- Email: [email protected]
Wheat seed proteins exhibit a complex mechanism of protein elasticity
Biochimica et Biophysica Acta 1548 (2001) 187^193 www.bba-direct.com
Wheat seed proteins exhibit a complex mechanism of protein elasticity Arthur S. Tatham a
a;
*, Larry Hayes b , Peter R. Shewry a , Dan W. Urry
c
Department of Agricultural Sciences, IACR-Long Ashton Research Station, University of Bristol, Long Ashton, Bristol BS41 9AF, UK b Bioelastics Research, Ltd., OADI Technology Center, 2800 Milan Court, Suite 386, Birmingham, AL 35211-6912, USA c Biological Process Technology Institute, 1479 Gortner Avenue, University of Minnesota, St. Paul, MN 55108-6106, USA Received 5 October 2000; received in revised form 30 May 2001; accepted 30 May 2001
Abstract Elastomeric proteins are found in a number of animal tissues (elastin, abductin and resilin), where they have evolved to fulfil a range of biological functions. All exhibit rubber-like elasticity, undergoing deformation without rupture, storing the energy involved in deformation, and then recovering to their initial state when the stress is removed. The second part of the process is passive, entropy decreasing when the proteins are deformed, with the higher entropy of the relaxed state providing the driving force for recoil. In plants there is only one well-documented elastomeric protein system, the alcohol-soluble seed storage proteins (gluten) of wheat. The elastic properties of these proteins have no known biological role, the proteins acting as a store for the germinating seed. Here we show that the modulus of elasticity of a group of wheat gluten subunits, when cross-linked by Q-radiation, is similar to that of the cross-linked polypentapeptide of elastin. However, thermoelasticity studies indicate that the mechanism of elastic recoil is different from elastin and other characterized protein elastomers. Elastomeric force, f, has two components, an internal energy component, fe , and an entropic component, fs . The ratio fe /f can be determined experimentally; if this ratio is less than 0.5 the elastomeric force is predominantly entropic in origin. The ratio was determined as 5.6 for the cross-linked high Mr subunits of wheat glutenin and near zero for the cross-linked polypentapeptide of elastin. Tensile stress must be entropic or energetic in origin, the results would suggest that elastic recoil in the wheat gluten subunits, in part, may be associated with extensive hydrogen bonding within and between subunits and that entropic and energetic mechanisms contribute to the observed elasticity. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Wheat; Glutenin; Elasticity
1. Introduction Elastic proteins have been characterized in detail from several animal groups: elastin from vertebrates [1,2], abductin from arthropods [3], byssus from mussel [4] and £agelliform [5] and dragline [6] silks from spiders. Although not related in sequence these proteins all consist of multiple domains, with the elastic
* Corresponding author. Fax: +44-1275-394281. E-mail address: [email protected] (A.S. Tatham).
domains being characterized by repeated sequence motifs rich in glycine, proline and hydrophobic amino acid residues. A similar amino acid composition has also been reported for resilin, which is present in bivalve molluscs, although no sequence data are available [1]. In all these proteins elastic recoil appears to result from a predominantly entropic mechanism, with a decrease in conformational entropy on stretching due to the restricted number of low energy conformations that the extended polypeptide chains can adopt. On removal of the stretching force, there is an increase in the number of conformations the
0167-4838 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 2 3 2 - 1
BBAPRO 36463 3-8-01
188
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
polypeptide chains can adopt, providing the free energy for elastic recoil [1,7]. In addition to possessing domains that can undergo deformation, the proteins must be cross linked (covalently or noncovalently) to form a network [1]. The behaviour of the proteins is therefore, in£uenced by the lengths and nature of their elastic domains and by the degree of cross-linking. The seed storage proteins of wheat account for about half the total proteins in the dry grain, but are unique among cereal proteins in that they form a continuous network when £our is wetted and kneaded to make dough. The network confers elastic properties to the dough which allow it to be processed into bread and other foods. Although gluten is a mixture of at least ¢fty proteins, one group of 3^6 proteins is largely responsible for the elastic properties. These high Mr subunits consist of three domains; with short non-repetitive N- and C-termini (88^104 and 42 residues, respectively) £anking a longer repetitive central domain (440^680 residues) [8]. The repetitive sequences are rich in glycine and proline residues, but, unlike other characterized elastomeric proteins, the remaining amino acid residues are mainly hydrophilic, most notably glutamine [8]. The central domain has been demonstrated to be rod-like in solution [9] while STM of the hydrated solid showed a spiral-like super-secondary structure [10]. Spectroscopic studies indicate that the structure of the central repetitive domain consists predominantly of L-reverse turns [11,12]. In contrast, the N- and Cterminal domains are globular and contain cysteine residues (2^5 and 1, respectively) for inter-chain cross-linking to other proteins, the high Mr subunits being only present in covalently bound polymers [8]. The high Mr subunits, therefore, possess the characteristics of a putative elastomer, with N- and C-terminal domains containing residues for covalent cross-linking and a central domain that can potentially undergo deformation. In this study we determine the modulus of elasticity of the high Mr subunits of wheat gluten using stress-strain studies and explore the mechanisms of elasticity by determining the temperature dependence of the elastomeric force. Comparisons with similar analyses of the cross-linked polypentapeptide of elastin demonstrate that the high Mr subunits exhibit a novel mechanism of elasticity.
