Tuning β-sheet peptide self-assembly and hydrogelation behavior by modification of sequence hydrophobicity and aromaticity

Peptide self-assembly leading to cross-β amyloid structures is a widely studied phenomenon because of its role in amyloid pathology and the exploitation of amyloid as a functional biomaterial. The self-assembly process is governed by hydrogen bonding, hydrophobic, aromatic π-π, and electrostatic Coulombic interactions. A role for aromatic π-π interactions in peptide self-assembly leading to amyloid has been proposed, but the relative contributions of π-π versus general hydrophobic interactions in these processes are poorly understood. The Ac-(XKXK)(2)-NH(2) peptide was used to study the contributions of aromatic and hydrophobic interactions to peptide self-assembly. Position X was globally replaced by valine (Val), isoleucine (Ile), phenylalanine (Phe), pentafluorophenylalanine (F(5)-Phe), and cyclohexylalanine (Cha). At low pH, these peptides remain monomeric because of repulsion of charged lysine (Lys) residues. Increasing the solvent ionic strength to shield repulsive charge-charge interactions between protonated Lys residues facilitated cross-β fibril formation. It was generally found that as peptide hydrophobicity increased, the required ionic strength to induce self-assembly decreased. At [NaCl] ranging from 0 to 1000 mM, the Val sequence failed to assemble. Assembly of the Phe sequence commenced at 700 mM NaCl and at 300 mM NaCl for the less hydrophobic Ile variant, even though it displayed a mixture of random coil and β-sheet secondary structures over all NaCl concentrations. β-Sheet formation for F(5)-Phe and Cha sequences was observed at only 20 and 60 mM NaCl, respectively. Whereas self-assembly propensity generally correlated to peptide hydrophobicity and not aromatic character the presence of aromatic amino acids imparted unique properties to fibrils derived from these peptides. Nonaromatic peptides formed fibrils of 3-15 nm in diameter, whereas aromatic peptides formed nanotape or nanoribbon architectures of 3-7 nm widths. In addition, all peptides formed fibrillar hydrogels at sufficient peptide concentrations, but nonaromatic peptides formed weak gels, whereas aromatic peptides formed rigid gels. These findings clarify the influence of aromatic amino acids on peptide self-assembly processes and illuminate design principles for the inclusion of aromatic amino acids in amyloid-derived biomaterials.     

Interaction of aromatic residues of proteins with nucleic acids. Circular dichroism studies of the binding of oligopeptides to poly(adenylic acid).

The binding of peptides containing lysyl and aromatic residues to poly(A) in its single-stranded form at pH 7 leads to a change of its circular dichroism (CD) spectrum, which is mainly due to the stacking of the aromatic amino acid with the bases of poly(A). Comparison is made between the binding of peptides having different primary structures which gives indications on the way the peptides bind to poly(A). A method is described which allows the calculation of the binding parameters from CD data. The magnitude of the association constant depends on the size of the aromatic ring and decreases in the order tryptophan greater than tyrosine greater than phenylalanine. The CD amplitude decreases linearly with the concentration of bound molecules. These results are discussed with respect to the role played by aromatic amino acids in complex formation between nucleic acids and proteins.     

Bulky aromatic amino acids increase the antibacterial activity of 15-residue bovine lactoferricin derivatives.

A model peptide, FKCRRWQWRMKKLGA, residues 17-31 of bovine lactoferricin, has been subjected to structure-antibacterial activity relationship studies. The two Trp residues are very important for antibacterial activity, and analogue studies have demonstrated the significance of the size, shape and aromatic character of the side chains. In the current study we have replaced Trp residues in the model peptide with bulky aromatic amino acids to elucidate further the importance of size and shape. The counterproductive Cys residue in position 3 was also replaced by these aromatic amino acids. The largest aromatic amino acids employed resulted in the most active peptides. The peptides containing these hydrophobic residues were generally more active against Staphylococcus aureus than against Escherichia coli, indicating that the bacterial specificity as well as the antibacterial efficiency can be altered by employing large hydrophobic aromatic amino acid residues.     

The role of tryptophan in the antibacterial activity of a 15-residue bovine lactoferricin peptide.

