The effect of sequence variations and structure on the cytolytic activity of melittin peptides

50

Biochimica et Biophysica Acta, 1157 (19ff3)50-54 © 1993 Elsevier Science Publishers B.V. All rights reserved 0304-4165/93/$06.00

BBAGEN 23784

The effect of sequence variations and structure on the cytolytic activity of melittin peptides Jerome A. Werkmeister, Alan Kirkpatrick, Julie A. McKenzie and Donald E. Rivett CSIRO, Division of Biomolecular Engineering, Parkville, Victoria (Australia)

(Received 3 August 1992)

Key words: Melittin; Peptide analog; Cytolyticactivity;Monoclonal antibody The importance of various amino acid residues in melittin for cytolytic function against mammalian cells was assessed by use of a monoclonal antibody to the C-terminal region, synthesis of peptide analogues and chemical modification of specific residues. A monoclonal anti-melittin antibody directed to the basic C-terminal region inhibited cytolytic activity. Consistent with this, deletion of one of the two Lys Arg sequences at the C terminal end of the peptide reduced cytolysis 8-fold, although significant activity was still present. A similar reduction in activity was also found with a synthetic analogue which had the reverse sequence to melittin. In contrast, when the last 6 residues of the C-terminal region were transferred to the N-terminus, a peptide with markedly reduced activity was obtained. Chemical modification of lysine and arginine residues of melittin indicated that lysine was only minimally important for functional activity compared with arginine which was essential. In particular, our results demonstrate that substitution of serine for lysine 7 has no significant effect on the activity of the peptide and suggest that this residue is important only in maintaining the amphipathic helix of the peptide.

Introduction Melittin, a cytolytic peptide found in the venom of the honey bee, consists of a single peptide chain of 26 amino acids. Residues 1-20 form an amphipathic ahelix of hydrophobic amino acids whilst residues 21-26 are charged (basic) and constitute the putative binding domain of the molecule. Melittin, in common with several other insect toxins, produces cell lysis either by forming ion channels which span m e m b r a n e bilayers or by forming other structures which increase m e m b r a n e permeability [1]. More recently it has been proposed that binding to m e m b r a n e proteins is involved [2]. The length of the amphipathic helical segment is important since it has been shown that analogues with shortened helical segments have poor lytic properties [3,4]. Early studies demonstrated that N-acetylation of all lysine residues and the N-terminal group of melittin had little effect on the lytic activity of the protein [5]. Current opinion on the relationship between the pep-

Correspondence to: Jerome A. Werkrneister, CSIRO, Division of Biomolecular Engineering, 343 Royal Parade, Parkville, Victoria 3052, Australia. Abbreviations: MAb, monoclonal antibody; ELISA, enzyme-linked immunosorbant assay; Pare, phenylacetamidomethyl; MTS, mesitylene-2-sulphonyl; MBS, rn-maleimidobenzoyl-N-hydroxysuccinimide ester; Acetyl OSu, acetyl-N-hydroxysuccinimideester.

tide sequence and the functional integrity of the molecule are conflicting. Kaiser and Kezdy [6] have shown that the presence of lysine residues were not critical for lysis of erythrocytes and have suggested that the ability to form a defined secondary structure is the key to the biological activity of the peptides. In contrast, the results of Gevod and Birdi [4] implied the involvement of Lys-7 as having a specific role on the hemolytic activity of melittin. This conclusion has been supported by recent work based on the synthesis of sequential peptides with single amino acid deletions and substitutions [7,8]. To understand the structural importance of lysine residues, particularly Lys-7, as well as other residues such as arginine, we synthesised a range of peptides and assessed these for cytolytic activity against mammalian cells, the human H M Y - 2 lymphoma cell line in vitro. In addition, specific amino acid modifications along with antibody inhibition studies were performed. Our results demonstrate that arginine but not lysine residues are essential for biological activity of melittin and more specifically that Lys-7 is not critical for cytolytic activity.

Experimental Materials

All solvents used were of reagent grade. Melittin, free of phospholipase, was obtained from Fluka Bio-

51 chemika, Switzerland. A Submaxillaris arginine specific protease (E.C. 3.4.21.40) [9] was obtained from Pierce (Rockford, IL, USA). Chromium-51 (350-600 mCi/mg) was supplied by Amersham Int. (Amersham, UK).

