Semisynthesis and in vitro photodynamic activity evaluations of halogenated and glycosylated derivatives of pheophorbide a

Article abstract of DOI:10.1002/ejoc.201500387

The present work focuses on the semisynthesis of halogenated and glycosylated derivatives of pheophorbide a (Pha). Because of the low reaction yields enocuntered en route to halogenated derivatives, we then focused only on the semisynthesis of glycosylated derivatives of Pha with the aim at enhancing the Pha specificity for cancer cells by introducing β-galactose moieties expected to bind gal-1. We applied LC-SPE-NMR/MS, to facilitate the direct identification of glycosylated derivatives. The transposition of these analytical methods to a preparative scale facilitated the isolation of glycosylated compounds in quantities sufficient to evaluate in vitro photodynamic efficacies. The in vitro growth inhibi-tory activity of semisynthesized compounds was then evaluated using the MTT colorimetric assay in the presence and absence of light. However, this pharmacological evaluation method seems to be unable to efficiently yield information about carbohydrate effects in relation to possible compound specificities for gal-1 overexpressed by B16F10 cancer cells.

Full text of DOI:10.1002/ejoc.201500387

FULL PAPER
DOI: 10.1002/ejoc.201500387
Semisynthesis and in Vitro Photodynamic Activity Evaluations of Halogenated
and Glycosylated Derivatives of Pheophorbide a
Ewa Cieckiewicz,*[a] Véronique Mathieu,[b] Luc Angenot,[a] Thierry Gras,[b]
Bieke Dejaegher,[c] Pascal de Tullio,[d] Bernard Pirotte,[d] and Michel Frédérich[a]
Keywords: Photosensitizers / Pheophorbide a / Halogenation / Glycosylation / Photodynamic activity / Photochemistry /
Cytotoxicity / Cancer
The present work focuses on the semisynthesis of halogen-
ated and glycosylated derivatives of pheophorbide a (Pha).
Because of the low reaction yields enocuntered en route to
halogenated derivatives, we then focused only on the semis-
ynthesis of glycosylated derivatives of Pha with the aim at
enhancing the Pha specificity for cancer cells by introducing
β-galactose moieties expected to bind gal-1. We applied LC-
SPE-NMR/MS, to facilitate the direct identification of glycos-
ylated derivatives. The transposition of these analytical
methods to a preparative scale facilitated the isolation of
glycosylated compounds in quantities sufficient to evaluate
in vitro photodynamic efficacies. The in vitro growth inhibi-
tory activity of semisynthesized compounds was then evalu-
ated using the MTT colorimetric assay in the presence and
absence of light. However, this pharmacological evaluation
method seems to be unable to efficiently yield information
about carbohydrate effects in relation to possible compound
specificities for gal-1 overexpressed by B16F10 cancer cells.
Natural pigment derivatives from chlorophylls and bac-
teriochlorophylls are widely used as starting compounds to
develop new efficient PSs.[2] Our previous bio-guided frac-
tionation of crude EtOAc extracts from Carpinus betulus
leaves led to the isolation of a photo-activatable derivative
of chlorophyll a, namely pheophorbide a (Pha).[3] Pha exhi-
bits good physicochemical properties, thus fulfilling PDT
requirements. However, like others, PSs currently used in
PDT, Pha suffers from some drawbacks.[4] The main feature
requiring improvement is that of cancer cell selectivity. To
improve selectivity for cancer cells, the PS may be conju-
gated to targeting agents, which are recognized by bio-
logical molecules overexpressed in cancer cells such as the
galectins. Galectins belong to a 15-member family of pro-
teins characterized by consensus amino acid sequences of
about 130 amino acids.[5] This consensus sequence corre-
sponds to a carbohydrate recognition domain that binds to
β-galactoside units. Galectins, and especially galectin-1 (gal-
1), are involved in mechanisms of tumor progression such
as cancer cell migration, proliferation, development of resis-
tance to apoptosis, establishment of metastatic potential
and immune escape.[6] Research has shown that gal-1 over-
expression in several human cancer cells directly correlates
to poor prognoses.[7] Another important problem of current
PSs is their long half-life.[8] This property of many PSs re-
quires PDT patients to stay in the dark as long as the PS is
present in their system. Consequently, vectorisation of PSs
with galactoside derivatives could represent an interesting
approach to improving cell selectivity. Moreover, in ad-
dition to participating in the process of cancer cell recogni-
tion, hydrophilic carbohydrate moieties may impart amphi-
Introduction
Surgery, chemotherapy and radiotherapy are current
mainstay treatments used in clinical oncology. In addition
to these cancer therapeutic modalities, photodynamic ther-
apy (PDT) is a new clinically approved antitumor treatment
alternative. PDT is a local treatment requiring the simulta-
neous presence of three key players: a photosensitizer (PS),
visible light with an appropriate wavelength (600–800 nm)
and molecular oxygen. The PS is activated by light of the
appropriate wavelength (corresponding to the PS maximal
absorption) leading to production of singlet oxygen and
other reactive oxygen species. These reactive oxygen species
induce photo-killing of cancer cells through direct effects
including apoptosis, necrosis or autophagy and through in-
direct effects leading to vascular damage events. These two
pathways lead to inflammatory responses, which can induce
systemic immune responses.[1]
[a] Laboratory of Pharmacognosy, University of Liège, Quartier de
l’Hôpital,
Avenue Hippocrate 15, B36, 4000 Liège, Belgium
E-mail: e.cieckiewicz@ulg.ac.be
http://www.cirm-ulg.be
[b] Laboratoire de Cancérologie et de Toxicologie Expérimentale,
Faculté de Pharmacie, Université Libre de Bruxelles,
Campus de la Plaine, 1050 Bruxelles, Belgium
[c] Laboratoire de Bioanalyse et Toxicologie Pharmaceutique,
Faculté de Pharmacie, Université Libre de Bruxelles,
Campus de la Plaine, 1050 Bruxelles, Belgium
[d] University of Liège, CIRM, Laboratory of Medicinal
Chemistry,
4000 Liège, Belgium
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201500387.
Eur. J. Org. Chem. 2015, 6061–6074
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6061

E. Cieckiewicz et al.
FULL PAPER
philic properties to Pha, which, by itself, is quite hydro- Semisynthesis of Halogenated Derivatives of
phobic. This putative increase of hydrophilicity leads us Pheophorbide a
also to envision faster clearance of a glycosylated Pha rela-
tive to nonglycosylated PSs, thus lifting current light restric-
Our synthetic approach (Figure 1), focused on haloge-
tions to those undergoing PS-related therapies.[8b] In the nating the 20-meso-position of the chlorin ring of Pha. We
past, many carbohydrate-conjugated PSs have been devel- considered halogenation of this position, because, com-
oped.[8,9] However, Pha has not yet been functionalized pared to the C-5- and C-10-meso-positions, C-20 was envi-
with carbohydrates possessing β-galactoside units (β-galact- sioned to be the most susceptible to electrophilic halogen
ose and β-lactose); these represent excellent lures for re- substitution.[11a] Halogenation at C-20 was realized with the
cognition and binding by gal-1 overexpressed by cancer hope of improving the efficiency of the PS by redshifting
cells.[10]
the absorption maxima observed in the visible spectra. This
Another limit of PDT involves poor light penetration effect was expected to improve the penetration of light nec-
through cancerous tissues.[1] The UV/Vis spectra of Pha is essary for PS activation (through cancerous tissues) by ef-
characterized by mean absorption peaks at 408 nm and fecting a change in activating wavelength.
666 nm. Notably, monochromatic light of 666 nm used to
Chlorination of the 20-position was based on, and
activate the Pha displays better tissue penetrating properties slightly adapted from, the protocol described by Wang and
than those used to activate other clinically approved PSs. co-workers[11c] for the synthesis of 20-chloropyropheophor-
However, longer wavelength light (Ͻ 800 nm) penetrates bide a. Pha, treated with NCS, gave desired compound 2.
deeper into tissues increasing PDT efficiency.[1] Accord- Nevertheless, because of the presence of many byproducts,
ingly, a bathochromic shift of absorption maxima could be the use of a reversed-phase Prep-HPLC purification step
expected to result from halogenation at the 20-meso-posi- followed by silica gel column chromatography were re-
tion of the chlorin ring of Pha.[11]
quired to generate 2 in sufficient purity. The structure of
In our study, we established a synthetic route to novel 20-chloro-Pha (Figure 2) was assessed principally on the
Pha derivatives with long wavelength absorbing properties basis of proton and carbon NMR spectra, which showed
in the visible spectrum as well as recognition/binding sites the disappearance of the initial CH-20 signal of Pha (Fig-
for gal-1. Our work took place in two stages. First, we con- ure 3). The remaining structure of 2 was similar to that of
sidered halogenation (with Cl, Br) at the 20-meso-position starting material Pha. The assignment of the chlorinated C-
of the chlorin ring in an attempt to improve the Pha activity 20 carbon atom was not directly achieved by carbon NMR
and to generate redshifted absorption maxima in the visible but was, instead, established by the HMBC signal at δ =
spectrum leading to enhanced light penetration through tis- 106.0 ppm superimposed with the peak representing C-15
sues.
(see Supporting Information). The HR-ESI-MS data con-
The second approach consisted of conjugating haloge- firmed the presence of one chlorine atom in 2 by exhibiting
nated Pha derivatives with carbohydrate-containing β-ga- a molecular peak at m/z = 627.23602 corresponding to
lactoside fragments (β-galactose and β-lactose) to generate C35H36ClN4O5 (calcd. 627.23742) and its associated isotope
gal-1-targeting halo-Pha analogs. This approach was envi- peak at m/z = 629.23169. The intensities of the molecular
sioned to exploit the Pha carboxylate as a readily modifi- and of the isotope peak were in the ratio of 3:1, respectively.
able handle.
The purity of 20-chloro-Pha was determined by HPLC and
MTT colorimetric assays were used to evaluate the in found to exceed 93%.
vitro photodynamic efficacy of the newly synthesized com-
Based on the successful results using NCS for chlorina-
pounds against several cancer cells including B16F10 mela- tion of the 20-meso-position of Pha, we attempted bromina-
noma cancer cells, which overexpress gal-1.
tion of Pha using NBS according to the procedure de-
scribed for the synthesis of 20-chloro-Pha. The desired com-
pound was successfully obtained but, once more, in small
quantities, thus complicating purification efforts. Indeed,
the isolation of 3 required several steps including a first
fractionation by Fractogel®, followed by dry column vac-
uum chromatography (DCVC) separation and subsequent
Prep-TLC under both normal and reversed-phase condi-
tions. The multiple purification steps required to generate
pure 3 greatly diminished the yield of 3. This explains why
Results and Discussion
Large-Scale Isolation of Pheophorbide a and Use as a
Starting Compound for the Semisyntheses
In order to realize all semisyntheses, the starting com- the photodynamic properties of 20-bromo-Pha have not yet
1
pound was primarily isolated from C. betulus leaves using been determined. The H NMR spectra of the compound
an in-house extraction method (for details see Exp. Sect.).[3] indicated, as they had for compound 2, the absence of the
However, Pha isolated in this manner only partially met our CH-20 signal based on comparisons to 1H and HSQC spec-
starting material needs. The difference between Pha quanti- tra of starting Pha. The small quantity of isolated 3 did not
ties needed and those initially provided by in-house meth- permit direct measurement of the carbon spectrum. In-
ods was made up by silica gel column purification [CH2Cl2/ stead, these data were elucidated on the basis of HMBC
MeOH (90:10)] of commercially available material.
and HSQC spectra acquisitions (see Supporting Infor-
6062
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 6061–6074