Bovine lactoferricin is a 25-residue antibacterial peptide isolated after gastric cleavage of the iron transporting protein lactoferrin. A 15-residue fragment, FKCRRWQWRMKKLGA of this peptide sustains most of the antibacterial activity. In this truncated sequence, the two Trp residues are found to be essential for antibacterial activity. The anchoring properties of Trp, as have been observed in membrane proteins, are believed to be important for the interaction of Trp containing antibacterial peptides with bacterial cell membranes. We have investigated the molecular properties which make Trp important for the antibacterial activity of the 15-residue peptide by replacing Trp with natural and unnatural aromatic amino acids. This series of peptides was tested for antibacterial activity against Echerichia coli and Staphylococcus aureus. We found that neither the hydrogen bonding ability nor the amphipathicity of the indole system are essential properties for the effect of Trp on the antibacterial activity of the peptides. Replacement of Trp with residues containing aromatic hydrocarbon side chains gave the most active peptides. We propose that aromatic hydrocarbon residues are able to position themselves deeper into the bacterial cell membrane, making the peptide more efficient in disrupting the bacterial cell membrane. From our results the size, shape and aromatic character of Trp seem to be the most important features for the activity of this class of Trp containing antibacterial peptides.     

Role of strong anchor residues in the effective binding of 10-mer and 11-mer peptides to HLA-A*2402 molecules

The binding capacity of one-hundred-and-seventy-two 8-mer to 11-mer peptides carrying HLA-A24 anchor residues to HLA-A^*2402 molecules was analyzed by using a HLA class I stabilization assay. Most (76.2%) of these peptides bound to HLA-A^*2402 molecules. These results confirmed previous findings that Tyr and Phe at P2 as well as Phe, Trp, Ile, and Leu at the C-terminus were main anchor residues for HLA-A^*2402. Tyr at P2 was a stronger anchor residue than Phe, while bulky aromatic hydrophobic residues Phe and Trp at the C-terminus are stronger anchors than aliphatic hydrophobic residues Ile and Leu. These results were also supported by an analysis using a panel of mutated 9-mer peptides at P2 and P9. Taken together, these results suggest that HLA-A^*2402 molecules have deep B- and F-pockets because they favor peptides carrying bulky aromatic hydrophobic residues at P2 and the C-terminus. The affinity of 8-mer peptides was significantly lower than that of 9-mer to 11-mer peptides, while there was no difference in affinity between 9-mer, 10-mer, and 11-mer peptides. The affinity of peptides carrying bulky aromatic hydrophobic residues at the C-terminus was higher than that of peptides carrying aliphatic hydrophobic residues in each of the 8-mer to 11-mer peptides, though the greatest difference in affinity was observed in 11-mer peptides. The strong interaction of side chains of these anchor residues with the corresponding pockets may permit the effective binding of 10-mer and 11-mer peptides to HLA-A^*2402 molecules.     

Mapping and analysis of protein sulfhydryl groups by modification with 2-bromoacetamido-4-nitrophenol

A method for identifying cysteine-containing peptides in proteins is presented using 2-bromoacetamido-4-nitrophenol (BNP) to introduce an easily detectable probe. The formation of a covalent bond between the protein sulfhydryl group and the acetamido moiety of BNP introduces a chromophore with an absorbance maximum at 410 nm. The modified protein can then be cleaved with appropriate proteases and the resulting peptides separated by chromatographic methods. Monitoring the effluent at a single wavelength (405 nm) provides a rapid and simple method of detecting and isolating only those peptides which contain cysteine residue(s). The nitrophenol derivative is stable under conditions required for protease cleavage. The reagent is therefore useful for locating cysteine-containing peptides in protein digests and can be used to explore the accessibility of different cysteines under a variety of conditions. The ease of modification, specificity of reaction, product stability, and simple detection of modified peptides make BNP ideal for investigation of cysteine residues.     

