Newly Synthesized Lipid–Porphyrin Conjugates: Evaluation of Their Self-Assembling Properties, Their Miscibility with Phospholipids and Their Photodynamic Activity In Vitro

Article abstract of DOI:10.1002/chem.201804865

Lipid–porphyrin conjugates are considered nowadays as promising building blocks for the conception of supramolecular structures with multifunctional properties, required for efficient cancer therapy by photodynamic therapy (PDT). The synthesis of two new lipid–porphyrin conjugates coupling pheophorbide-a (Pheo-a), a photosensitizer derived from chlorophyll-a, to either chemically modified lyso-phosphatidylcholine (PhLPC) or egg lyso-sphingomyelin (PhLSM) is reported. The impact of the lipid backbone of these conjugates on their self-assembling properties, as well as on their physicochemical properties, including interfacial behavior at the air/buffer interface, fluorescence and absorption properties, thermotropic behavior, and incorporation rate in the membrane of liposomes were studied. Finally, their photodynamic activity was evaluated on esophageal squamous cell carcinoma (ESCC) and normal esophageal squamous epithelium cell lines. The liposome-like vesicles resulting from self-assembly of the pure conjugates were unstable and turned into aggregates with undefined structure within few days. However, both lipid–porphyrin conjugates could be efficiently incorporated in lipid vesicles, with higher loading rates than unconjugated Pheo-a. Interestingly, phototoxicity tests of free and liposome-incorporated lipid–porphyrin conjugates demonstrated a better selectivity in vitro to esophageal squamous cell carcinoma relative to normal cells.

Full text of DOI:10.1002/chem.201804865

A Journal of
Accepted Article
Title: Newly synthesized lipid-porphyrin conjugates: evaluation of their
self assembling properties, their miscibility with phospholipids and
their photodynamic activity in vitro
Authors: Julien Massiot, Véronique Rosilio, Nada Ibrahim, Akihisa
Yamamoto, Valérie Nicolas, Oleg Konovalov, Motomu
Tanaka, and Ali Makky
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201804865
Link to VoR: http://dx.doi.org/10.1002/chem.201804865
Supported by

Chemistry - A European Journal
10.1002/chem.201804865
Newly synthesized lipid-porphyrin conjugates: evaluation of their
self assembling properties, their miscibility with phospholipids and
their photodynamic activity in vitro
a
a,
a
b
Julien Massiot , Véronique Rosilio Nada Ibrahim , Akihisa Yamamoto , Valérie
c
d
b,e
a
Nicolas , Oleg Konovalov , Motomu Tanaka , Ali Makky *
a
Institut Galien Paris Sud, Univ Paris-Sud, CNRS, Université Paris-Saclay, 92296 Châtenay-
Malabry, France
b
Center for Integrative Medicine and Physics, Institute for Advanced Study, Kyoto University,
6
c
06-8501 Kyoto, Japan
UMS IPSIT, Univ Paris-Sud, US 31 INSERM, UMS 3679 CNRS, 92290 Châtenay-Malabry,
France
d
European Synchrotron Radiation Facility (ESRF), Grenoble Cedex 9 38053, France
Physical Chemistry of Biosystems, Physical Chemistry Institute, University of Heidelberg,
e
6
*
9120 Heidelberg, Germany
Corresponding author: ali.makky@u-psud.fr
Abstract
Lipid-porphyrin conjugates are considered nowadays as promising building blocks for the
conception of supramolecular structures with multifunctional properties, required for efficient
cancer therapy by photodynamic therapy (PDT). Herein, we report on the synthesis of two new
lipid-porphyrin conjugates coupling pheophorbide-a (Pheo-a), a photosensitizer derived from
chlorophyll-a, to either chemically modified milk lyso-phosphatidylcholine (PhLPC) or egg lyso-
sphingomyelin (PhLSM). We investigated the impact of the lipid backbone of these conjugates
on their self-assembling properties, as well as on their physicochemical properties, including
interfacial behavior at the air/buffer interface, fluorescence and absorption properties,
thermotropic behavior and incorporation rate in the membrane of liposomes. Finally, their
photodynamic activity was evaluated on esophageal squamous cell carcinoma (ESCC) and
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
normal esophageal squamous epithelium cell lines. The liposome-like vesicles resulting from
self-assembling of the pure conjugates were unstable and turned to aggregates with undefined
structure within few days. However, both lipid-porphyrin conjugates could be efficiently
incorporated in lipid vesicles, with higher loading rates than unconjugated Pheo-a. Interestingly,
phototoxicity tests of free and liposome-incorporated lipid-porphyrin conjugates demonstrated a
better selectivity in vitro to esophageal squamous cell carcinoma relative to normal cells.
Introduction
Porphyrins are ubiquitous in geological and biological systems and are considered as the
[1]
most abundant pigments in nature . Depending on their chemical structure and on the nature of
the metal chelated in their ring, they can play diverse important biological functions such as
oxygen transport (heme), electron transport in cell respiration (cytochromes), photosynthesis
(
chlorophylls) and many other functions that are essentials for life [1]. Porphyrin pigments are
thus called the “colors of life,” since they are necessary to sustain key activities in nearly all
organisms [1]. Added to these properties, porphyrins have received considerable attention as
promising photosensitizers (PSs) for the treatment of small solid tumors by photodynamic
therapy (PDT). Indeed, this latter consists in the combination of a photosensitizer, oxygen and
harmless visible light at the appropriate wavelength to produce reactive oxygen species that can
oxidize several vital biomolecules in cancerous cells and subsequently lead to cell death.
