De novo design and structure-activity relationships of peptide emulsifiers and foaming agents

design and structure-activity relationships of peptide emulsifiers and foaming agents De novo

M. Enser*, G. B. Bloomberg and C. Brock AFRC Institute of Food Research, Bristol Laboratory, Langford, Bristol BSI8 7D Y, UK

and D. C. Clark AFRC Institute of Food Research, Norwich Laboratory, Colney Lane, Norwich NR4 7UA, UK

(Received 28 August 1989; revised 16 November 1989) A series of eight amphipathic peptides (8,11,15, 2 x 18, 22, 26, 29 amino acids in length) were designed to investigate the effects of amino acid composition, peptide length and secondary structure on surface activity assessed as emulsification and foaming activity. The potential for ~-helix formation at the hydrophobic/hydrophilic interface was maximized through the use of helix-forming amino acids, a relatively large hydrophobic surface of 200 ° of arc and ion pairs between basic and acidic amino acids on the hydrophilic surface. Emulsification activity increased rapidly between 11 and 22 residues as ~-helicity in aqueous solution increased. Despite their small size, the peptides produced exceptionally stable emulsions, compared with proteins. Foaming activity was enhanced by the presence of aromatic amino acids and the activity of the best peptide examined was superior to that of bovine serum albumin and fl-lactoglobulin, Keywords: Peptides; emulsifiers;foaming agents

Introduction Many proteins are surface-active and are widely used as emulsifiers and foaming agents, particularly in the food industry 1-4, although such functionalities are incidental and their exploitation and enhancement have occurred in a pragmatic manner. The basis of much of the surface activity of proteins in the amphipathic ~-helix, was originally described to explain the interaction between the proteins and lipids in the plasma lipoproteins 5. There have been extensive studies of the surface active properties of these proteins and model peptides based upon their amino acid sequences 6-8. In the food area most work has concentrated on native proteins or peptides derived from them. Although much of the work on lipoproteins and related model peptides is relevant to understanding the mechanisms of emulsion formation and stabilization, that is not their sole function in vivo and hence their activity in these respects may be constrained by other required functionalities. Furthermore, the surface of the lipid droplet in the plasma lipoproteins contains cholesterol and phosphatides and it appears that charge distribution on the amphipathic a-helices of the proteins is arranged such that the cationic groups are at the edge of the molecule where they can interact with the phosphate groups of the phospholipids 9. In systems not containing anionic surfactants, internal ionic interactions in the peptide may be more significant than external ones since they are potentially important components in the stabilization of an ~-helix. The surface activity of proteins, such as soy or albumin, used in the food industry can be * To whomcorrespondenceshould be addressed. Presented in part at BiologicallyEngineered Polymers Conference, Churchill College,Cambridge,31 July-2 August 1989 0141-8130/90/020118~)7 © 1990Butterworth& Co. (Publishers) Ltd 118 Int. J. Biol. Macromol., 1990, Vol. 12, April

enhanced by partial denaturation of the protein to expose the hydrophobic surfaces within the core during processing 1°. Following adsorption, surface denaturation of the protein, interaction with neighbouring molecules and multilayer formation can enhance the stability of the film 11,12. However, only parts of a protein molecule may consist of amphipathic a-helices or t-sheets with high surface activity so that a peptide designed to have a maximally stabilized structure should have a higher potential specific surface activity. The aim of our programme is to investigate the role of the primary and secondary structure of peptides in the formation and stabilization of oil/water emulsions and foams. Here we report initial studies on the design and functionality of peptides with respect to peptide amino acid sequence and length.

