NATURE BIOTECHNOLOGY VOL 18 FEBRUARY 2000 http://biotech.nature.com 225

cycle coordination between nucleus and cytoplasm and the use of in vitromatured ooctyes. Mol. Reprod. Dev. 47, 255–264 (1997).

5. Wilmut, I., Schnieke, A.E., McWhir, J., Kind, A.J. & Campbell, K.H.S. Viable off-spring derived from fetal and adult mammalian cells. Nature 385, 810–813 (1997).

6. Cheong, H.T., Takahashi, Y. & Kanagawa, H. Relationship between nuclearremodeling and subsequent development of mouse embryonic nuclei transferredto enucleated oocytes. Mol. Reprod. Dev. 37, 138–145 (1994).

7. Smith, L.C. Membrane and intracellular effects of ultraviolet irradiation withHoechst 33342 on bovine secondary oocytes matured in vitro. J. Reprod. Fertil.99, 39–44 (1993).

8. Evans, M.J. et al. Mitochondrial DNA genotypes in nuclear transfer-derivedcloned sheep. Nat. Genet. 23, 90–93 (1999).

9. Oldenbourg, R. A new view on polarization microscopy. Nature 381, 811–812(1996).

10. Inoue, S. Polarization optical studies of the mitotic spindle. I. The demonstrationof spindle fibers in living cells. Chromosoma 5, 487–500 (1953).

11. Hiramoto, Y. et al. Quantitative studies on the polarization optical properties ofliving cells. II. The role of microtubules in birefringence of the spindle of the seaurchin egg. J. Cell Biol. 89, 121–130 (1981).

12. Oldenbourg, R. Polarized light microscopy of spindles. Methods Cell Biol.. 61,175–208 (1999).

13. Keefe, D., Tran, P., Pellegrini, C. & Oldenbourg, R. Polarized light microscopy anddigital image processing identify a multilaminar structure of the hamster zonapellucida. Human Reprod. 12, 1250–1252 (1997).

14. Silva, C.P., Kommineni, K., Oldenbourg, R. & Keefe, D.L. The first polar bodydoes not predict accurately the location of the metaphase II meiotic spindle inmammalian oocytes. Fertil. Steril. 71, 719–721 (1999).

15. Wakayama, T. & Yanagimachi, R. Fertilisability and developmental ability ofmouse oocytes with reduced amounts of cytoplasm. Zygote 6, 341–346 (1998).

16. Peura, T.T., Lewis, I.M. & Trounson, A.O. The effect of recipient oocyte volume onnuclear transfer in cattle. Mol. Reprod. Dev. 50, 185–191 (1998).

17. Wakayama, T., Perry, A.C.F., Zuccotti, M., Johnson, K.R. & Yanagimachi, R. Full-term development of mice from enucleated oocytes injected with cumulus cellnuclei. Nature 394, 369–374 (1998).

18 Tsunoda, Y., Shioda, Y., Onodera, M., Nakamura, K. & Uchida, T. Differential sen-sitivity of mouse pronuclei and zygote cytoplasm to Hoechst staining and ultravi-olet irradiation. J. Reprod. Fertil. 82, 173–178 (1988).

19. Latham, K.E.L. & Solter, D. Transplantation of nuclei to oocytes and embryos.Methods Enzymol. Vol. 225, 719–732 (1993).

20. Liu, Z. & Foote, R.H. Effects of amino acids on the development of in-vitromatured in-vitro fertilized bovine embryos in a simple protein-free medium.Human Reprod. 10, 2985–2991 (1995).

with antimicrobial membrane peptides. The color changes in thesystem occur because of the structural perturbation of the lipidsfollowing their interactions with antimicrobial peptides. The assaywas also sensitive to the antibacterial properties of structurally andfunctionally related peptide analogs.