Porfimer sodium (HpD, Photofrin ®) was the earliest porphyrin derivative employed in PDT [2]
and was the first PS approved by FDA in 1995 for early stage lung cancer treatment [3]. Despite
its high singlet oxygen quantum yield [4] and its efficiency in the treatment of different cancers,
porfimer sodium has several drawbacks including weak light absorption in the phototherapeutic
[
5]
window and long-term cutaneous phototoxicity . A large variety of photosensitizers have been
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
developed afterwards in order to minimize these drawbacks. Nevertheless, only few of them,
such as temoporfin (m-THPC, Foscan®) and the benzoporphyrin derivative monoacid ring A
(BPD-MA, verteporfin, Visudyne®) have been approved as PDT agents for the treatment of head
and neck cancer [6] and age-related macular degeneration (AMD) , respectively. The poor water
solubility of most of these porphyrin derivatives and their tendency to aggregate under
physiological conditions are key limitations to the achievement of an efficient photodynamic
activity. In fact, the hydrophobic nature of most photosensitizers makes their intravenous
administration a difficult task. Furthermore, the monomeric state of PSs is required to maintain
their photophysical, chemical and biological properties [8]. In addition to their water solubility
issue, many photosensitizers display poor tumor selectivity [9]. In order to overcome these
drawbacks, several strategies have been adopted during the last years, including porphyrin
glycosylation [10], pegylation [11] and their incorporation into nanocarriers such as organic [12] and
inorganic nanoparticles[12c, 13], and liposomes [7, 14]. Among these strategies, embedding PSs in
liposomal bilayers seems to be the strategy of choice for several reasons. Indeed, liposomes are
composed of biocompatible, biodegradable materials and can be easily produced at industrial
scale, due to their simplicity and to lower investment costs compared to other nanoparticulate
systems [15]. Added to these advantages, several studies have shown the efficiency of liposomes
in improving the solubility and selectivity of PSs [16]. Indeed, the selective accumulation of
liposomes in tumors is at least partially related to the leaky tumor vasculature which allows
liposomes to extravasate across the leaky tumor vessels [17]. Despite these improvements,
liposomal delivery systems showed low loading efficiency of PSs within their lipid bilayers and
rapid clearance of PSs from the blood due to their transfer to serum components[16b, 18]. Thus, the
focus on PS drug delivery has shifted recently towards the development of new nanocarriers
composed of only one building block with self-assembly properties, allowing high PS payload,
[7]
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
and facilitating their clinical translation as well as their production at industrial scale [9]. To our
best of knowledge, only one kind of such materials has been reported based on lipid-porphyrin
conjugates. These building blocks have been synthesized by Gang Zheng’s group [19] who
discovered the efficiency of such compounds to self-assemble into liposome-like nanoparticles
named “porphysomes” and possessing multifunctional properties, including photothermal therapy
(
PTT), photodynamic therapy (PDT), photo-triggered drug release and photoacoustic imaging
PAI) [19-20]. Thanks to their organic nature, porphysomes are enzymatically biodegradable and
(
induce minimal acute toxicity during their retention in mice [19b]. These lipid-porphyrin
conjugates were synthesized through esterification of the sn-2 position of
lysophosphatidylcholine with either pyropheophorbide a or bacteriochlorophyll a [19b]. In
addition, such conjugates may exhibit different localization into cell organelles compared to free
porphyrins compounds[21]. Considering these advantages, we believe there is still considerable
room for the development of new lipid-porphyrin conjugates, the study of their self-assembling
properties and PDT efficiency. Thus, the aim of this work was to synthesize a new kind of lipid-
porphyrin conjugates based on various lipid backbones linked to a PS via a peptidic bond instead
of the ester bond used in porphysomes [19b, 20]. To do so, two new lipid-porphyrin conjugates were
synthesized by coupling Pheo-a, a photosensitizer derived from chlorophyll-a, to either
chemically modified milk lyso-phosphatidylcholine (lyso-PC) or egg lyso-sphingomyelin (lyso-
eSM). The physicochemical properties of these compounds and their self-assembling properties
were assessed, as well as their efficiency in vitro on esophageal squamous cell carcinoma (ESCC)
cell lines.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Material and methods
Chemicals
Pheophorbide a (Pheo-a, ≥ 95%, mixture of diastereomers, Mw = 592.69 g/mol) was purchased
from Frontier Scientific (Logan, UT), and 6-(Fmoc-amino)hexanoic acid (≥ 98%, Mw = 353.42
g/mol) from Novabiochem (Laufelfingen, Switzerland). N,N′-Dicyclohexylcarbodiimide (DCC, ≥
9
9%, Mw = 206.33 g/mol), HATU (≥ 97%, Mw = 380.23 g/mol), 4-(Dimethylamino)pyridine
DMAP, ≥ 99%, Mw = 122.17 g/mol), Dowex® 50WX8-100 ion exchange resin (hydrogen
form), N,N-Diisopropylethylamine (DIPEA, 99%, Mw = 129.24 g/mol), HEPES (99.5%, Mw =
38.31 g/mol), sodium chloride (NaCl, 99%, Mw = 58.44 g/mol), Ammonium molybdate(VI)
(
2
tetrahydrate (81-83%, Mw = 1235.86 g/mol), L-Ascorbic acid (99%, Mw = 176.12 g/mol), 0.65
mM Phosphorus standard solution, hydrogen peroxide (30 wt %), chloroform anhydrous (≥99%,
stabilized with amylenes) and methanolic hydrogen chloride (0.5N) were provided by Sigma (St.
Louis, MO, USA).
The phospholipids 1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine (16:0 Lyso PC, 99%,
4
95.63 g/mol), 1-stearoyl-2- oleoyl-sn-glycero-3-phosphocholine (SOPC, 99%, Mw = 788.14
g/mol), 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC, 99%, Mw = 790.15 g/mol), 1,2-
distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethyleneglycol)–2000]-
ammonium salt (DSPE-mPEG2000, 99%, Mw = 2805,497 g/mol) and egg sphingomyelin (Egg
SM, 99%, Mw = 710.965) were purchased from Avanti Polar Lipids (Alabaster, AL).
Chloroform, methanol and anhydrous N,N-dimethylformamide (DMF, 99.8% pure) were
analytical-grade reagents purchased from Carlo Erba (Val de Reuil, France). The ultrapure water
(γ = 72.2 mN/m at 22° C) used in all experiments was produced by a Millipore Milli-Q® Direct 8
water purification System, with a resistivity of 18.2 MΩ.cm.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Synthesis of compound 1
1
6:0 Lyso PC (150 mg, 0.3 mmol) and Fmoc-6-Ahx-OH (212 mg, 0.6 mmol) were mixed in 5 ml
of anhydrous chloroform and stirred until clear mixture was obtained. DMAP (110 mg, 0.9
mmol) and DCC (120 mg, 0.6 mmol) were added separately, in cold anhydrous chloroform.
Glass beads (2 mm, previously washed with ethanol, and dried under vacuum) were added and
the mixture was brought back to room temperature and sonicated for 8 hours. Temperature was
kept under 25 °C. Fmoc-6-Ahx-OH (50 mg, 0.15 mmol) was added after 2, 4 and 6 hours of
sonication. Once the 8 hours sonication were over, the mixture was stirred at room temperature
for an additional 12 hours. It was then incubated with DOWEX (Dowex® 50WX8 hydrogen
form) for 45 minutes to remove DMAP, filtered, and concentrated under vacuum until a white
precipitate appeared. The 1-2 ml mixture was then centrifuged for 5 min at 2000 g and the yellow
liquid crude mixture was purified by chromatography on silica gel (eluted with chloroform-
methanol-water 65:25:4, volume ratio; Rf = 0.35-0.4). Pure compound 1 (207 mg, white powder,
1
yield 83 %) was dried under vacuum, lyophilized overnight, and stored at -20 °C. NMR: H
NMR (CDCl
3
, 300 MHz) δ (ppm) 7.76 (d, 2H, J = 7.5 Hz), 7.61 (d, 2H, J = 7.1 Hz), 7.42 (m,
4
9
H), 5.46 (br s, 1H), 5.23 (m, 1H), 4.38-4.10 (br m, 7H), 3.96 (br t, 2H), 3.77 (br m, 2H), 3.32 (s,
13
H), 3.17 (m, 2H), 2.29 (m, 4H), 1.58 (m, 6H), 1.26 (br s, 26H), 0.89 (t, 3H, J = 6.5 Hz); C
, 75 MHz) δ (ppm) 167.59, 165.52, 144.04, 141.26, 127.64, 127.03, 125.14, 119.93,
0.72, 66.33, 59.37,54.37,47.29, 40.81, 34.08, 31.91, 29.71 (br), 29.36 29.18, 26.11, 24.89,
NMR (CDCl
3
7
2
+
+
2.68, 14.11. MS (ESI) for [C45
H
72
N O10P] ; calculated: 831.4925; observed: 831.4912.