Effects of i and i+3 residue identity on cis-trans isomerism of the aromatic(i+1)-prolyl(i+2) amide bond: implications for type VI beta-turn formation

Cis-trans isomerization of amide bonds plays critical roles in protein molecular recognition, protein folding, protein misfolding, and disease. Aromatic-proline sequences are particularly prone to exhibit cis amide bonds. The roles of residues adjacent to a tyrosine-proline residue pair on cis-trans isomerism were examined. A short series of peptides XYPZ was synthesized and cis-trans isomerism was analyzed. Based on these initial studies, a series of peptides XYPN, X = all 20 canonical amino acids, was synthesized and analyzed by NMR for i residue effects on cis-trans isomerization. The following effects were observed: (a) aromatic residues immediately preceding Tyr-Pro disfavor cis amide bonds, with K(trans/cis)= 5.7-8.0, W > Y > F; (b) proline residues preceding Tyr-Pro lead to multiple species, exhibiting cis-trans isomerization of either or both X-Pro amide bonds; and (c) other residues exhibit similar values of K(trans/cis) (= 2.9-4.2), with Thr and protonated His exhibiting the highest fraction cis. beta-Branched and short polar residues were somewhat more favorable in stabilizing the cis conformation. Phosphorylation of serine at the i position modestly increases the stability of the cis conformer. In addition, the effect of the i+3 residue was examined in a limited series of peptides TYPZ. NMR data indicated that aromatic residues, Pro, Asn, Ala, and Val at the i+3 residue all favor cis amide bonds, with aromatic residues and Asn favoring more compact phi at Tyr(cis) and Ala and Pro favoring more extended phi at Tyr(cis). D-Alanine at the i+3 position particularly disfavors cis amide bonds.     

蛋白质及多肽的C端分析

本文采用（异）硫氰酸化学法,以乙酸酐为活化试剂,四丁铵硫氰酸为偶联试剂,溴甲基萘为烷化试剂,将蛋白质或多肽的Ｃ端依次转化并裂解为ＡＴＨ－氨基酸,在２５４ｎｍ进行检测。分析了若干种天然的、基因工程表达的蛋白质或多肽,测定了不同长度的Ｃ端序列,并确证了Ｃ端的修饰及突变情况。为蛋白质和多肽的完整性、均一性,基因工程产品及合成多肽的质控提供了重要的Ｃ端信息。目前,可在１￣２ｎｍｏｌ水平上测定了３￣５个Ｃ端     

The effects of hydrophobic mismatch between phosphatidylcholine bilayers and transmembrane alpha-helical peptides depend on the nature of interfacially exposed aromatic and charged residues

In this study, we investigated the extent to which different aromatic and positively charged side chains, which often flank transmembrane segments of proteins, can influence lipid-peptide interactions. Model systems consisting of phosphatidylcholine and hydrophobic alpha-helical peptides with different flanking residues were investigated. The peptides were incorporated in relatively thick and in relatively thin lipid bilayers to create a peptide-bilayer hydrophobic mismatch, and the compensating effects on lipid structure were analyzed. When relatively long with respect to the thickness of the bilayer, the peptides that are flanked by the aromatic side chains, Trp, Tyr, and Phe, all induce a significant ordering of the lipid acyl chains, while the peptides flanked by the charged residues Lys, Arg, and His do not. However, when the peptides are relatively short with respect to the thickness of the bilayer, their effect on lipid organization does not depend primarily on their aromatic or charged character. Peptides flanked by Trp, Tyr, Lys, or (at low pH) His residues are effective in inducing mismatch-relieving cubic and inverted hexagonal phases, while analogues flanked by Phe, Arg, or (at neutral pH) His residues cannot induce an inverted hexagonal phase. The different responses to mismatch might reflect the different interfacial affinities of the residues that were investigated.     

Hydrophobicity modulation as a tool for the purification of histidyl peptides

A methodology is described for purification of histidyl peptides based on the changes in hydrophobicity induced by the specific and reversible modification of histidine residues by ethoxyformic anhydride. The mixture of modified peptides is subjected to HPLC; peptides containing modified histidines are detected at 242 nm, recovered, deethoxyformylated, and rechromatographed under the same conditions, then being detected at 220 nm. This procedure allows their isolation free from contaminants.     

Peptide hairpins with strand segments containing alpha- and beta-amino acid residues: cross-strand aromatic interactions of facing Phe residues

