`
`'H Nuclear-Magnetic-Resonance Studies
`of the Molecular Conformation of Monomeric Glucagon
`in Aqueous Solution
`
`Chris BOESCH, Arno BUNDI, Max OPPLIGER, and Kurt WUTHRICH
`Institut fur Molekularbiologie und Biophysik, Eidgenossische Technische Hochschule, Zurich-Honggerberg
`
`(Received May 26, 1978)
`
`Dilute aqueous solutions of glucagon were investigated by high-resolution 'H nuclear magnetic
`resonance at 360 MHz. Monomeric glucagon was found to adopt predominantly an extended
`flexible conformation which contains, however, a local non-random spatial structure involving the
`fragment - Phe-22 - Val-23 - Gln-24 - Trp-25 - . This local conformation is preserved in the partial
`sequence 22 - 26 and could thus be characterized in detail. Two interesting conclusions resulted
`from these experiments. One is that the local spatial structure in the fragment 22-25 of glucagon
`is identical to that observed in the fragment 20 - 23 of the human parathyroid hormone. Secondly,
`the backbone conformation in the C-terminal fragment of glucagon in solution must be different
`from the a-helical structure observed in single crystals of glucagon. These new structural data are
`analyzed with regard to relationships with glucagon binding to the target cells.
`
`Glucagon is a polypeptide hormone which consists
`of one linear peptide chain with 29 amino acid residues
`[I]. The biological function of glucagon was found to
`be related with specific binding to a plasma membrane
`receptor, which results in stimulation of adenylate
`cyclase [2]. To gain more detailed insight into the
`structurc-function relations, the molecular conforma-
`tion of glucagon was extensively investigated, both in
`single crystals and in solution. In the single crystal
`X-ray structure, glucagon trimers were observed,
`where the individual molecules adopt a mainly a-heli-
`cal conformation [3]. Glucagon in solution was in-
`vestigated by various spectroscopic techniques, such
`as circular dichroism [4 - 71, optically detected mag-
`netic resonance [8] and nuclear magnetic resonance
`(NMR) [9- 111. Two observations stand out among
`the results of the solution studies. One is that the
`conformational properties of glucagon manifested in
`the spectral parameters depend strongly on the solution
`conditions. Thus it was reported that glucagon in
`freshly prepared dilute aqueous solution adopts pre-
`dominantly a flexible 'random coil' form, while under
`different solution conditions or simply after prolonged
`standing of the solutions more highly structured
`aggregates were observed [4 - 81. Conformational
`changes are apparently also induced by interactions
`
`Abbreviaticms. NMR, nuclear magnetic resonancc; Z, benzyl-
`oxycarbonyl protecting group.
`
`of glucagon with lipids or detergents [12,13]. Secondly,
`the descriptions of the solution conformations are
`throughout in the terminology of circular dichroism
`measurements, i.e. random coil structures and those
`with varying a-helix of 8-pleated sheet contents were
`distinguished [4-91.
`In view of the pronounced conformational poly-
`morphism shown by numerous investigations under
`different experimental conditions, it appears particu-
`larly important to complement the single-crystal X-ray
`structure analysis by detailed conformational studies
`in well-defined solutions. High-resolution NMR is the
`method of choice for obtaining a niany-parameter
`characterization of the solution conformation of poly-
`peptides [14]. In the earlier N M R studies of glucagon
`[9 - 111 the interpretation of the spectral data was,
`however, complicated by aggregation in the relatively
`concentrated peptide solutions or by additives used
`to prevent aggregation. In the present study the high
`sensitivity of Fourier transform 'H NMR spectro-
`scopy at 360 MHz was used to investigate dilute
`aqueous solutions of glucagon.
`At the outset of this investigation we noticed that
`the amino acid sequences of glucagon [l] and human
`parathyroid hormone [15] both contained the penta-
`peptide fragment -X-Val-Gln-Trp-Leu, with X stand-
`ing for phenylalanine in glucagon and for arginine
`in the parathyroid hormone. An extensive investiga-
`tion of the human parathyroid hormone had shown
`
`Page 1
`
`NPS EX. 2051
`CFAD v. NPS
`IPR2015-00990
`
`
`
`210
`
`that this particular pentapeptide fragment adopts a
`local non-random spatial structure, which is preserved
`also in the pentapeptide partial sequence [16,17]. Here
`the pentapeptide Phe-Val-Gln-Trp-Leu was synthe-
`sized, its solution conformation determined by 'H
`NMR and compared with the spatial structure of the
`corresponding pentapeptide fragment in the intact
`glucagon molecule. This strategy for the present study
`appeared promising with regard to both comparison
`with the corresponding experiments with parathyroid
`hormone [16,17] and the implication from earlier
`investigations that the above peptide fragment is
`essential for binding of glucagon to its receptor [18].