2
Synthesis of compound 2 (PhLPC)
Compound 1 (133 mg, 0.16 mmol) was dissolved into 4 mL of DMF anhydrous, 2 mL of DIPEA
and was stirred for 5 hours at room temperature to complete the full Fmoc deprotection. Pheo-a
(95 mg, 0.16 mmol) and HATU (75 mg, 0.2 mmol) were combined in 4 ml of anhydrous DMF,
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
stirred for 1 hour at room temperature under Argon, in the dark, and then added to the
deprotected compound 1. The mixture was stirred in the dark at room temperature, under Argon,
for 24 hours. DMF was then removed under vacuum. The crude was resuspended in minimum
amount of chloroform and purified by chromatography on silica gel (eluted with chloroform-
methanol-water 65:25:4, volume ratio; Rf = 0.5). Compound 2 was obtained (120 mg, dark-green
1
powder, yield 63 %). NMR: H NMR (DMSO-d
6
, 400 MHz) δ (ppm) 9.34 (s, 1H), 8.93 (s, 1H),
8
4
3
3
0
1
1
7
.80 (s, 1H), 7.86 (m, 1H), 6.40 (s, 1H), 6.15 (m, 1H), 6.03 (bd, 1H, J = 11.5 Hz), 5.01 (br s, 1H),
.58 (d, 1H, J = 6.9 Hz), 4.19 (d, 2H, J = 9.3 Hz), 4.05 (br m, 3H), 3.86 (s, 3H), 3.72 (br m, 2H),
.50 (m, 5H), 3.27 (br s, 5H), 3.13 (s, 9H), 2.91 (br m, 2H), 2.80 (s, 3H), 2.11 (br m, 6H), 1.81 (d,
H, J = 6.9 Hz), 1.60 (m, 2H), 1.42 (m, 5H), 1.31 (m, 2H), 1.08 (m, 2H), 1.06-0.85 (br s, 26H),
13
.706 (t, 3H, J = 7.1 Hz), 0.17 (s, 1H); C NMR (DMSO-d , 100 MHz) δ (ppm) 189.18, 173.05,
6
72.52, 172.23, 171.36, 169.27, 161.90, 154.43, 150.00, 148.70, 144.57, 141.30, 137.05, 135.72,
35.10, 131.89, 128.62 (CH), 128.26, 122.80, 105.21, 104.20 (CH), 96.48 (CH), 93.70 (CH),
0.57 (CH), 65.47, 64.30, 62.60, 62.33, 58.30, 53.10 (3xCH
3
), 52.62 (CH ), 51.30 (CH), 49.41
3
(CH), 38.26, 33.22, 32.47, 31.12, 28.78, 28.51, 28.24, 25.71, 24.24, 24.10, 22.82 (CH
3
), 21.94,
1
8.31, 17.13 (CH
3
), 13.77 (CH
3
), 11.75 (CH
3
), 11.53 (CH
3
), 10.36 (CH
3
). MS (MALDI-TOF)+
+
for [C65
H
96
N
6
O
12P] ; calculated: 1183.48; observed: 1183.66.
Synthesis of compound 3
Egg SM (400 mg, 0.56 mmol) was dissolved in anhydrous methanolic hydrogen chloride (40 mL,
.5 M) in a sealed vessel, and stirred at 55 °C for 7 days. The crude mixture was then dried under
0
vacuum, resuspended in minimum amount of chloroform-methanol (9:1), and purified by
chromatography on silica gel (eluted with chloroform-methanol-water 65:25:4, volume ratio; Rf
1
=
0.1). Compound 3 was obtained (165 mg, white powder, yield 65 %). NMR: H NMR (MeOD,
3
00 MHz) δ (ppm) 5.96-5.86 (m, 1H), 5.51 (dd, 1H, J = 15.2, 6.6 Hz), 4.33 (m, 2H), 4.08 (br m,
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
3
,
H), 3.71 (br s, 2H), 3.40 (br s, 1H), 3.27 (s, 9H), 2.11 (m, 2H), 1.45 (br m, 2H), 1.30 (br s, 22H)
1
3
0.91 (t, 3H, J = 6.6 Hz); C NMR (MeOD, 75 MHz) δ (ppm) 137.29 (CH), 128.25 (CH), 70.66
), 33.45, 33.07, 30.80, 30.66, 30.46, 30.17, 23.74, 14.46
P] ; calculated: 465.3457; observed: 465.3456).
(CH), 67.31, 56.92 (CH), 54.79 (3xCH
3
+
+
(CH
3
); MS (ESI) for [C23
H
50
N
2
O
5
Synthesis of compound 4 (PhLSM)
Pheo-a (160 mg, 0.27 mmol) and HATU (100 mg, 0.27 mmol) were mixed in 4 ml of anhydrous
DMF and stirred for 1 hour at room temperature under Argon, in the dark. Compound 3 (115 mg,
0
.25 mmol) was dissolved in anhydrous DMF with 0.5 ml of DIPEA, and then added to the Pheo
a - HATU mixture, and stirred in the dark, at room temperature and under Argon for 24 hours.
DMF was then removed under strong vacuum. The crude was resuspended in minimum amount
of chloroform and purified by chromatography on silica gel (eluted with chloroform-methanol-
ammonia 70:30:4, volume ratio; Rf = 0.3). Compound 4 was obtained (78 mg, white powder,
1
yield 30 %). NMR: H NMR (DMSO-d
6
, 400 MHz) δ (ppm) 9.51 (s, 1H), 9.11 (s, 1H), 8.82 (s,
1
1
3
1
1
1
6
1
1
H), 7.96 (m, 1H), 6.61 (s, 1H), 6.38 (s, 1H), 6.18 (d, 1H, J = 18 Hz), 6.06 (m, 1H), 5.42 (br m,
H), 5.26 (br m, 1H), 4.53 (br s, 1H), 4.05-4.01 (br m, 3H), 3.87 (br m, 1H), 3.83 (s, 3H), 3.57 (s,
H), 3.52 (s, 2H), 3.44-3.40 (br, m, 4H), 3.31 (s, 3H), 3.11 (s, 9H), 2.95 (s, 3H), 2.14 (br, m, 2H),
1
3
.75 (s, 3H), 1.59-1.61 (br, m, 3H), 1.50 (s, 3H), 1.20-0.50 (br, m, 27H); C NMR (DMSO-d ,
6
00 MHz) δ (ppm) 189.34, 173.24, 169.59, 155.52, 154.57, 145.02, 141.55, 137.23, 136.04,
35.36, 132.07, 132.01, 131.03, 130.72, 128.80, 128.41, 123.09, 105.11, 104.46, 96.57, 93.79,
9.90, 65.51, 58.42, 54.53, 53.21, 52.66, 51.42, 49.76, 31.17, 30.88, 28.69, 22.82, 21.98, 21.84,
+
+
8.38, 17.28, 13.87, 11.93, 11.66, 10.60; MS (MALDI-TOF) for [C58
H
84
N
6
O P] ; calculated:
9
039.60; observed: 1039.58.
