Lipidation and Glycosylation of a T Cell Antigen Receptor (TCR) Transmembrane Hydrophobic Peptide Dramatically Enhances in Vitro and in Vivo Function

Biochimica et Biophysica Acta 1763 (2006) 879–888 www.elsevier.com/locate/bbamcr

Lipidation and glycosylation of a T cell antigen receptor (TCR) transmembrane hydrophobic peptide dramatically enhances in vitro and in vivo function

Michael A. Amon a, Marina Ali a, Vera Bender a, Yiu-Ngok Chan b, Istvan Toth c, ⁎ Nicholas Manolios a,

a Department of Rheumatology, Westmead Hospital, Westmead. NSW. 2145, Australia b School of Pharmacy, Griffith University, Gold Coast, QLD, Australia c School of Molecular and Microbial Sciences, School of Pharmacy, The University of Queensland, Brisbane QLD. 4072, Australia Received 14 March 2006; received in revised form 19 April 2006; accepted 26 April 2006 Available online 10 May 2006

Abstract

A T cell antigen receptor (TCR) transmembrane sequence derived peptide (CP) has been shown to inhibit T cell activation both in vitro and in vivo at the membrane level of the receptor signal transduction. To examine the effect of sugar or lipid conjugations on CP function, we linked CP to 1-aminoglucosesuccinate (GS), N-myristate (MYR), mono-di-tripalmitate (LP1, LP2, or LP3), and a lipoamino acid (LA) and examined the effects of these compounds on T cell activation in vitro and by using a rat model of adjuvant-induced arthritis, in vivo. In vitro, antigen presentation results demonstrated that lipid conjugation enhanced CP's ability to lower IL-2 production from 56.99%±15.69 S.D. observed with CP, to 12.08%±3.34 S.D. observed with LA. The sugar conjugate GS resulted in only a mild loss of in vitro activity compared to CP (82.95%± 14.96 S.D.). In vivo, lipid conjugation retarded the progression of adjuvant-induced arthritis by approximately 50%, whereas the sugar conjugated CP, GS, almost completely inhibited the progression of arthritis. This study demonstrates that hydrophobic peptide activity is markedly enhanced in vitro and in vivo by conjugation to lipids or sugars. This may have practical applications in drug delivery and bioavailability of hydrophobic peptides. © 2006 Elsevier B.V. All rights reserved.

Keywords: Peptides; Lipopeptides; Peptide conjugation; Transmembranes; T cell antigen receptor

1. Introduction Administration being the eagerly anticipated HIV fusion inhibitor peptide, enfuvirtide (Fuzeon®, formerly T-20). This peptide has The diagnosis and treatment of clinical diseases with peptides been designed to bind to a structural intermediate of gp41's fusion are quickly becoming a reality with many peptides already being process, thereby inhibiting HIV-1 fusion and subsequent infection applied in the clinical setting. Two diagnostic kits presently of CD4+ T cells [3]. A 25-amino-acid peptide isolated from the available employing peptides as their major effector are Cone Snail (Conus Magus), termed ω-conotoxin (ziconotide), has technetium Tc99m-apcitide for acute deep vein thrombosis, recently entered the market as a non-addictive treatment for pain which binds with high specificity and affinity to activated relief in cancer and HIV patients. ω-Conotoxin acts by selectively platelets [1], and indium In-111 pentetreotide (OctreoScan®) [2] inhibiting calcium channels [4] and has been shown to be 1000 which is capable of scintographic localization of primary and timesmoreactivethanmorphineinanimalmodels[5].Another metastatic neuroendocrine tumors bearing somatostatin receptors. peptide on the market is glatiramer acetate (Copaxone, copoly- To date, relatively few peptide drugs are available commercially, mer-1), which has been approved in the treatment for multiple with one of the most recently approved by the US Food and Drug sclerosis. Glatiramer acetate is a peptide mimic of myelin basic protein which consists of fixed ratios of alanine, lysine, glutamic ⁎ Corresponding author. Tel.: +61 2 9845 8099; fax: +61 2 9845 8317. acid, and tyrosine and is likely to inhibit multiple sclerosis by E-mail address: [email protected] (N. Manolios). switching the immune response from the pro-inflammatory Th1

