the R-helical
`content of the peptide was not altered). Interestingly, Ser5-
`Ala was also found to exhibit a reduced Emax for GLP-2
`receptor activation, consistent with findings that Thr5 is
`important for production of a (cid:226)-turn in the glucagon
`molecule, and is required for glucagon receptor activation
`(35).
`The Phe-Ile residues at positions 22 and 23 are also
`completely retained in GLP-2, GLP-1, and exendin-4, and
`are conservatively substituted by Phe-Val in glucagon and
`GIP (Figure 1) (15, 36-38). Alanine substitution at either
`of these positions in GLP-1 (24, 25) and GLP-2 (present
`study) diminishes receptor activation, and it has been
`suggested that this sequence is important for the conformation
`of GLP-1. Similarly, the Trp-Leu sequence at positions 25
`and 26 in GLP-2 is completely conserved in GIP, GLP-1,
`exendin-4, and GLP-2, and alanine substitution of these
`
`6
`
`
`
`8894 Biochemistry, Vol. 39, No. 30, 2000
`
`amino acids in both GLP-1 (24, 25) and GLP-2 (this study)
`reduced receptor activation in vitro.
`In summary, we define here the importance of specific
`amino acid residues for GLP-2 receptor binding and activa-
`tion in vitro. The results of these studies may be useful for
`future attempts designed to generate more potent agonists
`of GLP-2. Furthermore, the observation that several GLP-2
`derivatives exhibit normal binding but markedly reduced
`receptor activation suggests that these molecules should be
`examined for their ability to function as GLP-2 antagonists
`both in vitro and in vivo.
`
`ACKNOWLEDGMENT
`
`We acknowledge the assistance of F. Wang and A. Izzo,
`and the laboratory of Dr. A. Davidson, University of Toronto,
`for technical assistance with the CD scans.
`
`REFERENCES
`
`1. Mojsov, S., Heinrich, G., Wilson, I. B., Ravazzola, M., Orci,
`L., and Habener, J. F. (1986) J. Biol. Chem. 261, 11880-
`11889.
`2. Dhanvantari, S., Seidah, N. G., and Brubaker, P. L. (1996)
`Mol. Endocrinol. 10, 342-355.
`3. Cheeseman, C. I., and Tsang, R. (1996) Am. J. Physiol.
`Gastrointest. LiVer Physiol. 271, G477-G482.
`4. Wojdemann, M., Wettergren, A., Hartmann, B., and Holst, J.
`J. (1998) Scand. J. Gastroenterol. 33, 828-832.
`5. Drucker, D. J., Ehrlich, P., Asa, S. L., and Brubaker, P. L.
`(1996) Proc. Natl. Acad. Sci. U.S.A. 93, 7911-7916.
`6. Tsai, C.-H., Hill, M., and Drucker, D. J. (1997) Am. J. Physiol.
`272, G662-G668.
`7. Tsai, C.-H., Hill, M., Asa, S. L., Brubaker, P. L., and Drucker,
`D. J. (1997) Am. J. Physiol. 273, E77-E84.
`8. Chance, W. T., Foley-Nelson, T., Thomas, I., and Balasubra-
`maniam, A. (1997) Am. J. Physiol. 273, G559-G563.
`9. Scott, R. B., Kirk, D., MacNaughton, W. K., and Meddings,
`J. B. (1998) Am. J. Physiol. 275, G911-G921.
`10. Boushey, R. P., Yusta, B., and Drucker, D. J. (1999) Am. J.
`Physiol 277, E937-E947.
`11. Drucker, D. J., Yusta, B., Boushey, R. P., Deforest, L., and
`Brubaker, P. L. (1999) Am. J. Physiol. 276, G79-G91.
`12. Munroe, D. G., Gupta, A. K., Kooshesh, P., Rizkalla, G.,
`Wang, H., Demchyshyn, L., Yang, Z.-J., Kamboj, R. K., Chen,
`H., McCallum, K., Sumner-Smith, M., Drucker, D. J., and
`Crivici, A. (1999) Proc. Natl. Acad. Sci. U.S.A. 96, 1569-
`1573.
`13. Yusta, B., Somwar, R., Wang, F., Munroe, D., Grinstein, S.,
`Klip, A., and Drucker, D. J. (1999) J. Biol. Chem. 274, 30459-
`30467.