(b) The arginine residues of melittin were blocked by treating the peptide, in 0.3 M NaOH solution, with an excess of phenylglyoxyl. In both modifications, excess reagent was removed by dialysis prior to assay.

Peptide synthesis

Monoclonal antibody production and characterisation

The following peptides were prepared and are compared with melittin:

12-wk-old female B A L B / c mice were immunized as previously described [11] by two intraperitoneal injections of 10/zg melittin emulsified in Complete Freund's Adjuvant and Incomplete Freund's Adjuvant, respectively. Three weeks after the second immunisation, mice were boosted with 20 ~g melittin intravenously and then hybridomas were prepared using NS-1 cells [12,11]. Antibodies with specificity for melittin were identified by an enzyme-linked immunosorbant assay (ELISA) [11], using rabbit anti-mouse Immunoglobulin coupled to alkaline phosphatase (EC 3.1.3.1) (Sigma Chemical Co., St. Louis, MO). Monoclonal antibodies were also assessed for their ability to block cytolytic activity of melittin against mammalian cells. One antibody was selected (5C3/Dll-MEL) which bound strongly to melittin by ELISA and which inhibited activity in a 51Cr-release lysis assay.

(Melittin):

G I G A V L K V L T T L I S W I K R K R Q Q

(Peptide-l): Q Q T T (Peptide-2): G I L I (Peptide-3): K R T T (Peptide-4): G Z L I (Peptide-5): G I L I

R K R K I W S I L L V K L V A G I G (3 A V L S V L T T S WI K R K R Q a G I G A V G L P A L I S W I G A V L K V L T T S WI K R K R G A V t. K V L T T S W I K R

G L P A A P L G G L P A L K V L G L P A G L P A

Peptides 1, 2 and 3 were synthesised onto a phenylacetamidomethyl (Pam) resin, with the first C-terminal Boc-protected amino acid attached by the Merrifield solid phase technique [10], using an Applied Biosysterns 430A Peptide Synthesiser. The protecting group used for the amino terminus was Boc. Threonine and serine side chains were protected by benzyl groups. The lysine side chain was protected by 2-chlorobenzyloxycarbonyl groups and arginine by mesitylene-2sulphonyl (MTS) groups. The complete peptides were cleaved from the resin with anhydrous hydrogen fluoride to yield the peptides as the free carboxylic acids. Purification was performed by standard reverse-phase HPLC. Peptides 4 and 5 were obtained by digesting melittin with the Submaxillaris protease in 0.1 M aqueous NaHCO 3 and incubated for 3 h at 37°. Semi-preparative reverse-phase HPLC was used to isolate peptides 4 and 5. All peptides were confirmed by amino acid analysis which was performed on a Waters analyser. Reversephase HPLC (Vydac 218TP column) was performed using a Perkin Elmer series 4 solvent delivery system and monitored using a Hewlett-Packard 1040A photodiode array detector.

Chemical modification of melittin The following derivatives of melittin were also prepared to examine the importance of lysine and arginine residues on the cytolytic nature of the peptide: (a) The N-terminal amino group and the E-amino groups of the lysine residues were blocked by treating melittin, in a 1 : 3 solution of dioxan and 0.3 M NaOH, with an excess of m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS) or acetyl-N-hydroxysuccinimide ester (Acetyl OSu).