0167-4889/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamcr.2006.04.013 880 M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888 to the anti-inflammatory Th2 response [6]. Other commercially resulting in T cell activation and an immune response [25].An available peptides, such as the spleen peptide preparation Poly- important aspect of TCR structure and assembly is the presence erga [7], Prograf (tacrolimus or FK506) [8], or Thymosin-α [9] of at least one charged amino acid within each subunits trans- are also aimed at various forms of immunoregulation. Many other membrane domain [26]. The invariant CD3 chains each have peptides are in various phases of clinical trial and are expected to one negative charge within their transmembrane domains, be on the market in the near future against various cancers [10] whereas the TCR subunits have positive charges, with the such as the 11-amino-acid peptide P-108 against breast cancer TCRα possessing two, and the TCRβ with one. These trans- [11], with many more being developed [12]. membrane charges have been shown to be an essential requ- A number of difficulties must be overcome prior to the wid- irement for proper TCR assembly. Call et al. have recently espread clinical usage of peptides. These include peptide solub- proposed the TCR is composed of four separate dimers, the ility, mode of administration, and the subsequent delivery of the TCRαβ antigen recognition subunit, and the CD3δε,CD3γε, peptide in its active form (reviewed by Ali and Manolios [13]). In and CD3ζζ signalling molecules [26]. In this model, the basic most cases, peptides translocate across the cell membrane either transmembrane lysine residue of the TCRα interacts with the by a receptor-mediated energy-dependent process or by passive two acidic aspartate residues of the CD3δε,andtheTCRα's diffusion [14]. In the latter case, this is generally facilitated by basic arginine residue interacts with the 2 acidic aspartate increasing the lipophilicity of peptides [15]. Lipid conjugation residues of the CD3ζζ dimer. The TCRβ's single basic residue can be performed with various lipids employing many different (lysine) was shown to interact with the acidic glutamate and strategies. For example, Schelhaas et al. [16] have utilized enzy- aspartate residues of the CD3γε dimer [26]. This results in a me-labile protecting groups and a metal sensitive allyl ester to TCR structure consisting of three separate trimers character- successfully lipidate peptides corresponding to the C terminus of ized by the interactions between one basic and 2 acidic residues the N-Ras protein. A discussion on the development of various to produce TCRα-CD3δε,TCRα-CD3ζζ,andTCRβ-CD3γε protecting group techniques used in peptide lipidation is furthered trimers [26]. inareviewbyPetersetal.[17].MoraandTam[18] describe a CP was synthesized as a direct copy of the nine amino acid conjugation technique employing branched lysines and a tripa- sequence within the TCR-α transmembrane region (G–L–R–I– lmitoyl S-glycerine cysteine to create a lipidated multiple antigen L–L–L–K–V), containing the two basic residues, arginine and peptide. A reversible lipidization technique, as described by Wang lysine. This peptide was subsequently shown to inhibit T-cell- et al., has resulted in an anti-diuretic peptide's activity being in- mediated responses both in vitro and in vivo by inhibiting IL-2 creased 250-fold in a rat model [19]. Tris-linkers have also been production in T cells following antigen recognition and by con- used to conjugate 1, 2, or 3 lipid groups to peptides while avoiding ferring therapeutic protection in rats with adjuvant-induced art- structural isomerization [20]. Lipoamino acids, as used by Toth et hritis [24]. Based on current evidence, CP acts at the cell surface al. [21], are also being synthesized to increase the membrane membrane level to disrupt TCR signal transduction between the permeability of various peptides. antigen recognizing TCR-α/β and the signal transduction chains The therapeutic success of peptides combined with advances in of the CD3 subunits [27,28] by a non-receptor-mediated process. biotechnology is hailing a paradigm shift in peptide applicaion, Surface plasmon resonance (SPR) has been used to determine drug design, and discovery. Whereas drugs were once derived from CP's affinity for membranes [29] and shown to bind irreversibly the purification of natural compounds, it is now possible to design to model membrane bilayers composed of either dimyristoyl- and synthesize new compounds based on structure–function phosphatidylcholine or dimyristoyl-phosphatidylglycerol. relationships. Otvos et al. [22] have designed a potent anti-micro- CP is poorly soluble when tested in vitro requiring solu- bial peptide statistically based on native proline-rich sequences bilisation in DMSO or acetic acid. Heating, shaking, soni- from known antibacterial peptides followed by an optimisation cation, and repeated lyophilization improve solubility, but step involving the replacement of non-productive residues. Ano- solubility remains poor. To overcome some of these difficul- ther example is the neuraminic acid derivative, zanamivir (Re- ties, Gollner et al. [30] emulsified CP in a cream and applied it lenza). Using 3D crystallography, it was possible to design a new to skin with effective results. In this paper, Gollner et al. drug able to sterically inhibit the active site of the influenza showed that topical application of CP was able to locally inhibit neuraminidase binding site leading to potent anti-influenza activity the elicitation of contact sensitivity in mice when applied as [23]. Hence, drugs may be able to be designed based on their ability either a cream or when DNA encoding CP was injected into the to disrupt the tertiary structure of proteins, either by the competi- skin. Similarly, in an ingenious means of drug delivery Mahnke tive/steric inhibition of active sites or by the inhibition of receptor et al. have demonstrated that transduction of dendritic cells by activation as is the case with the T cell receptor mimic peptide retroviral agents containing CP constructs were able to secrete termed core peptide (CP) [24]. CP in a localized and antigen-specific manner capable of in- The TCR is composed of at least seven transmembrane hibiting experimental allergic encephalomyelitis in a rodent proteins that are responsible for recognizing antigens presented model [31]. Although these papers showed the clinical rele- in the context of MHC-encoded proteins on the surface of vance and potential of CP, a more practical and readily avail- antigen-presenting cells. The TCR α/β chains are responsible able delivery method still needs to be developed. for antigen recognition, while the invariant CD3 chains (γ, δ, ε, This study examines the influence of lipid and/or sugar con- and ζ) are responsible for transmitting antigen recognition jugations on hydrophobic peptide (CP) function and correlates signals from the TCRα/β subunit through to the cytoplasm, this with the analogues behaviour in model membranes using M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888 881

SPR. The effectiveness of these analogues in vivo using the rat mass spectroscopy (MS). The synthesis of the LA and GS was accomplished auto- adjuvant arthritis model was examined. matically by a stepwise solid-phase procedure on a MBHA Novabiochem resin (substitution 0.48 mmol/g resin) and their primary structure verified by MS and NMR. The lipoamino acid synthesis and the 1-amino-glucosesuccinate synthesis were coup- 2. Materials and methods led to the resin peptides manually. The synthesis of these compounds has previously been described [35,36]. Briefly, peptide construction was achieved using a 4-fold 2.1. Cells used excess of preformed symmetrical anhydride of N-Boc amino acids in dichloromethane/N-methylpyrrolidone. The protecting groups were Boc groups for the α-amino- The following cell lines were used. 2B4.11, a murine T cell hybridoma that termini. For all residues, the first coupling was done with the preformed symmetrical expresses a complete antigen receptor on the cell surface and produces IL-2 anhydride in dichloromethane, a second coupling in dichloromethane/N-methylpyr- following antigen recognition (cytochrome c). CTLL cells, an IL-2-dependant T rolidone. After the second coupling, deprotection of the N-termini was performed in cell line, for conventional biological IL-2 assays. The B cell hybridoma line 65% TFA in dichloromethane (20 ml for 1 min, then another 20 ml for 10 min). The LK35.2 (LK, I-Ek bearing), which acts as the antigen-presenting cell. The deprotected resin peptide was neutralised with 10% di-isopropylethylamine in hybridomas were grown in T cell medium (RPMI 1640 medium containing 10% dichloromethane. The peptide was removed from the resin support with the high HF fetal calf serum (FCS), glutamine (2 mM), beta-mercaptoethanol (0.002%), method (1.5 ml cresol, 1.5 ml thiocresol, 20 ml HF) to yield the crude peptide, which penicillin/streptomycin (50 U/ml). was precipitated with ethyl acetate and redissolved in 6 M guanidine HCl–0.1 M Tris solution (20 ml). 2.2. Antibodies Peptide purification was done on a Vydac C18 5 RAC column. HPLC grade acetonitrile (Aldrich) and water were filtered through a membrane filter and 145-2C11, a hamster anti-mouse-CD3ε antibody [32], was purified from degassed with helium prior to use. All HPLC solvents contained 0.1% trifluo- supernatant in laboratory. H57-597, an antibody directed against the TCR Cβ roacetic acid. Analytical separation was achieved with a solvent gradient begin- FG loop of the TCR-β chain [33,34] (Becton Dickinson Pharmingen Catalogue ning with 0% acetonitrile, increasing constantly to 60% acetonitrile at 30 min, #553167 for plate-bound stimulation and Santa Cruz Catalogue #sc-18904 l for staying at this concentration for 20 min and decreasing steadily to 0% acetonitrile soluble antibody stimulation). for 10 min at a constant flow of 0.7 ml/min. For preparative separation, a TSK- GEL preparative C18 column was used. Separation was achieved with a solvent 2.3. Peptides/lipopeptide synthesis gradient beginning with 0% acetonitrile, increasing constantly to 60% acetonitrile at 180 min, staying at this concentration for 60 min and decreasing steadily to 0% acetonitrile for 30 min at a constant flow of 7 ml/min. The gradient Fig. 1 shows the structure of each peptide conjugate. Peptides were synthesized by was controlled by two microprocessor-controlled Gilson 302 single piston manual solid phase peptide synthesis, using side-chain protected Fmoc-conjugated pumps. Compounds were detected with a Holochrome UV–VIS detector at amino acids. Peptides were then purified using Vydac C18 reverse-phase column on an 214 nm (analytical) and 230 nm (preparative). Chromatographs were recorded Agilant 1100 series HPLC and analysed by mass spectroscopy. Lipopeptides were with an LKB 2210 single channel chart recorder. synthesized using the Tris-linked technology as previously described by Whittaker Since the 2-aminododecanoic acid was used as a racemate, the synthesized et al. [20]. Briefly, the C-terminal valine of protected CP was linked to Gly-Tris– – – LA was a diastereomeric mixture. palmitate, dipalmitate and tripalmitate, followed by purification using a silica Myristoylated-CP (MYR) was purchased from Auspep (Melbourne Australia) column and organic eluants (chloroform and methanol) and the products analysed by with a purity greater than 90%. CP and its lipid conjugates were dissolved in 100% DMSOtostockconcentrationsof10mM,5mM,and2mMandweredilutedinmedia to final concentrations of 50 μM, 25 μM, and 10 μM in 0.5% DMSO for each experiment. As such, each lipid conjugate was compared to a 0.5% DMSO control. Stock GS was dissolved in distilled water to stock concentrations of 1 mM, 0.5 mM, 0.2 mM, and 0.1 mM and were diluted in media to a final concentration of 50 μM, 25 μM, 10 μM, and 5 μM. GS results are expressed relative to diluent control.