Experimental Reagents were purchased from either Sigma Chemical Co. or Aldrich Chemical Co. as the highest grade of purity supplied. Fmoc amino acids for peptide synthesis were purchased from Milligen/Biosearch Division, Millipore (UK) Ltd. For the circular dichroism measurements, surface chemically pure water was prepared by steam distillation from alkaline permanganate solution ~3 and used throughout. Peptide synthesis

The peptides were synthesized on a semi-automatic Cambridge Research Biochemicals Pepsynthesizer 2 by the Fmoc-polyamide continuous flow method 14 using manufacturer's protocols with slightly modified flow

Peptide emulsifiers and foaming agents: M. Enser et al. times. The first amino acids were coupled to Pepsyn®KA resin as the symmetrical anhydride in the case of serine, and as the pentafluorophenyl ester in the case of lysine. The side chain protection used was Ser(But), Glu(OBu t) and Lys(Boc). Peptides were cleaved from the resin using trifluoroacetic acid/ethanedithiol/anisole (90:5:5) for 3 h, filtered, and the solution diluted with water. They were washed with two small portions of diethyl ether and lyophilized. The residue was dissolved in a solution of 10% glacial acetic acid and re-lyophilized to give a fluffy white precipitate of peptide. Amino acid analysis (Table 2) was performed on an LKB 4400 amino acid analyser. H.p.l.c. was performed on a Waters 600E gradient system fitted with a Phase Separations 25cm x20mm i.d. column using water (0.1%) trifluoroacetic acid [TFA]: acetonitrile(0.1% TFA) 20-80% over 50min with u.v. detection.

Circular dichroism measurements Peptide samples were prepared by dissolving lyophilized powder in a small volume of 20 mM Na phosphate pH 7.0. After frequent whirlimixing, the samples were centrifuged (using an MSE microcentaur centrifuge) for 5 min at 11 500 g. The concentration of the peptide in the supernatant was measured spectrophotometrically using an absorption coefficient for tryptophan of 5000 1 mol-1 cm-1. Samples for circular dichroism (c.d.) were prepared by dilution with phosphate buffer, distilled water and trifluoroethanol (TFE) as required, such that the final concentration of buffer was 20 mM phosphate pH 7.0. In addition, the peptide concentration was maintained at 0.5mg/ml unless specified otherwise in an effort to minimize errors arising from the concentrationdependent aggregation and helix content. Far u.v. circular dichroism spectra (190-260 nm) were measured using a Jasco J600 spectropolarimeter and data were recorded on-line using an IBM-PC. The spectra presented are an average of two scans, recorded at 10 nm min-1, using a silica quartz demountable cell of 0.1-mm pathlength. An instrument sensitivity of _ 20 mdeg full scale was routinely used, along with a 4 s time constant 15. The secondary structure content of c.d. spectra, recorded with peptide solutions of known concentration was analysed using the CONTIN program 16 on a VAX computer. The program fits the data with the c.d. spectra of 16 proteins of known 3D structure. The results of the best fit were output as fractions of the three structural components, or-helix, fl-sheet and random coil. Instrument calibration was regularly checked during the course of the work using ammonium dl0-camphorsulphonate 17 and d ( - )-pantalactone 1s

Emulsification measurements Emulsification activity indices (EAIs) and stability were determined according to the method of Pearce and Kinsella 19 with slightly modified conditions. Five ml of peptide at 2 mg/ml in 10 mM phosphate buffer at pH 5.0 and pH 7.0 were blended with 5 ml corn oil in a 25 ml measuring cylinder at 22 700 rev min-1 for 20 s using a Kinematica Polytron (giving a 1:1 oil/water ratio. This is referred to as the oil-in-water emulsion). After degassing and gentle stirring, a 10/A aliquot was diluted to 10ml and the u.v. absorbance read at 500nm. This was

repeated with 0.1% SDS present. The EAI was calculated as 2T/~bC where T is the turbidity and ~b is the volume fraction of the dispersed phase and C is the weight of protein per unit volume 19. Measurements with SDS were to test for flocculation of the oil globules. Emulsion stability was measured as the time from emulsification to the first appearance of an oil phase at the top of the emulsion, indicating the start of breakage.