The incporation of beta-amino acid residues into the strand segments of designed beta-hairpin leads to the formation of polar sheets, since in the case of beta-peptide strands, all adjacent carbonyl groups point in one direction and the amide groups orient in the opposite direction. The conformational analysis of two designed peptide hairpins composed of alpha/beta-hybrid segments are described: Boc-Leu-betaPhe-Val-(D)-Pro-Gly-Leu-betaPhe-Val-OMe (1) and Boc-betaLeu-Phe-betaVal-D-Pro-Gly-betaLeu-Phe-betaVal-OMe (2). A 500-MHz 1H-NMR (nuclear magnetic resonance) analysis in methanol supports a significant population of hairpin conformations in both peptides. Diagnostic nuclear Overhauser effects (NOEs) are observed in both cases. X-ray diffraction studies on single crystals of peptide 1 reveal a beta-hairpin conformation in both the molecules, which constitute the crystallographic asymmetric unit. Three cross-strand hydrogen bonds and a nucleating type II' beta-turn at the D-Pro-Gly segment are observed in the two independent molecules. In peptide 1, the betaPhe residues at positions 2 and 7 occur at the nonhydrogen-bonding position, with the benzyl side chains pointing on opposite faces of the beta-sheet. The observed aromatic centroid-to-centroid distances are 8.92 A (molecule A) and 8.94 A (molecule B). In peptide 2, the aromatic rings must occupy facing positions in antiparallel strands, in the NMR-derived structure.Peptide 1 yields a normal "hairpin-like" CD spectrum in methanol with a minimum at 224 nm. The CD spectrum of peptide 2 reveals a negative band at 234 nm and a positive band at 221 nm, suggestive of an exciton split doublet. Modeling of the facing Phe side chains at the hydrogen-bonding position of a canonical beta-hairpin suggests that interring separation is approximately 4.78 A for the gauche+ gauche- (g+ g-) rotamer. A previously reported peptide beta-hairpin composed of only alpha-amino acids, Boc-Leu-Phe-Val-D-Pro-Gly-Leu-Phe-Val-OMe also exhibited an anomalous far-UV (ultraviolet) CD (circular dichroism) spectrum, which was interpreted in terms of interactions between facing aromatic chromophores, Phe 2 and Phe 7 (C. Zhao, P. L. Polavarapu, C. Das, and P. Balaram, Journal of the American Chemical Society, 2000, Vol 122, pp. 8228-8231).     

Reversed-phase high-pressure liquid chromatographic tryptic peptide mapping for the comparison and study of monoclonal antibodies.

We have examined and optimized several parameters to generate efficient, high-resolution, high-information tryptic peptide maps of monoclonal antibodies and their Fab fragments, without separating the H and L chains, using reversed-phase high-pressure liquid chromatography. Use of a high protease-to-substrate ratio with optimized digestion time and HPLC gradient conditions led to a reproducible mapping of the reduced, carboxymethylated Fab fragments of two antibodies. The technique was then used to screen Fab lots for batch-to-batch consistency, and for examining the effect of 10 mM cysteine on papain cleavage of whole antibody. The technique was modified by labeling cysteine with chromophoric analogues of iodoacetamide instead of radiolabeled iodoacetamide, resulting in a three-dimensional peptide map. With multiwavelength detection, this consisted of simultaneous observation of all chromophores at 214 nm, those with aromatic residues at 280 nm, and those with cysteine at 422 nm, without collecting and counting each peak to identify cysteine-containing peptides.     

Conformation of peptides constructed from achiral amino acid residues Aib and DeltaZPhe: computational study of the effect of L/D- Leu at terminal positions

Conformational studies of the peptides constructed from achiral amino acid residues Aib and Delta(Z)Phe (I) Ac-Aib-Delta(Z)Phe-NHMe (II), and Ac-(Aib-Delta(Z)Phe)(3)-NHMe; peptides III-VI having L-Leu or D-Leu at either the N- or the C-terminal position and of peptides VII-X having Leu residues in different enantiomeric combinations at both the N- and the C-terminal positions in peptide II have been studied to design the peptide with the required helical sense. Peptide II, as expected, adopts degenerate left- and right-handed helical structures. It has been shown that the peptides IV and VI having D-Leu at either the N or the C terminus can be realized in the right-handed helical structure with the phi,psi values of -20 degrees and -60 degrees for the Aib/Delta(Z)Phe residues. L-Leu and D- Leu at both the terminals in peptides VII and VIII, respectively, have hardly any effect as both the left- and the right-handed structures are found to be degenerate. Peptides III and IX can be realized in right- and left-handed helical structures, respectively, in solvents of low polarity whereas peptides V and X are predicted to be in the right-handed helical structures stabilized by carbonyl-carbonyl interactions without the formation of hydrogen bonds. The conformational states with the phi,psi values of 0 degrees and -85 degrees in peptide V are characterized by rise per residue of 2.03 A, rotation per residue of 117.5 degrees , and 3.06 residues per turn. In all peptides having Leu residue at the N terminus, the methyl moiety of the acetyl group is involved in the CH/pi interactions with the Cepsilon--Cdelta edge of the aromatic ring of Delta(Z)Phe (3) and the amino group NH of Delta(Z)Phe is involved in the NH/pi interactions with its own aromatic ring. The CH(3) groups of the Aib residues are also involved in CH/pi interactions with the i + 1th and i + 3th Delta(Z)Phe's aromatic side chains.     