2.4. Antigen presentation assay

To establish the ability of CP conjugates to inhibit Tcell activation the antigen presentation model was used. In this system, cytochrome c-specific mouse hybridoma T cells were incubated with antigen-presenting cells, cytochrome c, and the conjugate of interest. The amount of IL-2 produced during this overnight stimulation was quantitated using the IL-2-dependant CTLL-2 cell line. This has been previously described by Manolios et al. [37]. Briefly, mouse T cell 2B4.11 hybridoma (5×104) were cultured in microtiter wells with LK35.2 antigen-presenting B cells (5×104) and 50 μM pigeon cytochrome c. After 20 h, 100 μl of assay supernatant was removed and assayed for the presence of IL-2 in serial 2-fold dilutions. CTLL cells (1–2×104) were then cultured in the supernatant for 20 h, pulsed with 0.5 μCi [3H] thymidine for 5 h and harvested on glass fibre filters (Packard Biosciences catalogue number 6005412) using a cell harvester (TiterTek). A Packard Biosciences Top Count NXT microplate scinti- llation and luminescence counter was used to measure [3H] thymidine uptake. Peptide conjugates were all tested at 50 μM, 25 μM, and 10 μM in triplicate per assay. Each assay was performed at least three times.

2.5. Viability and proliferation assay

2B4.11 T cells (1×105) were cultured in microtiter wells in the presence of the peptide conjugates at 50 μM. After 20 h, 30 μlwasremovedandcellsassessedfor viability by trypan blue staining. The remaining culture was then pulsed with 0.5 μCi Fig. 1. Structure of synthesized conjugates. [3H] thymidine for 5 h, harvested, and quantitated for [3H] thymidine uptake. 882 M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888

2.6. CD3 and TCR α/β cross-linking (bound antibody) 2.12. Adjuvant-induced arthritis model

96-well flat bottom microplate (Becton Dickinson, Australia) were coated Female Wistar rats, aged 8–10 weeks, were injected subcutaneously at the base with 10 μg/ml of 145-2C11 (laboratory purified) or H57-597 (Becton Dickinson of the tail with 1 mg of heat-killed Mycobacterium tuberculosis suspended in 100 μl Pharmingen Catalogue #553167) in 0.1%BSA/PBS overnight at 4 °C. Plates of squalene. One day after the first signs of arthritis appeared (usually around 10 to were then washed with media prior to the addition of 1×105 2B4.11 cells in the 12 days later), rats were subcutaneously treated for 4 consecutive days with 6 mg of presence of each peptide conjugate (at 50 μM and 0.5% DMSO) in 250 μl. Each peptide suspended in squalene (CP, n=5; LA, n=5; LP3, n=5; MYR, n=4), or reaction was performed in triplicate. Experiments were then continued as per the dH2O(GS,n=3), or 1.25 mg cyclosporin (n=5)indH2O (5 mg/kg). Controls were antigen presentation experiment described above. injected with squalene alone (n=5). Arthritis severity was scored daily using 5 parameters: rat weight, number of arthritic joints, and by using digital calipers, paw 2.7. CD3 and TCRα/β cross-linking (soluble antibody) thickness, paw width, ankle width were measured. All procedures were performed under general anaesthesia with isofluorane, and analgesics (buprenorphine) were given to relieve pain. Paw thickness, paw width, and ankle width, measurements T cell hybridoma, 2B4.11 cells (1×105) were cultured in 96-well, round were then combined and averaged for each group to give an overall assessment of bottom plates in the presence of LK cells (5×104), 0.5 μg/ml of soluble 145-2C11 the severity of arthritis for each peptide conjugate tested. (laboratory purified) or H57-597 (Santa Cruz Catalogue #sc-18904L), and each peptide conjugate (50 μM and 0.5% DMSO) to a final volume of 250 μl. Plates were incubated for 24 h at 37 °C with 5% CO2. Supernatants were then collected 3. Results and analysed for IL-2 content. All conjugates were tested simultaneously in triplicate wells, and each experiment was performed on three separate occasions. 3.1. Effect of CP conjugations on T cell antigen activation 2.8. PMA and ionomycin stimulation Fig. 2 shows the average IL-2 production by T cells in the PMA (40 ng/ml) and ionomycin (5 μg/ml) were added to 2B4 cells (5×104) presence of each CP conjugate relative to a control of 0.5% DMSO. and incubated with each CP-conjugate at 50 μM in 0.5% DMSO in 96-well, All but one of the lipid conjugated CP analogues (LP2) proved to be round bottom microtiter plates for 24 h at 37 °C with 5%CO2 in a total volume of more effective than CP (84.86±30.95 vs. 56.95 ±15.69% IL-2 μ 250 l. Supernatants were then collected and analysed for IL-2 production. All production at 50 μM, mean±SD, respectively). The water-soluble CP conjugates were tested simultaneously in triplicate wells, and each experiment was performed on three separate occasions. GS proved capable of inhibiting IL-2 production relative to diluent controls (82.95±14.96% vs. 100±14.92%, respectively; 2.9. Staphylococcal enterotoxin A activation P=0.028). This inhibition was significantly less compared to CP (82.95±14.96% vs. 56.95±15.69%, respectively; P=0.002). The T cell hybridoma, 2B4.11 (5×104), were incubated with LK 35.2 cells mono-palmitate conjugate of CP (LP1, 50 μM)resultedinno 4 (5×10 ) in the presence of 100 ng/ml of Staphylococcal enterotoxin A and quantifiable IL-2 being produced. This was due to its toxic effect on μ 50 M of each peptide overnight at 37 °C/5% CO2 in round bottomed 96-well cells (see section below on proliferation and viability). LP3 demo- plates. T cell stimulation was assessed by quantifying IL-2 production as described above. nstrated a 1.5 times increase in inhibition when compared to CP0 at 50 μM, resulting in 39.43±11.58% IL-2 production (P=0.016). 2.10. Preparation of small unilamellar vesicles (SUV) Myristoylation also resulted in a significant improvement on CP's ability to inhibit IL-2 production (P=2.6×10−4) to 28.85±9.04% Dimyristoyl phosphatidyl choline (DMPC) and dimyristoyl phosphatidyl IL-2 production. The most effective improvements on CP's ability glycerol (DMPG) were dissolved in dry chloroform and chloroform/methanol (2:1) to inhibit IL-2 production were LA (P=3.0×10−7), resulting in respectively to give 10 mg/ml solutions. These solutions were mixed in appropriate 12.08±3.34% IL-2 production when tested at 50 μM. LA was also proportions and subsequently evaporated under reduced pressure. The resulting lipid film was dried overnight in vacuo. Lipids were then hydrated by resuspending shown to be the most effective inhibitor of IL-2 production at in HEPES buffer for 60 min at 34 °C to give a 0.5 M concentration in respect of 25 μM and 10 μM (data not shown). All conjugations including CP phospholipids. The solution was then sonicated in an ultrasonic bath for 20 min. demonstrated dose responsiveness (data not shown). Eight cycles of freeze/thawing was followed by extrusion through polycarbonate filters, first 100 nm (21 times), and then 50 nm (21 times) pore diameter, using a Lipo-fast apparatus (Avestin, Ottawa, Canada). The resulting SUVs were used immediately. DMPC and DMPG were used in 70:30 ratio (DMPC:DMPG) to approximate the proportion found in vertebrate cell membranes.