`
`MATERIALS AND METHODS
`Synthesis of Phe- Val-Gln- Try-Leu
`The N-protected tripeptide Z-Gln-Trp-Leu ( Z ,
`benzyloxycarbonyl) was obtained as a gift from Dr W.
`Rittel (Ciba-Geigy AG, Basel). The corresponding
`free tripeptide was obtained by hydrogenation with
`palladium on coal. The protected dipeptide Z-Phe-
`Val was purchased from Bachem AG (Liestal). The
`two peptides were coupled by the N-hydroxysuccin-
`imide active ester technique [19]. The solvent was then
`removed at room temperature in the high vacuum and
`methanol was added to the residue. The precipitate
`was filtered off and washed several times with 2-pro-
`panol followed by drying in the vacuum. The protected
`pentapeptide was homogeneous by the criteria of thin-
`layer chromatography in different systems, with RF
`= 0.81 in l/l chloroform/methanol. The free penta-
`peptide was obtained through removal of the benzyl-
`oxycarbonyl protecting group by hydrogenation with
`palladium on coal with methanol as the solvent. The
`catalyst was filtered off and the filtrate evaporated
`to dryness. Subsequently the peptide was twice lyo-
`philised from water. The product was homogeneous
`by the criteria of thin-layer chromatography in differ-
`ent systems, with RF = 0.31 in 6/4 chloroform/meth-
`anol. 'H NMR at 360 MHz showed that the product
`contained the expected amino acid composition.
`
`Preparation of the NMR Samples
`The synthetic partial sequence Phe-22 - Val-23 -
`Gln-24 - Trp-25 - Leu-26 of glucagon was studied in
`0.05 M solution in 'H20 or in a mixed solvent of 90
`H20/10 % 'HzO.
`Bovine glucagon was obtained from Calbiochem
`and used without further purification. In view of the
`previously described pronounced dependence of the
`glucagon conformation on the solvent medium [4 - 131,
`particular care was taken with the sample preparation.
`To prevent aggregation, the lowest possible peptide
`
`Glucagon Conformation in Aqueous Solution
`
`concentration for high-resolution 'H NMR was em-
`ployed, i.e. either 0.1 mM or 0.05 mM, and the ionic
`strength was kept low. For chemical shift measure-
`ments, some experiments were repeated with and
`without addition of an internal reference. The p2H
`values were adjusted by the addition of a trace of
`K 0 2 H or 2HC1 solution. pH-meter readings in 'Hz0
`solution are reported without correction for isotope
`effects [20]. p2H and temperature were selected so
`that according to previous reports monomeric forms
`of glucagon should prevail in both basic [6] and acidic
`[21] solution (Fig. 1). All the measurements were
`completed within 24 h after sample preparation. With
`these precautions, identical 'H NMR spectra could
`be reproduced in different experiments.
`Besides preventing aggregation of glucagon, special
`care in the sample preparation was also required with
`regard to optimizing the ratio of the relative intensities
`of the solute and solvent resonances. 2H20 of isotope
`purity of 99.979% was obtained from the Eidgenos-
`sisches Institut fur Reaktorforschung (Wurenlingen).
`For the crucial experiments, glucagon was twice lyo-
`philized from 2H20 to replace the labile protons with
`deuterium, and the final solution was handled under
`a dry nitrogen atmosphere, carefully degassed and
`sealed for the NMR measurements.
`It should be pointed out that the high-resolution
`'H NMR spectra present a sensitive means to monitor
`the equilibrium between monomeric and aggregated
`forms of glucagon. In solvent media where aggregation
`is expected [4- 71, pronounced line broadening was
`observed in the NMR spectra, e.g. after prolonged
`standing of the solutions [7].