Phospholipid–PDA vesicles were prepared through sonication ofthe aqueous mixtures of the phospholipid and 10,12-tricosadiynoicacid (4:6 mole ratio) at a temperature of ∼ 70°C, followed by slow cool-ing overnight and irradiation at 254 nm (ref. 2). The vesicles exhibitedan intense blue color due to the alternating conjugated triplebond/double bond PDA backbone3. Figure 1 shows a portion of a 96-well plate containing polymerized dimiristoylphosphocholine(DMPC)–PDA vesicle solutions into which identical quantities of themembrane-associated antibiotic peptides melittin4, magainin II5, andalamethicin6, as well as the soluble neuropeptide hormone PBAN7,were added separately. These results demonstrate that addition of themembrane peptides led to color changes within the vesicle solutions.Peptides that are not expected to bind cellular membranes do notinduce detectable colorimetric transitions, as shown, for example, in amixture consisting of the vesicles and the neuropeptide PBAN (Fig.1E). No colorimetric responses (negative controls) have been detectedfor various other soluble peptides and proteins that generally appear inphysiological solutions, such as albumin. The different degrees of colorchanges induced by each peptide most likely depend upon their dis-tinctive modes of interactions with the DMPC/PDA vesicle interface.

Previous studies have revealed that biological processes leadingto structural perturbations at the PDA vesicle interface, includingligand–receptor recognition8, pH changes9, and enzymatic catalysis2,are responsible for the blue-red transitions occurring in the vesicleassemblies. Specifically, the color changes are caused by strainsinduced in the conjugated backbone of the PDA vesicles followingrearrangements of the pendant side chains of the PDA molecules10,11.A similar mechanism most likely accounts for the colorimetric tran-sitions that occur in the phospholipid–PDA vesicles following addi-tion of the membrane peptides. Binding of the peptides gives rise toextensive molecular reorganization within the phospholipiddomains8,12, affecting the structure of the surrounding PDA matrix,which results in the blue-red color changes.

We carried out further experiments to evaluate the sensitivity ofthe technique and the dependence of the color change upon the typeof phospholipids included within the PDA matrix. Titration curvesin Figure 2A depict the colorimetric responses (CR) of the phospho-lipid–PDA vesicles as a function of the concentration of melittin.The CRs were calculated from the relative blue and red absorbancesdetected in the UV-visible spectra, and provide a quantitative valueof the blue-red transitions8. The results shown in Figure 2A indicatethat addition of melittin induced color changes in PDA assembliesincorporating DMPC, dimiristoylphosphoethanolamine (PE), andin mixtures of PE and dimiristoylphosphatidylglycerol (PG), andcardiolipin. Low CR was detected upon addition of melittin to purePDA vesicles (no incorporated phospholipids). This residual CR,due to nonspecific interactions between the PDA vesicles and melit-tin, was treated as a “background” in the analysis of the assay results(see discussion below).

The titration curves (Fig. 2A) demonstrate the applicability ofthe assay using a variety of lipid components. It is important to notethat PE, PG, and cardiolipin, rather than phosphocholine deriva-tives, are the primary constituents of bacterial membranes13. As thisfigure also indicates, the highest colorimetric sensitivity to melittinis found in PE–PDA vesicles—a result that is most likely due to thesmaller headgroup of PE, which facilitates easier interaction of thepeptide with the lipid interior.

The titration curves of melittin (Fig. 2A) reveal that the phos-pholipid–PDA vesicles could detect melittin even in micromolarconcentrations (e.g., a CR of <20% would be perceived as a distinct

A colorimetric assay forrapid screening ofantimicrobial peptidesSofiya Kolusheva, Laurent Boyer, and Raz Jelinek

Department of Chemistry, Ben Gurion University of the Negev, Beersheva84105, Israel.