Foaming measurements Peptide samples were prepared by addition of 0.9 ml distilled water to the lyophilized powder. After occasional agitation, ammonia vapour was puffed onto the surface of the suspension. Finally, 0.1 ml of phosphate buffer was added to the solution. After frequent whirlimixing, the samples were centrifuged (using an MSE microcentaur centrifuge) for 5 min at 11 500 g. The supernant was then loaded onto a Sephadex G-15 column (0.9×30cm) pre-equilibrated with 0.1 M phosphate pH 7.0. Fractions containing the peptide were pooled and centrifuged at 11 500 g for a further 5 min. The concentration of soluble peptide was estimated spectrophotometrically using an absorption coefficient for tryptophan of 50001 mol-1 cm- 1 at 280 nm. These samples were prepared in this manner to ensure that all the peptide was fully in solution so that their foaming properties could be analysed directly using a microconductivity apparatus first described by Wright and Hemmant 2°. The apparatus consisted of a column containing platinum electrodes which were flush with the walls of the column. The column itself was enclosed within a water jacket which was thermostatically regulated during the course of the experiment at 20°C. The foam was generated by sparging 2 ml of peptide solution contained within the column with oxygen-free nitrogen at a pressure of 131 kPa, which entered the base of the column through a 15-pm Coulter orifice. Sparging was continued until the foam had risen 6 mm above the electrodes. The conductivity data were recorded using a Philips PW9505 conductivity meter coupled to a chart recorder and BBC microcomputer. Initial data analysis was undertaken using the parameters identified by Wright and Hemmant 2°. In their comparison of conductimetric and whipping methods in the study of protein foams, they found that the maximum conductivity (Co) observed immediately after sparging had stopped, correlated with foam expansion and C1/2/Co, where C1/2 corresponds to the conductivity 0.5 min from the maximum and had been shown to correlate with foam stability.

Results Eight peptides (Table I) were designed using helical net 21 and helical wheel 22 diagrams as graphical aids. Peptides 1-7 progressively increase in length but carry on the same sequence and peptide 8 is a variation of peptide 4 (18 amino acids in length) where the leucine residues are replaced by tryptophan residues. The aim was to obtain maximum stability of the peptides at the oil/water interface as an amphipathic helix. Peptide 8 (modelled from peptide 4) was designed to enhance foaming activity. Three major areas were considered in the peptide design: (1) factors which stabilize or-helices; (2) the overall shape of the molecule for emulsion stability; and (3) the

Int. J. Biol. Macromol., 1990, Vol. 12, April

119

Peptide emulsifiers and foaming agents: M. Enser et al. Table 1 Sequences of designed synthetic peptides Reference

No. of amino acids

Sequence

Possible helical turns

Peptide Peptide Peptide Peptide Peptide Peptide Peptide Peptide

8 11 15 18 22 26 29 18

SWAEALSK SWAEALSKLLS SWAEALSKLLESLAK SWAEALSKLLESLAKALS SWAEALSKLLESLAKALSELAK SWAEALSKLLESLAKALESLAKALES SWAEALSKLLESLAKALSELAKALESLLK SWAEAWSKWWESWAKAWS

2 3 4 5 6 7 8 5

1 2 3 4 5 6 7 8

I

I

I

J.L u I I

L...__

1, ol

I

o.ol sE. Figure 1 An example of a helical net diagram drawn for peptide 4. The dotted lines linking residues in successive turns of the helix indicate potential salt bridges which were incorporated with the aid of such diagrams

hydrophobic/hydrophilic balance of amino acids (which was kept close to 50/50). For category 1 several features were incorporated which are known to stabilize a-helices. First, a distinct hydrophobic/hydrophilic boundary was defined with the hydrophobic region occupying 200 ° of the circumference of the helix. The amino acids in the hydrophobic domain contribute to the stabilization of the overall configuration by hydrophobic interactions. Leucine and alanine were chosen as the hydrophobic amino acids since they have a high propensity for occurring in ~-helices 23. Second, the hydrophilic amino acids lysine and glutamic acid were chosen so that when placed at positions i and i + 4 in the peptide, helix stabilization by the formation of an intramolecular G l u - ... Lys ÷ salt bridge is incorporated (Figure I). The 0t-helix is also stabilized by electrostatic interactions of oppositely charged groups with either end of the overall helix dipole 24'2s. Although it was not possible to incorporate this feature in conjunction with the other considerations, a glutamic acid residue was placed at position 4 from the amino terminus thus allowing some possible interaction. A tryptophan residue was also incorporated because of its properties as a strong u.v. chromophore. The other main area considered in peptide design was the shape of the molecule. The helical wheel representation of