`
`NMR Measurements
`360-MHz Fourier transform 'H NMR spectra
`were recorded on a Bruker HX 360 spectrometer.
`Under steady-state conditions the solvent resonance
`in the carefully deuterated and degassed samples could
`be almost completely suppressed [14]. An acceptable
`signal-to-noise ratio in 0.1 mM glucagon solutions
`was thus obtained with accumulation of 5000- 10000
`scans. For spin decoupling experiments, double-
`resonance difference spectra were recorded as de-
`scribed previously [22]. Chemical shifts are relative to
`internal sodium 3-trimethylsilyl-(2,2,3,3-2H~)propio-
`nate.
`
`RESULTS
`Selected 'H NMR parameters of the synthetic
`partial sequence 22-26 of glucagon are listed in
`Table 1. The parameters which are not shown in the
`table were found to be essentially identical to the
`corresponding random coil values [20]. In particular,
`
`Page 2
`
`
`
`C. Boesch, A. Bundi, M. Oppliger, and K. Wiithrich
`
`21 1
`
`Table 1. H N M R parameters .for valine and tryptophan in the partial sequences 22-26 of glucagon and 20-24 of human parathyroid
`hormone flS] and in intact glucagon
`The assignment of the methyl resonances is arbitrary; y A is always the resonance at higher field. Glucagon 1-29 values were measured
`
`for monomeric glucagon at two different p2H values, i.e. p2H = 2.4 and 10.8, t = 37°C. Chemical shifts 6 are given * 0.01 ppm, spin-
`spin coupling constants J * 0.5 Hz. Glucagon 22 -26 and parathyroid hormone values were measured at p2H 7.0, t = 30 "C, 6 * 0.002 pprn,
`
`J
`
`0.2 Hz
`
`Peptide
`
`Glucagon 1 - 29
`Glucagon 22 - 26
`Parathyroid hormone 20 - 24
`Random coil value [20]
`
`a Not measured.
`
`6
`
`yAC&
`
`PPm
`-
`0.75
`0.683
`0.742
`0.942
`
`Y BCH3
`
`Jao
`. _ _ _ ~ ~ _
`aCH
`
`3 J a p ~
`
`3 J a p ~
`
`0.84
`0.802
`0.845
`0.969
`
`_._____
`3.88
`3.942
`4.065
`4.184
`
`~
`
`Hz
`~ ~ ~ - ~ _ _ _
`~.
`7.9
`a
`-
`-
`8.8
`5.1
`8.0
`7.7
`5.3
`8.8
`7.8
`6.9
`6.0
`
`~~
`
`V a l - 2 3
`
`I
`
`'r
`
`n"
`
`I
`4
`6 (ppm)
`Fig. 1. 360-MHz Fourier transform ' H N M R spectra of 0.1 mM solutions of glucagon in 'HzO, t = 37°C. (A) p2H = 10.8. (B) p2H = 2.4.
`(C) Double-resonance difference spectrum obtained as the difference of two spectra recorded with and without spin-decoupling irradiation
`at 1.93 ppm. The chemical shifts of the spin system of Val-23 are also indicated
`
`I
`3
`
`I
`?
`
`I
`1
`
`I
`8
`
`I
`7
`
`"
`
`there was no evidence for intramolecular hydrogen
`bonding in the pH dependence of the amide-proton
`chemical shifts in H20 solution of the pentapeptide [23].
`Outstanding among the conformation-dependent
`spectral features of the glucagon fragment 22 - 26
`(Table 1) is the high-field shift of one of the y-methyl
`resonances of Val-23. In the 'H NMR spectrum of
`glucagon, a high-field-shifted methyl doublet reso-
`nance is also readily seen (Fig. 1). The spin decoupling
`experiment of Fig. 1 C, where irradiation at 1.93 ppm
`caused the simultaneous collapse of two methyl
`doublets at 0.75 and 0.84 ppm and a weaker doublet
`resonance at 3.88 ppm, unambiguously showed that
`this high-field methyl line was part of a valine A3BsMX
`spin system [14]. Since glucagon contains only a single
`
`valyl residue in position 23 [l], the 'H NMR param-
`eters of Val-23 had thus been obtained also for intact
`glucagon (Table 1).