Halogenated and Glycosylated Derivatives of Pheophorbide a
Figure 1. Semisynthetic strategy used to obtain compounds 2–7. (i) NCS/N2/anhydrous CH2Cl2; (ii) NBS/N2/anhydrous CH2Cl2;
(iii) lac(OAc)7NH2, TEA, BOP; (iv) CH3ONa, anhydrous MeOH/CH2Cl2, HOAc; (v) gal(OAc)4NH2, TEA, BOP; (vi) CH3ONa, anhy-
drous MeOH/CH2Cl2, HOAc.
mation). Bromination at C-20 was confirmed by MS analy-
sis. The HR-ESI-MS of 3 was characterized by a molecular
ion peak at m/z
=
671.18616 corresponding to
C35H36BrN4O5 (calcd. 671.18691) and an isotope peak of
almost equal intensity at m/z = 679.1839, suggesting the
presence of one bromine atom in structure 3.
The UV/Vis spectra of 20-chloro-Pha (674 nm) and 20-
bromo-Pha (676 nm) (Figure 4) revealed the expected ba-
thochromic shifts in the red range of visible spectra com-
pared to the UV/Vis spectrum of Pha (666 nm). This benefi-
cial photophysical effect may be due to the electron-with-
drawing character of halogen atoms introduced at the con-
jugated C-20 atom. This small difference in absorption
maxima of both 20-halogenated Pha derivatives can be ex-
plained by the increased bulkiness characteristic of the
bromide moiety relative to that of the chloride; this likely
increases the strain of the chlorin π-rings in the bromide
Figure 2. Chemical structure of halogenated derivatives of Pha with
its assignments. X = Cl for 2; X = Br for 3.
Eur. J. Org. Chem. 2015, 6061–6074
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6063

E. Cieckiewicz et al.
FULL PAPER
1
Figure 3. H NMR spectra of 20-chloro-Pha. The arrow shows the disappearance of the 20-CH signal.
Figure 4. UV/Vis spectra of Pha (in black), 20-chloro-Pha (in blue) and 20-bromo-Pha (in red) showing bathochromic shifts of halogenated
compounds. Vis (MeOH) {λmax [nm] (log [ε/m–1 cm–1])}: Pha: 666 (3.74), 607 (3.08), 535 (3.10), 504 (3.16), 406 (4.13); 20-chloro-Pha: 674
(3.99), 617 (3.18), 547 (3.48), 516 (3.35), 410 (4.37); 20-bromo-Pha: 676 (3.99), 618 (3.15), 548 (3.45), 517 (3.32), 410 (4.29).
analog relative to its chlorinated congener. These redshifted only 630 nm providing light propagation (through tissues)
absorption maxima of 20-chloro-Pha and 20-bromo-Pha of less than 3 mm.
may translate into slightly improved light penetration (due
Due to the complexity involved in generating the desired
to the difference in required wavelengths of ca. 10 nm) rela- 20-halogenated Pha derivatives, we altered our initial chemi-
tive to starting Pha. Notably, such improved photophysical cal strategy to involve glycosylation of these derivatives; con-
properties are not negligible since the first clinically ap- sequently, we then to focused on semisynthesis of Pha deriva-
proved PS (Photofrin®) has an absorption maximum of tives bearing pendant carbohydrates for gal-1 targeting.
6064
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 6061–6074

Halogenated and Glycosylated Derivatives of Pheophorbide a
Semisynthesis of Glycosylated Derivatives of
Pheophorbide a
traditional isolation and purification procedures failed to
furnish pure products in quantities sufficient for complete
identification. Indeed, all semisynthetic trials were per-
Since our objective was to equip Pha with binding speci- formed on small amounts of Pha (10 mg) because of the
ficity for gal-1 overexpressed by cancer cells, we introduced, low quantities of Pha available. Consequently, crude resi-
at the C-17c position of Pha, different carbohydrate moie- dues for each reaction were analyzed by LC-SPE-NMR/MS
ties possessing a β-galactoside unit; installation of this unit to identify compounds of interest. Compounds represented
was expected to enable gal-1 binding. In this context, the by each major peak observed in each chromatogram were
β-galactose and the β-lactose moieties were modified to trapped on SPE cartridges and, following elution, were ana-
protect the hydroxy groups and to introduce a primary lyzed by 1H NMR (see Figure S1 of Supporting Infor-
amine at the hemiacetal position, thus enabling amide for- mation).
1
mation with the Pha C-17c carboxylate. The synthesis of 1-
The H NMR spectra of acetylated intermediates 4 and
β-amino-lactose heptaacetate (12) and 1-β-amino-galactose 6 displayed characteristic chemical shifts of Pha and ad-
tetraacetate (15) was achieved in four steps using a known ditional signals at δ ≈ 5 ppm attributable to carbohydrate
procedure[8a] (Figure 5).
protons. Because of the solvent suppression sequence, in the
The 1-β-amino peracetylated carbohydrates were conju- region of δ = 2 ppm, we could not observe the acetyl group
gated to the carboxylic acid group of Pha by means of methyl groups on relevant carbohydrates; these methyl moi-
amide formation in the presence of triethylamine (TEA) eties normally appear at δ ≈ 2 ppm. HR-ESI-MS analyses
and (benzotriazol-1-yloxy)tris(dimethylamino)phosphon- of acetylated precursors 4 and 6 revealed the expected mo-
ium hexafluorophosphate (BOP) as the coupling reagent to lecular formulas. 1H NMR spectroscopic data of final com-
give corresponding intermediates 4 and 6.[9a] These synthe- pounds 5 and 7 indicated that the Pha skeleton was slightly
ses were first monitored by TLC (CH2Cl2/MeOH, 19:1) to modified at the exocyclic ring V. The main alterations ob-
observe the disappearance of starting compound and then served by NMR spectroscopy included: (i) disappearance of
by hyphenated analytical methods including liquid the sharp singlet (at δ = 6.30 ppm) due to H-13b, and
chromatography (LC), solid-phase-extraction (SPE), cou- (ii) the appearance of a singlet (at δ = 3.81 ppm for com-
pled with NMR spectroscopy. Off-line HR-ESI-MS analy- pound 5 and at δ = 3.74 ppm for compound 7) attributable
ses enabled data set acquisitions required for preliminary to H-15c and of a multiplet (at δ = 5.33 ppm for compound
structural identifications. Hyphenated analytical methods 5 and at δ = 5.30 ppm for compound 7) due to H-15a.
allowed us to optimize the reaction conditions suggested by
The HR-ESI-MS analyses of deacetylated compounds 5
Pandey et al.[9a] The original combination of LC-SPE- and 7 confirmed the probable modifications of the Pha skel-
NMR/MS facilitated direct preliminary structure elucida- eton since recorded MS spectra failed to afford the expected
tions of quantity-limited semisynthetic products, whereas data.
Figure 5. Top: Synthesis of 1-β-amino-lactose heptaacetate (12) from β-lactose (8). Bottom: Synthesis of 1-β-amino-galactose tetraacetate
(15) from 1-α-bromo-galactose tetraacetate (13). (i) Ac2O, NaOAc; (ii) HBr/HOAc; (iii) NaN3, Bu4NHSO4; (iv) H2, Pd/C.
Eur. J. Org. Chem. 2015, 6061–6074
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6065

E. Cieckiewicz et al.
FULL PAPER
Fast NMR spectroscopic data analyses of semisynthetic
The structures of 4 and 6 isolated by Prep-HPLC were
compounds were sufficient to identify the main signals of confirmed by comparing their NMR spectroscopic data
the Pha skeleton and those of newly installed carbo- (1H, 13C, COSY, HSQC, HMBC data are presented in Sup-
hydrates. To elucidate the entire structures of the isolated porting Information) with those of the starting Pha (Fig-
compounds, extensive NMR studies were required.
ure 6). The β-orientation of each carbohydrate was deter-
To confirm the structures of acetylated precursors 4 and mined by the large coupling constant of H-1Ј (J Ͼ 9 Hz).[12]
6 and to elucidate the structural changes incorporated The detailed NMR assignments of peracetylated com-
within the Pha skeleton of final compounds 5 and 7, com- pounds 4 and 6 are compiled in Supporting Information.
pounds 4–7 were isolated on large scale by Prep-HPLC The structures of the peracetylated carbohydrate derivatives
using a method adapted from earlier analytical scale chro- of Pha were confirmed by HR-ESI-MS data. The MS
matograms.
analyses of precursors 4 and 6 showed the expected [M +
Figure 6. Top: Chemical structure of 4 along with its assignments. Bottom: Chemical structure of 6 along with its assignments.
6066 Eur. J. Org. Chem. 2015, 6061–6074
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Halogenated and Glycosylated Derivatives of Pheophorbide a
H]+ ion peaks at m/z = 1210.47192 and 922.38702, respec- show correlations between confidently assigned carbon
tively. Compound 4 was obtained in 30% yield and its atoms and corresponding protons of carbohydrates, thereby
HPLC purity estimated to exceed 90%. Compound 6 was suggesting linkage of a galactose or lactose moiety to 5 and
synthesized in 28% yield, and its purity, based on HPLC, 7, respectively. The β-orientation of each carbohydrate moi-
exceeded 92%.
ety is established by the large coupling constant of H-1Ј (J
Isolated intermediates 4 and 6, bearing hydroxy-pro- Ͼ 7 Hz).[12]
tected carbohydrates were deacetylated using sodium meth-
oxide to afford final compounds 5 and 7, respectively.
1H NMR spectroscopic data of these deacetylated com-
pounds were in accordance with our previous LC-SPE-
NMR observations. The absence of the singlet of H-13b
and the additional presence of a methylene doublet of dou-
blets (corresponding to H-15a at δ = 5.33 and 5.30 ppm for
compounds 5 and 7, respectively) and a sharp methyl sing-
let (corresponding to H-15c at δ = 3.81 and 3.74 ppm for
compounds 5 and 7, respectively) suggest the opening of
the exocyclic ring V in both cases. This hypothesis was cor-
roborated by interpretation of the 2D NMR spectral data
(13C, COSY, HSQC and HMBC) for compounds 5 and 7
as compiled in the Supporting Information.
13C NMR spectroscopic data of 5 and 7 indicated the
absence of the carbonyl ketonic group located at the 13a-
position of the exocyclic ring V normally present δ ≈
189 ppm. HMBC experiments with these final compounds
allowed identification of correlations between the ad-
ditional 15a-H and 15c-H proton signals, which were iden-
tified by HSQC and 13C NMR spectra as methylene and
methyl groups, respectively. HMBC spectra of both final
compounds indicated correlations of the methylene group
at C-15a with three quaternary carbon atoms correspond-
ing to C-15, C-14, C-16 and with one carbonyl moiety at
C-15b. These same spectra displayed clear correlations be-
tween the methyl group at C-15c and the carbonyl group at
C-15b suggesting the presence of an ester group at this posi-
tion. HMBC correlations between the methyl group at C-
13b and the carbonyl group at C-13a were also consistent
with the presence of an ester at the 13a-position. These as-
signments confirmed that the exocyclic ring V was modified
during deacetylation of peracetylated Pha carbohydrate de- Figure 8. Top: Chemical structure of final compound 5 with its
assignments and HMBC correlations of methyl groups at 12a, 13b,
rivatives using sodium methoxide. Opening of the exocyclic
15c and methylene group at 15a. Bottom: Chemical structure of
ring V, under these conditions, has been previously dis-
final compound 7 with its assignments and HMBC correlations of
methyl groups at 12a, 13b, 15c and methylene group at 15a.
cussed by Wang et al.,[11c] and the envisioned mechanism of
ring opening is depicted in Figure 7.
The remaining NMR assignments of compounds 5 and
The connection between the H-1Ј of the carbohydrate
7 are in agreement with their respective structures and are and Pha through an amide bond was demonstrated by a
detailed in Supporting Information (Figure 8). However, it weak COSY correlation of 7 as measured in CDCl3 with
is difficult to assign protons of the carbohydrate moieties, trace CD3OD. However, this connection was not revealed
because these signals are superimposed on Pha skeleton sig- in NMR spectra of 5, which were recorded in CD3OD with
nals. HSQC spectra of deacetylated compounds 5 and 7 trace CDCl3 (ensuring complete dissolution). These par-
Figure 7. Postulated mechanism of ring cleavage for exocyclic ring V.
Eur. J. Org. Chem. 2015, 6061–6074
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6067