Surface pressure measurements
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Surface pressure-molecular area isotherms (-A) of pure components or their mixtures with
DSPC were recorded using a thermostated KSV-Nima Langmuir film balance (Biolin Scientific,
2
Finland), composed of a teflon trough (775.75 cm ) equipped with two 145 mm Delrin barriers.
16
Pure components or mixtures in a chloroform/methanol (9:1) solution (4.0 x 10 molecules)
were spread onto the aqueous buffer solution (10 mM HEPES, 150 mM NaCl, pH = 7.4). After
deposition, the solvents were allowed to evaporate for 15 min before compression of the
monolayer at a rate of 5.0 Ų/molecule/min. All experiments were performed at 22.1 ± 0.7°C and
the results reported are mean values of at least three measurements. From the surface
pressure−area data, the surface compressional moduli K of monolayers were calculated, using Eq
(1) with A the molecular area and dπ, the surface pressure change:
푑휋
퐾 = −퐴 ( )
(Eq. 1)
푑ꢀ
푇
The excess free energy of mixing (∆GEXC) of Pheo-a derivatives and DSPC was calculated
according to Eq. (2):
휋
퐺퐸푥푐 = ∫ ꢁ퐴 − 푋 퐴 − 푋 퐴 ꢂꢃꢄ (Eq. 2)
∆
12
1
1
2 2
0
where A12, A
1
and A are the experimental molecular areas of the binary mixture and pure
2
compounds, respectively. X
1
and X are the molar fractions of the phospholipid and the
2
photosensitizer, respectively. ∆Gexc values were plotted as a function of the monolayer
−
1
composition, for surface pressures of 5, 10, 15, 20, 25 and 30 mN.m .
X-ray reflectivity experiments (XRR) at the air/buffer interface
XRR experiments were carried out at the beamline ID10B of the European Synchrotron
Radiation
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Facility (ESRF, Grenoble). The samples were irradiated with a monochromatic synchrotron beam
with an energy of 8 keV (λ = 1.55 Å ). The XRR experiments were performed on monolayers of
Pheo-a derivatives spread on the surface of HEPES buffer (HEPES 10 mM, KCl 150 mM, pH
7
.4) and compressed to a surface pressure of 30 mN/m. The film balance was kept in a He
atmosphere during the measurement to minimize the radiation damage. XRR was measured with
a linear detector (Vantec-1, Bruker AXS, USA). After subtraction of the diffuse intensity
background (at α
f
≠ α
), the specular reflectivity was analyzed using the Parratt formalism [22] with
i
a genetic minimization algorithm implemented in the MOTOFIT software package [23]
.
Differential scanning calorimetry (DSC)
DSC measurements were carried out using a DSC Diamond Perkin-Elmer apparatus. Four scans
of consecutive heating and cooling cycles between -10 °C and 15 °C were recorded to make sure
that the thermal equilibrium was reached. Different scan rates were recorded, 5 °C/min (for the
first two cycles), 2 °C/min and 1 °C/min. An empty pan was used as a reference. In addition,
before each scan, a 2 min isotherm was recorded at the initial temperature to ensure that the
samples were at thermal equilibrium. Multilamellar suspensions were prepared by hydration of a
film made of SOPC-photosensitizer (97.5-2.5 mol %) with 45 µL of HEPES buffer (hydration
rate of 90%). For each sample, a total mass of ~ 15 mg was placed in hermetically sealed
aluminum pans. Samples were prepared in triplicate to check the reproducibility. To monitor the
photooxidation of SOPC caused by the embedded photosensitizers, 50 µL of lamellar phase
suspensions (5 mg SOPC) were illuminated for 14 min before starting the thermal measurements.
Illumination was done with a homemade lamp composed of 4 Philips TL fluorescent tubes
covered by a flat diffusing glass plate and fitted with an orange filter (λ ~ 520–680 nm with a λmax
=
590 nm) at a fluence of 2 J/cm2,[24]. Calibration was carried out with pure cyclohexane (>
9
9.9% purity, 6.7 °C melting temperature) [25]. Data were collected and processed using Pyris
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
thermal analysis software (version 9.1). Phospholipid transition onset temperatures (Ton) were
determined from the intercept of the baseline with the tangent to the left side of the peak.
Preparation and characterization of liposomes and self-assembled structures
Liposomes incorporating Pheo-a derivatives were prepared by the thin lipid film hydration
method [26] followed by extrusion of the vesicles suspension. In brief, a mixture of DSPC (95
mol%), DSPE-mPEG2000 (2.5 mol%) and the studied photosensitizer (2.5 mol%) was prepared
in chloroform:methanol (9:1 v/v). After removing the organic solvent under vacuum at 45 °C, the
resulting film was rehydrated with 1 ml of DPBS to get a final 5 mM concentration of lipids. The
mixture was vortexed and sonicated at 60 °C for 5 min. The suspension was then extruded 19
times through a 200 nm pore-sized polycarbonate membrane, while maintaining the temperature
at 80 °C. The self-assembled structures were prepared following the same procedure after
hydration of the lipid-porphyrin conjugate dry film with HEPES buffer. The hydrodynamic
diameter and the zeta potential were measured by dynamic light scattering (DLS) (Nano ZS90,
Malvern). All measurements were carried out at 25 °C. The mean diameter of the vesicles was
1
80 ± 10 nm, and their zeta potential was slightly negative (Table S3). The PS content in the
liposome bilayers was evaluated by measuring the absorption of each liposomal sample, after
disruption by addition of a methanol/THF mixture. PS loading efficiency (%) was determined as
previously described [24]
.
Cryo-TEM
The self-assembled structures made of lipid-porphyrin conjugates were deposited on perforated
carbon-coated, copper grid (TedPella, Inc) which was immediately plunged into a liquid ethane
bath cooled with liquid nitrogen (180 °C) and then mounted on a cryo holder [27]. Cryo-
Transmission electron microscopy (TEM) measurements were then performed using a JEOL
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
2
200FS (JEOL USA, Inc., Peabody, MA, U.S.A.) working under an acceleration voltage of 200
kV (Institut Curie). Electron micrographs were recorded by a CCD camera (Gatan, Evry, France).
Cell culture
The immortalized esophageal squamous cell line HET-1A, used as a model for normal
esophageal squamous epithelium, was purchased from the American Type Culture Collection
(
Manassas, VA) [28]. The human esophageal squamous cell carcinoma [29] Kyse-30 was obtained
from Sigma Aldrich. Cells were grown in Roswell Park Memorial Institute (RPMI) 1640 medium
supplemented with 10% fetal bovine serum (FBS), 100 IU/ml penicillin and 100 μg/ml
streptomycin (GIBCO, Invitrogen) in a humidified incubator with 5% CO at 37 °C. Cells were
2
passaged every three days using 0.25% trypsin/ethylenediamine tetraacetic acid (EDTA) when
confluence was at 70 to 80%.