2. Material and methods 2.1. Protein puri¢cation High Mr subunits of wheat glutenin were prepared from white endosperm £our of wheat cv. Mercia as reported previously [13]. The cysteine residues were not alkylated, allowing disulphide bonds to form when reducing agent was removed. 2.2. Cross-linking Five-hundred mg of subunits were dissolved in 4 M urea, 0.1% (v/v) 2-mercaptoethanol (20 ml). The sample was dialysed against water and centrifuged for 4 h at 2000 rpm. The sample was transferred to the top of a cross-linking mould, consisting of a centrifuge tube with a polypropylene insert, and centrifuged into the mould at 6000 rpm for 4 h. Cross-linking was performed at the Auburn University Nuclear Science Center by exposing the sample to 20 Mrad Q-radiation. A similar ¢lm was prepared from the synthetic polypentapeptide, poly(GVGVP), of elastin. 2.3. Stress^strain and thermoelasticity measurements The custom built stress^strain apparatus consisted of a rigidly mounted Statham model UC-2 transducing cell with a UL4^0.5 load cell accessory to record the force data. Using the required excitation and conditioning circuits, the output of the UC-2 transducer was recorded on the x-axis of an x^y chart recorder. The sample was elongated at constant rate using a Velmex model B2509C5 Unislide, driven by a variable speed motor through a Merton Instruments speed reducer. The position of the moving holder was recorded on the y-axis of the x^y recorder using a BLH linear displacement transducer and appropriate excitation and conditioning circuits. The thickness of the sample was measured before and after being placed in the clamps. One clamp was attached to the UC-2 transducer and the other to the moving platform. The x-axis scale was measured as L/Li where L is the displacement of the moving end from the sample. The cross-sectional area was used to calculate kPa from the recorded trace. A water-jacketed cylinder allowed samples to be
BBAPRO 36463 3-8-01
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
189
immersed in water during the stress^strain and thermoelasticity experiments. 3. Results 3.1. Measurement of Young's Modulus A ¢lm consisting of cross-linked subunits was initially tested by stretching to breaking point. Breaking occurred at approx. 50% extension (Fig. 1) which is similar to the initial behaviour reported for ¢lms consisting of chemically synthesized and cross-linked (PPP) (sequence GVGVP) of elastin [14]. However, extensions of over 200% have been observed using similarly cross-linked higher molecular mass (Mr approx. 100 000) poly(GVGVP) which has been chemically and microbially prepared. Young's modulus (or elastic modulus), Ym, is a measure of the ability of a material to withstand change in length, under tension or compression. It is only meaningful in the range where stress is proportional to the strain and the material returns to its original state when the force is removed. Comparison of the stress^strain curves of the cross-linked subunits and PPP showed similar behaviour (Fig. 2), with little hysteresis up to 20% extension. The stress:strain ratio, Ym, was calculated at 20% extension and 20³C as Ym stress=strain F =A v L=Li m=v Li cLi g=A 1
Fig. 1. Stress^strain curve for cross-linked high Mr subunits of wheat glutenin.
Fig. 2. Stress^strain curves for cross-linked high Mr subunits of wheat glutenin (999) and cross-linked poly(GVGVP) (- - -), showing hysteresis.
Where m = 2.24 g (force transducer calibrated by hanging a 5-g mass); g = 980 cm/s2 ; Li = initial length, 1.46 cm; L = ¢nal length, 1.75 cm; A, the cross-sectional area = 0.010 cm2 (width 0.40 cm and thickness 0.025 cm). Using these values, Young's modulus was calculated as approx. 1.1U102 kPa for both materials. 3.2. Measurement of elastomeric force The cross-linked subunit and PPP ¢lms were stretched to a ¢xed length (L) and the elastomeric force (f) measured as a function of temperature. Ln(f/T) was then plotted vs. T to obtain the slope Dln(f/T)/DT (Fig. 3). The two materials behaved quite di¡erently with increasing temperature, resulting in two di¡erent curves. The PPP of elastin undergoes a rapid increase in elastomeric force over the 25^ 35³C range, above 40³C being £at. This has been linked with coacervation behaviour with increasing temperature, this is described as an inverse temperature transition with hydrophobic intermolecular interactions become the dominant feature on association to form the coacervate [14]. This contrasts with the behaviour of the high Mr subunits, which show a gradual decrease in elastomeric force with increasing temperature. For this reason the slope was measured at 20^30³C for the subunits and 40^60³C for the PPP of elastin. Elastomeric force (f) has two components, an internal energy component (fe ) and an entropic com-
BBAPRO 36463 3-8-01
190
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
the plot of ln(f/T (K)) corrected for changes in cross-sectional area (A) vs. temperature, values of fe /f were determined from the slope of the plots at 20^30³C for the subunits and 40^60³C for the PPP multiplied by the absolute temperature. For the subunits fe /f = slope (30.0187)x3298 K (25³C) = 5.6. The PPP gave a slope very close to zero over the 40^60³C temperature range, indicating that the force has a large entropic component with fe /f 6 0.5. In contrast, the high value exhibited by the wheat glutenin subunits would suggest that the entropic component (fs ) is not the predominant component of the elastomeric force. Fig. 3. Thermoelastic plot of cross-linked high Mr subunits of wheat glutenin (R) and cross-linked poly(GVGVP) (b).
ponent (fs ). The conformational entropy component arises from the stressing of the polymer chains to a less probable state, with recoil resulting from an increase in entropy on removal of the stress. The internal energy component is derived from the direct distortion of the chemical bonds (changes in internal energy) rather than changes in conformational entropy. This ratio can be determined from the slope in Fig. 3, using the equation: f e =f 3T D lnf =T=D TV;L;n
2
the sample is held at ¢xed length (L) under conditions of constant volume, V, and constant composition, n [15]. However, it is di¤cult to experimentally hold V and n constant while varying T, particularly when T is often greater that 300 K, as required for a meaningful determination of the slope. An approximation has, therefore, been developed for conditions of constant pressure, P, constant length, L, and for an elastomer in equilibrium, eq, with a solvent [16]: f e =f 3T D lnf =T=D TP;L;eq 3 L eq T= K 3 V i =V 31 3 Where Leq is the thermal expansion coe¤cient ( = (Dln[f/T]/DT)P;L;eq ), K is the fractional increase in length, L/Li , where Li is the initial length and L the length at ¢xed extension, and Vi and V are the elastomer volumes at length Li and L, respectively. For elastin between, in the temperature range 50^ 70³C, Leq is approximately zero [7]. From Fig. 3,
4. Discussion The high Mr subunits are known to form extensive disulphide bonded networks in vivo and in dough [8]. However, when extracted under reducing conditions and reoxidized they did not form a ¢lm that was stable enough for mechanical testing. Extraction of the high Mr subunits from the glutenin polymers and subsequent reoxidation presumably did not allow formation of su¤cient intermolecular disulphide bonds to form a stable ¢lm. Gamma radiation was therefore used to introduce cross-links, following the same procedure as used for the synthetic PPP of elastin, which was studied as the control. The modulus of elasticity determined for the crosslinked subunits, 1.1U102 kPa, is similar to the modulus of the cross-linked PPP of elastin [14]. Rupture of the cross-linked HMW subunits occurs at approx. 50% extension, whereas rupture of the cross-linked PPP occurred (dependent on moisture content) at slightly higher extensions (50^80%). In contrast, aortic elastin only ruptures at approx. 400% extension [14]. These results, therefore, demonstrate that the HMW subunits are elastic. However, because it is not possible to determine the precise numbers and positions of cross-links introduced by gamma irradiation the precise elastic force that was measured may not be identical to that exhibited in gluten. For example, cross-linking in the native protein occurs predominantly between cysteine residues in the N- and C-terminal domains, while cross-linking by irradiation may have occurred throughout the proteins, affecting their elasticity.