120

Int. J. Biol. Macromol., 1990, Vol. 12, April

Figure 2 An example of a helical wheel diagram drawn for peptide 4 which facilitated the design of amphipathic structures. The hydrophobic and hydrophilic regions of the peptide can be seen clearly peptide 4 (Figure 2) shows large hydrophobic side chains (leucine) flanked by smaller hydrophobic side chains (alanine) on the hydrophobic face. The hydrophilic face comprises equal numbers of positively (Lys ÷) and negatively (Glu-) charged amino acids with large side chains, flanked by small polar but uncharged (Ser) side chains. When such a molecule orientates itself at an interface, the large hydrophobic chains will extend into the apolar phase and the large hydrophilic chains will extend into the polar phase. The smaller alanine and serine side chains are close to the phase interface. This 'oval-shaped' molecule should permit efficient packing at the oil/water interface thus increasing its effectiveness as an emulsifier.

Peptide purity Peptides were checked for purity by amino acid analysis (Table 2) and h.p.l.c. In all cases they were of 90% or greater purity which was considered sufficient for this initial study.

Characterization of solution properties of the peptides The secondary structure composition of the peptides was examined by far-u.v.c.d. Typical spectra for samples of varying chain length at a concentration of 0.15 mg ml-1 are shown in Figure 3. The spectrum of peptide 2 (11 residues) shows features typical of a sample

Peptide emulsifiers and foaming agents: M. Enser et al. Table 2 Amino acid analysis of peptides 1-7 Reference

Ser

Glu

Ala

Leu

Lys

Peptide 1 Peptide 2 Peptide 3 Peptide 4 Peptide 5 Peptide 6 Peptide 7

1.9(2) 3.1 (3) 2.9 (3) 4.1 (4) 3.9 (4) 4.8 (5) 4.8 (5)

112(1) 1.0(1) 2.3 (2) 2.2 (2) 3.2 (3) 3.9 (4) 4.2 (4)

2.1 (2) 2.1 (2) 3.0 (3) 4.2 (4) 5.0 (5) 5.9 (6) 6.3 (6)

0.9(1) 3.1 (3) 3.0 (4) 5.0 (5) 5.9 (6) 7.0 (7) 8.9 (9)

0.9(1) 1.0(1) 2.0 (2) 1.9 (2) 3.2 (3) 3.0 (3) 4.3 (4)

I0.0--

contained some a-helix, including peptides 1 and 2 (8 and 11 residues). On average the or-helical content of the samples in 50% T F E was increased by approximately 10-15% over that observed in aqueous solution. The conformation of peptide 8 was also examined by c.d. This peptide gave an unusual spectrum, with a small positive peak at 230 nm and a relatively weak negative minimum at 222 nm, which increased in negative intensity to a minimum at approximately 208 nm. The C O N T I N analysis of this spectrum returned a structural composition of 90% r-sheet and 10% 0t-helix and was almost unaffected by the addition of TFE. These results may be explained by contributions in the 220 nm region from the B E transition of the indole ring of tryptophan 26. Indeed, simple tryptophan derivatives such as N-acetyltryptophanmethylamide do have c.d. bands at 220 nm. Thus, a far-u.v, spectrum containing contributions from tryptophan and or-helix could produce a spectrum with features that are similar to those displayed by peptide 8.

Emulsification studies

S.O--

The emulsification properties of the peptides were examined at pH 5.0 and pH 7.0 using a turbidimetric assay 19 and the results obtained for emulsifying activity index and emulsion stability are presented in Figures 4 and 5 where they are plotted as a function of peptide length. Peptides 1 and 2 showed no emulsification activity

0.0

Table 3 The chain length dependence of helix content of the peptides -s.o

I 200

260

Reference

No. of amino acids

Helix % (aqueous)

Helix % (50% TFE)

Far-u.v.c.d. spectra of amphipathic peptides of varying chain length. (O) peptide 2, (O) peptide 3, (A) peptide 4, ( + ) peptide 6

Peptide 1 Peptide 2 Peptide 3 Peptide 4 Peptide 5 Peptide 6 Peptide 7

8 11 15 18 22 26 29

0.2 nil 18.3 27.3 37.5 37.1 47.4

11.3 15.8 27.9 38.9 43.8 56.9 64.2

180

I

I

220

240

Wavelength

(nm)