E. Cieckiewicz et al.
FULL PAPER
ticular NMR conditions were justified by the amphiphilic that the bulkiness of the heptaacetylated lactose moiety
character of final compounds 5 and 7. could decrease the in vitro growth inhibitory activity of Pha
HR-ESI-MS analyses of final compounds 5 and 7 (Table 1) in the absence of light.
yielded the expected [M + H]+ ion peaks at m/z = 948.42424
The IC50 values of final compounds 5 and 7 were gen-
and 786.37142, respectively. Compound 5 was obtained in erally higher than those of Pha for all cancer cells under
40% yield, and its HPLC purity was estimated at Ͼ99%. study (Table 1). A possible exception to this involves the
Compound 7 was synthesized in 45% yield with an HPLC activity of 5 against Hs683 cells since the IC50 values for
purity evaluated at Ͼ98%.
both 1 and 5 against this cell line exceed 100 µm. These
Finally, the UV/Vis spectra of intermediate compounds observations suggest that compounds 5 and 7 display
(4 and 6) and final compounds (5 and 7) were very similar weaker in vitro growth inhibitory activity under dark condi-
to those of Pha, although small hypsochromic effects are tions than is noted with Pha. Decreases in the in vitro
evident.
growth inhibitory activity of 5 and 7 under dark conditions
inspire us to postulate a structure-activity relationship in
which decreased dark bioactivity may result from the bulki-
ness of the carboxylic acid group or the opening of exocy-
clic ring V. Theoretical increases of the polarity of final
compounds 5 and 7 may explain these data also, because
these compounds are not able to easily enter the cancer
cells.
Importantly, the PDT profiles of semisynthetic PSs 4–7
are superior to those of Pha if we remember that one of the
main beneficial characteristics of an “ideal PS” is low dark
toxicity. Conversely, these results suggest that the carboxylic
acid group and/or exocyclic ring V are important for light-
independent bioactivities of Pha and related congeners.
Assessments of in Vitro Photodynamic Activity
Pha and semisynthetic compounds 2, 4–7 were evaluated
for their in vitro anticancer activities in the presence and
absence of activating light against B16F10, Hs683, MCF-7
and U373 cancer cells using the MTT colorimetric assay.
B16F10[13] as well as Hs683,[14] U373[15] and MCF-7[16] can-
cer cells express gal-1 with variable rates. Notably, we envi-
sioned that gal-1 should bind the β-galactose moieties of
compounds 5 and 7. Based on these data, we discuss dark
and light-dependent in vitro growth inhibitory activities for
each semisynthetic compound; our in vitro data suggests
that each compound exerts a different mechanism of action
in the dark versus in the light.
Light-Dependent in Vitro Growth Inhibitory Activity
In Vitro Growth Inhibitory Activity in the Dark
To determine in vitro growth inhibitory photosensitizing
To evaluate in vitro growth inhibitory activity in the activity of all semisynthetic compounds, the cells were
dark, cells were treated with a range of PS concentrations treated with a range of PS concentrations (10–8 to 10–4 m),
(10–8 to 10–4 m), and untreated cancer cells were kept in the and untreated cancer cells were exposed to light with a con-
dark without any specific light exposure. Under these dark stant intensity of 0.7 J/cm2. Generally speaking, all semi-
conditions, 20-chloro-Pha (2) was found to express a phar- synthetic compounds displayed greater photocytotoxicity
macological profile similar to that noted for Pha. An excep- than the corresponding dark controls, as reflected by signif-
tion to this trend was noted when using the Hs683 cancer icantly lower IC50 values observed in Table 2 relative to
cell line (Table 1).
Concerning the peracetylated intermediates, heptaacetyl- study are characterized by photoactivation properties.
ated lactose conjugate 4 did not show any significant ac- Pha displayed the optimal light-induced in vitro growth
those seen in Table 1. In other words, all compounds under
tivity against Hs683, U373 and MCF-7 cancer cells at con- inhibitory effectiveness against all cancer cells under study,
centrations lower than 100 μm. This same compound compared to semisynthetic compounds. However, other de-
showed a moderate activity in the absence of light on rivatives showed good light-induced in vitro growth inhibi-
B16F10 melanoma cells (Table 1). Tetraacetylated galactose tory effectiveness against all cancer cells under study, albeit
conjugate 6 exhibited IC50 values on par with those found not as effective as Pha; the one exception to this is the light-
using Pha, except when considering the Hs683 cancer cell triggered activity of 2 with MCF-7 cells, which appears to
line. The IC50 values of intermediate compounds suggest be a slightly more potent PS than Pha. Notably, compounds
Table 1. Determination of in vitro growth inhibitory activities under dark conditions for Pha derivatives 2, 4–7. *Chemical structures are
compiled in Figure 1.
Compound*
IC50 in vitro growth inhibitory concentration [μm] under dark conditions
Mouse cancer cell line
B16F10
Human cancer cell lines
MCF-7
Hs683
U373
Pha (1)
8
Ͼ 100
31
Ͼ 100
46
Ͼ 100
59
23
25
Ͼ 100
22
47
46
33
29
Ͼ 100
43
81
43
20-chloro-Pha (2)
10
34
10
31
28
17c-aminoacylheptaaceto-lactose-Pha (4)
17c-aminoacyltetraaceto-galactose-Pha (6)
17c-amino-lactose-acycloPha (5)
17c-amino-galactose-acycloPha (7)
6068
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 6061–6074

Halogenated and Glycosylated Derivatives of Pheophorbide a
Table 2. Determination of light-dependent in vitro growth inhibitory activities for Pha derivatives 2, 4–7. *Compound structures are
detailed in Figure 1. n.d.: not determined.
Compound*
Light-dependent IC50 in vitro growth inhibitory concentration [μm]
Mouse cancer cell line
Human cancer cell lines
MCF-7
B16F10
Hs683
U373
Pha (1)
0.5
n.d.
n.d.
0.8
3.2
3.0
1.5
1.5
4.6
1.5
3.2
4.1
0.6
0.5
2.0
1.1
2.8
3.4
0.5
1
4.5
2.0
2.3
2.7
20-chloro-Pha (2)
17c-aminoacylheptaaceto-lactose-Pha (4)
17c-aminoacyltetraaceto-galactose-Pha (6)
17c-amino-lactose-acycloPha (5)
17c-amino-galactose-acycloPha (7)
2 and 4 displayed interesting dark/light ratios of activity in abling its use as a starting material for more advanced glyc-
association with the MCF-7 cancer cell line. osylated derivatives. LC-SPE-NMR/MS facilitated struc-
Also, in a fashion similar that was seen under dark con- ture elucidations of intermediate and final glycosylated Pha
ditions, chlorinated compound 2 did not show appreciable derivatives. In this study we described, for the first time, the
differences between its IC50 values and those of nonhaloge- semisynthesis of Pha carbohydrate derivatives characterized
nated Pha under light conditions (Table 2); chlorination of by opening of the exocyclic ring V during the last reaction
Pha appeared to have little, if any, effect on the photo-acti- step. The MTT colorimetric assays performed with or with-
vated activities of the parent compound. The in vitro out PS photo-activation revealed that Pha chlorination does
growth inhibitory photosensitizing activities of peracetyl- not have a significant influence on dark and light in vitro
ated intermediates 4 and 6 were similar, but slightly higher, activities relative to Pha. These in vitro assays indicate that
than for Pha (Table 2); this diminished phototoxicity may all semisynthetic compounds 2, 4–7 display photosensitiz-
be explained by modification of the Pha carboxylic acid ing efficacy. The photodynamic properties of glycosylated
group. However, unlike the results of dark condition reac- derivatives, especially of compound 4, are superior to Pha;
tions (Table 1), the precise carbohydrate structure tethered they display lower inhibitory activities in the absence of
to the Pha core structure appears to be not a determinant light but display interesting light-dependent in vitro growth
for the photosensitizing activity. Indeed, compounds 4 and inhibitory activities. Unfortunately however, our in vitro re-
6 displayed similar IC50 values, indicating that the carbo- sults do not highlight any interactions between the β-galact-
hydrate moiety imparts little influence on in vitro growth ose moiety of semisynthetic compounds and gal-1. Accord-
inhibitory photosensitizing activities.
ingly, we are now developing other pharmacological models
Final compounds 5 and 7 displayed slightly higher IC50 designed to unveil possible gal-1-dependent cancer cell spe-
values than those of starting Pha against all cancer cells cificities of semisynthetic compounds 5 and 7.
studied (Table 2). The less potent activities of 5 and 7 may
be related, as under dark conditions, either to: (i) the carbo-
hydrate linkage to the Pha carboxylate, (ii) exocyclic ring V
scission, or (iii) the increased polarity of these agents rela-
tive to Pha.
Experimental Section
Instrumentation: UV/Vis spectra were recorded with a Hitachi U-
2910 spectrophotometer in MeOH solution. NMR spectra were
measured with a Bruker® Avance II spectrophotometer equipped
Concerning the expected specificity of final compounds
5 and 7 for gal-1, under both dark and light conditions,
these in vitro assays did not allow us to delineate any inter-
actions between the β-galactose moiety of these carbo-
hydrate derivatives and gal-1 expressed by the cancer cells
under study. The lack of convincing results concerning the
specificity 5 and 7 for gal-1 under dark and light MTT con-
ditions invites us to consider further in vivo assays or in
vitro co-culture assays that specifically differentiated gal-
1-dependent effects. For instance, we envision the possible
development of assays using cancer cells overexpressing gal-
1 and corresponding cell lines with impaired gal-1 expres-
sion in order to evaluate the affinity of glycosylated com-
pounds 5 and 7 for gal-1.
1
with a cryoprobe operating at 500 MHz for H and at 125.7 MHz
for 13C, in CDCl3 or CD3OD solution (Euriso top®, Gyf-sur-
Yvette, France) with TMS as an internal reference. MS analyses
were performed using a Thermo Scientific LTQ Orbitrap XL mass
spectrometer (Thermo Fisher Scientific, Bremen, Germany) with
ESI source in positive mode. The semisyntheses were monitored by
thin-layer chromatography (TLC) plates on precoated silica gel F254
(Merck®, Germany) plates. After development, TLC plates of
carbohydrates were sprayed with naphthoresorcinol reagent (VWR)
(2 mL of concentrated sulfuric acid in 25 mL of ethanol containing
100 mg of naphthoresorcinol) and heated at 110 °C for 10 min. Pre-
parative TLC plates were performed on freshly prepared TLC
plates using silica gel 60 PF254. Plates were activated prior to use
by heating at 120 °C during 1 h. Prep-TLC in reversed phase were
performed on TLC silica gel 60 RP-18 F254s from Merck® (Ger-
many). Silica gel Lichroprep Si 60 (40–63 μm, Merck®) was used
for column chromatography. N-chlorosuccinimide (NCS), N-bro-
mosuccinimide (NBS), triethylamine (TEA), (benzotriazol-1-yl-
Conclusions
We semisynthetically generated 20-bromo-Pha and 20-
chloro-Pha. 20-Chloro-Pha was isolated in quantities suf-
oxy)tris(dimethylamino)phosphonium
(BOP), palladium on carbon (Pd/C), glacial acetic acid and β-lact-
hexafluorophosphate
ficient for evaluation of in vitro PDT efficacy, but not en- ose were from Sigma Aldrich® (Belgium). 1-α-Bromo-galactose tet-
Eur. J. Org. Chem. 2015, 6061–6074
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6069