BBAPRO 36463 3-8-01
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
The mechanism of elastic recoil in elastin and natural rubber is predominantly entropic in origin, with fe /f ratios below 0.5. However, the ratio determined for the high Mr subunits was 5.6, indicating that the mechanism of elasticity is not predominantly entropic in origin. As the ratio is greater than 1, another mechanism may be contributing to the total elastomeric force in addition to the internal and entropic energy components and/or the assumptions made in the Flory derivation may not hold for the high Mr subunits of glutenin. Two di¡erent mechanisms have been proposed to account for the elasticity of elastin. The ¢rst is based on rubber theory in which elastin is considered to be a network of random chains of high entropy. Stress orders the chains and decreases the entropy, providing the driving force for recoil [7]. The second is based on the work of Urry and co-workers [16] who have carried out biophysical studies on K-elastin and synthetic peptides corresponding to repeated sequences found within the K-elastin sequence. Based on these studies they have proposed that the entropic elasticity derives from a L-spiral structure, a spiral that consists predominantly of L-turns. The L-turns regularly repeat on a helical (spiral) axis and act as spacers between the turns of the spiral, suspending the connecting chain segments in a relatively kinetically free state. The peptide moieties of the suspended segments are able to undergo large amplitude rocking motions, called librations, which decrease in amplitude on stretching, resulting in a decrease in entropy of the system and providing the driving force for recoil. In both models the restoring force is entropic in origin. Internal energy mechanisms are found in materials such as steel, where extension is only elastic to about 1%, elasticity arising from short-range inter-atomic forces. In proteins such internal energy mechanisms result from the stressing of chemical bonds, but are unlikely to be predominant, as they would lead to bond cleavage and lower elasticity than demonstrated in Fig. 1. Possible explanations for the di¡erent mechanisms of the high Mr subunits and other elastomers may relate to their amino acid compositions and/or their patterns of hydrogen bonding, either to water or adjacent protein subunits. The Flory derivation assumes no structural changes over the temperature range studied, as the assumption holds for a random
191
chain polymer network. Evidence of a fundamental di¡erence in mechanisms between elastin and the high Mr subunits comes from the e¡ects of heat. In elastin, heating results in contraction and the exclusion of water as the L-turns fold to optimize hydrophobic contacts between turns of the spiral, whereas heating of the high Mr subunits results in the absorption of water [17]. Thus, in Fig. 3 the reduction in force with increasing temperature for the high Mr subunits may result from inter and intramolecular hydrogen bonds being broken and replaced with subunit-water interactions. On heating the central repetitive domains of the subunits structure, to conformations with higher contents of L-turns [18]. Rubber elasticity theory assumes that the polymer units do not interact with each other and that there are no solvent-polymer interactions to consider, but these interactions are important in the high Mr subunits. Thus, rubber theory may be applicable to elastin and other predominantly entropic systems as they do not interact to a high degree, but is not completely applicable to the high Mr subunit system as it does not take into account hydrogen bonding. Therefore, the calculations of the slope are not strictly meaningful as hydrogen bonds within the system break as the temperature is raised. Such complex systems maybe present in other biological elastomeric materials which have less hydrophobic characteristics than elastin. The repeats present in the central repetitive domains of the high Mr subunits of wheat glutenin are based on three motifs (hexapeptides (PGQGQQ), nonapeptides (GYYPTSP/LQQ) and tripeptides (GQQ)), the domains are rich in proline, glycine and hydrophilic residues, predominantly glutamine [8]. L-Turns are the predominant structural feature, in solution, of these domains and a L-spiral structure has been hypothesized to form [11,12], as in elastin. Glutamine residues account for about 35 mol% of the total residues of the high Mr subunits, glycine about 20 mol% and proline 10 mol%. The repetitive domain is characterized by high levels of mobility in the presence of water, their structure consisting of L-turns and intermolecular L-sheets in proportions that vary with water content [19]. Evidence from spectroscopic studies has led to the suggestion that hydrogen bonding between the repetitive regions contributes to the elastomeric properties of the hydrated
BBAPRO 36463 3-8-01
192
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
protein network [20]. A `loop and train' model has been proposed for gluten elasticity [20]. Elasticity arising from regions of the protein that interact (`trains', associated with L-sheet formation) and mobile regions (`loops', hydrated L-turn structures). On stretching, the network initially deforms in the loop regions and then in the train regions. The restoring force derives from an entropic component associated with the conformational freedom of the loops, and from the enthalpy of hydrogen bond formation in the train regions; the entropy loss from the interprotein hydrogen bonds being o¡set, in part, by the increased entropy of the water released [20]. This ability to form intermolecular hydrogen bonds contrasts with the behaviour of the hydrophobic repeats of elastin. The results may indicate, therefore, that the elastomeric force observed for the high Mr subunits has a number of components. It is di¤cult to envisage that an entropic mechanism (resulting from the extension of the protein