1

1. Fluorescence of HNP-2

Figure S1. Fluorescence emission spectrum of HNP-2 in water at concentrations of 0.4 µM – 39 µM. λex = 280 nm.

2. Liposome Stability in the Presence of HNP-2

2.1 Membrane Fusion

Fluorescence resonance energy transfer (FRET) was used to determine whether HNP-2 was causing liposome fusion under the conditions employed for the other analyses. The FRET experiment was carried out using POPG/DOPC liposomes with 0.5 mol% of each of the donor N-(7-nitro-2,1,3-benzodiazol-4-yl)-phosphatidylethanolamine (NBD-PE) and the acceptor N-(Lissamine Rhodamine B sulfonyl)- phosphatidylethanolamine (Rh-PE). If liposomes composed of these two species are excited at around 460 nm (the excitation wavelength of NBD-PE), emission is observed from both species (λem NBD-PE = 540 nm, λem Rh-PE = 590 nm). If these liposomes are mixed with unlabelled liposomes and a species added which causes fusion or exchange between the labelled and unlabelled liposomes, an increase in the NBD-PE emission and decrease in the Rh-PE emission will be observed due to dilution.1

HNP-2 was added to a 3:1 mixture of unlabelled 50:50 (w/w) POPG/DOPC liposomes and NBD-PE/Rh-PE labelled POPG/DOPC liposomes to give the same peptide:lipid ratio as in the peptide binding experiments, and 60 fluorescence spectra recorded at 5 minute intervals (λex 450 nm, λem 470–650 nm) over 5 hours (Figure S2). A spectrum was also recorded of the same ratio of liposomes to which a 0.1% solution of Triton X-100 had been added to disrupt the membranes. A baseline

increasing [HNP-2]

2

spectrum of non-labelled liposomes in the absence of HNP-2 and triton X-100 was subtracted from all spectra.

Figure S2. Time series FRET spectrum of HNP-2 with unlabelled and labelled POPG/DOPC liposomes. HNP-2 (15 µM) was added to 3:1 (v/v) unlabelled 50:50 (w/w) POPG/DOPC liposomes and 50:50 (w/w) POPG/DOPC liposomes with 0.5 mol% of each of Rh-PE and NBD-PE (1.4 mM total lipid concentration) and fluorescence analyses recorded over 5 hours at 5 minute intervals. λex = 450 nm.

The ratio of NBD-PE fluorescence to Rh-PE fluorescence was calculated to see if it was changing over time using Equation S1:

!!!!! !!!!!! !

! (S1)

where (F1)t is the fluorescence of the HNP-2/liposome mixture at time t, (F1)0 is the fluorescence in the absence of HNP-2, (F2)t is the fluorescence of the HNP-2/liposome mixture following the addition of Triton X-100 and (F2)0 is the fluorescence following Triton X-100 addition in the absence of HNP-2. Fluorescence values were obtained as the ratio of the emission at 580 nm and 531 nm. The fluorescence change resulting from HNP-2 activity, (F1−F2)t, was calculated at each time point. As shown in Figure S3, no change was observed in the plot of the ratio of NBD-PE fluorescence to Rh- PE fluorescence over five hours after the addition of HNP-2 to the liposomes, confirming that there was no exchange of lipids between the labelled and unlabelled liposomes and therefore no fusion occurring.

0

1

2

3

4

5

6

7

8

9

10

470 490 510 530 550 570 590 610 630 650

Inte

nsity

Wavelength / nm

0min

50min

100min

150min

200min

250min

300min

3

Figure S3. Ratio of NBD-PE fluorescence to Rh-PE fluorescence against time, calculated using Equation 4.16 from time series fluorescence analysis of HNP-2 (15 µM) with 3:1 (v/v) unlabelled 50:50 (w/w) POPG/DOPC liposomes and 50:50 (w/w) POPG/DOPC liposomes with 0.5 mol% of each of Rh-PE and NBD- PE (1.4 mM total lipid concentration) recorded over 5 hours at 5 minute intervals. λex = 450 nm.