E. Cieckiewicz et al.
FULL PAPER
raacetate standard was purchased from Apollo Scientific® (United
Kingdom). Pha used for semisynthesis was obtained from Wako®
(Osaka, Japan) (purity ranging from 73% to 91%) and Livchem®
7 min, 89% of A; 21 min, 100% of A; 30 min, 85% of A. The major
fragment containing the compound of interest was purified on sil-
ica gel column with an isocratic mode of CH2Cl2/MeOH (19.5:0.5).
(Frankfurt, Germany) (purity Ͼ 90%). All solvents were obtained 1H NMR (500 MHz, CDCl3): δ = 9.57 (s, 1 H, mesoH-10), 9.56 (s,
from Fisher Chemical® (USA). Solvents with analytical grade were
purchased from Merck® (Germany). Some reactions were realized
under anhydrous conditions: glassware was dried in an oven at
100 °C, solvents with analytical grade were dried on molecular si-
eves of 4 Å, and the reactions were realized under nitrogen.
1 H, mesoH-5), 7.93 (dd, J = 17.8, 11.3 Hz, 1 H, H-3a), 6.29 (dd,
J = 11.3, 1.6 Hz, 1 H, H-3bcis), 6.24 (s, 1 H, H-13b), 6.17 (dd, J =
17.8, 1.6 Hz, 1 H, H-3btrans), 4.78, (q, J = 7.1 Hz, 1 H, H-18), 4.17
(J = 9.0, 3.2 Hz, 1 H, dd, H-17), 3.86 (s, 3 H, H-13d-COOMe),
3.69 (m, 5 H, H-12a-Me, H-8a-CH2), 3.59 (s, 3 H, H-2a-Me), 3.25
(s, 3 H, H-7a-Me), 2.62–2.54 (m, H-17a), 2.48–2.39 (m, H-17b),
2.28–2.21 (m, H-17a), 2.17–2.08 (m, H-17b), 1.70 (t, J = 7.8 Hz, 3
H, H-8b-Me), 1.62 (d, J = 7.1 Hz, 3 H, H-18a-Me), –1.87 (s, 2 H,
NH) ppm. 13C NMR (125 MHz, CDCl3): δ = 189.4 (CO, C-13a),
175.6 (COOH, C-17c), 171.1 (C, C-19) 169.7 (COOCH3, C-13c),
154.1 (C, C-6), 152.1 (C, C-14), 148.8 (C, C-9), 145.0 (C, C-8),
139.9 (C, C-3), 139.7 (C, C-11), 138.1 (C, C-4), 137.2 (C, C-7),
134.0 (C, C-1), 132.0 (C, C-2), 130.4 (C, C-12), 130.0 (C, C-13),
129.6 (CH, C-3a), 124.8 (CH2, C-3b), 106.0 (C, C-15, C-20), 104.1
(CH, C-10), 99.5 (CH, C-5), 65.3 (CH, C-13b), 53.0 (COOCH3, C-
13d), 51.1 (CH, C-17), 50.0 (CH, C-18), 33.6 (CH2, C-17a), 30.0
(CH2, C-17b), 20.5 (CH3, C-18a), 19.5 (CH2, C-8a), 17.4 (CH3, C-
2a), 17.2 (CH3, C-8b), 12.2 (CH3, C-12a), 11.3 (CH3, C-7a) ppm.
HR-ESI-MS: [M + H]+ ion peak at m/z = 627.23602 for
C35H36O5N4Cl (calcd. 627.23742) and isotope peak at m/z =
629.23169. UV/Vis (MeOH): λmax (log [ε/m–1 cm–1]) = 674 (3.99),
617 (3.18), 547 (3.48), 516 (3.35), 410 (4.37) nm. Compound 2 was
synthesized in 57% yield, and its purity was evaluated by HPLC
to be Ͼ 93%.
Plant Material: The samples of C. betulus leaves were collected in
the spring period during several years (Table 3) from the same trees
in Sart-Tilman (Liège). All collected samples were identified by
Prof. M. Frederich. Voucher specimens were deposited in the Her-
barium of the Laboratory of Pharmacognosy, University of Liège.
The leaves were air-dried at room temperature in a ventilated oven
and then pounded using an electrical grinder.
General Procedure of Extraction and Purification of Pheophorbide a
from EtOAc Extract of C. betulus Leaves: Powdered leaves of C.
betulus were extracted with EtOAc as described previously.[3] We
applied two different methods to isolate Pha in large quantities
from crude EtOAc extract of C. betulus leaves. First, this extract
was separated by silica gel column chromatography eluting with
isocratic mode of CH2Cl2/MeOH (18:2) affording a mean fraction
containing Pha. Another way to obtain a fraction enriched with
Pha was the liquid/liquid extraction of this crude extract.[3] Briefly,
Na2CO3 solution at pH 11 was used to ionize acid compounds of
the crude EtOAc extract dissolved in CH2Cl2. Ionized compounds
were gathered and neutralized with glacial acetic acid. The acid
compounds were then extracted with CH2Cl2. The main fractions
from both methods were injected into a Prep-HPLC (Armen In-
strument AP 100 equipped with Büchi fraction collector C-660) on
LichroPrep RP-18 (25–40 μm, Merck). The compound of interest
was isolated using a binary solvent system with a flow rate of
30 mL/min: solvent A (ACN in 0.1% formic acid) and solvent B
(water in 0.1% formic acid) (0 min, 85% of A; 7 min, 89% of A;
21 min, 100% A; 30 min, 85% of A). Pha was eluted between the
13th and 14th minute corresponding to a black eluate. The starting
compound was then purified by silica gel column chromatography
using an isocratic system consisting of CH2Cl2/MeOH (18:2).
Semisynthesis of 20-Bromopheophorbide a (3): The title compound
was obtained by the same method performed to synthesize the 20-
chloro-Pha but using NBS as brominating reagent. The compound
was isolated in several steps. First, the mixture containing the com-
pound of interest was fractioned on Fractogel® resin, eluting with
EtOAc. The main fraction was then purified by dry column vac-
uum chromatography (DCVC). To achieve this technique of sepa-
ration, the fraction was dissolved in EtOAc and was pre-adsorbed
on Celite® 545 (Sigma Aldrich). The Celite® was dried under re-
duced pressure in a rotary evaporator and was placed on the top
of the column filled with silica gel 60 (15–40 μm, Merck). The col-
umn was eluted with increasing gradients of EtOAc and MeOH:
100% of EtOAc (50 mL), 98% of EtOAc (50 mL), 96% of EtOAc
Synthesis of 20-Halo Derivatives of Pheophorbide a
Semisynthesis of 20-Chloropheophorbide a (2): NCS (2 equiv.) was (50 mL), 94% of EtOAc (50 mL). The eluted fractions were moni-
added to a solution of Pha (8.43ϫ10–5 mol) in dry CH2Cl2.[11c]
The reaction occurred under anhydrous conditions and was moni-
tored by TLC using CH2Cl2/MeOH (19:1) as mobile phase. After
the disappearance of the starting compound, the mixture was
poured into water and extracted with CH2Cl2. The organic phase
was dried with MgSO4, and the solvents were evaporated to dryness
under vacuum. The residue was fractionated by Prep-HPLC (Ar-
men Instrument AP 100 equipped with Büchi fraction collector C-
660) on LichroPrep RP-18 (25–40 μm, Merck) in following gradi-
ent mode using mobile phase A (ACN in 0.1% formic acid) and
mobile phase B (water in 0.1% formic acid): 0 min, 85% of A;
tored by TLC, and the fractions with similar profiles were gathered
to main fractions. The fraction of interest eluted with EtOAc/
MeOH (19.2:0.8) was purified again by Prep-TLC using the same
solvent system. The compound isolated by normal-phase TLC was
analyzed by reversed TLC with ACN plus a few drops of formic
acid as mobile phase. These reversed-phase chromatographic condi-
tions indicated that the compound of interest was composed of 3
main components, which were isolated by RP-18 Prep-TLC under
these same conditions. 20-Bromo-Pha was characterized by a reten-
tion factor (Rf) of 0.3 under the last TLC conditions. 1H NMR
(500 MHz, CDCl3): δ = 9.61 (s, 1 H, mesoH-10), 9.59 (s, 1 H, me-
Table 3. List of collected leaf samples of C. betulus used for the isolation of Pha.
Samples of C. betulus leaves
Harvesting date
April 19, 2010
Harvesting place
GPS coordinates
Yield [%]
5.1
LP2010
Sart Tilman (Liège)
5°33_57.40_E
50°34_43.41_N
5°33_57.40_E
50°34_43.41_N
5°33_57.40_E
50°34_43.41_N
LP2011
LP2012
April 25, 2011
April 16, 2012
Sart Tilman (Liège)
Sart Tilman (Liège)
5.0
5.1
6070
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 6061–6074