2.11. Binding analysis by surface plasmon resonance (SPR)

SPR was carried out on a BIAcore 2000 instrument using Pioneer Sensor Chip L1 and HEPES (HBS-N, Biacore, Upssala, Sweden) as running buffer. The chip surface was cleaned with 40 mM octyl glycoside (30 μl, 10 μl/min) followed by running buffer (100 μl, 10 μl/min). Liposomes (SUV) were injected (100 μl, 5 μl/ min) giving approximately 8000 response units (RU). 10 mM NaOH (40 μl, 10 μl/ min) removed any multilamellar vesicles from the surface, which was followed by Fig. 2. The influence of CP conjugates on IL-2 production induced by antigen 10 mM glycine pH 2.2 (10 μl, 10 μl/min) before injecting the peptide of interest activation. Antigen presentation assay results showing the %IL-2 produced, as (100 μl, 5 μl/min). After peptide injection, the dissociation stage was 1200 s. compared to 0.5% DMSO control, in the presence of CP, LA, MYR, Regeneration of the sensor chip was achieved with 40 mM β-octyl glycoside lipopeptides (LP1, LP2, LP3) and GS. GS is calculated relative to a diluent (30 μl, 10 μl/min). All SPR experiments were run at 25 °C, and all analyses were control set to 100% (control not shown). Conjugates were tested in triplicate per performed using the BIAevaluation software (Biacore, Upssala, Sweden). All experiment, and each experiment was performed on at least three separate peptides were prepared at 50 μM and 0.5% DMSO in HEPES. occasions. All samples presented at 50 μM. M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888 883

3.2. Effect of CP conjugation on T cell proliferation and viability

The effect of CP conjugates on T cell proliferation was assessed by [3H]-thymidine uptake (data not shown). The parent compound, CP, resulted in a proliferation response of 79.93%±10.48 S.D. relative to controls. Besides LP1, which resulted in a dramatic drop on cell proliferation to only 1.09%±0.96 S.D. proliferation relative to control, the only conjugate resulting in a significant reduction in cell proliferation when compared to CP was LP3 at 60.05%± 10.36 S.D. proliferation relative to DMSO (P=0.008 relative to CP). Neither, MYR (74.46%±19.24 S.D.) nor LA (88.66%± Fig. 4. The influence of CP conjugates on IL-2 production induced by CD3 5.64 S.D.) resulted in significant changes in cell proliferation cross-linking. Percent IL-2 production by T cells stimulated with CD3 cross- when compared to CP (PN0.05). linking by soluble 2C11. Conjugates were tested at 50 μM and results calculated Cell proliferation assays were coupled with trypan blue staining with respect to a control of 0.5% DMSO. All conjugates were tested simultaneously in triplicate wells, and each experiment was performed three to determine the CP conjugate's ability to disrupt the T cell mem- times. brane and thereby affect the cell's viability. The only conjugate to result in a percent viability below 90% relative to the DMSO control was LP1, which resulted in a viability of 14.72%±9.89 S. (data not shown). The results are expressed in terms of percent D. CP presented with a viability of 98.39%±1.69 S.D.; LP3 gave IL-2 produced compared to a control of 0.5% DMSO. With 93.64% ±2.85 S.D.; MYR resulted in 95.35%±3.44 S.D.; and LA soluble, as with plate-bound 2C11, CD3 stimulation in the demonstrated a viability of 94.97%±2.38 S.D. Taken together, the presence of CP at 50 μM did not inhibit IL-2 production proliferation and viability studies demonstrate that LP1 is cyto- (120.06%±19.42 S.D. and 152.31%±24.82 S.D. respectively). toxic, whereas LP3, MYR, and LA do not inhibit IL-2 production CD3 stimulation in the presence of MYR at 50 μM also did not due to cytotoxicity. inhibit IL-2 production with 122.04%±27.04 S.D. using bound antibody, and 102.38%±23.26 S.D. using soluble antibody. LP3 3.3. Effect of CP conjugations on IL-2 production following T and LA were the only conjugates to demonstrate an ability to cell stimulation with: inhibit IL-2 production. When treated with LA, IL-2 production after CD3 cross-linking was decreased to 66.83%±25.11 S.D. 3.3.1. PMA and ionomycin (P=1.71×10− 3) with immobilized 2C11, and 78.30%±17.13 S. In an attempt to understand if the mechanism of action of these D. (P=0.012) with soluble 2C11 at 50 μM. LP3 at 50 μM was compounds was different to that proposed for CP alone, exper- able to inhibit IL-2 production to 76.09%±24.81 S.D. (P=0.014) iments similar to those performed by Wang et al. were performed when immobilized antibody was used but was unable to inhibit [28]. To activate the T cell downstream from the TCR, 2B4 cells IL-2 production when soluble antibody was used as a stimulant were stimulated with PMA and ionomycin. No conjugate demon- (111.94%±20.38 S.D.). strated any ability to inhibit IL-2 production, in fact, all conjugates tested herein resulted in higher amounts of IL-2 production 3.3.3. Staphylococcal enterotoxin A (SEA) activation than the DMSO control, as shown in Fig. 3. This method activates T cells by cross-linking the MHC molecule of the antigen-presenting cell (LK 35.2) to the T cell's 3.3.2. CD3 cross-linking TCR-β (Vβ) chain, thereby inducing T cell activation [38,39]. 2B4 cells were next stimulated with CD3 cross-linking anti- The results of these experiments are shown in Fig. 5. In this bodies to bypass the TCRα/β antigen recognition subunits. Fig. 4 model, MYR was not able to inhibit IL-2 production, resulting in shows the results of cross-linking with soluble 2C11. The same experiments were repeated with immobilized, plate-bound 2C11

Fig. 5. The influence of CP conjugates on IL-2 production induced by the superantigen SEA. Percent IL-2 production by T cells stimulated with SEA. Fig. 3. Percent IL-2 production by T cells following PMA and ionomycin Conjugates were tested at 50 μM and results calculated with respect to a control stimulation in the presence of CP conjugates. Conjugates were tested at 50 μM of 0.5% DMSO. All conjugates were tested simultaneously in triplicate wells, and results calculated with respect to a control of 0.5% DMSO. and each experiment was performed three times. 884 M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888

114.11%±31.18 S.D., and while LP3 was able to inhibit IL-2 production marginally at 81.30%±26.10 S.D., statistical significance was not reached (P=0.07). CP was able to inhibit IL-2 production to 54.76%±8.87 S.D. (P=1.48×10− 7), and LA was able to bring IL-2 production to 37.04%±3.84 S.D. (P=1.0×10−10) relative to the controls.