Cytolytic assay A standard 51Cr-release lysis assay was carried out against the human lymphoblastoid cell line, HMy-2 [13] to determine the cytolytic activity of the peptides relative to melittin [12]. The cells were labelled with 51Cr by the addition of 100/zCi 51Cr to a 1.0 ml suspension of cells (1.0 × 106/1111) in Dulbecco's Modified Essential (DME) culture medium containing 10% foetal calf serum, 5 × 10-5M 2-mercaptoethanol, penicillin (100 U / m l ) and streptomycin (100/zg/ml). The cells were incubated at 370C for 1 h, washed three times in culture medium, then resuspended to 5-104 cells/ml and 100/zl of cells were added to each well of a 96 well V-bottomed Linbro microtitre plate (Flow Laboratories, Virginia, USA). The peptides were dissolved in medium and a 50/zl aliquot of serial 2-fold dilutions was added to the cells using triplicate wells per sample. The ceils were incubated at 370C in 5% CO 2 for 3 hrs. After incubation the cells were centrifuged and the amount of 51Cr released was measured on a gamma counter by removing and counting 100/~I of the supernatant from each well. The mean percent specific lysis was calculated from: (test-background)/(maximumbackground). The background 51Cr release was the mean supernatant count for target cells cultured in the absence of peptides and the maximum release was the mean count when target cells had been resuspended prior to supernatant harvesting. The spontaneous release from labelled target cells (100 × background/ maximum) was always around 5-7% and the SEM of each triplicate was always less than 1-2% of the response.

52 I00-

Results and Discussion

Monoclonal antibodies were produced against melittin and, of these, 5C3/Dll-MEL, was found to cause a significant decrease in cytolytic activity against 51Crlabelled HMY-2 cells (Fig. 1). At limiting cytolytic concentrations of melittin (2.5 and 5/zg/ml), the antibody produced 90-95% inhibition in activity. The antibody 5 C 3 / D l l - M E L was shown to be directed against the basic C-terminal region of melittin by ELISA. The MAb reacted strongly with melittin (mean absorbance, 2.47) and with peptide-4 (mean absorbance, 2.20), but less with the truncated 22 amino acid peptides-2 and -5 (mean absorbances, 0.61 and 0.69 respectively). Consistent with this reactivity, 5 C 3 / D l l - M E L had little effect on the cytolytic activity of peptide-2 (Fig. 1). The cytolytic activities of peptides-1 to -5 as assessed by 51Cr release assay are shown in Fig. 2 compared with melittin. Similar results were found with an erythrocyte haemolytic assay. Removal of the last two residues, Gln.Gln (peptide-4), was associated with only a minimal 2-fold reduction in cytolytic activity. On the other hand, the activities of both the 22 residue peptides-2 and -5 and of peptide-1 ('reverse melittin'), although still significantly cytolytic, were reduced around 8-fold. The activity of peptide-3 was most severely impaired and was only cytolytic (28%) at the highest concentration tested (80/zg/ml). 100"

80

.t/l

•

o~

s

;0

o

60

t~0

20

12

2.s

20

Peptide concentration (pg/ml) Fig. 1. Inhibition of melittin activity with a monoclonal anti-melittin antibody directed to the C-terminus. Monoclonal antibody 5C3/DI1-MEL, at a concentration of 10 ~g/ml, was preincubated for 30 minutes with various concentrations of melittin (0) or peptide-2 (zx) and cytolytic assays were performed for 3 hrs against SiCr-labelled HMy-2 ceils as described in the materials and methods. Cytolytic dose-response curves for melittin (e) and peptide-2 ( • ) in the presence of a control irrelevant antibody are shown. All values represent means of triplicate samples and the standard error of each mean was within 1%.

80

/ t~

60

~0

0'6

1'2

2'5

5

10

20

z,O

80

Peptide concentration (pg/m[) Fig. 2. Comparison of cytolytic activity of melittin and various synthetic and enzyme-derived peptides. Cytolytic activity of melittin (e), peptide-1 (©), peptide-2 (A), peptide-3 (A), peptide-4 (m) and peptide-5 (D) at various concentrations in a 3 hr 5]Cr-release assay against HMy-2 cells. Peptides-1 to -5 are described in Materials and Methods. All values represent means of triplicate samples and the standard error of each mean was within 1%.