Cytotoxicity and phototoxicity studies
Cells were seeded into 96-well plates (4000 cells, 100 μL cell culture medium per well) and
incubated overnight in the humidified incubator. On the day of experiments, free porphyrin
derivatives (in DMSO) or DSPC-porphyrin derivatives liposomes (in PBS buffer) at different
concentrations were added to the wells in the dark. Each well contained a final volume of 200 µl
of full medium. The final porphyrin concentrations ranged from 0 to 5 µM. Cells were incubated
again for 24 hours to ensure full internalization of the porphyrin derivatives. The following day,
the culture medium was replaced with fresh one. Cells were then either incubated in dark for
cytotoxicity tests or illuminated for 14 min for phototoxicity assessment. Cells illumination was
carried out at the bottom of the culture plates with orange light in sterile conditions [24] and cells
were incubated again for additional 72 hours. The cell viability was then determined by the MTT
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
assay. Briefly, MTT was added to each well at the final concentration of 0.5 mg/mL in full
medium and incubated at 37 °C with 5% CO for 1h30. The medium was then removed, and the
2
blue formazan product formed was dissolved in 200 µL DMSO. After 5 min shaking, the optical
density (OD) at 570 nm of each well was measured using an ELISA plate reader (LT-5000 MS,
Labtech). For each plate, each concentration was analyzed in triplicate.
Analysis of PSs intracellular distribution with confocal laser scanning microscopy (CLSM)
5
Cells (2 x 10 cells) were deposited on 25 mm glass cover slips housed in 6-well plates and left to
grow for 24h with 5% CO at 37 °C. Cells were then incubated in full RPMI medium for 24h
2
with the appropriate treatment (free PSs in DMSO or embedded into liposomes) and then rinsed
with fresh full media. Prior to imaging, cells were incubated 15 min with 200 nM of
MitoTracker® Green FM (ThermoFisher scientific, Invitrogen) in DPBS at 37 °C with 5% CO
After washing twice with DPBS, cells were incubated for 10 min with 10 µg/mL WGA Alexa
Fluor 555 (ThermoFisher scientific, Invitrogen) in DPBS at 37°C with 5% CO . Cells were rinsed
2
.
2
twice with DPBS, and cover slides were transferred to the confocal microscope chamber,
supplemented with full culture medium. Samples were then imaged with an inverted Leica TCS
SP8 microscope gated-STED (Leica, Germany) using a HC PL APO CS2 63x/1.40 oil immersion
objective lens. The instrument was equipped with a 405 nm diode for porphyrin excitation, and a
WLL Laser (490 nm excitation wavelength for MitoTrackerGreen and 555 nm for AlexaFluor
5
55). Far red, green and red fluorescence emission were collected respectively with a 650-800
nm, a 505-550 nm and a 560-630 nm wide emission slits under a sequential mode.
The statistical co-localization analysis of the photosensitizers with mitochondria and cell
membrane was performed using ImageJ statistical plugin JACoP [30]. JACoP is a commonly used
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
tool for the calculation of colocalisation coefficient such as Manders’ Co-localization Coefficient
MCC)[30] 31. MCC was calculated for more than 20 cells.
(
Results and discussion
Synthesis of the lipid-porphyrin conjugates
Two lipid-porphyrin conjugates with different lipid backbones were synthesized (Scheme 1). One
backbone was based on sn-1-palmitoyl lysophosphatidylcholine. This lipid was modified by the
introduction of 6-(Fmoc-amino)hexanoic acid via direct acylation of the secondary alcohol
groups at sn-2 position using sonication in the presence of glass beads, where the reaction is
believed to take place. Such procedure aims to avoid intramolecular acyl migration as
demonstrated previously by Rosseto et al.[31] and Oneill et al.[32]. Afterwards, the amino group
was deprotected and followed by attachment of Pheo-a using HATU as coupling reagent to give
PhLPC (compound 2, yield 65%).
The second lipid backbone, which is based on a Lyso-eSM was prepared by acidic hydrolysis of
egg sphingomyelin (N-hexadecanoyl-D-erythro-sphingosylphosphorylcholine) in anhydrous
methanolic hydrogen chloride at 50 °C following the same procedure as Bitmman et al. [33]. The
mild acidic hydrolysis allowed the preparation of Lyso-eSM with low extent of C-3 epimerization
compared to the conventional hydrolysis methods. The Pheo-a was then coupled via peptidic
coupling using the same procedure as for PhLPC, to yield compound 4 (PhLSM, yield 30%).
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Scheme 1. Synthesis route for the porphyrin-lipid conjugates PhLPC (A) and PhLSM (B)
Characterization of the self-assembling and photophysical properties of the lipid-porphyrin
conjugates:
The ability of the synthesized compounds to self-assemble into organized structures similar to
those reported for porphysomes was assessed after hydration of films made of PhLPC or PhLSM.
The extruded suspensions were then analyzed by dynamic light scattering (DLS) and cryo-
electron microscopy (Cryo-TEM). Interestingly, these suspensions were monodisperse (PdI <
0
.2) and exhibited an average size of approximately 200 nm (Figure S5). Cryo-TEM micrographs
revealed that both lipid-porphyrin conjugates could self-assemble into liposome-like structures
with a dense bilayer of lipid-porphyrin conjugates surrounding an aqueous core (Fig 1C-D). The
thickness of the bilayers was approximately 4-5 nm for both compounds, similar to that of
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
ordinary phospholipid bilayers [34]. However, whereas self-assembled PhLPC showed spherical
shape, PhLSM ones exhibited ovoid shape with undulated bilayer. The impact of PhLPC and
PhLSM vesicles on their photophysical properties was studied by recording their absorption and
fluorescence spectra before and after their solubilization in HEPES buffer/methanol/THF (0.2 :
0
.8 : 1 mL) mixture. (Fig. 1 E-H)
Fig. 1 Schematic representation of (A) PhLPC and (B) PhLSM. Cryo-electron micrographs of self-assemblies made
of pure (C) PhLPC and (D) PhLSM in HEPES buffer. Absorbance and fluorescence spectra of PhLPC (E, G) and
PhLSM (F, H) vesicles, respectively before (solid line) and after (dashed line) their solubilization in
HEPES/MeOH/THF (0.2, 0.8, 1 mL) mixture. The insets in (G) and (H) correspond to the quenched fluorescence
spectra of PhLPC and PhLSM in buffer respectively.