Received 19 December 1999; accepted 4 January 2000

The increased resistance of various bacteria toward availableantibiotic drugs has initiated intensive research efforts into identi-fying new sources of antimicrobial substances. Short antibioticpeptides (10–30 residues) are prevalent in nature as part of theintrinsic defense mechanisms of most organisms and have beenproposed as a blueprint for the design of novel antimicrobialagents1. Antimicrobial peptides are generally believed to kill bacte-ria through membrane permeabilization and extensive pore-for-mation1. Assays providing rapid and easy evaluation of interactionsbetween antimicrobial membrane peptides and lipid bilayers couldsignificantly improve screening for substances with effectiveantibacterial properties, as well as contribute to the elucidation ofstructural and functional properties of antimicrobial peptides.Here we describe a colorimetric sensor in which particles composedof phospholipids and polymerized polydiacetylene (PDA) lipidswere shown to exhibit striking color changes upon interactions

TECHNICAL REPORTS

© 2000 Nature America Inc. • http://biotech.nature.com©

200

0 N

atu

re A

mer

ica

Inc.

• h

ttp

://b

iote

ch.n

atu

re.c

om

color change, even with the naked eye). Such concentrationsapproach values employed in conventional antibacterial assays. Itshould be noted that the use of dedicated, more sensitive UV-visiblemeasuring apparatus (e.g., an ELISA reader) would probably furtherimprove the sensitivity of the assay.

Highly positively charged biomolecules such as polylysines or his-tone14 induce blue-red color changes as well. However, the colorimet-ric transitions in these cases are due to nonspecific, electrostatic inter-actions of the molecules with the negatively charged PDA interface12.The color transitions induced by nonspecific binding could be differ-entiated from the chromatic responses corresponding to biologicalassociation of antibiotic membrane peptides with the phospholipids,through the use of pure PDA vesicles (i.e., not containing phospho-lipids) as controls. The analysis is shown in Figure 2B–D.

Figure 2B,C depicts the UV-visible spectra of histone and melit-tin in the presence of pure PDA vesicles, and that of PE–PDA vesi-cles, respectively. The UV-visible spectra of histone (Fig. 2B), forexample, do not register a significant difference between solutions ofpure PDA and PE–PDA. However, Figure 2C clearly shows thatmelittin induces a greater colorimetric response in PE–PDA vesicles,compared to pure PDA vesicles. The “effective,” biologically mean-

ingful colorimetric response, which is solely due to association of theexamined peptide with the phospholipid domains, can be easilyobtained through subtraction of the “background” CR, acquired inthe presence of pure PDA vesicle solutions, from the CR detected inphospholipid–PDA vesicles. This is shown schematically in Figure

2D. The effective CR (∆CR) of melittin is significantlyhigher than the corresponding value of histone,because the color changes induced with histone arealmost all due to nonspecific interactions.

Figure 3 depicts color changes induced by analogsof melittin, magainin, and alamethicin, in which thenative amino acid sequences have been modified at sin-gle positions. Identification of relationships betweencolor changes induced in the phospholipid–PDA sys-tem by peptide analogs, and the biological activities ofthese analogs, is a critical requirement for applicationof the assay in high-throughput screening of antibacte-rial peptide libraries, and its use as a physiological diag-nostic tool. Furthermore, comparisons between thecolors induced by native peptide sequences, andanalogs in which residues have been omitted or substi-tuted in positions having prominent structural andfunctional roles, could illuminate specific structuralparameters contributing to peptide–membrane inter-actions.

Indeed, in all three peptides examined, differences incolor changes were observed between the native pep-tides and analogs in which particular residues have beenmodified. Two melittin analogs, for example, which sig-nificantly reduced antibacterial activities exhibit com-pared to native melittin15,16, were found to induce amore pronounced red-orange colors in the vesicle solu-tion (Fig. 3A-3 and A-4, respectively). The two analogs,one in which Lys7 has been substituted with leucine(K7L-melittin), the other in which Trp19 has beenomitted (∆W19-melittin), are unable to fold into the