Theories on the proposed mode of action of melittin can be grouped under two major hypotheses [1]. In one model, melittin forms aqueous channels by spanning the membrane and forming channel oligomers, exposing its hydrophobic side to the lipid and its hydrophilic side to the aqueous pore. In the second model, melittin produces lysis by disrupting the phospholipid structure of the membrane. Neither model should be affected by the direction of polarity of the molecule and it was therefore surprising to find partial loss of activity with peptide-1 and almost complete loss of activity with peptide-3 (Fig. 2). It could be argued that the results of Blondelle and Houghten [7,8] and Habermann and Kowallek [5] show that the position of the tryptophan in relation to the basic sequence is crucial to activity. The loss of activity of peptide-3 could thus be explained by the tryptophan residue being at the wrong end of the helical sequence relative to the basic sequence. This is in agreement with fluorescence studies of tryptophan 19 which show that the tryptophan penetrates to only a minimal degree into the membrane [141. It is difficult to explain why the reverse sequence of melittin, peptide-1, should have a reduction in activity compared to melittin as the relative positions of the residues are the same. In this context it is interesting to note that cecropins have the same general design as peptide-1. Cecropins are a family of broad spectrum antibacterial peptides derived predominantly from insects. These peptides have a structure that resembles

53 100"

.~_ N ~o

- iif 20

06

1-2 2"5 5 10 20 t~0 80 Pepfide concentration (pg/ml)

Fig. 3. The effect of chemical modificationof lysine and arginine residues on the cytolyticactivity of melittin. Lysine residues were modified by acetylation (o) or by succinylation(D) and arginine residues were modified with phenylglyoxyl(A) and assessed for cytolytic activity as above along with untreated melittin (e). All solvents used in the chemicalmodificationshad no deleteriouseffect on lysis of the HMy-2 cells. All values represent means of triplicate samples and the standard error of each mean was within 1%.

melittin but with a reversed polarity, the N-terminal sequence is strongly basic while the C-terminus is hydrophobic [15]. Boman et al. [16] demonstrated that a hybrid peptide containing the first 13 residues of cecropin and the first 13 residues of melittin was 100-fold more active than cecropin against S t a p h y l o c o c c u s aureus. We have recently reported that peptide-1 does have antibacterial activity [17] and we would predict that peptide-3 would also possess bactericidal activity. Blondelle et al. [8] have indicated that the C-terminal segment must only be sufficiently cationic to bind to the phospholipid head groups of the membrane if lysis is to ultimately occur, a finding which is also supported by the present study. However, there is considerable confusion as to the importance of the lysine and arginine residues within this region. Removal of the sequence lys.arg.gln.gln, from the Cterminal region leads to a significant reduction of activity although the truncated peptide-5 is still capable of initiating cell lysis. This finding, along with the antibody inhibition studies (Fig. I), provides direct evidence to support the hypothesis of Blondelle et al. [8] that only a minimal basic C-terminus is required. When the lysine and N-terminal glycine residues were substituted with rn-maleimidobenzoyl or acetyl groups, only a small decrease in cytolytic activity was observed (Fig. 3), which indicates that the free amino groups on these residues are not essential for lytic activity. These results are in agreement with other

findings showing no effects of N-acetylation of melittin on hemolysis of erythrocytes [5], but are in conflict with data obtained from succinylation where a loss of activity was reported [5,18]. The loss of activity on succinylation has been claimed to be due to the formation of intramolecular salt bridges [19] or to the absence of a positive charged residue at position 7 in the helix [7,8]. Our results are in agreement with the interpretation drawn by Hider et al. [19] and indicate that a positive charge at position 7 has no particular effect on the activity of the peptide. Chemical modification strongly indicates that unmodified lysine residues are not essential for activity, both acetylation and maleimidobenzoylation resulting in only a 2-fold reduction in lysis (Fig. 3). In these experiments and others using blocking with t-butyloxycarbonyl groups (data not shown), derivitisation of the melittin also led to a decrease in solubility which may account for the partial decrease in activity. However Dufton et al. [20] have also shown a similar decrease in melittin's aggregating potency [21] upon acetylation of its amino groups. The significance of the lysine residue at position 7 has also been investigated directly by comparison of the lytic activity of two 22 residue peptides, peptide-2 and peptide-5 which differ only in that the lysine at position 7 in peptide-2 is replaced by a serine residue. The results show that there is no significant difference in their ability to lyse the mammalian cells (Fig. 2). In the experiments of Blondelle and Houghten [7,8], a marked reduction of activity was found as a result either of deletion of Lys-7 or its replacement with a leucine residue; both of these modifications will change the relative positions of hydrophilic and hydrophobic residues. In the present study, we have maintained the relative hydrophilic nature of position 7 by the serine substitution and so the relative orientation of the other residues throughout the helix should not be disturbed maintaining its amphiphilic nature. Our results are also supported by the findings of Kaiser and Kezdy [6] in which modified cytolytic peptides with unaltered amphipathic helical structures were active against erythrocytes. These authors also retained tryptophan at position 19 and their results give further support to the importance of a tryptophan residue adjacent to the basic sequence. Habermann and Kowallek [5] converted the lysine residues of melittin to homoarginine residues and reported an increase in lytic activity. Our results demonstrate the importance of arginine residues in that blocking these residues with phenylglyoxyl, resulted in complete inhibition of cytolytic activity (Fig. 3). These results suggest that a basic sequence at the C-terminus is not in itself sufficient for binding to the cell surface, but that at least one arginine residue is required for lytic activity against mammalian ceils. It is possible that the presence of arginine in melittin and its absence in