3.3.4. TCR cross-linking Tcells were activated by cross-linking the TCRβ through its Cβ FG loop [40] using the H57-597 antibody in either soluble or immobilized forms. Fig. 6 shows the results obtained using soluble antibody with data expressed in terms of%IL-2 produced compared to the DMSO control. In both soluble and bound antibody activation models, only LA was able to inhibit IL-2 production significantly, producing 74.54%±23.20 S.D. (P=0.01) when the antibody is plate bound, and 75.95% ±11.66 S.D. (P=0.00038) when the antibody is administered soluble. Although MYR was Fig. 7. Binding of CP and conjugations to model membranes. Each conjugate was tested at 50 μM (0.5% DMSO) on membranes composed of DMPC: DMPG at able to inhibit IL-2 production when the cells were activated using 70:30. LP1: –••–,LP2:––••,LP3:•••,MYR: ,LA:–•–,CP:—,and bound antibody to 91.40%±20.07 S.D., statistical significance GS: ––. was not reached (P=0.3). Similarly, as shown in Fig. 6, there was no statistically significant inhibition with MYR using soluble membrane surface. MYR's final RU value was approximately antibody. All other peptides tested resulted in %IL-2 production 4000 RU. The carbohydrate conjugate GS showed an adsorption levels equal to or greater than that observed by the DMSO control. of about 1000 RU but rapidly dropped to approximately 300 RU as soon as the dissociation phase began. Of the Tris conjugates 3.4. Effect of CP conjugations on model membrane binding (reprinted with permission from Ali et al., copyright 2005, Ame- rican Chemical Society [41]), the strongest binding was observed To correlate function with the effect of conjugation on CP's with LP1. LP1's association levelled at roughly 18400 RU, and ability to bind to model lipid membranes, 50 μMofeachcon- upon dissociation, dropped to 17500 RU, which was about ten jugate was passed over membranes composed of DMPC and times more than that observed for CP (1700 RU). LP2 resulted in DMPG in a 70:30 ratio and analysed by biosensor technology 6000 RU bound after the dissociation phase, and LP3 resulted in (Fig. 7). All conjugates were added to membranes at time 0 and the least binding at roughly 3500 RU remaining after dissociation. were dissociated at 1200 s. In this study, CP presented with a In all cases, lipopeptides could not be removed from the bound response unit (RU) value of 5200 RU at the end of the association lipid, suggesting that at least part of the reason for their increased phase, declining to 1700 RU after dissociation. Interestingly, LA, effectiveness in vitro is likely due to an improved ability to remain which was the most active conjugate in vitro, showed a slow but bound to lipid membranes when compared to CP [41]. strong binding curve with no decrease in RU during the dissociation phase. The final RU observed for LA was 2900 RU, 3.5. Therapeutic effect of peptide conjugations on a rat model which could not be removed, implying it was firmly bound to the of adjuvant-induced arthritis lipid layer. MYR showed similar results with a slow association phase (compared to CP), and an inability to be removed from the To determine whether or not the in vitro activities of these peptide conjugations carried over to in vivo systems, rats with adjuvant-induced arthritis were given subcutaneous injections of each conjugate (excluding LP1 and LP2) to determine the effectiveness of each conjugate. Six different criteria were utilized to determine the severity of arthritis. These included digital caliper measurements of paw thickness, paw width, and ankle width, rat weight and the number of arthritic joints were assessed daily. Measurements were normalized by setting the measurement from the day before arthritis was observed to 100% and calculating all subsequent measurements based on this value. A statistical analysis comparing the slopes (progression) obtained by plotting condition (such as ankle thickness or paw width) on the y-axis vs. Fig. 6. The influence of CP conjugates on IL-2 production induced by soluble day on the x-axis, of each of the six criteria showed that all con- β antibody TCR- cross-linking. Percent IL-2 production by T cells stimulated jugations reduced the progression of arthritis significantly with TCRβ cross-linking with soluble H57-597 antibody. Conjugates were b tested at 50 μM and results calculated with respect to a control of 0.5% DMSO. (P 0.05) in at least one of the six criteria with the exception of All conjugates were tested simultaneously in triplicate wells, and each MYR, which approached significance in the ankle thickness experiment was performed three times. criteria (P=0.055).Fig. 8 shows the overall severity of arthritis in M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888 885