As shown in Figure 1 and table 1, lipid-porphyrin conjugates exhibited similar absorption and
fluorescence spectra to that of Pheo-a in organic solvents (Figure S6). This result indicates that
linking Pheo-a to the lipid backbones did not induce any change in the photophysical properties
of the PS when the PS-lipid conjugates were in their monomeric state. Conversely,
nanoassemblies of both compounds showed several interesting features. First, absorption spectra
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
of both lipid-porphyrin vesicles revealed a broadening of porphyrin Soret and Qmax-bands with a
significant red shift of approximately 12 nm for the latter. Compared to pure Pheo-a aggregates in
buffer (Fig. S6), the Soret and Qmax-band of the lipid-porphyrin vesicles were sharper, indicating
that they formed more organized aggregates within their vesicular structure. The extent of the
intermolecular interaction between lipid-porphyrin conjugates within the dense bilayers of the
lipid-porphyrin vesicles were further analyzed by investigating their fluorescence quenching. As
shown in the insets to figures 1G and 1H, the fluorescence spectra of the vesicles were
extensively quenched, compared to the corresponding monomers. Interestingly, the full
fluorescence intensity of lipid-porphyrin conjugates could be efficiently restored with
approximately 1000-fold intensity increase, after solubilization of the lipid-porphyrin conjugates
in organic solvent. Similar behavior has been described by Lovell et al.[19b] for nanoassemblies
made of pyro-lipids (pyropheophorbide-a linked to Lyso-PC via ester bond) that could be used as
efficient photothermal and photoacoustic agents for tumor thermal ablation and photoacoustic
imaging [19b]. Taken together, the strong quenching of fluorescence emission and the red shifted
absorption of the studied compounds are proof of the strong intermolecular interactions existing
between chromophores within the lipid-porphyrin vesicles. However, despite these interesting
properties, these vesicles were not stable and formed larger aggregates of undefined structure
within few days (Figure S7). These results are in line with those obtained by Zheng’s group [19b,
2
0]
with pyro-lipids assemblies to which addition of DSPC, cholesterol and DSPE-PEG was
necessary to retain efficiently an encapsulated hydrophilic cargo, and to promote higher stability
of the vesicles [19b, 20, 35]. In fact, the bilayer instability of the lipid-porphyrin conjugates can be
due to the mismatch between the length of the alkyl chain in sn-1 position and the adjacent
porphyrin, leading to an inadequate packing parameter for bilayer stability. The packing
parameter, defined as P = v/al, where v is the hydrocarbon chain volume, a is the area of the
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
polar headgroup and l the length of the hydrocarbon chain [36], is useful to determine the
preferential organization of a surfactant at high concentration in a liquid medium. Phospholipids
form bilayers because their P value usually lies between 0.5 and 1 [36]. In the case of PhLSM, the
porphyrin is grafted in the vicinity of the polar headgroup thus increasing its polar group area,
and subsequently decreasing P to a value lower than 0.5 (conical or truncated conical shape).
Conversely for PhLPC, the porphyrin is conjugated to the hydrophobic chain in sn-2, which may
induce an increase in the hydrophobic volume leading to a P value higher than 1 (inverted conical
shape) [37]. It has also been shown that when lipid molecules with complementary shapes are
associated together, the value of P becomes additive resulting in the formation of intermediate
blocks that can form a stable bilayer [37]. So, since the two studied lipid-porphyrin conjugates
cannot form stable bilayers on their own, they could be mixed with unmodified phospholipids to
counterbalance the effect of the length mismatch between alkyl chains.
Monomers
Compounds
Pheo-a
411
PhLPC
410
PhLSM
406

ε
max (Soret) [nm]
-1
-1
9.9 x 104
667 (21)
3.4 x 104
8.8 x 104
667 (20)
4.1 x 104
8.9 x 104
667 (21)
3.9 x 104
soret [M .cm ]

max (Q) [nm]
-1
-1
ε
Q
[M .cm ]
Vesicles or aggregates

max (Soret) [nm]
386
392
404
-1
-1
3.1 x 104
674 (39)
0.7 x 104
5.3 x 104
679 (30)
2.5 x 104
3.8 x 104
676 (33)
1.6 x 104
ε

soret [M .cm ]
max (Q) [nm]
-1
-1
ε [M .cm ]
Q
Table 1: Soret band, Q-band and the corresponding absorption coefficient (ε) of monomeric (after vesicles
solubilization in HEPES buffer/methanol/THF (0.2 : 0.8 : 1 mL) mixture) and aggregated forms (in HEPES buffer)
of Pheo-a, PhLPC and PhLSM respectively. Values in brackets are the bandwidths (nm) at half height.
Interfacial behavior of lipid-porphyrin conjugates
To further investigate the effect of the organization of the lipid-porphyrin conjugates on their
self-assembling properties, we studied the interfacial behavior of the two compounds at the air-
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
buffer interface using a Langmuir trough. The -A isotherms for Pheo-a and the lipid-porphyrin
conjugates (PhLPC and PhLSM) spread at the air-buffer interface are shown in figure 2 A and the
main characteristics are summarized in table 2. The data show that the three compounds formed
stable monolayers and reveal interesting differences between them. The influence of the
attachment of Pheo-a to the lyso-eSM backbone or to the sn-2 aliphatic chain of lyso-PC is
readily apparent on the graph, showing significant differences in isotherm shape, surface pressure
and molecular area at collapse between compounds. The surface pressure corresponding to the
lateral pressure in membranes or phospholipid vesicles is close to 30 mN/m [38]. At this surface
2
pressure, the molecular area of Pheo-a (A30) was 53 Å . Considering the approximate dimensions
of Pheo-a as determined using the Visual Molecular Dynamics (VMD) software [39], the
experimental interfacial molecular area of the PS was much smaller than that expected for a
2
Pheo-a molecule lying flat on the surface (~ 156 Å ). Conversely, the experimental A30 value for
Pheo-a was in good agreement with the calculated value of the surface area of a Pheo-a molecule
2
with carboxylic and ester groups facing the air/water interface (~ 55 Å ). This arrangement would
result from the formation of closely packed films of Pheo-a molecules, controlled by strong
attractive π-π interactions between the rings of neighboring molecules.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Figure 2. (A) π-A isotherms and the corresponding (B) compressional modulus for pure Pheo-a, PhLPC, and PhLSM
spread at the air-buffer interface.
Both lipid-porphyrin conjugates formed more expanded monolayers with larger molecular areas
2
2
at the surface pressure onset (A ) for PhLPC (116 Å ) and PhLSM (94 Å ) compared to Pheo-a
0
2
(
80 Å ). Surprisingly, while the molecular area expansion for PhLSM was maintained even at
higher surface pressure, that of PhLPC at 30 mN/m was significantly reduced as compared to
pure Pheo-a. In addition, while Pheo-a and PhLSM showed similar collapse surface pressure and
similar -A isotherms profile, PhLPC exhibited a completely different isotherm profile, with
dissimilar shape, surface pressure and molecular area from those of the two other compounds.
Such behavior could be explained by different molecular arrangements of the studied molecules
at the air/buffer interface. Apparently, grafting Pheo-a to the lyso-SM backbone did not affect its
interfacial arrangement at the air/water interface and the area expansion of its isotherm could be
only explained by the presence of the PC headgroup in the vicinity of the chromophore.
However, the longer C6 carbon chain linker of PhLPC, bearing the Pheo-a at its extremity would
provide more flexibility to the attached chromophore to adopt distinct local orientation at
different surface pressure. Indeed, the onset of the surface pressure for the PhLPC is observed at
2
a larger molecular area ( ~116 Å ) than the other compounds due to the presence of the acyl chain
bearing Pheo-a at the interface. Upon further monolayer compression, the molecular area of
2
PhLPC decreased significantly to reach a collapse molecular area of ~50 Å which lies between
that of Pheo-a and PhLSM. However, it should be noticed that this molecular area is smaller than
2
that of monounsaturated phospholipids (A
c
~ 60-70 Å ). Such behavior could be explained by the
reorientation of the Pheo-a to align with the sn-1 C16 carbon chain with subsequent solubilization
of molecules into the subphase during the compression. To gain further insight into the structural
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
characteristics of the lipid-porphyrin conjugates, the compressibility moduli of their isotherms
were calculated and plotted as a function of surface pressure. As shown in Figure 2B and Table 2,
a similar compressional modulus range is revealed for Pheo-a and PhLSM with a Kmax value
approaching ~ 200 mN/m. However, PhLPC exhibits a much lower value (~ 80 mN/m).