2.2 Dynamic light scattering (DLS)

DLS was used to determine of the average size of the liposomes before and after the addition of HNP-2, which would show the effects of any fusion, aggregation, micellization or increase in size due to the presence of the peptide. Liposome dispersions were analysed in the absence of HNP-2, diluted to the same concentration as those in the presence of HNP-2, and in the presence of HNP-2 after each 5 hour LD analysis. All DLS measurements were recorded at 21 °C regardless of the temperature of the LD analysis by averaging over 74 measurements to obtain the intensity weighted mean hydrodynamic diameter (z-average diameter). The polydispersity index (PdI) is a measure of the width of the distribution of sizes recorded. The PdI can be used to give the % polydispersity (%Pd) (Equation S2), which gives a clearer indication of the width of the distribution.

%Pd = 100 PdI (S2)

The data (Table S1) show that there was no real change in the diameter of the liposomes before and after the addition of HNP-2.

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 50 100 150 200 250 300

Rat

io (F

1-F 2

) t / (

F 1-F

2)0

Time / Minutes

4

Table S1. Diameter and polydispersity measurements of liposomes before and after the addition of HNP-2.

System Diameter / nm Average PdI %Pd

POPG/DOPC, Blank 119 0.067 26

POPG/DOPC, HNP-2, 30 °C 110 0.078 28

POPG/DOPC, HNP-2, 40 °C 113 0.071 27

POPG/DOPC, HNP-2, 53 °C 111 0.082 29

DMPG/DMPC, Blank 83 0.193 44

DMPG/DMPC, HNP-2, 17 °C 92 0.258 51

DMPG/DMPC, HNP-2, 33 °C 82 0.188 43

DOPS/DOPC, Blank 122 0.077 28

DOPS/DOPC, HNP-2, 30 °C 115 0.070 27

3. Phase Transition Temperature of DMPG/DMPC membranes

3.1 DLS

DLS experiments were carried out using a Zetasizer Nano ZS (Malvern Instruments Ltd., Malvern, UK) with a 3 mm pathlength cell. Liposomes were all prepared by extrusion through polycarbonate membranes with 100 nm diameter pores as described in the main article. Before the phase transition analyses were carried out, the temperature inside the cuvette was checked to ensure the accuracy of the temperature setting. For the determination of the phase transition temperature, 50 µL aliquots of liposome solutions (2 mg/ml, H2O) were heated from 10 °C to 40 °C or 60 °C at 0.5 °C intervals. The samples were equilibrated for 5 minutes before the start of the analyses, then for 1 minute at each temperature. The attenuator, sample position and number of repeat analyses (within a range of 15–17) were automatically optimised at each temperature. A fresh aliquot of liposome solution was used for each set of measurements. Diameter measurements before and after the addition of defensin were averaged from 71 individual measurements at 21 °C on the same instrument.

This followed a method developed by N. Michel et al. which uses changes in liposome size and shape caused by a lipid phase transition to determine the transition temperature.2 This method has the advantage of requiring small quantities of liposomes and can therefore be performed on dispersions that have been prepared for binding analyses. Figure S4 shows plots of the intensity weighted mean hydrodynamic diameter (z- average diameter) of liposomes composed of DMPG, DMPC and 50:50 (w/w) DMPG/DMPC lipids against temperature. The main transition of DMPG was observed between 20.8 °C and 23.1 °C. The second transition is attributed to the presence of sodium ions in the sample. The DMPG was supplied as a sodium salt and the presence of sodium ions is known to split the phase transition of lipids due to the different transition temperatures of the salt adducted and protonated species, with the intensity and position of secondary peaks relative to the transition of the protonated species dependant on the concentration of the sodium ions.3 The main transition (Tm) for DMPC was observed from the DLS analysis

5

between 22.2 °C and 24.0 °C. The liposomes composed of a 50:50 (w/w) mixture of DMPC and DMPG underwent a phase transition between 18.6 °C and 25.0 °C, again with a second transition due to the presence of sodium ions.

Figure S4. Intensity weighted mean hydrodynamic diameter (z-average diameter) measurements of liposomes composed of DMPG (blue), DMPC (red) and 50:50 (w/w) DMPG/DMPC recorded between 10 °C and 40 °C. All liposomes were extruded through polycarbonate membranes with pores of 100 nm diameter during preparation.