`14. Rothenberg, M. E., Eilertson, C. D., Klein, K., Zhou, Y.,
`Lindberg, I., McDonald, J. K., Mackin, R. B., and Noe, B. D.
`(1995) J. Biol. Chem. 270, 10136-10146.
`
`DaCambra et al.
`
`15. Heinrich, G., Gros, P., Lund, P. K., Bentley, R. C., and
`Habener, J. F. (1984) Endocrinology 115, 2176-2181.
`16. Bell, G. I., Sanchez-Pescador, R., Laybourn, P. J., and Najarian,
`R. C. (1983) Nature 304, 368-371.
`17. Irwin, D. M., and Wong, J. (1995) Mol. Endocrinol. 9, 267-
`277.
`18. Drucker, D. J., Shi, Q., Crivici, A., Sumner-Smith, M., Tavares,
`W., Hill, M., Deforest, L., Cooper, S., and Brubaker, P. L.
`(1997) Nat. Biotechnol. 15, 673-677.
`19. Brubaker, P. L., Crivici, A., Izzo, A., Ehrlich, P., Tsai, C.-H.,
`and Drucker, D. J. (1997) Endocrinology 138, 4837-4843.
`20. Unson, C. G., Macdonald, D., Ray, K., Durrah, T. L., and
`Merrifield, R. B. (1991) J. Biol. Chem. 266, 2763-2766.
`21. Unson, C. G., Macdonald, D., and Merrifield, R. B. (1993)
`Arch. Biochem. Biophys. 300, 747-750.
`22. Unson, C. G., Wu, C.-R., Fitzpatrick, K. J., and Merrifield,
`R. B. (1994) J. Biol. Chem. 269, 12548-12551.
`23. Unson, C. G., Wu, C.-R., and Merrifield, R. B. (1994)
`Biochemistry 33, 6884-6887.
`24. Adelhorst, K., Hedegaard, B. B., Knudsen, L. B., and Kirk,
`O. (1994) J. Biol. Chem. 269, 6275-6278.
`25. Gallwitz, B., Witt, M., Paetzold, G., Morys-Wortmann, C.,
`Zimmermann, B., Eckart, K., Fsˇlsch, U. R., and Schmidt, W.
`E. (1994) Eur. J. Biochem. 225, 1151-1156.
`26. Hareter, A., Hoffmann, E., Bode, H. P., Goke, B., and Goke,
`R. (1997) Endocr. J. 44, 701-705.
`27. Browne, C. A., Bennett, H. P., and Solomon, S. (1981)
`Biochemistry 20, 4538-4546.
`28. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R.
`J. (1951) J. Biol. Chem. 193, 265-275.
`29. Drucker, D. J. (1999) Trends Endocrinol. Metab. 10, 153-
`156.
`30. Drucker, D. J., Deforest, L., and Brubaker, P. L. (1997) Am.
`J. Physiol. 273, G1252-G1262.
`31. Kieffer, T. J., McIntosh, C. H. S., and Pederson, R. A. (1995)
`Endocrinology 136, 3585-3596.
`32. O’Harte, F. P., Mooney, M. H., and Flatt, P. R. (1999) Diabetes
`48, 758-765.
`33. Deacon, C. F., Knudsen, L. B., Madsen, K., Wiberg, F. C.,
`Jacobsen, O., and Holst, J. J. (1998) Diabetologia 41, 271-
`278.
`34. Unson, C. G., and Merrifield, R. B. (1994) Proc. Natl. Acad.
`Sci. U.S.A. 91, 454-458.
`35. Unson, C. G., Gurzenda, E. M., Iwasa, K., and Merrifield, R.
`B. (1989) J. Biol. Chem. 264, 789-794.
`36. White, J. W., and Saunders, G. F. (1986) Nucleic Acids Res.
`14, 4719-4730.
`37. Chen, Y. E., and Drucker, D. J. (1997) J. Biol. Chem. 272,
`4108-4115.
`38. Tseng, C. C., Jarboe, L. A., Landau, S. B., Williams, E. K.,
`and Wolfe, M. M. (1993) Proc. Natl. Acad. Sci. U.S.