Ranking the affinity of aromatic residues for carbon nanotubes by using designed surfactant peptides.

A series of surfactant peptides were created to evaluate the affinity of aromatic AAs for single-walled carbon nanotubes in the absence of complications from peptide folding or self-association. Each surfactant peptide has a lipidlike architecture, with two Lys residues at the C-terminus as a hydrophilic head, five Val residues to form a hydrophobic tail, and the testing AA at the N-terminus. Raman and CD spectroscopic studies reveal that the surfactant peptides have a large unordered structural component which is independent of peptide concentration, suggesting that the peptides undergo minimal association under experimental conditions, thus removing this interference from interpretation of the peptide/carbon nanotube interactions. A lack of peptide self-association is also indicated by sedimentation equilibrium ultracentrifugation results. Optical spectroscopy of the peptide/carbon nanotube dispersions indicate that among the three aromatic AAs, tryptophan has the highest affinity for carbon nanotubes (both bundled and individual states) when incorporated into a surfactant peptide, while the Tyr-containing peptide is more selective for individual carbon nanotubes. Phe has the lowest overall affinity for carbon nanotubes. Raman spectra of dispersions made with SPF, SPY and SPW display similar types of nanotubes dispersed, although differences in the relative nanotube populations are observed by optical spectroscopy.     

Conformational changes in monoamine oxidase A in response to ligand binding or reduction

The structure of monoamine oxidase B revealed three aromatic amino acid residues within contact distance of the flavin cofactor and a large number of aromatic residues in the substrate binding site. Circular dichroism (CD) spectroscopy can detect alterations in the environment of aromatic residues as a result of ligand binding or redox changes. CD spectra of MAO A indicate that a small inhibitor such d-amphetamine perturbs the aromatic residues very little, but binding of the larger pirlindole (2,3,3a,4,5,6-hexahydro-8-methyl-1H-pyrazino[3,2,1-j,k]carbazole hydrochloride) causes spectral changes consistent with the alteration of the environment of tyrosine and tryptophan residues in particular. Reduction of the flavin cofactor induces large enhancement of the CD signals in the aromatic region (260-310 nm). When covalent modification of the flavin by clorgyline accompanies reduction, the perturbation is even greater. In contrast to the static picture offered by crystallography, this study reveals changes in the aromatic cage on ligand binding and suggests that reduction of the cofactor substantially alters the environment of aromatic residues presumably near the flavin. In addition, the covalently modified reduced MAO A shows significant differences from the substrate-reduced enzyme.     

Characterization of the peptide-binding specificity of Mamu-B*17 and identification of Mamu-B*17-restricted epitopes derived from simian immunodeficiency virus proteins

The SIV-infected rhesus macaque is an excellent model to examine candidate AIDS virus vaccines. These vaccines should elicit strong CD8(+) responses. Previous definition of the peptide-binding motif and optimal peptides for Mamu-A*01 has created a demand for Mamu-A*01-positive animals. We have now studied a second MHC class I molecule, Mamu-B*17, that is present in 12% of captive-bred Indian rhesus macaques. The peptide-binding specificity of the Mamu-B*17 molecule was characterized using single substitution analogs of two Mamu-B*17-binding peptides and libraries of naturally occurring sequences of viral or bacterial origin. Mamu-B*17 uses position 2 and the C terminus of its peptide ligands as dominant anchor residues. The C terminus was found to have a very narrow specificity for the bulky aromatic residue W, with other aromatic residues (F and Y) being only occasionally tolerated. Position 2 is associated with a broad chemical specificity, readily accommodating basic (H and R), bulky hydrophobic (F and M), and small aliphatic (A) residues. Using this motif, we identified 50 peptides derived from SIV(mac)239 that bound Mamu-B*17 with an affinity of 500 nM or better. ELISPOT and intracellular cytokine-staining assays showed that 16 of these peptides were antigenic. We have, therefore, doubled the number of MHC class I molecules for which SIV-derived binding peptides have been characterized. This allows for the quantitation of immune responses through tetramers and analysis of CD8(+) function by intracellular cytokine-staining assays and ELISPOT. Furthermore, it is an important step toward the design of a multiepitope vaccine for SIV and HIV.     