For comparison, liposomes composed of 50:50 (w/w) POPG/DOPC were also analysed by DLS between 10 °C and 60 °C to determine whether they underwent any size changes. It was not expected that any phase transitions would be observed, as POPG and DOPC have phase transition temperatures of -20 °C and -2 °C respectively.4 As shown in Figure S5, the POPG/DOPC liposomes underwent only a gradual increase in diameter of approximately 6 nm between 10 °C and 35 °C, unlike the DMPG, DMPC and DMPG/DMPC liposomes, which decreased in size by 6–15 nm over temperature ranges of 2 °C to 7 °C during phase transition. The slight increase in size of the DOPC liposomes is likely to have been caused by thermal fluctuations.

Figure S5. Intensity weighted mean hydrodynamic diameter (z-average diameter) measurements of liposomes composed of 50:50 (w/w) POPG/DOPC recorded between 10 °C and 40 °C.

80

85

90

95

100

105

110

115

120

10 15 20 25 30 35 40

Z-Av

erag

e D

iam

eter

/ nm

Temperature / °C

DMPG

DMPC

Mixture

109

110

111

112

113

114

115

116

117

10 15 20 25 30 35 40 45 50 55 60

Z-Av

erag

e D

iam

eter

/ nm

Temperature / °C

6

Table S2. Phase transition temperatures of liposome dispersions determined by DLS

Liposome Composition Main Transition / °C Secondary Transition / °C

DMPG 20.8–23.1 29.9–34.9

DMPC 22.2–24.0 -

DMPG/DMPC (50:50 w/w) 18.6–25.0 28.1–31.3

2.2 Differential Scanning Calorimetry (DSC)

DSC analysis of DMPG was carried out using a PerkinElmer Pyrus 1 calorimeter (PerkinElmer, Cambridge, UK). Solid DMPG (1.204 mg) was sealed in an aluminium pan with a drop of water and taken through three heating and cooling cycles between 0 and 60 °C, heating at 4 °C/minute.

Analysis of DMPG produced a thermogram with two transitions and a shoulder, due to the presence of sodium ions, the main transition being observed at 22.5 °C (Figure S6). The literature value for the main phase transitions of DMPG and DMPC in the absence of added salt is 23 °C.4 On the basis of these results, CD and LD analyses of the binding of HNP-2 to 50:50 (w/w) DMPG/DMPC liposomes were carried out at 17 °C and 33 °C, above and below the phase transition temperature the lipid mixture.

Figure S6. Differential scanning calorimetry heating scan of DMPG (1.204 mg) sodium salt in water. ΔH = 14.2 kJ mol-1.

On the basis of these results, CD and LD analyses of the binding of HNP-2 to 50:50 (w/w) DMPG/DMPC liposomes were carried out at 17 °C and 33 °C, above and below the phase transition temperature the lipid mixture.

33.9

34

34.1

34.2

34.3

34.4

34.5

16 18 20 22 24 26 28

Hea

t Flo

w E

ndo

Up

/ mW

Temperature / °C

19.9 °C

22.4 °C

7

4. Circular dichrosim (CD) time series

CD measurements were carried out on a Jasco J-810 (CD) spectropolarimeter and processed using the Spectra Analysis program (version 1.5, Jasco, Great Dunmow, UK). All experiments were carried out using a 1 cm quartz cell.

Figure S7. CD time series following HNP-2 addition to A: POPG/DOPC, 30 °C; B: POPG/DOPC, 40 °C; C: POPG/DOPC, 53 °C; D: DOPS/DOPC, 30 °C; E: DMPG/DMPC, 17 °C; and F: DMPG/DMPC, 33 °C. Panels G-L correspond to A-F respectively and show the ellipticity at 222 nm. In all cases, HNP-2 (15 µM) was added to 50:50 (w/w) liposomes (1.4 mM).

In all cases, changes in the CD spectra were insignificant over course of the experiments.