`In Fig. 1 spectra of glucagon at two different p2H
`values are presented to show that solutions of mono-
`meric species were obtained at basic and acidic p2H.
`The spectra at p2H 10.8 and 2.4 both contain narrow
`lines characteristic of the monomeric peptide. The
`high-field-shifted methyl doublet at 0.75 ppm is seen
`at both p2H values, and a result similar to that of
`Fig. 1 C was also obtained at p2H 10.8. While small
`differences between the molecular conformations at
`the two p2H values cannot be excluded, it should be
`pointed out that most of the differences between the
`spectra A and B in Fig. 1 are a direct consequence of
`
`Page 3
`
`
`
`212
`
`Glucagon Conformation in Aqueous Solution
`
`Fig. 2. Two sputiul structuras of the fragment - Phe-22- Vul-23 - Gh-24 - Trp-25-
`in glucagon constructed with Luhquip molecular niodek.
`(A) Conformation proposed on the basis of the NMR data for glucagon in aqueous solution. For the residues 23-25,
`the structure is
`idcntical to that of the corresponding fragment 21 -23 in human parathyroid hormone [17]. (B) cc-Helical conformation. DiITererent parts of
`the structure are identified with the following numbers: I, Phc-22; 2, r-proton of Val-23; 3 , b-methine proton of Vul-23; 4 and 5,
`y-methyls of Val-23; 6 , Gln-24; 7, cc-proton of Trp-25; X, /i-methylene group of TrP-25; 9, indole ring of Trp-25
`
`the protonation of the ionizeable groups in glucagon
`[l]; e.g. in the low field region, the titration shifts of
`the histidine and tyrosine resonances lead to a some-
`what different appearance of the two spectra.
`
`DISCUSSION
`Comparison of the partial sequences glucagon
`22 - 26 and parathyroid hormone 20 - 24 in Table 1
`shows that the two peptides have nearly identical
`NMR parameters, except that the resonances of valine
`are shifted to even higher field in the glucagon frag-
`ment. The structural implications of the high-field
`shifts of the valine resonances and the vicinal spin-
`spin couplings in Table 1 were previously discussed in
`detail, and it was shown that the NMR parameters
`define a unique spatial arrangement of the tripeptide
`fragment -Val-Gln-Trp- [17]. The NMR data now
`show that this tripeptide sequence must adopt very
`similar spatial structures in glucagon and in human
`parathyroid hormone.
`A molecular model of the structure proposed for
`the fragment 22-25 of glucagon in solution is shown
`in Fig. 2A. For the residues - Val-23 - Gln-24 - Trp-
`25-,
`the structure is identical to that in the corre-
`sponding sequence 21 - 23 of human parathyroid
`hormone [17]. The a proton and the y-methyl groups
`of Val-23 are located close to the indole ring plane
`of Trp-25, which explains why they are shifted to high
`field relative to the random coil positions (Table 1).
`The spatial arrangements of the Val-23 and Trp-25
`side-chains relative to the peptide backbone corre-
`spond to those predicted by the vicinal coupling con-
`stants in Table 1. The arrangement of the Phe-22
`side-chain in the model of Fig. 2A is compatible with
`
`the observed relative chemical shifts of corresponding
`proton resonances in the parathyroid hormone and
`glucagon peptides. In the partial glucagon sequence
`22-26, where Phe-22 is at the N-terminus and hence
`quite mobile, it produces relatively small upfield ring
`current shifts for the a proton and both methyls of
`Val-23 (Table 1). That only the c( proton of Val-23
`is sizeably shifted upfield relative to the position in
`parathyroid hormone (Table 1) would then seem to
`imply that in intact glucagon Phe-22 is more rigidly
`fixed in the position shown in Fig.2A than in the
`partial sequence 22-26.
`In addition to its compati-
`bility with the NMR data, the structure of Fig.2A
`is also satisfactory in that it includes extensive con-
`tacts between hydrophobic groups, which appear to
`be the main factor for stabilizing this conforma-
`tion [17].