Figure 3

containing a mixture of both r-sheet (presumably intermolecular) and random coil. In contrast, peptide 3 (15 residues) shows clear indications of the presence of helix in the sample, with a sub-minimum in the 218 nm region and a minimum at 205-208 nm (Figure 3). However the spectrum has low negative intensity, signifying that s-helix is not the predominant structural component. Peptides 4 and 6 (18 and 26 residues) exhibit progressively more negative intensity in the a-helix region showing that the contribution of this component increases with chain length. Quantitative assessment of the secondary structure content of the spectra was made using the C O N T I N programme 16. The results are summarized in Table 3. Under the prevailing experimental conditions, the longest peptide examined, peptide 7 (29 residues), contained the highest level of or-helix (approx. 50%). The effect of reduced solvent polarity on peptide structure was examined by the addition of T F E to the samples. The rational behind these experiments was to investigate the possible induction of structure into the peptides upon adsorption at the oil/water or air/water interface. The at-helix content of the peptides in the presence of 25% and 50% T F E was determined by the C O N T I N program and the results obtained at 50% T F E are summarized in Table 3. Under these conditions, all of the peptides

Percentage helix was determined by the CONTIN method (ref. 16)

ot

0.3-

E .120

E

ILl v .100

O u')

0.2

+

~A

+

BSA pHS

~7

O c t~

' 80

~c_

- 6o

.>

0.1 ¸ - 40

O .Q

20

0.0

g 0

0 I0

20

30

E

W

Peptide length (amino acids)

Figure 4 The variation in emulsion activity index of peptides 1-8 with length assayed at two pHs. [(e) pH 7, ([]) pH 5.] The index for BSA is included to demonstrate the slightly superior activity of several of the synthetic peptides. Peptide 8 is shown at the point on the x-axis corresponding to its length (18 amino acids), marked with an asterisk (*)

Int. J. Biol. Macromol., 1990, Vol. 12, April

121

Peptide emulsifiers and foamin 9 ayents: M. Enser et al. 50

40' .~

30.

~0

20.

>', (U C~

10' + BSApH7 + BSApH5_ !

10

20

30

Pepfide length (amino acids)

Figure 5 The variation in emulsion stability of the peptides series 1-7 with length at two pHs. r(¢) pH 7, (t3) pH 5.] BSA is indicated to demonstrate the the markedly superior stability of several of the synthetic peptides. Peptide 8 is not shown since it was inactive in emulsion formation at pH7, and at pH5 the oil-in-water emulsion broke after 1 day at all. Peptide 3 showed activity at pH 7.0 only, where it stabilized the oil in water emulsion for 28 days. Peptides 5, 6 and 7 showed superior emulsifying activity indices (EAIs) compared with that of BSA which is one of the best protein emulsifiers. All peptides that showed emulsification activity showed superior emulsion stabilization properties at pH 7.0. Peptide 5 stabilized the oil-in-water emulsion for 34 days which was over eight times longer than BSA. At pH 5.0, superior emulsion stabilization properties were shown by peptides 5, 6 and 7. Each of these peptides stabilized the oil-in-water emulsion for 16-19 times longer (32-39 days) than BSA (2 days). Each emulsion, once broken, could be reformed by reprocessing with the polytron, indicating a remarkable resistance to over-processing compared with natural proteins. At pH 5.0, peptide 8 showed similar behaviour to peptide 4 which is of the same length, however at pH 7.0 no emulsification activity was observed with the former sample.

it had improved foamability (i.e. was more effective at encapsulating the gas) compared with the protein. The maximum conductivity achieved by peptide 8 was significantly greater than /%lactoglobulin, indicating that a greater volume of liquid bridged the electrode plates in this sample. This could be explained by either the average bubble size being smaller in the case of the peptide foam or that the films in this foam were thicker. Further work is required to identify the precise reason. The subsequent drop in conductivity resulted from the rapid drainage of bulk liquid from the foam. The rate of this drainage appeared to be faster for the protein. After several hundred seconds the conductivity reached a pseudoplateau, the level and persistence of which appeared to correlate with the observed stability of the foam. The peptide maintained a higher conductivity in this region than the protein, and was judged to be more stable. Visual inspection of the foams formed confirmed that the bubble size distribution was more uniform in the peptide sample and this foam persisted for many hours. In contrast, the protein foam tended to have less uniform bubble sizes, coarsened quite rapidly and ultimately collapsed over a similar time period.