chains to less favourable conformations) does not contribute to elasticity, given the structure of the subunits and their similarity to elastin. An internal energy component, through stressing of chemical bonds, as found in metals is problematic, as it would result in higher Young's moduli than found experimentally and low extensibility. Considering the high propensity for glutamine to form hydrogen bonds, whether a `loop' and `train' model is invoked or not, hydrogen bonding could be invoked as contributing to the elastomeric mechanism. The £agelliform silks, which form the capture threads of spiders webs, contain repeated sequences similar to those of elastin and can extend up to 200% before rupture, giving trapped insects little solid structure to push against [5]. In contrast the dragline silks, which form the framework of the web and dropping line, can extend to about 30% before rupture and are more highly cross-linked and contain di¡erent repetitive sequences to £agelliform silks [6]. The pentapeptide repeats in the dragline silks (GPGQQ and YGPGG) contain glutamine residues, as do those of the high Mr subunits of wheat glutenin. Studies of contracted silks indicate that the entropic component accounts for about 85% of the force. Some of the energy of the insect impacting on the web is dissipated as heat and not recovered
as elastic recoil, therefore the insect is unlikely to be catapulted out of the web and the energy converted to heat is not available to rupture the web [21]. The pentapeptide repeats of the £agelliform and dragline silks are predicted to form L-spirals [22] and water is essential for their elasticity, implying the spirals are held together by hydrophobic and hydrogen bonding. Their mechanism of elasticity may, therefore, have similarities to those of both elastin and the high Mr subunits of wheat. Thus, the repetitive sequences in silks and high Mr subunits appear to form similar secondary structures, which contribute to elasticity by their interactions. The advantage of an entropic mechanism of elastomeric force is that it depends on the decrease in the numbers of accessible states on extension and can be considered ideal for biological systems. Entropic elasticity provides for a more durable elastomer, with elastic ¢bres lasting the lifetime of an individual [23]. In evolutionary terms several unrelated proteins have evolved to achieve elasticity by a similar entropic mechanism, where energy is required to stress the proteins and relaxation is non-energy-consuming. However, in the dragline silks a purely entropic mechanism would not ful¢l the requirements of the material, so that a contribution from an internal energy mechanism and/or interaction through hydrogen bonds involving glutamine residues would be adventitious. The high Mr subunits have had no evolutionary pressure to develop elastic mechanisms. Seed storage proteins act solely as a store of carbon, nitrogen and sulphur for the developing seedling, with the main evolutionary pressures being for e¤cient packaging for storage and mobilization on germination. The high Mr subunits clearly provide e¤cient storage of nitrogen due to their high contents of glutamine, which their compact spiral structure may facilitate packaging within the protein bodies of the developing grain. Thus the elastomeric properties appear to have no biological signi¢cance and both these and their elastic mechanisms appear to have developed as a fortuitous consequence of the evolution of the proteins by reiteration of sequence repeats. Their elastomeric properties are, however, important in determining the ability of wheat £our to be mixed into doughs, which are expanded into ¢ne foams and baked to form leavened breads.
BBAPRO 36463 3-8-01
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
Acknowledgements IACR receives grant-aided support from the Biotechnology and Biological Sciences Research Council of the United Kingdom. References [1] J.M. Gosline, Rubber Chem. Technol. 60 (1987) 417^438. [2] Z. Indik, N. Ornstein-Goldstein, P. Sheppard, N. Anderson, J.C. Rosenbloom, L. Peltonen, J. Rosenbloom, Proc. Natl. Acad. Sci. USA 84 (1987) 5680^5684. [3] Q. Cao, Y. Wang, H. Bayley, Curr. Biol. 7 (1997) R677^ R678. [4] K.J. Coyne, X.X. Qin, H. Waite, Science 277 (1997) 1830^ 1832. [5] C.Y. Hayashi, R.V. Lewis, J. Mol. Biol. 275 (1988) 773^784. [6] P.A. Guerrette, D.G. Ginzinger, B.H.F. Weber, J.M. Gosline, Science 272 (1996) 112^115. [7] C.A.J. Hoeve, P.J. Flory, Biopolymers 13 (1974) 677^686. [8] P.R. Shewry, A.S. Tatham, Biochem. J. 267 (1990) 1^12. [9] J.M. Field, A.S. Tatham, A. Baker, P.R. Shewry, FEBS Lett. 200 (1987) 76^80. [10] M.J. Miles, H.J. Carr, T.J. McMaster, K.J. I'Anson, P.S. Belton, V.J. Morris, J.M. Field, P.R. Shewry, A.S. Tatham, Proc. Natl. Acad. Sci. USA 88 (1992) 76^80.
193
[11] A.S. Tatham, A.F. Drake, P.R. Shewry, J. Cereal Sci. 11 (1990) 189^200. [12] A.A. van Dijk, L.L. van Wijk, A. van Vliet, P. Harris, E. van Swieten, G.I. Tesser, G.T. Robillard, Protein Sci. 6 (1997) 637^648. [13] B.A. Marchylo, J.E. Kruger, D.W. Hatcher, J. Cereal Sci. 9 (1988) 113^130. [14] D.W. Urry, K. Okamoto, R.D. Harris, C.F. Hendrix, M.M. Long, Biochemistry 15 (1976) 4083^4089. [15] P.J. Flory, A. Ci¡eri, C.A.J. Hoeve, J. Polym. Sci. 45 (1960) 235^236. [16] D.W. Urry, J. Protein Chem. 1 (1988) 1^34. [17] P.S. Belton, I.J. Colquhoun, J.M. Field, A. Grant, P.R. Shewry, A.S. Tatham, J. Cereal Sci. 19 (1994) 115^121. [18] S.M. Gilbert, N. Wellner, P.S. Belton, J.A. Green¢eld, G. Siligardi, P.R. Shewry, A.S. Tatham, Biochim. Biophys. Acta 1479 (2000) 135^146. [19] P.S. Belton, I.J. Colquhoun, J.M. Field, A. Grant, P.R. Shewry, A.S. Tatham, N. Wellner, Int. J. Biol. Macromol. 17 (1995) 74^80. [20] P.S. Belton, J. Cereal Sci. 29 (1999) 103^107. [21] J.M. Gosline, M.E. DeMont, M.W. Denny, Endeavour 10 (1986) 37^43. [22] C.Y. Hayashi, N.H. Shipley, R.V. Lewis, Int. J. Biol. Macromol. 24 (1999) 271^275. [23] D.W. Urry, Sci. Am. 272 (1995) 64^69.