54 the basic 'cell recognition' sequence of cecropins is the reason why cecropins can lyse bacterial cells but not mammalian cells. The concept that a portion of the melittin molecule is involved in cell surface recognition is supported by the work of Portlock et al. [2]. References 1 Bernheimer, A.W. and Rudy, B. (1986) Biochim. Biophys. Acta 864, 123-141. 2 Portlock, S.H., Clague, M.J. and Cherry, R.J. (1990) Biochim. Biophys. Acta 1030, 1-10. 3 Dawson, C.R., Drake, A.F., Helliwell, J. and Hider, R.C. (1978) Biochim. Biophys. Acta 510, 75-86. 4 Gevod, V.S. and Birdi, K.S. (1984) Biophys. J. 45, 1079-1083. 5 Habermann, E. and Kowallek, H. (1970) Hoppe-Seyler's Z. Physiol. Chem. 351, 884-890. 6 Kaiser, E.T. and Kezdy, F.J. (1983) Proc. Natl. Acad. Sci. USA 80, 1137-1143. 7 Blondelle, S.E. and Houghten, R.A. (1991) Biochemistry 30, 4671-4678. 8 Blondelle, S.E. and Houghten, R.A. (1991) Peptide Res. 4, 12-18. 9 Schenkein, I., Levy, M., Franklin, E.C. and Frangione, B. (1977) Arch. Biochem. Biophys. 182, 64-70.

10 Barany, G. and Merrifield, R.B. (1980) in The Peptides, Vol. 2, Gross, E. and Meinhofer, J. (eds.), pp. 1-284, Academic Press, New York. 11 Werkmeister, J.A. and Ramshaw, J.A.M. (1991) Biochem. J. 274, 895-898. 12 Werkmeister, J.A., Triglia, T., Andrews, P. and Burns, G.F. (1985) J. Immunol. 135, 689-695. 13 Boux, H.A., Raison, R.L., Walker, K.Z., Hayden, G.E. and Basten, A. (1983) J. Exp. Med. 158, 1769-1774. 14 DeBony, J., Dufourcq, J. and Clin, B. (1979) Bioehim. Biophys. Acta 552, 531-534. 15 Boman, H.G. and Hultmark, D. (1987) Ann. Rev. Microbiol. 41, 103-126. 16 Boman, H.G., Wade, D., Boman, I.A., Wahlin, B. and Merrifield, R.B. (1989) FEBS Lett. 259, 103-106. 17 Rivett, D.E., Kirkpatrick, A., Macreadie, I.G., McKenzie, J.A., Raison, R.L., Weston, K.M. and Ward, A.C. (1990) J. Protein Chem. 9, 350-351. 18 Manjunatha Kini, R. and Evans, H.J. (1989) Int. J. Pept. Protein Res, 34, 277-286. 19 Hider, R,C., Khander, F. and Tatham, A.S. (1983) Biochim. Biophys, Aeta 728, 206-214. 20 Dufton, M,J., Hider, R.C. and Cherry, R.J. (1984) Eur. Biophys. J. 11, 17-24. 21 Clague, M.J. and Cherry, R.J. (1989) Biochim. Biophys. Acta 980, 93-99.