Halogenated and Glycosylated Derivatives of Pheophorbide a
soH-5), 7.93 (dd, 1 H, J = 17.9, 11.2 Hz, H-3a), 6.31 (dd, J = n.d., with MgSO4. The crude product 10 was recrystallized from EtOH
1 H, H-3bcis) 6.25 (dd, J = n.d., 1 H, H-3btrans), 6.25 (s, 1 H, H-
in good yields (86–92%) and was used without further purification
13b), 4.87 (m, 1 H, J = 7.1 Hz; H-18), 4.15 (m, J = 8.8 Hz, 1 H, because of its poor stability. 1H NMR (500 MHz, CDCl3): δ = 6.53
H-17), 3.87 (s, 3 H, H-13d-COOMe), 3.69 (m, 5 H, H-12a-Me, H- (d, J1Ј,2Ј = 4.0 Hz, 1 H, 1Ј-H), 5.56 (t, J2Ј,3Ј = 9.9 Hz, 1 H, 3Ј-H),
8a-CH2), 3.62 (s, 3 H, H-2a-Me), 3.27 (s, 3 H, H-7a-Me), 2.63–2.40
(m, H-17a), 2.36–2.10 (m, H-17b), 1.71 (t, J = 7.0 Hz, 3 H, H-8b- = 7.8, J2ЈЈ,3ЈЈ = 10.4 Hz, 1 H, 2ЈЈ-H), 4.97 (dd, J2ЈЈ,3ЈЈ = 10.4, J3ЈЈ,4ЈЈ
Me), 1.60 (d, J = 7.1 Hz, 3 H, H-18a-Me), –1.71 (s, 2 H, NH) = 3.5 Hz, 1 H, 3ЈЈ-H), 4.77 (dd, J1Ј,2Ј = 4.0, J2Ј,3Ј = 9.9 Hz, 1 H,
ppm. 13C NMR (125 MHz, CDCl3): δ = 189.4 (CO, C-13a), 173.0 2Ј-H), 4.53–4.48 (m, 2 H, 1ЈЈ-H, 5Ј-H), 4.23–4.06 (m, 4 H, 6ЈЈ-H,
5.36 (dd, J3ЈЈ,4ЈЈ = 3.5, J4ЈЈ,5ЈЈ = 1.2 Hz, 1 H, 4ЈЈ-H), 5.13 (dd, J1ЈЈ,2ЈЈ
(COOH, C-17c), 172.2 (C, C-19) 169.8 (COOCH3, C-13c), 161.5
(C, C-16), 154.1 (C, C-6), 148.7 (C, C-14), 151.6 (C, C-9), 145.0 (C,
C-8), 140.6 (C, C-11), 140.0 (C, C-4), 137.3 (C, C-7), 136.7 (C, C-
1), 136.3 (C, C-3), 132.7 (C, C-2), 130.4 (C, C-12), 129.5 (CH, C-
6Ј-H), 3.91–3.84 (m, 2 H, 4Ј-H, 5ЈЈ-H), 2.16 (s, 3 H, COCH3), 2.14
(s, 3 H, COCH3), 2.10 (s, 3 H, COCH3), 2.07 (s, 3 H, COCH3),
2.07 (s, 3 H, COCH3), 2.06 (s, 3 H, COCH3), 1.97 (s, 3 H, COCH3)
ppm. HR-ESI-MS: [M + NH4]+ ion peak at m/z = 716.13933 corre-
3a), 129.0 (C, C-13), 125.0 (CH2, C-3b), 106.0 (C, C-15), 104.2 sponding to C26H36BrO17 (calcd. 716.14014).
(CH, C-10), 99.7 (CH, C-5), 65.6 (CH, C-13b), 52.7 (COOCH3, C-
Synthesis of 1-β-Azido-Peracetylated Carbohydrates 11, 14: 1-α-
13d), 51.0 (CH, C-17), 52.1 (CH, C-18), 29.2 (CH2, C-17a), 25.0
(CH2, C-17b), 20.5 (CH3, C-18a), 19.4 (CH2, C-8a), 18.1 (CH3, C-
2a), 17.4 (CH3, C-8b), 12.1 (CH3, C-12a), 11.2 (CH3, C-7a) ppm.
HR-ESI-MS: [M + H]+ ion peak at m/z = 671.18616 for
C35H36O5N4Br (calcd. 671.18691) and isotope peak at m/z =
679.1839. UV/Vis (MeOH): λmax (log[ε/m–1 cm–1]) = 676 (3.99), 618
(3.15), 548 (3.45), 517 (3.32), 410 (4.29) nm. Compound 3 was syn-
thesized in 10% yield.
Bromo-lactose heptaacetate (10) (1.43ϫ10–3 mol, 1 g) was dis-
solved in a biphasic solution composed of EtOAc and saturated
NaHCO3 solution. Sodium azide (5 equiv.) and tetrabutylammo-
nium hydrogensulfate (Bu4NHSO4, 1 equiv.) were added to the
mixture. The reaction mixture was stirred for 8 h and then stopped
by addition of CH2Cl2. The organic phase was washed twice with
water and then concentrated under reduced pressure after drying
with MgSO4. The title compounds 11 and 14 were purified by a
silica gel column chromatography eluting with hexane/EtOAc
(10:10). The same conditions were used to check the collected frac-
tions by TLC. Developed TLC plates were sprayed with naphtho-
resorcinol reagent. The isolated compound 11 was recrystallised
from EtOH in good yields (78–90%). Compound 14 was obtained
according to the same procedure also in good yields (74–86%). 11:
1H NMR (500 MHz, CDCl3): δ = 5.35 (d, J3ЈЈ,4ЈЈ = 3.5 Hz, 1 H,
4ЈЈ-H), 5.21 (t, J2Ј,3Ј = 9.2 Hz, 1 H, 3Ј-H), 5.10 (dd, J1ЈЈ,2ЈЈ = 7.8;
General Procedure Used to Synthesize 1-β-Amino-Peracetylated
Carbohydrates: 1-β-Amino-lactose heptaacetate (12) and 1-β-
amino-galactose tetraacetate (15) were obtained using the same
synthetic scheme adapted from the methods described by Zheng et
al.,[8a] but each step was slightly adapted. 1-β-Amino-lactose hep-
taacetate (12) was synthesized from β-lactose (8), and 1-β-amino-
galactose tetraacetate (15) was obtained starting from 1-α-bromo-
galactose tetraacetate (13) (Figure 5).
J2ЈЈ,3ЈЈ = 10.4 Hz, 1 H, 2ЈЈ-H), 4.96 (dd, J2ЈЈ,3ЈЈ = 10.4, J3ЈЈ,4ЈЈ
=
Synthesis of 1-β-Lactose Octaacetate (9): Lactose octaacetate (9)
was prepared using acetic anhydride and sodium acetate[17] starting
from β-lactose (8) (0.0146 mol, 5 g). To protect the hydroxy groups,
1 equiv. of sodium acetate was used to acetylate one hydroxy func-
tion. The amounts of acetic anhydride corresponded to a quarter
of the mass of acetic anhydride required for the acetylation of all
hydroxy groups. The reaction mixture was heated at 100 °C for 5 h
and was then poured into ice/water. The aqueous phase was ex-
tracted twice with CH2Cl2. The combined organic layers were dried
with MgSO4 and then concentrated in vacuo. Compound 9 was
recrystallized from EtOH in almost quantitative yield (93%) and
was used without further purification. 1H NMR (500 MHz,
3.5 Hz, 1 H, 3ЈЈ-H), 4.86 (t, J2Ј,3Ј = 9.2 Hz, 1 H, 2Ј-H), 4.63 (d,
J1Ј,2Ј = 8.7 Hz, 1 H, 1Ј-H), 4.51 (m, 2 H, 1ЈЈ-H, 6aЈ-H), 4.12 (m, 3
H, 6bЈ-H, 6ЈЈ-H), 3.87 (t, J = 6.88 Hz, 1 H, 5ЈЈ-H), 3.83 (t, J3Ј,4Ј
=
9.5 Hz, 1 H, 4Ј-H), 3.72 (m, 1 H, 5Ј-H), 2.15 (s, 3 H, COCH3), 2.14
(s, 3 H, COCH3), 2.07 (s, 3 H, COCH3), 2.07 (s, 3 H, COCH3),
2.05 (s, 6 H, 2ϫ COCH3), 1.97 (s, 3 H, COCH3) ppm. HR-ESI-
MS: [M + Na]+ ion peak at m/z = 684.18468 corresponding to
C26H35N3O17 (calcd. 684.18642). 14: 1H NMR (500 MHz, CDCl3):
δ = 5.43 (dd, J3Ј,4Ј = 3.5; J4Ј,5Ј = 1.4 Hz, 1 H, 4Ј-H), 5.16 (dd, J1Ј,2Ј
= 8.7, J2Ј,3Ј = 10.4 Hz, 1 H, 2Ј-H), 5.04 (dd, J2Ј,3Ј = 10.4, J3Ј,4Ј
=
3.5 Hz, 1 H, 3Ј-H), 4.60 (d, J1Ј,2Ј = 8.7 Hz, 1 H, 1Ј-H), 4.17 (m, 2
H, 6Ј-H), 4.02 (td, J4Ј,5Ј = 1.4, J5Ј,6Ј = 6.5 Hz, 1 H, 5Ј-H), 2.17 (s,
3 H, COCH3), 2.09 (s, 3 H, COCH3), 2.06 (s, 3 H, COCH3), 1.99
(s, 3 H, COCH3) ppm. HR-ESI-MS: [M + NH4]+ ion peak at m/z
= 391.14542 corresponding to C14H19N3O9 (calcd. 391.14650).
CDCl3): δ = 5.67 (d, J1Ј,2Ј = 8.2 Hz, 1 H, 1Ј-H), 5.35 (d, J3ЈЈ,4ЈЈ
=
3.4 Hz, 1 H, 4ЈЈ-H), 5.25 (t, J2Ј,3Ј = 9.2 Hz, 1 H, 3Ј-H), 5.11 (dd,
J1ЈЈ,2ЈЈ = 7.9, J2ЈЈ,3ЈЈ = 10.4 Hz, 1 H, 2ЈЈ-H), 5.05 (dd, J1Ј,2Ј = 8.2,
J2Ј,3Ј = 9.2 Hz, 1 H, 2Ј-H), 4.95 (dd, J2ЈЈ,3ЈЈ = 10.4, J3ЈЈ,4ЈЈ = 3.4 Hz,
1 H, 3ЈЈ-H), 4.48 (d, J1ЈЈ,2ЈЈ = 7.9 Hz, 1 H, 1ЈЈ-H), 4.45 (m, 1 H,
6Јa-H), 4.17–4.06 (m, 3 H, 6bЈ-H, 6ЈЈ-H), 3.88 (m, 1 H, 5ЈЈ-H), 3.84
(dd, J3Ј,4Ј = 9.2, J4Ј,5Ј = 4.1 Hz, 1 H, 4Ј-H), 3.76 (dd, J4Ј,5Ј = 4.1,
J5Ј,6Ј = 10.88 Hz, 1 H, 5Ј-H), 2.15 (s, 3 H, COCH3), 2.12 (s, 3 H,
COCH3), 2.10 (s, 3 H, COCH3), 2.07 (s, 3 H, COCH3), 2.05 (s, 3
H, COCH3), 2.04 (s, 3 H, COCH3), 2.03 (s, 3 H, COCH3), 1.97 (s,
3 H, COCH3) ppm.
Synthesis of 1-β-Amino-Peracetylated Carbohydrates 12, 15: The ti-
tle compound 12 was obtained by catalytic hydrogenation of 11
(0.98ϫ10–3 mol, 650 mg) with Pd/C (10%) in dry MeOH. The re-
action mixture was stirred under a pressure of ca. 10 bars for 2 h.
After filtration through Celite®, MeOH was evaporated to dryness.
The final compound 12 was crystallized from EtOH in good yields
(79–87%) and was used without further purifications. Compound
15 was obtained according to the same procedure also in good
yields (75–85%). 12: 1H NMR (500 MHz, CDCl3): δ = 5.35 (dd,
J3ЈЈ,4ЈЈ = 3.4; J4ЈЈ,5ЈЈ = 1.2 Hz, 1 H, 4ЈЈ-H), 5.23 (t, J2Ј,3Ј = 9.4 Hz, 1
Synthesis of 1-α-Bromo-Lactose Heptaacetate (10): A solution of
33% HBr in acetic acid (5 mL) was added to a cooled solution of
β-lactose octaacetate (9) (1.47ϫ10–3 mol, 1 g) dissolved in a mini- H, 3Ј-H), 5.11 (dd, J2ЈЈ,3ЈЈ = 10.4; J1ЈЈ,2ЈЈ = 7.8 Hz, 1 H, 2ЈЈ-H), 4.96
mum volume of CH2Cl2. The reaction was monitored by TLC with
hexane/EtOAc (10:10) and naphthoresocinol as spraying reagent.
The reaction mixture was stirred for 2 h and then diluted in a bi-
phasic solution composed of CH2Cl2 and saturated NaHCO3 un-
der ice-cold conditions. The organic phase was washed twice with
water and then concentrated under reduced pressure after drying
(dd, 1 H, J2ЈЈ,3ЈЈ = 10.4; J3ЈЈ,4ЈЈ = 3.4 Hz, 3ЈЈ-H), 4.73 (t, J2Ј,3Ј
9.4 Hz, 1 H, 2Ј-H), 4.47 (m, J1ЈЈ,2ЈЈ = 7.8 Hz, 2 H, 1ЈЈ-H, 6aЈ-H),
4.11 (m, 1Ј-H, 6bЈ-H, 6ЈЈ-H, 4 H), 3.87 (dt, J4ЈЈ,5ЈЈ = 1.2; J5ЈЈ,6ЈЈ
6.7 Hz, H, 5ЈЈ-H), 3.73 (t, J3Ј,4Ј = 9.4 Hz, 1 H, 4Ј-H), 3.59 (m, 1 H,
5Ј-H), 2.15 (s, 3 H, COCH3), 2.13 (s, 3 H, COCH3), 2.07 (s, 6 H,
COCH3), 2.05 (s, 3 H, COCH3), 2.04 (s, 3 H, COCH3), 1.97 (s, 3 H,
=
=
Eur. J. Org. Chem. 2015, 6061–6074
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6071