-9000

-7000

-5000

-3000

-1000

1000

3000

5000

210 230 250 270 290 310

[Θ] /

deg

cm

2dm

ol-1

103

Wavelength / nm

0 min

50 min

100 min

150 min

200 min

250 min

300 min

-1500

-1000

-500

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250 300 350

[Θ] /

deg

cm

2dm

ol-1

103

Time / Minutes

-10000

-8000

-6000

-4000

-2000

0

2000

4000

6000

210 230 250 270 290 310[Θ

] / d

eg c

m2

dmol

-110

3

Wavelength / nm

0 min

50 min

100 min

150 min

200 min

250 min

300 min

0

500

1000

1500

2000

2500

3000

3500

4000

0 50 100 150 200 250 300 350

[Θ]/

deg

cm

2dm

ol-1

103

Time / Minutes

-10000

-8000

-6000

-4000

-2000

0

2000

4000

6000

210 230 250 270 290 310

[Θ] /

deg

cm

2dm

ol-1

103

Wavelength / nm

0 min

50 min

100 min

150 min

200 min

250 min

300 min

-2000

-1000

0

1000

2000

3000

4000

0 50 100 150 200 250 300 350

[Θ] /

deg

cm

2dm

ol-1

103

Time / Minutes

-8000

-6000

-4000

-2000

0

2000

4000

210 230 250 270 290 310

[Θ] /

deg

cm

2dm

ol-1

103

Wavelength / nm

0 min

50 min

100 min

150 min

200 min

250 min

300 min

-2500

-2000

-1500

-1000

-500

0

500

1000

1500

2000

0 50 100 150 200 250 300 350

[Θ] /

deg

cm

2dm

ol-1

103

Time / Minutes

-10000

-8000

-6000

-4000

-2000

0

2000

4000

210 230 250 270 290 310

[Θ] /

deg

cm

2dm

ol-1

103

Wavelength / nm

0 min

50 min

100 min

150 min

200 min

250 min

300 min

-3000

-2000

-1000

0

1000

2000

3000

0 50 100 150 200 250 300 350

[Θ] /

deg

cm

2dm

ol-1

103

Time / Minutes

-5000

-4000

-3000

-2000

-1000

0

1000

2000

3000

4000

5000

210 230 250 270 290 310

[Θ] /

deg

cm

2dm

ol-1

103

Wavelength / nm

0 min

50 min

100 min

150 min

200 min

250 min

300 min

-500

0

500

1000

1500

2000

2500

3000

0 50 100 150 200 250 300 350

[Θ] /

deg

cm

2dm

ol-1

103

Time / Minutes

A B C

D E F

G H I

J K L

8

5. Semi-log plots for LD evolution

Fig. S8. Semi-log plots of the evolution of the LD signal over 5 h following the addition of HNP-2 (14 µM) to 50:50 (w/w) POPG/DOPC, DOPS/DOPC or DMPG/DMPC liposomes (1.4 mM). A) POPG/DOPC, 40 °C at 198 nm; B) POPG/DOPC, 40 °C at 230 nm; C) DMPG/DMPC, 17 °C at 215 nm; D) DMPG/DMPC, 30 °C at 194 nm.

-8.2

-8

-7.8

-7.6

-7.4

-7.2

-7

0 50 100 150 200 250 300

ln(L

D /

AU

)

time / min

-8.9

-8.8

-8.7

-8.6

-8.5

-8.4

-8.3

-8.2

0 50 100 150 200 250 300

ln(–

LD /

AU

)

time / min

-6.35

-6.3

-6.25

-6.2

-6.15

-6.1

-6.05

-6

0 50 100 150 200 250 300

ln(L

D /

AU

)

time / min

‐9

‐8.5

‐8

‐7.5

‐7

‐6.5

‐6

0 50 100 150 200 250 300

ln(L

D /

AU

)

time / min

A (POPG/DOPC, 40 °C, 198 nm)

B (POPG/DOPC, 40 °C, 230 nm)

C (DMPG/DMPC, 17 °C, 215 nm)

D (DMPG/DMPC, 33 °C, 194 nm)

9

6. Residuals of the least-squares fitting of LD data to the 3-state kinetic model

The plots of the residuals showed some noise, but were all linear or close to linear across the time period analysed, showing a good fit across the range of the data.