BBAPRO 36463 3-8-01
Wheat seed proteins exhibit a complex mechanism of protein elasticity Arthur S. Tatham a
a;
*, Larry Hayes b , Peter R. Shewry a , Dan W. Urry
c
Department of Agricultural Sciences, IACR-Long Ashton Research Station, University of Bristol, Long Ashton, Bristol BS41 9AF, UK b Bioelastics Research, Ltd., OADI Technology Center, 2800 Milan Court, Suite 386, Birmingham, AL 35211-6912, USA c Biological Process Technology Institute, 1479 Gortner Avenue, University of Minnesota, St. Paul, MN 55108-6106, USA Received 5 October 2000; received in revised form 30 May 2001; accepted 30 May 2001
Abstract Elastomeric proteins are found in a number of animal tissues (elastin, abductin and resilin), where they have evolved to fulfil a range of biological functions. All exhibit rubber-like elasticity, undergoing deformation without rupture, storing the energy involved in deformation, and then recovering to their initial state when the stress is removed. The second part of the process is passive, entropy decreasing when the proteins are deformed, with the higher entropy of the relaxed state providing the driving force for recoil. In plants there is only one well-documented elastomeric protein system, the alcohol-soluble seed storage proteins (gluten) of wheat. The elastic properties of these proteins have no known biological role, the proteins acting as a store for the germinating seed. Here we show that the modulus of elasticity of a group of wheat gluten subunits, when cross-linked by Q-radiation, is similar to that of the cross-linked polypentapeptide of elastin. However, thermoelasticity studies indicate that the mechanism of elastic recoil is different from elastin and other characterized protein elastomers. Elastomeric force, f, has two components, an internal energy component, fe , and an entropic component, fs . The ratio fe /f can be determined experimentally; if this ratio is less than 0.5 the elastomeric force is predominantly entropic in origin. The ratio was determined as 5.6 for the cross-linked high Mr subunits of wheat glutenin and near zero for the cross-linked polypentapeptide of elastin. Tensile stress must be entropic or energetic in origin, the results would suggest that elastic recoil in the wheat gluten subunits, in part, may be associated with extensive hydrogen bonding within and between subunits and that entropic and energetic mechanisms contribute to the observed elasticity. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Wheat; Glutenin; Elasticity
1. Introduction Elastic proteins have been characterized in detail from several animal groups: elastin from vertebrates [1,2], abductin from arthropods [3], byssus from mussel [4] and £agelliform [5] and dragline [6] silks from spiders. Although not related in sequence these proteins all consist of multiple domains, with the elastic
* Corresponding author. Fax: +44-1275-394281. E-mail address: [email protected] (A.S. Tatham).
domains being characterized by repeated sequence motifs rich in glycine, proline and hydrophobic amino acid residues. A similar amino acid composition has also been reported for resilin, which is present in bivalve molluscs, although no sequence data are available [1]. In all these proteins elastic recoil appears to result from a predominantly entropic mechanism, with a decrease in conformational entropy on stretching due to the restricted number of low energy conformations that the extended polypeptide chains can adopt. On removal of the stretching force, there is an increase in the number of conformations the
0167-4838 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 2 3 2 - 1
BBAPRO 36463 3-8-01
188
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
polypeptide chains can adopt, providing the free energy for elastic recoil [1,7]. In addition to possessing domains that can undergo deformation, the proteins must be cross linked (covalently or noncovalently) to form a network [1]. The behaviour of the proteins is therefore, in£uenced by the lengths and nature of their elastic domains and by the degree of cross-linking. The seed storage proteins of wheat account for about half the total proteins in the dry grain, but are unique among cereal proteins in that they form a continuous network when £our is wetted and kneaded to make dough. The network confers elastic properties to the dough which allow it to be processed into bread and other foods. Although gluten is a mixture of at least ¢fty proteins, one group of 3^6 proteins is largely responsible for the elastic properties. These high Mr subunits consist of three domains; with short non-repetitive N- and C-termini (88^104 and 42 residues, respectively) £anking a longer repetitive central domain (440^680 residues) [8]. The repetitive sequences are rich in glycine and proline residues, but, unlike other characterized elastomeric proteins, the remaining amino acid residues are mainly hydrophilic, most notably glutamine [8]. The central domain has been demonstrated to be rod-like in solution [9] while STM of the hydrated solid showed a spiral-like super-secondary structure [10]. Spectroscopic studies indicate that the structure of the central repetitive domain consists predominantly of L-reverse turns [11,12]. In contrast, the N- and Cterminal domains are globular and contain cysteine residues (2^5 and 1, respectively) for inter-chain cross-linking to other proteins, the high Mr subunits being only present in covalently bound polymers [8]. The high Mr subunits, therefore, possess the characteristics of a putative elastomer, with N- and C-terminal domains containing residues for covalent cross-linking and a central domain that can potentially undergo deformation. In this study we determine the modulus of elasticity of the high Mr subunits of wheat gluten using stress-strain studies and explore the mechanisms of elasticity by determining the temperature dependence of the elastomeric force. Comparisons with similar analyses of the cross-linked polypentapeptide of elastin demonstrate that the high Mr subunits exhibit a novel mechanism of elasticity.