E. Cieckiewicz et al.
FULL PAPER
COCH3) ppm. HR-ESI-MS: [M + H]+ ion peak at m/z = 636.21301 Tetraacetylated Galactose-Pha 6: 1H NMR (500 MHz, CDCl3): δ
corresponding to C26H37NO17 (cal. 636.21397). 15: 1H NMR
(500 MHz, CDCl3): δ = 5.40 (d, J = 2.9 Hz, 1 H, 4Ј-H), 5.05 (m, 2
= 9.53 (s, 1 H, mesoH-10), 9.39 (s, 1 H, mesoH-5), 8.54 (s, 1 H,
mesoH-20), 8.00 (dd, J = 17.8, 11.6 Hz, 1 H, H-3a), 6.31 (s, 1 H,
H, 3Ј-H, 2Ј-H), 4.16 (d, J1Ј,2Ј = 7.9 Hz, 1 H, 1Ј-H), 4.10 (m, 2 H, H-13b), 6.30 (m, 1 H, H-3bcis), 6.19 (dd, J = 11.6, 1.1 Hz, 1 H, H-
6Ј-H), 3.9 (t, J = 6.5 Hz, 1 H, 5Ј-H), 2.16 (s, 3 H, COCH3), 2.09 3btrans), 6.12 (d, J = 9.3 Hz, 1 H, 17d-CONH), 5.36 (d, J = 3.3 Hz,
(s, 3 H, COCH3), 2.06 (s, 3 H, COCH3), 1.99 (s, 3 H, COCH3) 1 H, H-4Ј), 5.15 (t, J = 9.3 Hz, 1 H, H-1Ј), 5.06 (dd, J = 10.3,
ppm. HR-ESI-MS: [M + H]+ ion peak at m/z = 348.12818 corre-
sponding to C14H21NO9 (calcd. 348,12946).
3.3 Hz, 1 H, H-3Ј) 4.93 (t, J = 9.3 Hz, 1 H, H-2Ј), 4.00 (m, J =
n.d., 1 H, H-6Ј), 3.95 (m, 1 H, H-5Ј), 3.86 (s, 3 H, H-13d-COOMe),
3.69 (m, 5 H, H-12a-Me, H-8a-CH2), 3.41 (s, 3 H, H-2a-Me), 3.24
(s, 3 H, H-7a-Me), 2.68–2.61 (m, H-17a), 2.40–2.27 (m, H-17b),
2.40–2.27 (m, H-17a), 2.05 (s, 3 H, COCH3 of gal), 1.99 (s, 3 H,
COCH3 of gal), 1.98 (s, 3 H, COCH3 of gal), 1.93 (s, 3 H, COCH3
of gal), 1.91–1.84 (m, H-17b),1.80 (d, J = 7.3 Hz, 3 H, H-18a-Me),
1.70 (t, J = 7.6 Hz, 3 H, H-8b-Me), –1.62 (s, 2 H, NH) ppm. 13C
NMR (125 MHz, CDCl3): δ = 189.8 (CO, C-13a), 172.4 (CONH,
C-17c), 172.2 (C, C-19), 172.4 (COCH3 of gal), 170.3 (COCH3 of
gal), 170.0 (COCH3 of gal), 169.7 (COCH3 of gal), 169.7 (CO-
OCH3, C-13c), 160.7 (C, C-16), 155.8 (C, C-6), 151.0 (C, C-14),
149.6 (C, C-9), 145.3 (C, C-8), 142.1 (C, C-1), 138.0 (C, C-11),
136.6 (C, C-4), 136.4 (C, C-7), 136.3 (C, C-3), 132.0 (C, C-2), 129.1
(C, C-13), 129.1 (C, C-12), 129.0 (CH, C-3a), 122.9 (CH2, C-3b),
105.2 (C, C-15), 104.5 (CH, C-10), 97.6 (CH, C-5), 93.0 (CH, C-
20), 78.4 (CH, C-1Ј), 72.2 (CH, C-5Ј), 70.7 (CH, C-3Ј), 68.3 (CH,
C-2Ј), 67.0 (CH, C-4Ј), 64.7 (CH, C-13b), 61.0 (CH2, C-6Ј), 52.9
(COOCH3, C-13d), 51.1 (CH, C-17), 50.0 (CH, C-18), 33.1 (CH2,
C-17a), 29.8 (CH2, C-17b), 23.1 (CH3, C-18a), 20.8 (CH3, COCH3
of gal), 20.6 (CH3, COCH3 of gal), 20.5 (CH3, COCH3 of gal),
20.5 (CH3, COCH3 of gal), 19.5 (CH2, C-8a), 17.5 (CH3, C-8b),
12.1 (CH3, C-12a), 12.1 (CH3, C-2a), 11.3 (CH3, C-7a) ppm. HR-
ESI-MS: [M + H]+ ion peak at m/z = 922.38702 corresponding to
C49H55N5O13 (calcd. 922.38746). UV/Vis (MeOH): λmax
(log [ε/m–1 cm–1]) = 664 (4.07), 607 (3.45), 535 (3.48), 505 (3.52), 407
(4.42) nm. Compound 6 was synthesized in 28% yield, and its pu-
rity was evaluated by HPLC to be Ͼ 92%.
General Method To Introduce 1-β-Amino-Peracetylated Carbo-
hydrates 12, 15 at 17c-Position of Pheophorbide a with an Amide
Bond: This reaction was realized according to a method given in
the literature but with some adaptations.[8a] The reaction occurred
under anhydrous conditions. Pha (8.43ϫ10–5 mol, 50 mg) was dis-
solved in dry CH2Cl2 (2 mL) basified with TEA (10 equiv.). BOP
(1.5 equiv.) and 1-β-amino-peracetylated carbohydrate (2.5 equiv.)
were then added to the solution containing Pha. The reaction mix-
ture was stirred at room temperature for 4 h and was then diluted
with CH2Cl2. The organic phase was washed twice with water and
dried with MgSO4. The combined organic layers were concentrated
in vacuo and subjected to investigation by LC/SPE or by Prep-
HPLC.
1
Heptaacetylated Lactose-Pha 4: H NMR (500 MHz, CDCl3): δ =
9.52 (s, 1 H, mesoH-10), 9.39 (s, 1 H, mesoH-5), 8.55 (s, 1 H, me-
soH-20), 8.00 (dd, J = 17.9, 11.5 Hz, 1 H, H-3a), 6.30 (m, 2 H, H-
13b, H-3bcis), 6.19 (dd, J = 11.5, 1.4 Hz, 1 H, H-3btrans), 6.05 (d, J
= 9.3 Hz, 1 H, 17d-CONH), 5.32 (d, J = 3.4 Hz, 1 H, H-4ЈЈ), 5.24
(dd, J = 10.2, 3.4 Hz, 1 H, H-3Ј), 5.12 (t, J = 9.3 Hz, 1 H, H-1Ј)
5.07 (dd, J = 10.4, 7.9 Hz, 1 H, H-2ЈЈ), 4.90 (dd, J = 10.4, 3.4 Hz,
1 H, H-3ЈЈ), 4.67 (d, J = 10.2 Hz, 1 H, H-2Ј), 4.39 (d, J = 7.9 Hz,
1 H, H-1ЈЈ), 4.35 (d, J = 11.4 Hz, 1 H, H-6Ј), 4.09 (m, 1 H, H-6ЈЈ),
4.06–4.12 (m, J = n.d., 1 H, H-6Ј), 3.86 (s, 3 H, H-13d-COOMe),
3.82 (m, 1 H,5ЈH), 3.69 (m, 6 H, H-12a-Me, H-8a-CH2), 3.67 (m,
1 H, 4ЈH), 3.66 (m, 1 H, 5ЈЈH), 3.40 (s, 3 H, H-2a-Me), 3.24 (s, 3
H, H-7a-Me), 2.68–2.59 (m, H-17a), 2.40–2.26 (m, H-17a), 2.40–
2.26 (m, H-17b), 1.92–1.87 (m, H-17b), 2.15 (s, 3 H, COCH3 of
lac), 2.03 (s, 3 H, COCH3 of lac), 2.01 (s, 3 H, COCH3 of lac), 2.00
(s, 3 H, COCH3 of lac), 2.00 (s, 3 H, COCH3 of lac), 1.97 (s, 3 H,
COCH3 of lac), 1.93 (s, 3 H, COCH3 of lac), 1.80 (d, J = 7.3 Hz,
3 H, H-18a-Me), 1.70 (t, J = 7.7 Hz, 3 H, H-8b-Me), –1.64 (s, 2
H,NH) ppm. 13C NMR (CDCl3, 125 MHz): δ = 189.7 (CO, C-13a),
172.3 (CONH, C-17c), 172.2 (C, C-19), 171.4 (COCH3 of lac),
170.3 (COCH3 of lac), 170.3 (COCH3 of lac), 170.1 (COCH3 of
lac), 170.1 (COCH3 of lac), 169.3 (COCH3 of lac), 169.3 (CO-
OCH3-13c), 168.9 (COCH3 of lac), 160.8 (C, C-16), 155.8 (C, C-
6), 151.0 (C, C-14), 149.6 (C, C-9), 145.3 (C, C-8), 142.2 (C, C-1),
138.0 (C, C-11), 136.6 (C, C-4), 136.4 (C, C-7), 136.3 (C, C-3),
132.0 (C, C-2), 129.1 (C, C-12), 129.1 (CH, C-3a), 129.0 (C, C-13),
General Method for Deacetylation of Peracetylated Compounds 4
and 6: The reaction occurred under anhydrous conditions. The de-
acetylation was carried out under basic condition using 0.5 equiv.
of CH3ONa (1 m) to deprotect one acetate function. The peracetyl-
ated carbohydrate compounds 4 (4.13ϫ10–5 mol, 50 mg) or 6
(5.42ϫ10–5 mol, 50 mg) were dissolved in a mixture of dry CH2Cl2/
MeOH (2:1), and then the appropriate amount of CH3ONa was
added.[8a] The reaction mixture was stirred at 0 °C for 45 min. Gla-
cial acetic acid was added to neutralize the reaction mixture, which
was diluted with CH2Cl2. The organic phase was washed twice with
water and dried with MgSO4. The combined organic layers were
concentrated in vacuo and subjected to investigation by LC/SPE
or by Prep-HPLC.
1
122.9 (CH2, C-3b), 105.2 (C, C-15), 104.5 (CH, C-10), 100.9 (CH- Deacetylated Compound 5: H NMR (500 MHz, CD3OD/CDCl3):
C1ЈЈ), 97.6 (CH, C-5), 93.0 (CH, C-20), 78.0 (CH, C-1Ј), 76.0 (CH,
δ = 9.62 (s, 1 H, mesoH-10), 9.46 (s, 1 H, mesoH-5), 8.92 (s, 1 H,
C-5ЈЈ), 75.9 (CH, C-4Ј), 72.2 (CH, C-3Ј), 70.9 (CH, C-2Ј), 70.7 mesoH-20), 8.04 (dd, J = 17.9, 11.5 Hz, 1 H, H-3a), 6.32 (dd, J =
(CH, C-5Ј), 70.6 (CH, C-3ЈЈ), 68.9 (CH, C-2ЈЈ), 66.6 (CH, C-4ЈЈ), 11.5 Hz, 1 H, ND; H-3bcis), 6.13 (dd, J = 17.9 Hz, n.d., 1 H, H-
64.7 (CH, C-13b), 61.9 (CH2, C-6Ј), 60.8 (CH2, C-6ЈЈ), 52.9 (CO-
OCH3, C-13d), 51.1 (CH, C-17), 50.0 (CH, C-18), 33.1 (CH2, C- H-18), 4.50 (m, J = 9.7 Hz, H-17), 4.32 (d, J = 7.7 Hz, 1 H, H-
17a), 29.7 (CH2, C-17b), 23.1 (CH3, C-18a), 20.8 (CH3, COCH3 of 1ЈH), 4.27 (s, 3 H, H-13b-COOMe) 3.81 (s, 3 H, H-15c-COOMe),
3btrans), 5.33 (m, J = 11.5 Hz, 1 H, H-15a), 4.58 (m, J = 7.3 Hz,
lac), 20.7 (CH3, COCH3 of lac), 20.7 (CH3, COCH3 of lac), 20.6 3.68 (m, H-8a), 3.52 (s, 3 H, H-12a) 3.45 (s, 3 H, H-2a), 3.15 (s, 3
(CH3, COCH3 of lac), 20.6 (CH3, COCH3 of lac), 20.6 (CH3, H, H-7a-Me), 2.53–2.44 (m, H-17a), 2.30–2.21 (m, H-17b), 2.06–
COCH3 of lac), 20.5 (CH3, COCH3 of lac), 19.5 (CH2, C-8a), 17.4 1.99 (m, H-17a), 1.85–1.80 (m, H-17b), 1.77 (d, J = 7.3 Hz, 3 H,
(CH3, C-8b), 12.1 (CH3, C-12a), 12.1 (CH3, C-2a), 11.3 (CH3, C- H-18a-Me), 1.66 (t, J = 7.5 Hz, 3 H, H-8b-Me) ppm. 13C NMR
7a) ppm. HR-ESI-MS: [M + H]+ ion peak at m/z = 1210.47192
corresponding to C61H71N5O21 (calcd. 1210.47198). UV/Vis
(MeOH): λmax (log [ε/m–1 cm–1]) = 664 (4.16), 608 (3.52), 535 (3.55),
(125 MHz CD3OD/CDCl3): δ = 177.2 (CONH, C-17c), 175.9 (CO,
C-15b), 172.9 (C, C-19), 171.8 (COOCH3, C-13a), 169.7 (C, C-16),
169.3 (CH3, C-13b, COOCH3), 157.0 (C, C-6), 150.9 (C, C-9), 147.1
505 (3.58), 408 (4.51) nm. Compound 4 was synthesized in 30% (C, C-8), 141.5 (C, C-1), 138.0 (C, C-7), 138.0 (C, C-11), 137.3 (C,
yield, and its HPLC purity was estimated to Ͼ 90%. C-14), 137.0 (C, C-3), 136.9 (C, C-4), 132.9 (C, C-2), 131.1 (CH,
6072
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 6061–6074