`Besides the non-random structure of the fragment
`22-25
`(Fig. 2A), the population of which was esti-
`mated to be approximately 20% at ambient temper-
`ature (see the preceeding paper [I71 for details of the
`procedures used to estimate populations), the NMR
`data indicate that monomeric glucagon in aqueous
`solution adopts primarily a flexible extended ‘random
`coil’ form. This is in agreement with earlier spectro-
`scopic observations which indicated the occurrence
`of ‘random coil glucagon with some ordered structure
`near Trp-25’ [8,9]. When comparisons with these earlier
`studies are pursued in more detail, it should also be
`considered that different solvent systems were used
`and that the optical detection of magnetic resonance
`studies were done in frozen solutions at low temper-
`atures.
`An intriguing result of the NMR experiments is
`that the NMR parameters are incompatible with the
`helical glucagon structure in single crystals [3]. This
`
`Page 4
`
`
`
`C. Boesch, A. Bundi, M. Oppliger, and K. Wiithrich
`
`21 3
`
`is illustrated by the a-helical structure in Fig.2B. It is
`readily apparent that the C"-Cp bonds of Val-23
`and Trp-25 are located on the surface of the helix
`at an angle of approximately 180". For geometric
`reasons it would thus be impossible in a helical struc-
`ture that the resonances of Val-23 could be affected
`by the ring current field of Trp-25. [That the high-
`field shifts of the Val-23 resonances (Table 1) are
`mainly due to the indole ring current field was previ-
`ously demonstrated by comparison with a peptide
`analogue, where tryptophan was replaced by octa-
`hydrotryptophan [17].] As was discussed in the pre-
`ceding paper [17], it is also unlikely that the solution
`conformation of Fig. 2A occurs in an equilibrium
`with a sizeably populated helical structure.
`To relate the present results on the solution con-
`formation of glucagon with the biological function
`we have to consider a situation where glucagon occurs
`in an aqueous medium at physiological concentrations
`(approx. 0.01 nM), from where it has access to the
`target cells. On the basis of the X-ray structure of
`glucagon it was previously argued that while the
`hormone must, from the available evidence, occur as
`a monomeric random coil in the dilute solution, a
`helical conformation might be stabilized by hydro-
`phobic interactions with the binding sites on the
`receptor [3]. In this model glucagon in solution
`would retain no spatial structures which might func-
`tion as recognition sites for the receptor, and bind-
`ing would thus probably occur via an induced-
`fit mechanism. The NMR data described in the
`present paper now show that a local, quite rigid non-
`random spatial structure is contained in the solution
`conformation of monomeric glucagon, and there is
`no indication that this structure should not be pre-
`served at physiological concentrations. Interestingly,
`this local non-random structure is formed by a peptide
`fragment which was long ago found to be essential
`for binding of glucagon to its receptor [18]. The
`following would then appear to be an attractive
`scheme for receptor binding of glucagon.
`
`Solution
`
`coil forms
`
`t--- structure
`of Fig. 2A
`
`Monomeric random - Monomeric
`i
`
`Receptor
`Glucagon conformation
`stabilized by receptor t- - -
`- - - - -
`interactions (helical
`structure [3] or as yet
`unknown conformation)
`This scheme is based on the experimental observation
`that the dynamic ensemble of molecular structures,
`
`glucagon
`structure of
`
`to target cell
`
`which constitute the solution conformation of gluca-
`gon [24,25], contains a sizeable population, approxi-
`mately 20%, of the molecules in the form of Fig. 2A.
`The local rigid spatial structure of the fragment 23 - 25
`would function as the recognition site for binding to
`the target cells. While the initial binding of glucagon
`would thus be governed simply by the vertical equili-
`brium reaction in the above scheme, the overall reac-
`tion mechanism might be more intricate and involve
`successive conformational rearrangements of
`the
`bound polypeptide hormone. Such secondary reaction
`steps might present an effective regulatory mechanism ;
`e.g. glucagon could be accumulated on the target cells
`in the form of Fig. 2A followed by structural re-
`arrangements of the bound hormone, which would
`result in a reactive receptor . glucagon complex. It
`was previously reported that interactions with lipids
`or detergents induce global changes of the glucagon
`conforination [12,13]. To get more precise data on
`the structural properties of glucagon in environments
`resembling more closely that of the receptor-bound
`hormone, work is in progress to use high-resolution
`NMR for studies of the polypeptide conformation in
`lipidlwater interphases.
`
`Financial support by the Swiss National Science Foundation
`is gratefully acknowledged.
`
`REFERENCES
`
`1.