Discussion Proteins and detergents (or small molecule emulsifiers) stabilize interfaces by two distinct mechanisms 27. Most proteins form viscoelastic films at interfaces by forming protein-protein interactions with neighbouring molecules and it is this property rather than the reduction in surface tension that enables these molecules to produce stable emulsions and foams. The surface absorption of detergents causes a reduction in surface or interfacial tension which enhances the stability of foams or emulsions. The films formed have low surface viscosity and mechanical elasticity. Thus, if the film is subject to local stretching, surfactant concentration and surface tension gradients are produced. The equilibrium can be restored in these systems by either absorption of surfactant from bulk or lateral diffusion at the surface since the absorbed molecules do not interact to form a cohesive film. This is termed the Gibbs-Marangoni effect. Clearly, these two mechanisms are incompatible since one relies on interac-

Foamin9 studies The foaming experiments revealed that both foam expansion and stability increased with peptide length. Peptides 6 and 7 produced foams of comparable stability to those of the proteins investigated. Peptide 5 was not as good a performer as the proteins and the shorter peptides were even less effective. Stepwise conservative substitution of the amino acids present in peptide 4 resulted in the synthesis of a peptide that possessed considerably enhanced foaming properties. This peptide (peptide 8) was a derivative of peptide 4 which had Trp residues in place of Leu (Table I). Typical conductivity decay curves obtained from peptide 8 and fl-lactoglobulin are shown in Figure 6. Gas sparging was initiated at time zero (Figure 6). A few seconds elapsed before the foam rose to a sufficient height in the column to bridge the electrodes and produce a measurable conductivity signal. This signal increased until the sparging was stopped, when the foam reached a preset level above the electrode. Peptide 8 achieved this level in a shorter time than fl-lactoglobulin, signifying that

122

Int. J. Biol. Macromol., 1990, Vol. 12, April

3.0j

~

Peptide8 .....

o

13-Lactoglobulin

2.0

2

-

1.0tO

0

0.0

|

200

!

400

800

Time (See) Figure 6 Conductivity decay curves of foams of peptide 8 and fl-lactoglobulin. Sample concentration was 0.15 mg/ml

Peptide emulsifiers and foaming agents: M. Enser et al. tion (i.e. proteins) whilst the other depends on the surface absorbed species being free to diffuse (i.e. detergents). Both proteins and small molecule surfactants are present in most food systems. The equilibrium surface concentrations of the various components is determined by their surface activity and bulk concentration. In many cases the interfaces will be populated by both classes of components and when this occurs instability can result. The initial absorption of molecules is diffusion controlled and since small molecule surfactants are of lower molecular weight than proteins they tend to populate the freshly created interface first. Proteins diffuse slowly and after absorption undergo slow conformational changes in their tertiary and secondary structure (surface denaturation) to expose their hydrophobic core to the oil or air phase. Small decreases in interfacial/surface tension have been reported to occur over several tens of hours after protein absorption and have been attributed to changes in protein conformation. In this work we have examined the possibility of improving the performance of proteins in the stabilization of dispersed phase systems. Using de novo designed synthetic peptides of low molecular weight we have increased the bulk diffusion coefficient to facilitate the initial absorption and increase the rate of any conformational change at the surface since these molecules are too small to possess tertiary structure. Preliminary results obtained from drainage studies performed on air suspended peptide stabilized thin liquid films revealed that the peptides formed viscoelastic films similar to proteins but distinct from small molecule surfactants 2s. Thus, peptide interactions at the interface appear to be present and could be an important factor in the stabilization of interfaces by these molecules. Therefore, the performance of the peptides has been compared with proteins rather than small molecule surfactants. Simple rules have been used in the design and synthesis of a series of peptides, which adopt a structure in solution containing some amphipathic or-helix with inherent high surface activity, in order to promote the formation of highly stable emulsions and foams. Circular dichroism has been used to assess the extent of 0t-helix formation by peptides in solution and has beer~ correlated with their emulsifying activity. The results show that emulsifying activity is only present in samples which show some helical conformation in solution (i.e. peptides 3-7). The condition is met only when there is sufficient sequence to support [in theory], at least three turns of a-helix. The estimated values of ~t-helix content obtained from analysis of the c.d. spectra refer to the [average] conformation of the peptide arranged over many molecules. For example, it is not possible for individual molecules ofpeptide 3 to be 20% helical, since this corresponds to a maximum of 0.8 turns of or-helix in this molecule. It is more likely that the peptide is present in solution in a variety of conformations, probably with an equilibrium existing between the extremes of 0 and 2 turns of c~-helix. All the peptides show an increased or-helix formation in the presence of 50% TFE, indicating that they can undergo a conformational change when subjected to an alteration in solvent polarity. This implies that a conformational change could occur upon adsorption of the peptide at an oil/water er air/water interface. Given the small size of the peptides and the absence of tertiary structure, any change in structure would be very fast and it is unlikely to be a controlling factor in the stabilization