Halogenated and Glycosylated Derivatives of Pheophorbide a
1
C-3a), 130.9 (C, C-12), 129.0 (C, C-13), 123.4 (CH2, C-3b), 105.8 2.5 mm NMR capillaries. For each H NMR experiment, a double
(CH, C-1Ј), 103.8 (CH, C-10), 104.4 (C, C-15), 100.1 (CH, C-5), presaturation of solvents was realized to suppress the signals of
95.7 (CH, C-20), 55.1 (CH, C-17), 51.4 (CH, C-18), 40.2 (CH2, C- acetonitrile and water in order to obtain an appropriate signal-to-
15a), 34.5 (CH2, C-17a), 32.2 (CH2, C-17b), 24.2 (CH3, C-18a),
noise ratio. In the same perspective, the number of scans was ad-
21.0 (CH2, C-8a), 18.9 (CH3, C-8b), 13.1 (CH3, C-12a), 13.0 (CH3, justed for each 1H NMR experiment. After NMR analysis, the mix-
C-2a), 12.0 (CH3, C-7a) ppm. HR-ESI-MS: [M + H]+ ion peak at tures were concentrated to dryness and dissolved in MeOH for MS
m/z = 948.42353 corresponding to C48H61N5O15 (calcd. 948.42424).
analysis. HR-ESI-MS analyses were performed with a LTQ-
UV/Vis (MeOH): λmax (log[ε/m–1 cm–1]) = 661 (4.00), 498 (3.47), Orbitrap XL hybrid mass spectrometer (Thermo Fisher Scientific,
398 (4.57), 298 (3.96) nm. Compound 5 was obtained in 40% yield,
and its HPLC purity was estimated at Ͼ 99%.
Bremen, Germany). Data were acquired in positive ion mode using
full-scan MS with a mass range of m/z = 100–2000. All experimen-
tal data were acquired using daily external calibration prior to data
acquisition.
1
Deacetylated Compound 7: H NMR (500 MHz, CDCl3/ CD3OD):
δ = 9.68 (s, 1 H, mesoH-10), 9.54 (s, 1 H, mesoH-5), 8.75 (s, 1 H,
mesoH-20), 8.05 (dd, J = 17.9, 11.7 Hz, 1 H, H-3a), 7.27 [assigned
on basis of COSY spectrum] (n.d., 1 H, CONH), 6.35 (dd, J =
17.9, 3.6 Hz, 1 H, H-3bcis), 6.15 (dd, J = 11.7, 3.6 Hz, 1 H, H-
3btrans), 5.30 (m, 1 H, H-15a), 4.47 (m, H-17), 4.24 (s, 3 H, H-13b-
COOMe), 3.74 (s, 3 H, H-15c-COOMe), 3.78 (m, H-8a), 3.46 (s, 3
H, H-2a-Me), 4.65 (d, J = 8.9 Hz, 1 H, H-1ЈH), 3.56 (m, 6 H, H-
12a-Me), 3.58–3.52 (m, H-6Ј) [assigned on basis of HSQC spec-
trum], 3.29 (s, 3 H, H-7a-Me), 2.31–2.20 (m, H-17a); 2.30–2.21 (m,
H-17b), 1.92–1.84 (m, H-17b), 1.63–1.55 (m, H-17a), 1.72 (d, J =
7.1 Hz, 3 H, H-18a-Me), 1.72 (t, J = 10.8 Hz, 3 H, H-8b-Me) ppm.
13C NMR (125 MHz, CDCl3/CD3OD): δ = 174.6 (CONH, C-17c),
174.1 (CO, C-15b), 170.3 (C, C-19), 169.7 (COOCH3, C-13a), 166.5
(C, C-16), 155.0 (C, C-6), 148.8 (C, C-9), 145.2 (C, C-8), 139.5 (C,
C-1), 136.3 (C, C-7), 136.0 (C, C-11), 136.0 (C, C-14), 135.4 (C, C-
3), 134.9 (C, C-4), 130.8 (C, C-2), 129.2 (CH, C-3a), 128.9 (C, C-
12), 123.1 (C, C-13), 122.0 (CH2, C-3b), 102.2 (CH, C-10), 102.1
(C, C-15), 98.5 (CH, C-5), 93.5 (CH, C-20), 79.9 (CH, C-1Ј), 61.8
(CH2, C-6Ј), 53.0 (COOCH3, C-13b), 51.1 (CH, C-17), 50.0 (CH,
C-18), 38.6 (CH2, C-15a), 32.1 (CH2, C-17a), 30.0 (CH2, C-17b),
22.9 (CH3, C-18a), 19.5 (CH2, C-8a), 17.6 (CH3, C-8b), 12.2 (CH3,
C-12a), 11.9 (CH3, C-2a), 11.2 (CH3, C-7a) ppm. HR-ESI-MS of
7 showed an [M + H]+ ion at m/z = 786.37171 corresponding to
the molecular formula C42H51N5O10 (calcd. 786.37142). UV/Vis
(MeOH): λmax [log(ε/m–1 cm–1)] = 661 [3.86], 498 [3.35], 398 [4.43],
298 [3.82] nm. Compound 7 was synthesized in 45% yield with an
HPLC purity evaluated at greater than 98%.
Preparative HPLC Conditions: The analytical chromatographic
conditions used to isolate 4–7 by LC/SPE were adapted to Prep-
HPLC conditions in order to isolate these compounds of interest
in large amount. The Prep-HPLC used was a Varian PrepStar 218
coupled with a DAD detector set at 408 nm (DAD ProStar 335
UV/Vis) and equipped with a fraction collector (440LC). The crude
residues of reaction were fractionated on LichroPrep 100 RP-18
(12 μm, Merck). The peaks corresponding to compounds 4–7 were
collected manually based on chromatogram peak retention times.
HPLC Purity: The purities of 4–7 were determined by HPLC using
the same conditions as those used to isolate them by Prep-HPLC.
The purity of 2 was ascertained by the HPLC method used for
precursors 4 and 6. The purities were determined from all peak
areas recorded at 408 nm.
In Vitro Photodynamic Activity: We used 3 human cancer cell lines
(Hs683, MCF-7 and U373) and one mouse cancer cell line
(B16F10). The mouse B16F10 melanoma (ATCC number: CRL-
6475), and the Hs683 oligodendroglioma (ATCC number: HTB-
138) were obtained from the American Type Culture Collection
(ATCC, Manassas, VA, USA). The U373 glioblastoma (ECACC
code 08061901) was from the European Cell Culture Collection.
The MCF-7 breast cancer cell line (DSMZ code ACC115) was pur-
chased from the Deutsche Sammlung von Mikroorganismen und
Zellkulturen (DSMZ, Braunschweig, Germany). All cell lines were
maintained as monolayers in RPMI 1640 medium (Invitrogen,
Merelbeke, Belgium) supplemented with penicillin-streptomycin
(200 U/mL and 200 μg/mL), gentamycin (100 μg/mL), l-glutamine
(4 mm) and with 10% inactivated fetal calf serum (Invitrogen) in a
humid atmosphere at 37 °C and 5% CO2. We used previously re-
ported MTT colorimetric assays[3,18] to determine global in vitro
growth inhibitory values of each compound in the dark and after
photodynamic treatment. In order to observe PDT effects of our
semisynthetic compounds on cell viability, we adapted our MTT
conditions as follows. Cells were seeded in transparent 96-well cul-
LC/SPE/NMR-MS Conditions: The hyphenated LC/SPE/NMR
system was performed with an Agilent 1100 series LC system cou-
pled to an AV500 MHz spectrometer from Bruker BioSpin
equipped with a cryoprobe. The chromatographic separation of the
reaction mixture was carried out on Hypersil ODS C18 column
(250ϫ4.6 mm, i.d.; particle size 5 μm). The crude mixture contain-
ing peracetylated compounds 4 or 6 was analyzed in the following
gradient mode using mobile phase A composed of acetonitrile and
mobile phase B composed of water: 0 min, 85% of A; 7 min, 89%
of A; 21 min, 100% of A; 30 min, 85% of A. The crude mixtures ture plates. After 24 h of incubation, media of each 96-well plate
containing deacetylated compounds 5 or 7 were separated in iso-
cratic mode using a mixture of methanol/water (95:5). All chroma-
tographic analyses were carried out at room temperature. The UV
detector wavelength was set at 408 nm. The flow rate was delivered
at 1.0 mL/min. The peaks of interest were trapped with a Prospekt
2 SPE unit (Bruker, BioSpin and Spark) using HySphere GP car-
tridges (2 mm i.d., particle size 10–12 μm) by addition of water at
3 mL/min to the postcolumn eluate. The peaks of interest were
trapped manually. The crude residues were filtered through 0.45 μm
filters before injection. To obtain useful NMR spectra, 50 μL of
crude mixture at 1.5 mg/2 mL (MeOH) were injected using a Rheo-
dyne 7725i injection valve with a 100 μL injection loop. The loaded
cartridges were dried for 59 min under a stream of dry N2 gas. The
trapped compounds were then eluted with 290 μL CD3CN (Euriso
top®, Gyf-sur-Yvette, France) from the dried SPE cartridges to the
were removed, and attached cells were treated with a range of nine
concentrations of each semisynthesized compound (10–8 to 10–4 m)
Each tested compound was added (100 μL/well) in two different
96-well plates in order to determine dark and light-dependent ac-
tivity. After 2 h of incubation with the photosensitizer, both 96-well
plates seeded with the same photosensitizer were removed from the
incubator. One plate was kept in the dark and the other one was
illuminated from below with warm white light using 4 neon tubes
(Philips, TL-D, super 80, 15 W, 827, 2700 K corresponding to 655–
660 nm). For all illuminated experiments, the plates were irradiated
with a constant light fluent of 0.7 J/cm2. The light intensity was
checked before each MTT assay thanks to a luxmeter (elix®, LX-
6610). After illumination, plates were put back into the incubator
for 48 h. The medium was then removed, and the MTT solution
(0.5 mg/mL) was added to each well. After 4 h of incubation, the
Eur. J. Org. Chem. 2015, 6061–6074
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6073