continued

A B

C D

E F

G H

10

Figure S9. Residuals of the least-squares fitting of LD data to the 3-state kinetic model described in Section 4.2.6.2. A–C) 50:50 (w/w) POPG/DOPC 30 °C at 200, 225 and 245 nm; D–E) 50:50 (w/w) POPG/DOPC 40 °C at 230 and 245 nm; F–G) 50:50 (w/w) POPG/DOPC 53 °C at 206 and 228 nm; H–J) 50:50 (w/w) DOPS/DOPC 30 °C at 194, 209 and 224 nm; K–M) 50:50 (w/w) DMPG/DMPC 17 °C at 196, 215 and 235 nm and N–P) 50:50 (w/w) DMPG/DMPC 33 °C at 188, 194 and 198 nm.

I J

K L

M N

O

P

11

7. Dichrocalc calculations LD spectra were calculated for the peptide dimer oriented as in Fig. 10 in the main paper and for structures rotated by increments of 9° around either the x- or y-axis relative to the reference state. In order to take account of free rotation of the peptide about the z-axis (equivalent to the membrane normal), and the expectation that peptides could be located at any position around the cylindrical surface of a shear-deformed liposome (which aligns along the x-axis), a series of spectra obtained at 30° rotations about the z-axis and 30° tilt angles about the x-axis were averaged for each starting structure. It should be noted that our axis assignments correspond to those of Dichrocalc and differ from those used in the standard derivations for shear-aligned liposomes, where the z-axis is the long axis of the liposome. Due to the dependence of LD on the square of the cosine of the angle between a transition moment µ and the z-axis, the data generated by rotation of each of δ or ψ by π/2 radians could be used to calculate the LD spectrum over a complete cycle of 2π radians (Fig. 11A in the main paper).

Fig. S10. A graphical depiction of the Dichrocalc calculations. From an arbitrary reference HNP-2 orientation, incremental changes were made in the orientation of the peptide about the x- or y-axis. For each change, the resulting peptide orientation was submitted to Dichrocalc with the z-axis as the rotation axis and the x-axis as the tilt axis.

y

z

x

x

z

rotation axis

tilt axis

ψ"

δ"

rotation axis

ψ"

y

z

tilt axis

rotation axis

δ"

tilt axis

12

8. Parameters obtained from analysis of Dichrocalc data Table S3. Parameters derived from fitting Equations 7 and 8 (main article) to the Dichrocalc data.

λ / nm αi,δ / ° !λ,δ! / eÅ !!,!! / e2Å2 αi,ψ

/ ° !λ,ψ! / eÅ !!,!! / e2Å2 aλ

187 -15 1.037 1.093 3 1.095 0.963 -0.063 188 -15 1.129 0.974 2 0.975 1.050 0.076 189 -15 1.207 0.845 0 0.845 1.125 0.214 190 -15 1.269 0.710 -2 0.711 1.185 0.344 191 -15 1.312 0.576 -5 0.580 1.228 0.461 192 -14 1.334 0.449 -9 0.461 1.250 0.558 193 -14 1.334 0.337 -16 0.365 1.252 0.628 194 -14 1.311 0.248 -25 0.301 1.233 0.665 195 -14 1.268 0.186 -35 0.276 1.193 0.666 196 -14 1.206 0.148 -44 0.283 1.135 0.636 197 -14 1.128 0.125 -50 0.303 1.060 0.524 198 -14 1.036 0.112 -50 0.303 1.060 0.524 199 -15 0.935 0.103 -56 0.335 0.875 0.457 200 -15 0.827 0.097 -58 0.337 0.771 0.388 201 -16 0.718 0.092 -58 0.329 0.665 0.321 202 -17 0.611 0.089 -58 0.312 0.559 0.257 203 -18 0.508 0.086 -57 0.288 0.458 0.197 204 -20 0.414 0.084 -55 0.261 0.363 0.144 205 -23 0.330 0.083 -53 0.231 0.278 0.098

9. Parameters obtained from fitting Dichrocalc data to the experimental data Table S4. Parameters derived from fitting Equations 9 (main article) to the experimental data.