226 NATURE BIOTECHNOLOGY VOL 18 FEBRUARY 2000 http://biotech.nature.com

TECHNICAL REPORTS

Figure 3. Photograph of a portion of a 96-well plate depicting thecolors of DMPC–PDA solutions after additions of the following nativepeptides and peptide analogs. (A) Melittin analogs: 1, Control (nopeptide added to vesicle solution); 2, Native melittin; 3, K7L-melittin(Lys7 substituted with leucine); 4, ∆W19-melittin (tryptophan omittedat position 19); 5, K19E-melittin (Lys19 substituted with glutamicacid). (B) Magainin II analogs: 1, Control (no peptide added to vesiclesolution); 2, Native magainin II; 3, bK10E-magainin II (Lys10substituted with glutamic acid); 4, K10E,K11E-magainin II (Lys10 andLys11 substituted with glutamic acids); 5, F12W-magainin II (Phe12substituted with tryptophan). (C) Alamethicin analog: 1, Control(aqueous vesicle solution + 10 ml trifluoroethanol, TFE); 2, Nativealamethicin (dissolved in 10 ml TFE); 3, P12-alamethicin, dissolved in10 ml TFE (proline moved from position 14 to position 12 in thealamethicin sequence). All cells contained 100 µl solutions of 1 mMDMPC–PDA vesicles (4:6 mole ratio) and 2 mM Tris at pH 8.5. Peptideconcentrations were adjusted to 0.1 mM.

A

1 2 43 5

B

C

Figure 2. (A) Colorimetric response of vesiclesolutions titrated with melittin. The vesiclesexamined were as follows: A, Pure PDAvesicles (no phospholipids); B, PE–PG–PDA(3:1:6 mole ratio); C, PE–cardiolipin–PDA (3:1:6mole ratio); D, PC–PDA (4:6 mole ratio); E,PE–PDA (4:6 mole ratio). (B) UV-visible spectraof vesicle solutions, in the presence of histone(0.15 mM). The vesicles used were pure PDA(100% PDA) (broken line), and PE–PDA (4:6mole ratio) (solid line). (C) UV-visible spectra ofvesicle solutions, in the presence of histone(0.15 mM). The vesicles used were pure PDA(100% PDA) (broken line) and PE–PDA (4:6mole ratio) (solid line). (D) Graph depicting the“effective” colorimetric response (∆CR),defined as the difference between the CR ofthe PE–PDA vesicles and CR obtained in thepure PDA vesicle solution (see text).

B

C

D

A

Figure 1. Photograph of DMPC–PDA solutions at 27°C, afteraddition of the following peptides: (A) Control solution (no peptide);(B) Melittin; (C) Magainin II; (D) Alamethicin; and (E) PBAN. Each cellcontained 100 µl of 1 mM DMPC–PDA (4:6 mole ratio) and 2 mM Trissolution at pH 8.5. Total peptide concentrations were ∼ 0.1 mM.

A B C D E

© 2000 Nature America Inc. • http://biotech.nature.com©

200

0 N

atu

re A

mer

ica

Inc.

• h

ttp

://b

iote

ch.n

atu

re.c

om

NATURE BIOTECHNOLOGY VOL 18 FEBRUARY 2000 http://biotech.nature.com 227

active helix-loop-helix melittin conformation needed to completeinsertion and incorporation within the phospholipid domains15,16.Thus, these analogs most likely bind to the negatively charged PDAinterface through nonspecific electrostatic interactions involvingtheir lysine residues. Control experiments using 100% PDA vesicles(see discussion of Fig. 2B–D) confirmed the nonspecific nature of thebinding, indicating that the “effective” colorimetric responses ofK7L-melittin and ∆W19-melittin are negligible (S. Kolusheva, T.Shadal, & K. Jelinek, manuscript in preparation).

The pronounced color changes observed for K7L-melittin and∆W19-melittin stand in contrast to the color change induced by amelittin analog in which Lys19 has been substituted with glutamicacid (K19E-melittin). This analog induced a violet color almost iden-tical to that induced by native melittin (Fig. 3A-5). This result is con-sistent with biological data indicating that substitution of Lys19 inmelittin does not have a significant functional or structural effect16.