According to Davies and Rideal [40], the values of PhLSM and Pheo-a would correspond to the
liquid condensed state (100 mN/m < Kmax < 250 mN/m), while that of PhLPC would indicate a
monolayer in liquid-expanded state (Kmax < 100 mN/m). Thus, PhLPC formed a less organized
monolayer than the other studied compounds.
Monolayer
composition
A
30 (Å2)
at 30 mN/m
A
0
(Å2)
A
c
(Å2)

c
(mN/m)
K
max (mN/m)
Pheo-a
PhLPC
PhLSM
80
116
94
53
60
72
41
50
64
50.2
41.5
51.5
198.5
82.6
205.4
Table 2: Molecular area at surface pressure onset (A
0
), molecular Area (A30) at surface pressure of 30 mN/m,
molecular Area (Ac), surface Pressure (π
c
) at collapse, and maximal compressional modulus Kmax for compounds
monolayers.
Analysis of the fine structures of Pheo-a derivatives monolayers
To get a better understanding of the fine structures perpendicular to the plane of Pheo-a
derivatives monolayers, the specular X-ray reflectivity (XRR) was measured on monolayers
compressed to a surface pressure of 30 mN/m. Figure 3A, shows the XRR curves of Pheo-a,
PhLPC and PhLSM monolayers spread on HEPES buffer, fitted using a two-slab model. The
corresponding electron density profiles (휌) reconstructed from the best fit results (solid red lines
in figure 3A) along the z-axis are also shown in figure 3B. The thickness (d), electron density (휌)
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
and root mean square roughness (휎) of each interface are summarized in table 3. Pheo-a exhibited
total thickness dpheo-a of 15.7 Å. The hydrophobic core had a thickness of 9.5 Å and an electron
-
−3
density of 0.436 e × Å . Since these values are consistent with those reported for other
porphyrin monolayers [41], it is plausible that Pheo-a molecules take an upright orientation with
respect to the interface (Figure 3C). In fact, this result agrees well with the area per molecule and
compression modulus determined from -A isotherms (Figure 2). The thickness and electron
density of the hydrophobic regions of the PhLPC monolayer are dHC(PHLPC) 11.6 Å and 휌HC(PhLPC)
=
0.373 e- × Å−3 , respectively. Interestingly, the corresponding values for the PhLSM
-
−3
monolayer are dHC(PHLSM) 9.4 Å and 휌HC(PhLSM) of 0.391 e × Å , respectively. The 휌HC values of
both compounds are higher than those reported for saturated [42] or monounsaturated [43] alkyl
chains of phospholipids. This could suggest the presence of porphyrin core within the alkyl
chains. Although the thickness of hydrophobic region of PhLPC is larger than that of PhLSM, it
is notable that the total thickness of PhLPC (dPhLPC = 21.5 Å) is 2 Å thicker than that of PhLSM
(dPhLSM = 19.5 Å). This could be explained in terms of the conformational difference of the
porphyrin. In the case of PhLPC, sn-1 C16 carbon chain and porphyrins are aligned (Figure 3C,
middle) while such an alignement is sterically prohibited in the case of PhLSM. In order to
validate the fitting quality, the average number of electrons per molecule was calculated from the
fit and compared to that calculated from the chemical formula of the studied compounds. As
shown in table 3, there is a good agreement between the calculated number of electrons and the
theoretical one for Pheo-a and PhLSM. Conversely for PhLPC, there is a significant difference
between the number of electrons calculated from the molecular formula and that from the fit.
This discrepancy could be interpreted by the error in PhLPC molecular area evaluation, due to its
solubilization into buffer subphase.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Figure 3: (A) XRR curves of a Pheo-a derivatives monolayers at a surface pressure of 30 mN/m. The solid lines
represent the best model fits to the experimental data. The experimental errors are within the symbol size. (B) The
reconstructed electron density profiles along the Z-axis. (C) Schematic representation of the orientation of pheo-a
derivatives at the air/buffer interface.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Average
Theoretical
number of
e /molecule
number of
-
−3
d (Å)
ρ (e × Å
)
σ (Å)
-
e /molecule
-
from the fit
Pheophorbide-a
0.436 ± 0.007
0.372 ± 0.010
0.335
Hydrophobic core
Hydrophilic groups
Buffer
9.5 ± 0.3
6.2 ± 0.5

4.1 ± 0.1
4.1 ± 0.5
4.5 ± 0.6
3
4
5
42
62
46
314
638
560
PhLPC
Hydrophobic chains
Choline headgroup
Buffer
11.6 ± 0.5
9.8 ± 0.5

0.373 ± 0.001
0.369 ± 0.001
0.335
4.4 ± 0.8
3.6 ± 0.5
3.1 ± 0.1
PhLSM
Hydrophobic chains
Choline headgroup
Buffer
9.4 ± 0.6
10.1 ± 0.6

0.391 ± 0.001
0.387 ± 0.001
0.335
4.3 ± 0.1
4.6 ± 0.8
3.1 ± 0.1
Table 3: Best fit parameters for the XRR Results for Pheo-a derivatives monolayers at 30mN/m as presented
in Figure 3.
Miscibility of Pheo-a derivatives with phospholipids
We then evaluated the behavior of porphyrin derivatives when mixed with lipids in order to
determine if these compounds could form stable systems when incorporated in a liposomal
bilayer. To do so, we chose DSPC as phospholipid, and mixed it with increasing molar
percentage of lipid-porphyrin derivatives. The recorded π-A isotherms of pure DSPC and its
mixture with Pheo-a, PhLPC and PhLSM at the air/buffer interface are shown in Fig. 4 A, B and
C, respectively. Pure DSPC formed a condensed phase as previously reported [44]
.
DSPC/Pheo-a mixtures did not behave in the same manner when the molar percentage of Pheo-a
increased. Indeed, the isotherms for mixtures containing up to 10 mol% Pheo-a were shifted
toward larger molecular areas than that of pure DSPC, but remained still in between the
isotherms of the pure components up to 30 mN/m. However, at higher surface pressures, the
isotherms of DSPC/Pheo-a exhibited slight shoulders at approximately 33 mN/m, 40 mN/m and
4
4 mN/m for monolayers containing 2.5%, 5% and 10% of Pheo-a respectively. This behavior
could be related to the solubilization of some Pheo-a in the aqueous phase. Interestingly, the
isotherms for the mixtures with a Pheo-a content higher than 10% were shifted toward larger
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
molecular areas than that of the pure Pheo-a indicating unfavorable interactions between the two
components. A similar behavior was observed for mixed DSPC/PhLSM monolayers. However,
unlike DSPC/Pheo-a mixtures, the isotherms for DSPC/PhLSM monolayers exhibited shoulders
at higher surface pressure (around 40 mN/m). Furthermore, the isotherms were all intercalated
between those of the pure components, even at 50 mol% PhLSM, which would account for a
better miscibility of DSPC with PhLSM than with the free Pheo-a.