Recent studies have shown that lipid conjugation can increase the functional activity of peptides [46–48]. In agreement with these studies, lipoamino acid conjugation of CP enhanced CP function. As noted by Toth et al., it is possible to lipoconjugate a compound using lipoamino acids at the N-terminal while reta- ining the parent compound's biological activity [49]. While the lipoamino acid conjugate was the most active compound tested herein, the carbohydrate conjugated CP was surprisingly one of the least active. When CP was conjugated to glucosesuccinate, this compound failed to inhibit T cell activation in the antigen presentation assay. Intuitively, it was thought that an increase in CP's solubility by glycosylation would increase its bioavailability Fig. 8. The influence of CP conjugates as a treatment in the rat adjuvant-induced arthritis model. Rats were treated subcutaneously with 6 mg of peptide for 4 and hence its activity. However, the results were contrary to consecutive days starting one day after arthritis was diagnosed. Results are expectations and suggested that the requirements for an effective presented as the overall paw swelling averaged for all measurements over time drug delivery agent must take into account the site of action of the (days). 100% is set at Day 0 (the day arthritis was diagnosed). CP: —; Squalane: drug. If a drug is more effective in a hydrophilic environment such ––•• ––• –•– ••• –••– ; LP3: ; LA: ; MYR: ; Cyclosporin (Cyc): ; as the cytosol, then a more soluble carrier may be useful. Whereas GS: ––. if the drug is to act in a hydrophobic region such as a trans- each treatment group, calculated by averaging the percent swelling membrane domain, this aspect should be reflected in the carrier. In from every rat in each treatment group and plotting this over time this instance, the hydrophilic nature of the glucosesuccinate may (day). In agreement with results presented by Kurosaka et al. have sequestered the short hydrophobic CP from successfully (manuscript submitted), LP3 administered subcutaneously show- reaching its target or at least inhibited CP from binding to the ed no improvement on CP's ability to inhibit arthritis in this cell's membrane. In this respect, localization experiments with model. By contrast, MYR and LA both improved CP's ability to FITC-labelled peptides using confocal microscopy may be useful treat arthritis. Rats treated with GS showed a response similar to and is currently under investigation. Using surface plasmon reso- that observed for cyclosporin. nance experiments GS showed no quantifiable binding to model membrane bilayers (Fig. 7), supporting the above argument. 4. Discussion Conjugates were next tested for their ability to retain the parent compounds original site of biological activity. Similar to previous In this study, the effectiveness of various conjugations on an experiments with CP [28], the lack of activity from the conjugates insoluble hydrophobic peptide (CP) was examined. Changes in in the PMA/ionomycin activation and CD3ε cross-linking expe- CP's ability to inhibit T cell activation was observed following riments suggested that these conjugations were active upstream of each conjugation. Efficacy of lipid conjugation was dependent the CD3 complex, likely at the TCR α/β interface. To further test upon the type and number of lipid groups attached to CP. In the this, T cells were activated by either the superantigen SEA, which case of myristoylation, CP's ability to inhibit antigen activation was significantly increased in vitro. Myristoylation is a well- known post-translational modification used by cells to anchor proteins to the plasma membrane and includes examples such as myristoylated alanine-rich C kinase substrate and endothelial nitric oxide synthase [42]. The same can be said of palmitoylated CP conjugates, since many proteins are stably anchored to the cell membrane upon palmitoylation [43].AlthoughtheTris-lipidation employed here is different from the N-palmitoylation normally used by the cell, the structural similarity of the Tris molecule to glycerol allows it to mimic naturally occurring glycerides. In the Tris molecule, however, the central carbon atom is symmetrical and lends itself to derivatisation by avoiding structural isomerization, which can be a difficulty in the synthesis and purification of glycerol derivatives. Tris-lipidation has previously been proven successful in the delivery of methotrexate in animal models demonstrating reduced toxicity and improved bioavailability [44]. Fig. 9. A potential mode of action of CP. (a) The normal transmembrane charge The results of the antigen presentation assay, proliferation assay, interactions between the TCR α/β and the CD3εδ, CD3ζζ, and the CD3γε and viability studies for the Tris-conjugates have recently been chains essential in the formation of a functional T cell receptor. (b) Proposed published elsewhere [41], where it was argued that the differential transmembrane interactions between CP's arginine and lysine residues and those of the CD3δε and CD3ζζ chains. This model predicts that CP disrupts the results between each of the compounds was due to differential transmembrane charges between the TCRα and the CD3δε and CD3ζζ chains. localization within the cell, or their variable susceptibility to This results in a non-functioning T cell receptor consisting of only the TCRα, enzyme digestion as has been reported by other groups [24,45]. TCRβ, and the CD3γε subunits. 886 M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888

Table 1 The proposed first step in attempting to destabilize the TCR by Summary of CP activity in different activation models CP is membrane binding, followed by membrane insertion. To Activation IL-2 production determine the effect of conjugation on the ability of each peptide TCRαβ dependent Antigen − to bind to artificial membranes, each conjugate was examined SEA − using surface plasmon resonance. When tested with artificial TCR-β cross-linking + membrane bilayers composed of 70:30 DMPC:DMPG (Fig. 9), CD3 dependent CD3 cross-linking + each lipid conjugate was found to bind irreversibly. Extensive non TCR-CD3 dependent PMA/ionmycin + washing was incapable of removing bound peptide. With the Results of CP activity in each of the activation models tested were the following: exception of LP2, the irreversible binding observed for the lipid “−” represents an inhibition of IL-2 production to levels significantly below that presented by the DMSO controls (Pb0.05); and “+” representing IL-2 conjugations may be the reason for their increased ability to production equal to or greater than levels presented by DMSO controls. inhibit IL-2 production in the antigen presentation assay. The high binding of LP1, and its ability to enter membranes much more acts by cross-linking the TCRβ chain with the MHC of an anti- rapidly than its di- and tri-lipidated counterparts [54],ispoten- gen-presenting cell resulting in a similar signalling cascade as tially the reason for its cytotoxicity as observed previously [41]. observed with peptide antigens [38,50,51] or by the TCR-β- Surprisingly, GS was not capable of binding to artificial mem- specific H57-597 antibody [34,40]. Core peptide and LA were branes in this system. able to significantly inhibit IL-2 production during SEA To evaluate the clinical efficacy of these conjugates rats with activation (Pb0.05), LP3 showed partial inhibition (P=0.07), adjuvant-induced arthritis were treated with GS, LA, MYR, LP3, whereas MYR showed no inhibition, suggesting slightly different CP, and cyclosporin, for 4 consecutive days. Cyclosporin had a modes of action when compared to CP for LP3 and MYR. good effect on arthritis outcome. As expected, animals treated Activation through the TCRβ chain by either soluble or plate- with CP,LA, and MYR also had a good outcome and mirrored the bound H57-597 showed that only LAwas capable of significantly results obtained from the antigen presentation assay. Not expected inhibiting IL-2 production, albeit only to approximately 75% of was the ability of GS to abolish arthritis, such that only a few control, suggesting that these compounds associate more closely joints remained red, with minimal swelling. This discordance with the TCRα chain than the TCRβ chain as utilized in this raises an important point of this study in that loss of activity in model of activation in agreement with the hypothesized mode of vitro by conjugation (or at least partial loss) does not correlate CP action (Fig. 9). with loss of activity in vivo. It is possible that GS, being water- Sigalov recently proposed a model of T cell activation in soluble, enters the bloodstream with greater ease where serum which, depending upon the strength of the TCR–peptide–MHC proteases, or other enzymes, cleave the glucosesuccinate moiety interaction, the CD3ζζ, CD3δε and CD3γε signalling oligo- resulting in an active form of CP. Further work is required to mers dissociate differentially from the TCRαβ subunit [52]. determine the precise mechanism of action of GS. Although few These signalling dimers then interact with other TCRs to result in animals were employed in these experiments, statistically signi- higher order signalling complexes and an amplification of the ficant results were observed with CP conjugation. activation signal. Within this context, it has been proposed that CP In summary, conjugation of CP with lipids/sugars enhances inhibits TCRα interactions with its associated CD3 chains, activity without altering the site of action of peptide function. creating a state where the CD3δε and CD3ζζ dimers are “pre- Further assessment of these compounds as potential therapeutic dissociated” from the TCRαβ antigen recognition subunit [52]. agents is currently the focus of ongoing research within our Upon antigen activation, which requires a fully assembled TCR- laboratory. CD3 molecule, the CD3δε and ζζ subunits are already dissociated from the TCRα resulting in no activation. To highlight the agre- Acknowledgements ement between the experimental results and this model, Table 1 summarizes CP's ability to inhibit IL-2 production following each We express our sincere appreciation to Dr. Andrea Bolte for of the stimulation methods examined. Neither PMA/ionomycin the helpful discussions. M. A. Amon was supported by the activation or CD3 cross-linking require the CD3 chains to be Barbara Cameron Memorial Scholarship Sydney University, and associated with the TCRαβ, which may explain why CP has no M. Ali was supported by a Grant-in-aid from the Arthritis Fou- effect. By contrast, CP successfully inhibits IL-2 production ndation of Australia. A grant-in-aid from the Rebecca Cooper requiring complete CD3 association with the TCRαβ subunit Medical Research Foundation is appreciated. (antigen and SEA activation). However, CP is not capable of inhibiting IL-2 production when activation is through cross- References linking with a monoclonal antibody directed against the TCRβ chains. This may be because CP does not affect the interactions [1] S.M. Bates, J. Lister-James, J.A. Julian, R. Taillefer, B.R. Moyer, J.S. between the TCRβ and the CD3γε, which has been shown to Ginsberg, Imaging characteristics of a novel Technetium Tc 99m-labeled involve extracellular [34,40,53] as well as transmembrane charge platelet glycoprotein IIb/IIIa receptor antagonist in patients with acute deep interactions [26], making this interaction possibly harder to vein thrombosis or a history of deep vein thrombosis, Arch. Intern. Med. 163 (2003) 452–456. disrupt. These results taken together, suggest that CP activity may [2] S. Albertario, P. Forti, C. Bianchi, G. Morone, F. Tinozzi, P. Moglia, M. Abelli, be centred around the interactions between the TCRα and its M. Benedetti, C. Aprile, Radioguided surgery for gastrinoma: a case report, associated CD3δε and ζζ subunits (Fig. 9). Tumori 88 (2002) S41–S43. M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888 887