Differences between the color changes induced by native peptidesequences and their analogs have also been detected for magainin IIand alamethicin. As shown in Figure 3B-3 and B-4, no color changeswere induced by magainin II analogs in which either Lys10 or Lys11,or both, were replaced with negatively charged residues, such as glu-tamic acid. These results are consistent with the reduced antibacteri-al activities and lower membrane permeabilities of such analogscompared with native magainin II17,18. In contrast, the cell in posi-tion B-5 of Figure 3 shows that a magainin II analog in which Phe12was substituted with tryptophan does induce a color change, albeitmore moderate than native magainin II. Phe12 has a prominentfunctional role within the magainin II sequence18,19; however, substi-tution of this residue with tryptophan, which also contains an aro-matic group, had a relatively small effect upon the biological activityof the peptide19.

Similar colorimetric results are shown in Figure 3C, which depictsalamethicin and an alamethicin analog in which proline was movedfrom position 14 to position 12 in the sequence (P12-alamethicin).Again, the absence of a color change by P12-alamethicin correlatedwith published data indicating a significantly reduced functionalityof this analog, due to the prominent structural role of Pro14 (ref. 20).

The color changes induced in the phospholipid–PDA solutionsoccurred within seconds after addition of the peptides. The speed ofthe assay, combined with the simplicity of detection, would offer animportant advantage over conventional antibacterial assays, whichin general take much longer to complete. The assay is robust and canbe easily expanded to include a variety of membrane models; vesiclescan also incorporate glycolipids and proteins, as well as other mem-brane components (S. Kolusheva, T. Shadal, & K. Jelinek et al., man-uscript in preparation). This colorimetric assay can be applied forrapid screening of antibacterial peptide activities, and could providestructural and functional information on peptide–membrane inter-actions and mechanisms of membrane permeability.

Experimental protocolMaterials. Synthetic phospholipids (DMPC, DMPE, DMPG) and cardiolipinwere purchased from Avanti Polar Lipids (Alabaster, AL). Tricosadiynoic acidwas purchased from GFS Chemicals (Powell, OH). Both lipids were used aspurchased. Synthetic melittin (GIGAVLKVLTTGLPALISWIKRKRQQ), andmagainin II (GIGKFLHSAKKFGKAFVGEIMNS) were purified to >95%purity using reverse-phase high-pressure liquid chromatography (HPLC).Alamethicin (AibPAibAAibAQAibVAibGLAibPVAibAibEQF-OH, where Aibdenotes α-amino isobutyric acid) was purchased from Sigma (St. Louis,MO). The PBAN peptide (LSEDMPATPADQEMYQPDPEEMESRTRYF-

SPRL) was generously provided by Prof. Chaim Gilon (Hebrew University,Jerusalem).

Peptide analogs were synthesized by Alpha Diagnostics Inc. (San Antonio,TX) using standard Fmoc chemistry. Identity of peptides was confirmed bymass spectrometry and amino acid analysis. Peptides were purified to >90%using reverse-phase HPLC. P12-alamethicin analog was synthesized by AlphaDiagnostics using phenylalaninol-2-chlorotrityl resin and Fmoc chemistry.

UV-visible measurements. Samples were prepared by adding peptides to0.2 ml vesicle solutions at concentrations of 1 mM vesicles, 2 mM Tris. Allexperiments were done with the solutions at pH 8.5. Following addition ofthe peptides, the solutions were diluted to 1 ml and the spectra were acquired.All measurements have been carried out at 27°C on a Hewlett-Packard 8452Adiode-array spectrophotometer, using a 1 cm optical path cell.

The colorimetric response (CR) is defined14:CR = (PB0 - PBI) / PB0

where PB = Ablue / (Ablue + Ared) and A is the absorbance at either the “blue”component in the UV-visible spectrum (∼ 640 nm) or the “red” component(∼ 500 nm). Recall that “blue” and “red” refer to the visual appearance of thematerial, not its actual absorbance. PB0 is the red/blue ratio of the controlsample (without peptides), and PBI is the value obtained for the vesicle–pep-tide solutions.

AcknowledgmentsR.J. is grateful to the Israel–US Bi-National Science Foundation for financialsupport of this work. We are grateful to Professor C. Gilon (Hebrew University,Jerusalem) for a generous gift of the PBAN peptide, and Dr. D. Charych (ChironInc., Emeryville, CA) is acknowledged for helpful discussions.