System State δ / ° ψ / ° S

POPG/DOPC, 30 °C Pm1 172 174 100.8 Pm2 227 140 185.0

POPG/DOPC, 40 °C Pm1 163 177 114.7 Pm2 224 143 168.1

POPG/DOPC, 53 °C Pm1 45 25 2786.2 Pm2 224 207 869.1

DOPS/DOPC, 30 °C Pm1 163 187 142.0 Pm2 170 199 88.7

DMPG/DMPC, 33 °C Pm1 227 140 560.2 Pm2 178 123 33.8

13

Figure S11. Fits of the Dichrocalc data to the experimental data for A: POPG/DOPC, 30 °C; B: POPG/DOPC, 40 °C; C: POPG/DOPC, 53 °C; D: DOPS/DOPC, 30 °C; and E: DMPG/DMPC, 33 °C. The numeric suffix indicates the membrane-associated state: 1 for Pm1 and 2 for Pm2. The regression parameters from these fits are given in Table S4.

14

Figure S12. Simulated LD spectra generated using Dichrocalc for the HNP-2 dimer in orientations corresponding to each of the models in Fig. 9 in the main paper.

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

185 205 225 245 265 285

LD /

mol

-1dm

3cm

-1

wavelength / nm

Mulimeric PoreDimer PoreWedge

15

Figure S13. HNP-2 structures for the Pm1 state in POPG/DOPC membranes aligned according to the least fit parameters in Table S4. The alignments shown are averages from all the experiments in POPG/DOPC membranes (at 30 °C, 40 °C and 53 °C). Tyrosine residues are shown in magenta, arginine in blue, and tryptophan in orange. The plane of the membrane corresponds to the x-y plane, with the z-axis corresponding to the membrane normal.

16

Figure S14. HNP-2 structures for the Pm2 state in POPG/DOPC membranes aligned according to the least fit parameters in Table S4. The alignments shown are averages from all the experiments in POPG/DOPC membranes (at 30 °C, 40 °C and 53 °C). Tyrosine residues are shown in magenta, arginine in blue, and tryptophan in orange. The plane of the membrane corresponds to the x-y plane, with the z-axis corresponding to the membrane normal.

17

Figure S15. HNP-2 structures for the Pm1 state in DMPG/DMPC membranes aligned according to the least fit parameters in Table S4. Tyrosine residues are shown in magenta, arginine in blue, and tryptophan in orange. The plane of the membrane corresponds to the x-y plane, with the z-axis corresponding to the membrane normal.

18

Figure S16. HNP-2 structures for the Pm2 state in DMPG/DMPC membranes aligned according to the least fit parameters in Table S4. Tyrosine residues are shown in magenta, arginine in blue, and tryptophan in orange. The plane of the membrane corresponds to the x-y plane, with the z-axis corresponding to the membrane normal.

19

10. Surface potential of HNP-2

Figure S17. The electrostatic surface potential of HNP-2. Negative potential=red; positive potential=blue; contoured from −10 to 10 kT/e. A: the surface distal to the convex surface; B: the convex surface; C: face view; D: end view. This is the same as Fig. 12 in the main paper, without the surface transparency. References [1] V. Marchi-Artzner, T. Gulik-Krzywicki, M.-A. Guedeau-Boudeville, C.

Gosse, J.M. Sanderson, J.-C. Dedieu, J.M. Lehn, Selective Adhesion, Lipid Exchange and Membrane-Fusion Processes between Vesicles of Various Sizes Bearing Complementary Molecular Recognition Groups, ChemPhysChem, 2 (2001) 367–376.

[2] N. Michel, A.-S. Fabiano, A. Polidori, R. Jack, B. Pucci, Determination of phase transition temperatures of lipids by light scattering, Chem. Phys. Lipids, 139 (2006) 11–19.

[3] I.S. Salonen, K.K. Eklund, J.A. Virtanen, P.K. Kinnunen, Comparison of the effects of NaCl on the thermotropic behaviour of sn-1“ and sn-3” stereoisomers of 1,2-dimyristoyl-sn-glycero-3-phosphatidylglycerol, Biochim. Biophys. Acta, 982 (1989) 205–215.

[4] J.R. Silvius in Lipid Protein Interactions, ed. P. C. Jost and O. H. Griffth, John Wiley and Sons, New York, USA, 1982, Vol. 2, pp. 240–281.