2. Material and methods 2.1. Protein puri¢cation High Mr subunits of wheat glutenin were prepared from white endosperm £our of wheat cv. Mercia as reported previously [13]. The cysteine residues were not alkylated, allowing disulphide bonds to form when reducing agent was removed. 2.2. Cross-linking Five-hundred mg of subunits were dissolved in 4 M urea, 0.1% (v/v) 2-mercaptoethanol (20 ml). The sample was dialysed against water and centrifuged for 4 h at 2000 rpm. The sample was transferred to the top of a cross-linking mould, consisting of a centrifuge tube with a polypropylene insert, and centrifuged into the mould at 6000 rpm for 4 h. Cross-linking was performed at the Auburn University Nuclear Science Center by exposing the sample to 20 Mrad Q-radiation. A similar ¢lm was prepared from the synthetic polypentapeptide, poly(GVGVP), of elastin. 2.3. Stress^strain and thermoelasticity measurements The custom built stress^strain apparatus consisted of a rigidly mounted Statham model UC-2 transducing cell with a UL4^0.5 load cell accessory to record the force data. Using the required excitation and conditioning circuits, the output of the UC-2 transducer was recorded on the x-axis of an x^y chart recorder. The sample was elongated at constant rate using a Velmex model B2509C5 Unislide, driven by a variable speed motor through a Merton Instruments speed reducer. The position of the moving holder was recorded on the y-axis of the x^y recorder using a BLH linear displacement transducer and appropriate excitation and conditioning circuits. The thickness of the sample was measured before and after being placed in the clamps. One clamp was attached to the UC-2 transducer and the other to the moving platform. The x-axis scale was measured as L/Li where L is the displacement of the moving end from the sample. The cross-sectional area was used to calculate kPa from the recorded trace. A water-jacketed cylinder allowed samples to be
BBAPRO 36463 3-8-01
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
189
immersed in water during the stress^strain and thermoelasticity experiments. 3. Results 3.1. Measurement of Young's Modulus A ¢lm consisting of cross-linked subunits was initially tested by stretching to breaking point. Breaking occurred at approx. 50% extension (Fig. 1) which is similar to the initial behaviour reported for ¢lms consisting of chemically synthesized and cross-linked (PPP) (sequence GVGVP) of elastin [14]. However, extensions of over 200% have been observed using similarly cross-linked higher molecular mass (Mr approx. 100 000) poly(GVGVP) which has been chemically and microbially prepared. Young's modulus (or elastic modulus), Ym, is a measure of the ability of a material to withstand change in length, under tension or compression. It is only meaningful in the range where stress is proportional to the strain and the material returns to its original state when the force is removed. Comparison of the stress^strain curves of the cross-linked subunits and PPP showed similar behaviour (Fig. 2), with little hysteresis up to 20% extension. The stress:strain ratio, Ym, was calculated at 20% extension and 20³C as Ym stress=strain F =A v L=Li m=v Li cLi g=A 1
Fig. 1. Stress^strain curve for cross-linked high Mr subunits of wheat glutenin.
Fig. 2. Stress^strain curves for cross-linked high Mr subunits of wheat glutenin (999) and cross-linked poly(GVGVP) (- - -), showing hysteresis.
Where m = 2.24 g (force transducer calibrated by hanging a 5-g mass); g = 980 cm/s2 ; Li = initial length, 1.46 cm; L = ¢nal length, 1.75 cm; A, the cross-sectional area = 0.010 cm2 (width 0.40 cm and thickness 0.025 cm). Using these values, Young's modulus was calculated as approx. 1.1U102 kPa for both materials. 3.2. Measurement of elastomeric force The cross-linked subunit and PPP ¢lms were stretched to a ¢xed length (L) and the elastomeric force (f) measured as a function of temperature. Ln(f/T) was then plotted vs. T to obtain the slope Dln(f/T)/DT (Fig. 3). The two materials behaved quite di¡erently with increasing temperature, resulting in two di¡erent curves. The PPP of elastin undergoes a rapid increase in elastomeric force over the 25^ 35³C range, above 40³C being £at. This has been linked with coacervation behaviour with increasing temperature, this is described as an inverse temperature transition with hydrophobic intermolecular interactions become the dominant feature on association to form the coacervate [14]. This contrasts with the behaviour of the high Mr subunits, which show a gradual decrease in elastomeric force with increasing temperature. For this reason the slope was measured at 20^30³C for the subunits and 40^60³C for the PPP of elastin. Elastomeric force (f) has two components, an internal energy component (fe ) and an entropic com-
BBAPRO 36463 3-8-01
190
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
the plot of ln(f/T (K)) corrected for changes in cross-sectional area (A) vs. temperature, values of fe /f were determined from the slope of the plots at 20^30³C for the subunits and 40^60³C for the PPP multiplied by the absolute temperature. For the subunits fe /f = slope (30.0187)x3298 K (25³C) = 5.6. The PPP gave a slope very close to zero over the 40^60³C temperature range, indicating that the force has a large entropic component with fe /f 6 0.5. In contrast, the high value exhibited by the wheat glutenin subunits would suggest that the entropic component (fs ) is not the predominant component of the elastomeric force. Fig. 3. Thermoelastic plot of cross-linked high Mr subunits of wheat glutenin (R) and cross-linked poly(GVGVP) (b).
ponent (fs ). The conformational entropy component arises from the stressing of the polymer chains to a less probable state, with recoil resulting from an increase in entropy on removal of the stress. The internal energy component is derived from the direct distortion of the chemical bonds (changes in internal energy) rather than changes in conformational entropy. This ratio can be determined from the slope in Fig. 3, using the equation: f e =f 3T D lnf =T=D TV;L;n
2
the sample is held at ¢xed length (L) under conditions of constant volume, V, and constant composition, n [15]. However, it is di¤cult to experimentally hold V and n constant while varying T, particularly when T is often greater that 300 K, as required for a meaningful determination of the slope. An approximation has, therefore, been developed for conditions of constant pressure, P, constant length, L, and for an elastomer in equilibrium, eq, with a solvent [16]: f e =f 3T D lnf =T=D TP;L;eq 3 L eq T= K 3 V i =V 31 3 Where Leq is the thermal expansion coe¤cient ( = (Dln[f/T]/DT)P;L;eq ), K is the fractional increase in length, L/Li , where Li is the initial length and L the length at ¢xed extension, and Vi and V are the elastomer volumes at length Li and L, respectively. For elastin between, in the temperature range 50^ 70³C, Leq is approximately zero [7]. From Fig. 3,
4. Discussion The high Mr subunits are known to form extensive disulphide bonded networks in vivo and in dough [8]. However, when extracted under reducing conditions and reoxidized they did not form a ¢lm that was stable enough for mechanical testing. Extraction of the high Mr subunits from the glutenin polymers and subsequent reoxidation presumably did not allow formation of su¤cient intermolecular disulphide bonds to form a stable ¢lm. Gamma radiation was therefore used to introduce cross-links, following the same procedure as used for the synthetic PPP of elastin, which was studied as the control. The modulus of elasticity determined for the crosslinked subunits, 1.1U102 kPa, is similar to the modulus of the cross-linked PPP of elastin [14]. Rupture of the cross-linked HMW subunits occurs at approx. 50% extension, whereas rupture of the cross-linked PPP occurred (dependent on moisture content) at slightly higher extensions (50^80%). In contrast, aortic elastin only ruptures at approx. 400% extension [14]. These results, therefore, demonstrate that the HMW subunits are elastic. However, because it is not possible to determine the precise numbers and positions of cross-links introduced by gamma irradiation the precise elastic force that was measured may not be identical to that exhibited in gluten. For example, cross-linking in the native protein occurs predominantly between cysteine residues in the N- and C-terminal domains, while cross-linking by irradiation may have occurred throughout the proteins, affecting their elasticity.