The periplasmic bacterial molecular chaperone SurA adapts its structure to bind peptides in different conformations to assert a sequence preference for aromatic residues

The periplasmic molecular chaperone protein SurA facilitates correct folding and maturation of outer membrane proteins in Gram-negative bacteria. It preferentially binds peptides that have a high fraction of aromatic amino acids. Phage display selections, isothermal titration calorimetry and crystallographic structure determination have been used to elucidate the basis of the binding specificity. The peptide recognition is imparted by the first peptidyl-prolyl isomerase (PPIase) domain of SurA. Crystal structures of complexes between peptides of sequence WEYIPNV and NFTLKFWDIFRK with the first PPIase domain of the Escherichia coli SurA protein at 1.3 A resolution, and of a complex between the dodecapeptide and a SurA fragment lacking the second PPIase domain at 3.4 A resolution, have been solved. SurA binds as a monomer to the heptapeptide in an extended conformation. It binds as a dimer to the dodecapeptide in an alpha-helical conformation, predicated on a substantial structural rearrangement of the SurA protein. In both cases, side-chains of aromatic residues of the peptides contribute a large fraction of the binding interactions. SurA therefore asserts a recognition preference for aromatic amino acids in a variety of sequence configurations by adopting alternative tertiary and quaternary structures to bind peptides in different conformations.     

Identification of peptides containing tryptophan, tyrosine, and phenylalanine using photodiode-array spectrophotometry

The characteristic absorption spectra of aromatic amino acids between 240 and 310 nm were used to identify tryptophan, tyrosine, and phenylalanine-containing peptides. In acidic solution, the absorption spectra of these amino acids exhibit minima or maxima at 255, 270, and 286 nm. Based on these characteristics, the content of the aromatic amino acid in peptide can be estimated. For this study, 2 nmol of tryptic peptides from human apolipoprotein A-1 was separated by high-performance liquid chromatography using a reverse-phase column. The peptide fragments were monitored by a photodiode-array spectrophotometer. This new approach offers a rapid, simple, sensitive, and direct identification of peptides containing aromatic amino acids. Those containing Trp, which may be of interest for DNA sequencing and important in sequence analysis of proteins, can be selectively purified using this technique.     

Enhanced antitumour activity of 15-residue bovine lactoferricin derivatives containing bulky aromatic amino acids and lipophilic N-terminal modifications.

In a structure-antibacterial activity relationship study of a peptide fragment of bovine lactoferricin consisting of FKCRRWQWRMKKLGA (LFB 17-31), it was revealed that the two Trp residues were important for antibacterial activity. It has further been demonstrated that the size, shape and the aromatic character of the side chains were even more important than the Trp itself. In this study the antitumour effect of a series of LFB 17-31 derivatives are reported, in which the two Trp residues in position 6 and 8 were replaced with the larger non-coded aromatic amino acids Tbt, Tpc, Bip and Dip. The counterproductive Cys in position 3 was also substituted with these larger aromatic residues. In addition, the effect of introducing lipophilic groups of different size and shape in the N-terminal of the LFB 17-31 sequence was addressed. The resulting peptide derivatives were tested for activity against three human tumour cell lines and against normal human umbilical vein endothelial cells and fibroblasts. High antitumour activity by several of the peptides demonstrated that Trp successfully could be substituted by the bulky aromatic residues, and peptides containing the large and rigid Tbt residue in position 6 and/or 8 in LFB 17-31 were the most active candidates. The antitumour effect was even more increased by the Tbt-modified peptides when the three counterproductive amino acids Cys3, Gln7 and Gly14 were replaced by Ala. Enhanced antitumour activity was also obtained by modifying the N-terminal of LFB 17-31 with either long-chained fatty acids or bulky moieties. Thus, our results revealed that the size and shape of the lipophilic groups and their position in the peptide sequence were important for antitumour activity.