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`
`8.
`
`9.
`
`10.
`11.
`12.
`
`13.
`
`14.
`
`15.
`
`16.
`
`Bromer, W. W., Sinn, L. G. & Behrens, 0. K. (1957) J. Am.
`C'hem. Soc. 7Y, 2807 - 2810.
`Pohl, S. L., Birnbaumer, L. & Rodbell, M. (1969) Sciencc.
`(Wash. D.C.) 164, 566-567.
`Sasaki, K., Dockerill, S., Adamiak, D. A,, Tickle, I. J . & Blun-
`dell, T. (1975) Nature (Lond.) 257, 751 -757.
`Gratzer, W. B. & Beaven, G. H. (1969) J . Biol. Chem. 244,
`6675 - 6679.
`Gratzer, W. B., Creeth, J. M. & Beaven, G. H. (1972) Eur. .I.
`Bioclzrm. 31, 505 - 509.
`Formisano, S., Johnson, M. L. & Edelhoch, H. (1977) Proc.
`Nut1 Acud. Sci. U.S.A. 74, 3340-3344.
`Moran, E. C., Chou, P. J. & Fasman, G. D. (1977) Biochem.
`B i ~ p h j ) ~ . Res. Commun. 77, 1300-1306.
`Ross, J. B., Rousslang, K. W., Deranleau, D. A. & Kwiram,
`A. L. (1977) BiochPmistry, 16, 5398-5402.
`Gratzer, W. B., Beaven, G . H., Rattle, H. W. E. & Bradbury,
`E. M. (1968) Eur. J . Biochem. 3, 276-283.
`Patel, D. J. (1970) Macromolecules, 3, 448-449.
`Epand, R. M. (1972) J. Bid. Chrm. 247, 2132-2138.
`Schneider, A. B. & Edelhoch, H. (1972) J. Biol. Clzem. 247,
`4992 - 4995.
`Epand, R. M., Jones, A. J. S. & Sayer, B. (1977) Biochemistrj.,
`16,4360-4368.
`Wiithrich, K. (1 976) NMR in Biulogiccil Research : Peptides
`and Proteins, North Holland, Amsterdam.
`Brewer, H. B., Rittel, W., Littedike, T. & Arnaud, C. M. (1974)
`Am. J. Med. 56, 759-766.
`Bundi, A., Andreatta, R., Rittel, W. & Wiithrich, K. (1976)
`FEBS Lett. 64, 126 - 129.
`
`Page 5
`
`
`
`214
`
`C. Boesch, A. Bundi, M. Oppliger, and K. Wiithrich: Glucagon Conformation in Aqueous Solution
`
`17. Bundi, A,, Andreatta, R. & Wiithrich, K. (1978) Eur. J. Bio-
`chern. 91,201-208.
`18. Rodbell, M., Birnbaumer, L., Pohl, S. L. & Sundby, F. (1971)
`Proc. Natl Acad. Sci. U.S.A. 68, 909 - 91 3.
`19. Anderson, G. W., Zimmerman, J. E. & Callahan, F. M. (1964)
`J . Am. Chem. Soc. 86, 1839-1842.
`20. Bundi, A. & Wiithrich, K. (1978) Biopolymers, in the press.
`21. McBride-Warren, P. A. & Epand, R. M. (1972) Biochemistry,
`11, 3571 -3575.
`
`22. De Marco, A,, Tschesche, H., Wagner, G. & Wuthrich, K.
`(1977) Biophys. Struct. Mechanism. 3, 303 - 315.
`23. Bundi, A. & Wiithrich, K. (1978) Biopolymers, in the press.
`24. Wiithrich, K., Wagner, G. & Bundi, A. (1978) Proc. 11th
`Jerusalem Symposium on Nuclear Magnetic Resonance in
`Molecular Biology (Pullman. B., ed.) Reidel, Dordrecht, in
`the press.
`25. Wiithrich,K. & Wagner,G. (1978) Trends Biochem. Sci. 3,
`227 -230.
`
`C. Boesch, A. Bundi, M. Oppliger, and K. Wiithrich, Institut fur Molekularbiologie und Biophysik
`der Eidgenossischen Technischen Hochschule Zurich-Honggerberg,
`CH-8093 Zurich. Switzerland
`
`Page 6