of a disperse system. Therefore, it is unclear at this stage whether the Ilength] dependence of emulsifying activity on the length of the peptide is controlled by the presence of some or-helix in the peptide in aqueous solution--which may act as a catalyst for further 0t-helix induction upon adsorption. It is difficult to make definitive statements regarding the relationship between peptide length and emulsifying activity/stability from the preliminary data presented in this report. However, the longer peptides significantly outperform BSA and the threshold length at which this occurs is apparently pH-dependent (peptides 3-7 at pH 7.0 and peptides 5-7 at pH 5.0). The explanation for the pH dependence must be related to changes in charge on the peptide, but further work is required to confirm this observation. The foaming properties of the peptides were examined by conductivity. The conductivity decay curves contain information about a range of processes occurring in the foam including drainage, disproportionation and film rupture. It is not possible to delineate these individual processes from the data without a universal theory and independent measurements of the discrete physical processes. We have used conductivity to screen our synthesized peptides in order to identify the most effective foamers. A more thorourh characterization of the best performers will follow. The conductivity measured relates to the volume of solution in the foam that connects the two electrode plates in the conductivity cell. Other workers 2° who have used this method have assigned numerical values to foam expansion and stability. Reasonable correlation exists between this method and a standard whipping technique for the quantitation of protein foams. However, we have observed that the shape of the conductivity decay curves from simple detergents is highly concentration dependent. Thus, we restricted ourselves to qualitative analysis until we have a clearer understanding of the drainage of peptide foams studied by conductimetric methods. The foaming properties of the peptides improved with increasing chain length and helix content However, only the longer members of the series examined (i.e. peptides 6 and 7) produced foams of comparable stability and expansion to the proteins investigated by the conductrimetric method. The nature of the hydrophobic side chain does appear to be an important factor in the determination of foaming activity. The substitution of Leu by Trp (peptide 4 and 8) produced a significant improvement in foaming properties. The reason for this is unclear. The secondary structure of peptide 8 has been examined by c.d. but quantitative assessment of the helix content was complicated by the superimposition of a c.d. band from the BL transition of the indole ring with the peptide spectrum. Simple subtraction of the c.d. spectrum of N-acetyl tryptophanamide at a concentration similar to the side chain content ofpeptide 8 produced a spectrum which upon analysis appeared to contained less helix than peptide 4. Although the accuracy of structure determination by this empirical approach may be questionable, it would suggest that the differences in foaming behaviour of peptides 4 and 8 were not a consequence of their secondary structure in solution. In addition, the change in foaming properties does not appear to be linked to the hydrophobicity of the side chains since Trp is considerably less hydrophobic than Leu 29. At present the explanation for this effect eludes us but we are currently

Int. J. Biol. Macromol., 1990, Vol. 12, April

123

P e p t i d e emulsifiers a n d f o a m i n g a g e n t s : M . E n s e r et al.

considering that the stacking of indole rings of Trp could possibly have some influence on the foaming behaviour of peptide 8.

10 11 12

Conclusions Synthetic amphipathic a-helical peptides were shown to be highly effective emulsifiers and stabilizers of foams. Emulsion-forming activity showed a close correlation with a-helix content of the peptides in aqueous solution. The stabilization of foams is more pronounced with peptides containing hydrophobic amino acids with aromatic rather than aliphatic side chains.