The compression isotherms of DSPC-PhLPC mixtures exhibited sharp inflection followed by a
plateau region at ~ 43 mN/m. This sharp inflection is observed at a surface pressure close to that
of pure PhLPC collapse and lower than that of pure DSPC. This could be interpreted as a
demixing between the two compounds, with subsequent loss of PhLPC from the monolayer in the
aqueous subphase. Similar behavior has been observed by Kinnunen’s group who demonstrated
that addition of oxidized phospholipids to phospholipid monolayers induced phase separation and
their solubilisation in the subphase, with subsequent micelles formation due to the presence of
polar groups (either carbonyl or carboxylate) on the sn-2 chain [45]. In the case of PhLPC, the
driving force for the phase separation would be due to the strong attractive π-π interaction
between Pheo-a cores.
The isotherms of the mixtures were further analyzed in terms of compressibility modulus as
shown in Figures 4 D-F. In fact, for the three mixtures, adding a porphyrin or a porphyrin
derivative to a DSPC monolayer induced a decrease in their rigidity with a more significant
decrease in the presence of PhLPC, as expected from the low surface compressional modulus of
this conjugate. In addition, compared to DSPC/Pheo-a mixtures, the sharp inflection point
observed for the isotherms of PhLPC mixtures at high surface pressure (~ 40mN/m) is clearly
revealed by a minimum in K at ~ 44 mN/m, close to the  of pure PhLPC, followed by an
c
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
increase until the K values superimpose with those of pure DSPC. This is another indication of
the complete expulsion of PhLPC from the monolayers at high surface pressure. In order to
analyze quantitatively the thermodynamics of interaction between the binary mixtures, the excess
Exc
free energy of mixing (∆G ) was calculated up to a surface pressure of 30 mN/m. As inferred
from Figures 4G-I, the ∆GExc values of binary mixtures of DSPC-Pheo-a and DSPC-PhLPC at
different surface pressures are positive for the various PS molar fractions denoting repulsive
interactions between DSPC and PS compounds in the mixtures. However, the situation is
different for DSPC-PhLSM mixture. Indeed, this latter exhibited positive ∆GExc values up to 10%
PhLSM, which then decreased down to values close to 0 at 50%. In addition, it should be stressed
that ∆GExc values were the highest for DSPC-Pheo-a followed by DSPC-PhLPC and DSPC-
PhLSM, thus indicating that interactions in the mixed films containing lipid-porphyrin conjugates
were less repulsive than those in monolayers containing Pheo-a. Taken together, these results
indicate that the three Pheo-a derivatives cannot be homogeneously mixed with DSPC and tend
to segregate in the lipid monolayer.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
Figure 4. π-A isotherms of mixed monolayers of Pheo-a (A), PhLPC (B) and PhLSM (C) with DSPC at various
molar %. Their corresponding compressional modulus and excess free energy of mixing (∆Gexc) are shown in (D-F)
and (G-I) respectively.
The incorporation of lipid-porphyrin conjugates into liposomes
In a next step, we investigated the incorporation efficiency of Pheo-a derivatives into liposomes,
and evaluated their impact on their stability. DSPC liposomes doped with 2.5 mol% of DSPE-
PEG2000 and increasing molar percentages (2.5-20 mol %) of Pheo-a derivatives were prepared
and characterized. As shown in figure S8, the highest Pheo-a loading rate was 5 mol%, but the
liposomes increased in diameter and polydispersity. An important loss of material was also
observed on the polycarbonate membrane during extrusion (Figure S8). A higher loading
efficiency was achieved with PhLPC and PhLSM, with monodisperse vesicle suspensions and no
significant material loss during extrusion.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
We measured the fluorescence of DSPC liposomes incorporating the different PSs at 2.5 mol %.
As depicted in Figure S9, the three PSs showed fluorescence quenching. The highest values were
obtained for PhLSM and PhLPC. Such fluorescence quenching could be explained by the
aggregation of the PSs into organized patterns in the bilayer due to their high packing density.
Similar behavior has been observed by Gang Zheng’s group with other lipid-porphyrin
conjugates when incorporated within liposomal bilayers, and it was attributed to the formation of
J-aggregates [46]
.
Thermotropic behavior of phospholipid bilayers incorporating lipid-porphyrin conjugates
PS aggregation in the membrane of liposomes could be an issue for liposomal delivery of
photosensitizing agents for PDT applications. Indeed, the aggregation of PSs such as Pheo-a
induces a decrease in the quantum yield of triplet state ( ), thus reducing the quantum yield of
t
1
) [11, 47], which is responsible for the photodynamic activity. Therefore, in
singlet oxygen ( O
2
order to investigate the impact of the PSs incorporation percentage on their oxidative potential
upon illumination, we performed a calorimetric analysis of SOPC lamellar suspensions
incorporating different molar percentage of either PhLSM or PhLPC. SOPC was chosen in these
experiments because it contains an unsaturated alkyl chain that is a good substrate for the
[
24]
formation of lipid hydroperoxides
upon interaction with singlet oxygen during the
photodynamic reaction. The formation of such species can induce a phase separation within the
lipid matrix that can be easily detected by DSC. Although qualitative, such method allows the
determination of the optimal molar percentage of embedded PSs for efficient photodynamic
activity.
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201804865
First, the impact of PSs incorporation percentage on the phase behavior of the SOPC membrane
was assessed. Fig. 5 shows the heat capacity (C ) scans for the various lamellar suspensions.
p
Compared to pure SOPC which exhibited a sharp endothermic peak at ∼6 °C, both conjugates
caused a broadening of the main transition peak and a shift toward lower temperatures in a
concentration-dependent manner. This suggested the destabilization of the intermolecular
cooperativity of SOPC molecules. Such tendency was more pronounced for PhLSM, which could
be explained by its higher disordering effect on the hydrocarbon chains of phospholipids and/or
its higher incorporation efficiency than that of PhLPC.
DSC scans were also performed after illumination of the SOPC lamellar suspensions in the
conditions described in the experimental section. Compared to the pure SOPC sample, the
illumination of samples containing up to 2.5 mol % of either PhLPC or PhLSM induced dramatic
changes in the thermograms. The main transition peaks of both SOPC-PhLPC and SOPC-PhLSM
samples were broadened with a significant shift of Tonset towards lower temperatures and the
appearance of a second peak/shoulder at lower temperature. These results indicate the formation of
new phases upon illumination of the SOPC–PS systems, which could be related to the formation of
new chemical species within the lipid bilayer, as demonstrated in our previous work [24]. It should be
noticed, however, that the impact of the illumination was more pronounced in PhLPC samples than in
those of PhLSM. This could be explained by the longer spacer of PhLPC compared to PhLSM, which
would allow the Pheo-a moiety to be more deeply inserted into the lipid bilayer. Hence, the singlet
oxygen generated upon illumination would have greater probability to oxidize the unsaturated chains
within the lipid matrix.
This article is protected by copyright. All rights reserved.