[3] J.S. Cervia, M.A. Smith, Enfuvirtide (T-20): a novel human immunodeficiency [26] M.E. Call, J. Pyrdol, M. Wiedmann, K.W. Wucherpfennig, The organizing virus type 1 fusion inhibitor, Clin. Infect. Dis. 37 (2003) 1102–1106. principle in the formation of the T Cell receptor-CD3 complex, Cell 111 [4] E.W. McClesky, A.P. Fox, D.H. Feldman, L.J. Cruz, B.M. Olivera, R.W. Tsien, (2002) 967–979. D. Yoshikami, ω-Conotoxin: direct and persistent blockade of specific types of [27] X.M. Wang, J.T. Djordjevic, V. Bender, N. Manolios, T cell antigen receptor calcium channels in neurons but not muscle, Proc. Natl. Acad. Sci. U. S. A. 84 (TCR) transmembrane peptides colocalize with TCR, not lipid rafts, in surface (1987) 4327–4331. membranes, Cell. Immunol. 215 (2002) 12–19. [5] P. Proksch, R. Edrada, P. Ebel, Drugs from the seas—current status and [28] X.M. Wang, J.T. Djordjevic, N. Kurosaka, S. Schibeci, L. Lee, P. Williamson, microbiological implications, Appl. Microbiol. Biotechnol. 59 (2002) N. Manolios, T-cell antigen receptor peptides inhibit signal transduction within 125–134. the membrane bilayer, Clin. Immunol. 105 (2002) 199–207. [6] S. Dhib-Jalbut, Glatiramer acetate (Copaxone1) therapy for multiple sclerosis, [29] V. Bender, M. Ali, M. Amon, E. Diefenbach, N. Manolios, T cell antigen Pharmacol. Ther. 98 (2003) 129–265. receptor peptide–lipid membrane interactions using surface plasmon [7] J. Borghardt, B. Rosien, R. Gortelmeyer, S. Lindemann, M. Hartleb, M. resonance, The J. Biol. Chem. 279 (2004) 54002–54007. Klingmuller, Effects of a spleen peptide preparation as supportive therapy in [30] G. Gollner, G. Muller, R. Alt, J. Knop, A.H. Enk, Therapeutic application inoperable head and neck cancer patients, Arzneim.-Forsch. 50 (2000) of T cell receptor mimic peptides or cDNA in the treatment of T cell 178–184. mediated skin diseases, Gene Ther. 7 (2000) 1000–1004. [8] C.F.I. Shield, M.M. McGrath, T.F. Goss, Assessment of health-related quality [31] K. Mahnke, Q. Yingjie, J. Knop, A.H. Enk, Dendritic cells, engineered to of life in kidney transplant patients receiving tacrolimus (FK506)-based versus secrete a T-cell receptor mimic peptide, induce antigen-specific immuno- cyclosporine-based immunosuppression, Transplantation 64 (1997) suppression in vivo, Nat. Biotechnol. 21 (2003) 903–908. 1738–1743. [32] O. Leo, M. Foo, D.H. Sachs, L.E. Samelson, J.A. Bluestone, Identification [9] M.G. Mutchnick, K.L. Lindsay, E.R. Schiff, G.D. Cummings, H.D. Appelman, of a monoclonal antibody specific for a murine T3 polypeptide, Proc. Natl. R.R. Peleman, M. Silva, K.C. Roach, F. Simmons, S. Milstein, S.C. Gordon, Acad. Sci. U. S. A. 84 (1987) 1374–1378. M.N. Ehrinpreis, Thymosin [alpha]1 treatment of chronic hepatitis B: results [33] R.T. Kubo, W. Born, J.W. Kappler, P. Marrack, M. Pigeon, Characteriza- of a phase III multicentre, randomized, double-blind and placebo-controlled tion of a monoclonal antibody which detects all murine ab T cell receptors, study, J. Viral Hepatitis 6 (1999) 397–403. The J. Immunol. 142 (1989) 2736–2742. [10] A. Sehgal, Recent developments in peptide-based cancer therapeutics, Curr. [34] J.-h. Wang, K. Lim, A. Smolyar, M.-k. Teng, J.-h. Liu, A.G.D. Tse, J. Opin. Drug Discov. Dev. 5 (2002) 245–250. Liu, R.E. Hussey, Y. Chishti, C.T. Thomson, R.M. Sweet, S.G. Nathan- [11] H.-L. Wang, A. Kurtz, Breast cancer growth inhibition by delivery of the son, H.-C. Chang, J.C. Sacchettini, E.L. Reinherz, Atomic structure of an MDGI-derived peptide P108 Space, Oncogene 19 (2000) 2455–2460. ab T cell receptor (TCR) heterodimer in complex with an anti-TCR Fab [12] P.W. Latham, Therapeutic peptides revisited, Nat. Biotechnol. 17 (1999) fragment derived from a mitogenic antibody, The EMBO J. 17 (1998) 755–757. 10–26. [13] M. Ali, N. Manolios, Peptide delivery systems, Lett. Pept. Sci. 8 (2002) 289–294. [35] W.A. Gibbons, R.A. Hughes, M. Chralambous, A. Szeto, A.E. Aulabaugh, [14] P.S. Burton, R.A. Conradi, N.F.H. Ho, A.R. Hilgers, R.T. Borchardt, P. Maseagni, I. Toth, Synthesis, resolution and structural elucidation of How structural features influence the biomembrane permeability of lipidic amino acids and their homo- and hetero-oligomers. Liebigs Ann. peptides, J. Pharm. Sci. 85 (1996) 1336–1340. Chem. (1990) 1175–1183. [15] B. Drouillat, A.M. Hillery, G. Dekany, R. Falconer, K. Wright, I. Toth, Novel [36] B.P. Ross, S.E. DeCruz, T.B. Lynch, K. Davis-Goff, I. Toth, Design, synthesis liposaccharide conjugates for drug and peptide delivery, J. Pharm. Sci. 87 and evaluation of a liposaccharide drug delivery agent: application to the (1998) 25–30. gastrointestinal absorption of gentamicin, J. Med. Chem. 47 (2004) 1251–1258. [16] M. Schelhaas, E. Nagele, N. Kuder, B. Bader, J. Kuhlmann, A. Wirringhofer, H. [37] N. Manolios, S. Collier, J. Taylor, J. Pollard, L.C. Harrison, V. Bender, Waldmann, Chemoenzymatic synthesis of biotinylated ras peptides and their Immunomodulatory antigen receptor transmembrane peptides and their use in membrane binding studies of lipidated model proteins by surface effect on T-cell mediated disease, Nat. Med. 3 (1997) 84–88. plasmon resonance, Chem. Eur. J. 5 (1999) 1239–1252. [38] M.D. Baker, K.R. Acharya, Superantigens: structure–function relation- [17] C. Peters, M. Wagner, M. Volkert, H. Waldmann, Bridging the gap between ships, Int. J. Med. Microbiol. 293 (2004) 529–537. cell biology and organic chemistry: chemical synthesis and biological [39] L. Cuff, R.G. Ulrich, M.A. Olson, Prediction of the multimeric assembly of application of lipidated peptides and proteins. Naturwissenschaften 89 (2002) staphylococcal enterotoxin A with cell-surface protein receptors, The J. 381–390. Graphics Model. 21 (2003) 473–486. [18] A.L. Mora, J.P. Tam, Controlled lipidation and encapsulation of peptides as [40] Y. Ghendler, A. Smolyar, H.-C. Chang, E.L. Reinherz, One of the CD3e a useful approach to mucosal immunizations, The J. Immunol. 161 (1998) subunits within a T cell receptor complex lies in close proximity to the Cb 3616–3623. FG loop, J. Exp. Med. 187 (1998) 1529–1536. [19] J. Wang, D. Shen, W.-C. Shen, Preparation, purification, and characterization [41] M. Ali, M. Amon, V. Bender, N. Manolios, Hydrophobic transmembrane- of a reversibly lipidized desmopressin with potentiated anti-diuretic activity, peptide lipid conjugations enhance membrane binding and functional Pharm. Res. 16 (1999) 1674–1679. activity in T-cells, Bioconjug. Chem. 16 (2005) 1556–1563. [20] R.G. Whittaker, P.J. Hayes, V.J. Bender, A gentle method for linking Tris [42] H. Taniguchi, Protein myristoylation in protein–lipid and protein–protein to amino acids and peptides, Pept. Res. 6 (1993) 125–128. interactions, Biophys. Chemist. 82 (1999) 129–137. [21] I. Toth, A novel chemical approach to drug delivery: lipidic amino acid [43] M.-J. Bijlmakers, M. Marsh, The on-off story of protein palmitoylation, conjugates, J. Drug Target. 2 (1994) 217–239. Trends Cell Biol. 13 (2003) 32–42. [22] L.J. Otvos, J.D. Wade, F. Lin, B.A. Condie, J. Hanrieder, R. Hoffmann, [44] T. Lockett, W. Reilly, M. Manthey, X. Wells, F. Cameron, M. Moghaddam, Designer antibacterial peptides kill fluoroquinolone-resistant clinical J. Johnston, K. Smith, C. Francis, Q. Yang, R. Whittaker, Tris lipidation: a isolates, J. Med. Chem. 48 (2005) 5349–5359. chemically flexible technology for modifying the delivery of drugs and [23] M. von Itzstein, W.-Y. Wu, G.B. Kok, M.S. Pegg, J.C. Dyason, B. Jin, T.V. genes, Clin. Exp. Pharmacol. Physiol. 27 (2000) 563–567. Phan, M.L. Smythe, H.F. White, S.W. Oliver, P.M. Colman, J.N. Varghese, [45] F.H. Cameron, M.J. Moghaddam, V.J. Bender, R.G. Whittaker, M. Mott, D.M. Ryan, J.M. Woods, R.C. Bethell, V.J. Hotham, J.M. Cameron, C.R. T.J. Lockett, A transfection compound series based on a versatile Tris Penn, Rational design of potent sialidase-based inhibitors of influenza linkage, Biochim. Biophys. Acta, Biomembr. 1417 (1999) 37–50. virus replication, Nature 363 (1993) 418–423. [46] D. Avrahami, Y. Shai, Bestowing antifungal and antibacterial activities by [24] N. Manolios, S. Collier, J. Taylor, J. Pollard, L.C. Harrison, V. Bender, T- lipophilic acid conjugation to D,L-amino acid-containing antimicrobial cell antigen receptor transmembrane peptides modulate T-cell function and peptides: a plausible mode of action, Biochemistry 42 (2003) 14946–14956. T-cell mediated disease, Nat. Med. 3 (1997) 84–88. [47] N. Sal-Man, D. Gerber, Y. Shai, The identification of a minimal dimerization [25] N. Manolios, Structure and assembly of the TCR, Today's Life Sci. 4 (1992) motif QXXS that enables homo- and hetero-association of transmembrane 32–43. helices in vivo, The J. Biol. Chem. 280 (2005) 27449–27457. 888 M.A. Amon et al. / Biochimica et Biophysica Acta 1763 (2006) 879–888