13 14 15 16 17 18 19

References 1 2 3 4 5 6 7 8 9

124

Vananuvat, P. and KinseIla, J. E. J. Agric. Food Chem. 1975, 23, 613 Yasumatsu, K., Sawada, K., Moritaka, S., Misaki, M., Toda, J., Wada, T. and Ishii, H. Agr. Biol. Chem. 1972, 36, 719 Dziezak, J. D. Food Technol. 1988, 45, 171 Dickinson, E. Food Hydrocolloids, 1986, 1, 3 Segrest, J. P., Jackson, R. L., Morrisett, J. D. and Gotto, A. M. FEBS Lett. 1974, 38, 247 Chung, B. H., Anantharamaiah, G. M., Brouillette, C. G., Nishidu, T. and Segrest, J. P. d. Biol. Chem. 1985, 260, 10248 Davis, J. H., Clare, D. M., Hodges, R. S. and Bloom, M. Biochemistry, 1983, 22, 5298 Kanellis, P., Romans, A. Y., Johnson, B. J., Kercret, H., Chiovetti, R., Allen, T. M. and Segrest, J. P. J. Biol. Chem. 1980, 255, 11464 Anantharamaiah, G. M., Jones, J. L., Brouillette, C. G., Schmidt,

Int. J. Biol. M a c r o m o l . , 1990, Vol. 12, A p r i l

20 21 22 23 24 25 26 27 28 29

C. F., Chung, B. H., Hughes, T. A., Bhown, A. S. and Segrist, J. P. J. Biol. Chem, 1985, 260, 10248 Kinsella, J. F. J. Am. Oil Chemists" Soc. 1979, 56, 242 Phillips, M. C. Food Technol. 1981, 35, 50 Clark, D. C., Mackie, A. R., Smith, L. J. and Wilson, D. R. in "Food Colloids' (Eds R. D. Bee, J. Mingins and P. Richmond), Royal Society of Chemistry special publication No. 75, 1989, p 97 Clark, D. C., Mingins, J., Sloan, F. E., Smith, L. J. and Wilson, D. R. in 'Food Emulsions and Foams' (Ed. E. Dickinson), Royal Society Special Publication No. 58, 1987, p 110 Atherton, E. and Sheppard, R. C. d. Chem. Soc. Chem. Commun. 1985, 3, 165 Clark, D. C. and Smith, L. J. J. Agric. Food Chem. 1989, 37, 627 Provencher, S. W. and Glockner, J. Biochemistry, 1981, 20, 33 Konno, T., Meguro, H. Tuzimura, H. Anal. Biochem. 1975, 67, 226 Takakuwa, T.,Konno, T. andMeguro, H. Anal. Sci. 1975,1,215 Pearce, K. N. and Kinsella, J. E. d. Agric. Food Chem. 1987, 41, 361 Wright, D. J. and Hemmant, J. d. Sci. Food Agric. 1987, 41,361 Dunill, P. Biophys. J. 1968, 8, 865 Schiffer, M. and Edmunson, A. B. Biophys. J. 1967, 7, 121 Chou, P. Y. and Fasman, G. D. Adv. Enzymol. 1978, 47, 45 Shoemaker, K. R., Kim, P. S., York, E. J., Stewart, J. M. and Baldwin, R. L. Nature, 1987, 326, 563 Shoemaker, K. R., Kim, P. S., Brems, D. N., Marqusee, S., York, E. J., Chaiken, I. M., Stewart, J. M. and Baldwin, R. L. Proc. Natl. Acad. Sci. USA 1985, 82, 2349 Woody, R. W. 'Proceedings of the 2rid Intl. Conference on Circular Dichroism,' Budapest, 1987, p 38 Clark, D. C., Coke, M., Smith, L. J. and Wilson, D. R. in 'Foams: Physics, Chemistry and Structure' (Ed. A. J. Wilson), SpringerVerlag, 1989, 55 Clark, D. C. Unpublished data Zimmerman, J., Eliezer, N. and Simha, R. J. Theoret. Biol. 1968, 21, 170