1. Hancock, R.E.W. Host defense cationic peptides, Drugs 57, 469–483 (1999).2. Okada, S.Y., Jelinek. R. & Charych, D.H., Induced color change of conjugated

polymeric vesicles by interfacial catalysis of phospholipase A 2. Angew. Chem.Int. Edn Engl. 38, 655–659 (1999).

3. Tanaka, H., Gomez, M.A., Tonelli, A.E. & Thakur, M. Thermochromic phase transi-tions of a polydiacetylene, poly(ETCD), studied by high resolution solid state C13NMR Macromolecules 22, 1208–1215 (1989).

4. Dempsey, C.E. The actions of melittin on membranes. Biochim. Biophys. Acta1031, 143–161 (1990).

5. Zasloff, M., Martin, B., Chen, H.C. Antimicrobial activity of synthetic magaininpeptides and several analogues. Proc. Natl. Acad. Sci. USA 85, 910–913 (1988).

6. Jen, W.C. et al. The antibacterial activity of alamethicin and zervamicins. J. Appl.Bacteriol. 63, 293–298 (1987).

7. Raina, A.K. et al. Identification of a neuropeptide hormone that regulates sexpheromone production in female moths.Science 244, 796–799 (1989).

8. Charych, D.H., Nagy, J.O., Spevak, W. & Bednarski, M.D. Direct colorimetricdetection of a receptor–ligand interaction by a polymerized bilayer assembly.Science 261, 585–588 (1993).

9. Cheng, Q. & Stevens, R.C. Charge-induced chromatic transition of amino-acidderivatized polydiacetylene liposomes. Langmuir 14, 1974–1976 (1998).

10. Ringsdorf, H., Schlarb, B. & Venzmer, J. Molecular architecture and function ofpolymeric oriented systems: models for the study of organization, surface recog-nition, and dynamics of biomembranes. Angew. Chem. Int. Edn. Engl. 27,113–158 (1988).

11. Jelinek, R., Okada, S.Y., Norvez, S. & Charych, D.H. Interfacial catalysis by phos-pholipases at conjugated lipid vesicles: colorimetric detection and NMR spec-troscopy. Chem. Biol. 5, 619–629 (1998).

12. Okada, S., Peng, S., Spevak, W. & Charych, D.H. Color and chromism of polydi-acetylene vesicles. Accounts Chem. Res. 31, 229–239 (1998).

13. Voet, D. & Voet, J.G. Biochemistry. (Wiley, New York, NY; 1990).14. Satake, K. et al. J. Biol. Chem. 235, 2801 (1960).15. Perez-Paya, E., Houghten, R.A. & Blondelle, S.E. The role of amphipathicity in the

folding, self-association and biological activity of multiple subunit small proteins.J. Biol. Chem. 270, 1048–1053 (1995).

16. Blondelle, S.E. & Houghten, R.A Hemolyic and antimicrobial activities of thetwenty-four individual omission analogues of melittin. Biochemistry 30,4671–4678 (1991).

17. Matsuzaki et al. Interactions of an antimicrobial peptide, magainin 2, with outerand inner membranes of Gram-negative bacteria. Biochem. Biophys. Acta 1327,119–130 (1997).

18. Matsuzaki et al. Modulation of magainin 2-lipid bilayer interactions by peptidecharge. Biochemistry 36, 2104–2111 (1997).

19. Matsuzaki et al. Orientational and aggregational states of magainin 2 in phospho-lipid bilayers. Biochemistry 33, 3343–3349 (1994).

20. Dathe et al. Proline at position 14 of alamethicin is essential for hemolytic activi-ty, catecholamine secretion from chromaffin cells and enhanced metabolic activ-ity in endothelial cells. Biochem. Biophys. Acta 1370, 175–183 (1998).

TECHNICAL REPORTS

© 2000 Nature America Inc. • http://biotech.nature.com©

200

0 N

atu

re A

mer

ica

Inc.

• h

ttp

://b

iote

ch.n

atu

re.c

om