BBAPRO 36463 3-8-01
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
The mechanism of elastic recoil in elastin and natural rubber is predominantly entropic in origin, with fe /f ratios below 0.5. However, the ratio determined for the high Mr subunits was 5.6, indicating that the mechanism of elasticity is not predominantly entropic in origin. As the ratio is greater than 1, another mechanism may be contributing to the total elastomeric force in addition to the internal and entropic energy components and/or the assumptions made in the Flory derivation may not hold for the high Mr subunits of glutenin. Two di¡erent mechanisms have been proposed to account for the elasticity of elastin. The ¢rst is based on rubber theory in which elastin is considered to be a network of random chains of high entropy. Stress orders the chains and decreases the entropy, providing the driving force for recoil [7]. The second is based on the work of Urry and co-workers [16] who have carried out biophysical studies on K-elastin and synthetic peptides corresponding to repeated sequences found within the K-elastin sequence. Based on these studies they have proposed that the entropic elasticity derives from a L-spiral structure, a spiral that consists predominantly of L-turns. The L-turns regularly repeat on a helical (spiral) axis and act as spacers between the turns of the spiral, suspending the connecting chain segments in a relatively kinetically free state. The peptide moieties of the suspended segments are able to undergo large amplitude rocking motions, called librations, which decrease in amplitude on stretching, resulting in a decrease in entropy of the system and providing the driving force for recoil. In both models the restoring force is entropic in origin. Internal energy mechanisms are found in materials such as steel, where extension is only elastic to about 1%, elasticity arising from short-range inter-atomic forces. In proteins such internal energy mechanisms result from the stressing of chemical bonds, but are unlikely to be predominant, as they would lead to bond cleavage and lower elasticity than demonstrated in Fig. 1. Possible explanations for the di¡erent mechanisms of the high Mr subunits and other elastomers may relate to their amino acid compositions and/or their patterns of hydrogen bonding, either to water or adjacent protein subunits. The Flory derivation assumes no structural changes over the temperature range studied, as the assumption holds for a random
191
chain polymer network. Evidence of a fundamental di¡erence in mechanisms between elastin and the high Mr subunits comes from the e¡ects of heat. In elastin, heating results in contraction and the exclusion of water as the L-turns fold to optimize hydrophobic contacts between turns of the spiral, whereas heating of the high Mr subunits results in the absorption of water [17]. Thus, in Fig. 3 the reduction in force with increasing temperature for the high Mr subunits may result from inter and intramolecular hydrogen bonds being broken and replaced with subunit-water interactions. On heating the central repetitive domains of the subunits structure, to conformations with higher contents of L-turns [18]. Rubber elasticity theory assumes that the polymer units do not interact with each other and that there are no solvent-polymer interactions to consider, but these interactions are important in the high Mr subunits. Thus, rubber theory may be applicable to elastin and other predominantly entropic systems as they do not interact to a high degree, but is not completely applicable to the high Mr subunit system as it does not take into account hydrogen bonding. Therefore, the calculations of the slope are not strictly meaningful as hydrogen bonds within the system break as the temperature is raised. Such complex systems maybe present in other biological elastomeric materials which have less hydrophobic characteristics than elastin. The repeats present in the central repetitive domains of the high Mr subunits of wheat glutenin are based on three motifs (hexapeptides (PGQGQQ), nonapeptides (GYYPTSP/LQQ) and tripeptides (GQQ)), the domains are rich in proline, glycine and hydrophilic residues, predominantly glutamine [8]. L-Turns are the predominant structural feature, in solution, of these domains and a L-spiral structure has been hypothesized to form [11,12], as in elastin. Glutamine residues account for about 35 mol% of the total residues of the high Mr subunits, glycine about 20 mol% and proline 10 mol%. The repetitive domain is characterized by high levels of mobility in the presence of water, their structure consisting of L-turns and intermolecular L-sheets in proportions that vary with water content [19]. Evidence from spectroscopic studies has led to the suggestion that hydrogen bonding between the repetitive regions contributes to the elastomeric properties of the hydrated
BBAPRO 36463 3-8-01
192
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
protein network [20]. A `loop and train' model has been proposed for gluten elasticity [20]. Elasticity arising from regions of the protein that interact (`trains', associated with L-sheet formation) and mobile regions (`loops', hydrated L-turn structures). On stretching, the network initially deforms in the loop regions and then in the train regions. The restoring force derives from an entropic component associated with the conformational freedom of the loops, and from the enthalpy of hydrogen bond formation in the train regions; the entropy loss from the interprotein hydrogen bonds being o¡set, in part, by the increased entropy of the water released [20]. This ability to form intermolecular hydrogen bonds contrasts with the behaviour of the hydrophobic repeats of elastin. The results may indicate, therefore, that the elastomeric force observed for the high Mr subunits has a number of components. It is di¤cult to envisage that an entropic mechanism (resulting from the extension of the protein chains to less