[48] A. Majerle, J. Kidric, R. Jerala, Enhancement of antibacterial and lipo- [51] D.L. Woodland, R. Wen, M.A. Blackman, Why do superantigens care polysaccharide binding activities of a human lactoferrin peptide fragment by about peptides? Immunol. Today 18 (1997) 18–22. the addition of acyl chain, J. Antimicrob. Chemother. 51 (2003) 1159–1165. [52] A. Sigalov, Multi-chain immune recognition receptors: spatial organization [49] I. Toth, J.P. Malkinson, N.S. Flinn, B. Drouillat, A. Horvath, J. Erchegyi, and signal transduction, Semin. Immunol. 17 (2005) 51–64. M. Idei, A. Venetianer, P. Artursson, L. Lazorova, B. Szende, G. Keri, [53] N. Manolios, L.Z.G., The T cell antigen receptor beta chain interacts with Novel lipoamineo acid- and liposaccharide-based system for peptide the extracellular domain of CD3-gamma, Immunol. Cell Biol. 73 (1995) delivery: application for oral administration of tumor-selective somato- 532–536. statin analogues, J. Med. Chem. 42 (1999) 4010–4013. [54] V.J. Bender, F. Cameron, P. Hendry, T.J. Lockett, R.G. Whittaker, [50] T. Proft, J. Fraser, Superantigens: just like peptides only different, J. Exp. Delivery Studies using Peptide-Fatty Acid Conjugates, ESCOM, Lieden, Med. 187 (1990) 819–821. 1994.