favourable conformations) does not contribute to elasticity, given the structure of the subunits and their similarity to elastin. An internal energy component, through stressing of chemical bonds, as found in metals is problematic, as it would result in higher Young's moduli than found experimentally and low extensibility. Considering the high propensity for glutamine to form hydrogen bonds, whether a `loop' and `train' model is invoked or not, hydrogen bonding could be invoked as contributing to the elastomeric mechanism. The £agelliform silks, which form the capture threads of spiders webs, contain repeated sequences similar to those of elastin and can extend up to 200% before rupture, giving trapped insects little solid structure to push against [5]. In contrast the dragline silks, which form the framework of the web and dropping line, can extend to about 30% before rupture and are more highly cross-linked and contain di¡erent repetitive sequences to £agelliform silks [6]. The pentapeptide repeats in the dragline silks (GPGQQ and YGPGG) contain glutamine residues, as do those of the high Mr subunits of wheat glutenin. Studies of contracted silks indicate that the entropic component accounts for about 85% of the force. Some of the energy of the insect impacting on the web is dissipated as heat and not recovered
as elastic recoil, therefore the insect is unlikely to be catapulted out of the web and the energy converted to heat is not available to rupture the web [21]. The pentapeptide repeats of the £agelliform and dragline silks are predicted to form L-spirals [22] and water is essential for their elasticity, implying the spirals are held together by hydrophobic and hydrogen bonding. Their mechanism of elasticity may, therefore, have similarities to those of both elastin and the high Mr subunits of wheat. Thus, the repetitive sequences in silks and high Mr subunits appear to form similar secondary structures, which contribute to elasticity by their interactions. The advantage of an entropic mechanism of elastomeric force is that it depends on the decrease in the numbers of accessible states on extension and can be considered ideal for biological systems. Entropic elasticity provides for a more durable elastomer, with elastic ¢bres lasting the lifetime of an individual [23]. In evolutionary terms several unrelated proteins have evolved to achieve elasticity by a similar entropic mechanism, where energy is required to stress the proteins and relaxation is non-energy-consuming. However, in the dragline silks a purely entropic mechanism would not ful¢l the requirements of the material, so that a contribution from an internal energy mechanism and/or interaction through hydrogen bonds involving glutamine residues would be adventitious. The high Mr subunits have had no evolutionary pressure to develop elastic mechanisms. Seed storage proteins act solely as a store of carbon, nitrogen and sulphur for the developing seedling, with the main evolutionary pressures being for e¤cient packaging for storage and mobilization on germination. The high Mr subunits clearly provide e¤cient storage of nitrogen due to their high contents of glutamine, which their compact spiral structure may facilitate packaging within the protein bodies of the developing grain. Thus the elastomeric properties appear to have no biological signi¢cance and both these and their elastic mechanisms appear to have developed as a fortuitous consequence of the evolution of the proteins by reiteration of sequence repeats. Their elastomeric properties are, however, important in determining the ability of wheat £our to be mixed into doughs, which are expanded into ¢ne foams and baked to form leavened breads.
BBAPRO 36463 3-8-01
A.S. Tatham et al. / Biochimica et Biophysica Acta 1548 (2001) 187^193
Acknowledgements IACR receives grant-aided support from the Biotechnology and Biological Sciences Research Council of the United Kingdom. References [1] J.M. Gosline, Rubber Chem. Technol. 60 (1987) 417^438. [2] Z. Indik, N. Ornstein-Goldstein, P. Sheppard, N. Anderson, J.C. Rosenbloom, L. Peltonen, J. Rosenbloom, Proc. Natl. Acad. Sci. USA 84 (1987) 5680^5684. [3] Q. Cao, Y. Wang, H. Bayley, Curr. Biol. 7 (1997) R677^ R678. [4] K.J. Coyne, X.X. Qin, H. Waite, Science 277 (1997) 1830^ 1832. [5] C.Y. Hayashi, R.V. Lewis, J. Mol. Biol. 275 (1988) 773^784. [6] P.A. Guerrette, D.G. Ginzinger, B.H.F. Weber, J.M. Gosline, Science 272 (1996) 112^115. [7] C.A.J. Hoeve, P.J. Flory, Biopolymers 13 (1974) 677^686. [8] P.R. Shewry, A.S. Tatham, Biochem. J. 267 (1990) 1^12. [9] J.M. Field, A.S. Tatham, A. Baker, P.R. Shewry, FEBS Lett. 200 (1987) 76^80. [10] M.J. Miles, H.J. Carr, T.J. McMaster, K.J. I'Anson, P.S. Belton, V.J. Morris, J.M. Field, P.R. Shewry, A.S. Tatham, Proc. Natl. Acad. Sci. USA 88 (1992) 76^80.
193
[11] A.S. Tatham, A.F. Drake, P.R. Shewry, J. Cereal Sci. 11 (1990) 189^200. [12] A.A. van Dijk, L.L. van Wijk, A. van Vliet, P. Harris, E. van Swieten, G.I. Tesser, G.T. Robillard, Protein Sci. 6 (1997) 637^648. [13] B.A. Marchylo, J.E. Kruger, D.W. Hatcher, J. Cereal Sci. 9 (1988) 113^130. [14] D.W. Urry, K. Okamoto, R.D. Harris, C.F. Hendrix, M.M. Long, Biochemistry 15 (1976) 4083^4089. [15] P.J. Flory, A. Ci¡eri, C.A.J. Hoeve, J. Polym. Sci. 45 (1960) 235^236. [16] D.W. Urry, J. Protein Chem. 1 (1988) 1^34. [17] P.S. Belton, I.J. Colquhoun, J.M. Field, A. Grant, P.R. Shewry, A.S. Tatham, J. Cereal Sci. 19 (1994) 115^121. [18] S.M. Gilbert, N. Wellner, P.S. Belton, J.A. Green¢eld, G. Siligardi, P.R. Shewry, A.S. Tatham, Biochim. Biophys. Acta 1479 (2000) 135^146. [19] P.S. Belton, I.J. Colquhoun, J.M. Field, A. Grant, P.R. Shewry, A.S. Tatham, N. Wellner, Int. J. Biol. Macromol. 17 (1995) 74^80. [20] P.S. Belton, J. Cereal Sci. 29 (1999) 103^107. [21] J.M. Gosline, M.E. DeMont, M.W. Denny, Endeavour 10 (1986) 37^43. [22] C.Y. Hayashi, N.H. Shipley, R.V. Lewis, Int. J. Biol. Macromol. 24 (1999) 271^275. [23] D.W. Urry, Sci. Am. 272 (1995) 64^69.
BBAPRO 36463 3-8-01