E. Cieckiewicz et al.
FULL PAPER
[8]
[9]
a) X. Zheng, J. Morgan, S. K. Pandey, Y. Chen, E. Tracy, H.
Baumann, J. R. Missert, C. Batt, J. Jackson, D. A. Bellnier,
B. W. Henderson, R. K. Pandey, J. Med. Chem. 2009, 52, 4306–
4318; b) A. A. Aksenova, I. L. Sebiakin, A. F. Mironov, Bioorg.
Khim. 2000, 26, 126–129; c) D. V. Titov, M. L. Gening, Y. E.
Tsvetkov, N. E. Nifantiev, Russ. Chem. Rev. 2014, 83, 523–554;
d) X. Zheng, R. K. Pandey, Anticancer Agents Med. Chem.
2008, 8, 241–268; e) A. A. Aksenova, Yu. L. Sebyakin, A. F.
Mironov, Bioorg. Khim. 2001, 27, 145–150.
a) S. K. Pandey, X. Zheng, J. Morgan, J. R. Missert, T. H. Liu,
M. Shibata, D. A. Bellnier, A. R. Oseroff, B. W. Henderson,
T. J. Dougherty, R. K. Pandey, Mol. Pharm. 2007, 4, 448–464;
b) G. Zheng, A. Graham, M. Shibata, J. R. Missert, A. R.
Oseroff, T. J. Dougherty, R. K. Pandey, J. Org. Chem. 2001, 66,
8709–8716.
L. Ingrassia, I. Camby, F. Lefranc, V. Mathieu, P. Nshimyumu-
kiza, F. Darro, R. Kiss, Curr. Med. Chem. 2006, 13, 3513–3527.
a) G.-F. Han, J.-J. Wang, Y. K. Shim, J. Photosci. 2001, 8, 71–
73; b) K. Yasuhiko, T. Hitoshi, J. Porphyrins Phthalocyanines
1998, 2, 159–169; c) J. Wang, L. Zhao, J. Li, Chin. J. Org. Chem.
2009, 29, 1598–1605.
MTT solution was replaced by dimethyl sulfoxide. The absorbance
values corresponding to each experimental condition were quanti-
tatively measured by spectrophotometry (Biorad Model 680XR,
Nazareth, Belgium) using a wavelength at 570 nm and a reference
at 630 nm. The absorbance values of untreated cells were taken as
100% cell growth. For the PDT conditions, the untreated cells were
also illuminated. The IC50 values of each compound were deter-
mined under dark and illuminated conditions.
Acknowledgments
The authors sincerely wish to thank the laboratory of Pharmacog-
nosy from the Catholic University of Leuven and particularly Dr.
Marie-France Herent for her help in recording MS data. This work
was partly sponsored by the Fonds Léon Fredericq and by a grant
from Fonds National de la Recherche Scientifique (FNRS)
(3.4533.10 and T.0190.13). E. C. was a research follow from the
Fonds pour la formation à la Recherche dans l’Industrie et l’Agri-
culture (FRIA).
[10]
[11]
[12]
[13]
W. Seebacher, R. Weis, J. Jurenitsh, K. Rauchensteiner, E. Ha-
slinger, Monatsh. Chem. 2000, 131, 985–996.
V. Mathieu, M. Le Mercier, N. De Neve, S. Sauvage, T. Gras,
I. Roland, F. Lefranc, R. Kiss, J. Invest. Dermatol. 2007, 127,
2399–2410.
[1] P. Agostinis, K. Berg, K. A. Cengel, T. H. Foster, A. W. Girotti,
S. O. Gollnick, S. M. Hahn, M. R. Hamblin, A. Juzeniene, D.
Kessel, M. Korbelik, J. Moan, P. Mroz, D. Nowis, J. Piette,
B. C. Wilson, J. Golab, CA Cancer J. Clin. 2011, 61, 250–281.
[2] a) M. R. Detty, S. L. Gibson, S. J. Wagner, J. Med. Chem. 2004,
47, 3897–3915; b) R. K. Pandey, L. N. Goswami, Y. Chen, A.
Gryshuk, J. R. Missert, A. Oseroff, T. J. Dougherty, Lasers
Surg. Med. 2006, 38, 445–467.
[14] M. Le Mercier, S. Fortin, V. Mathieu, I. Roland, S. Spiegl-
Kreinecker, B. Haibe-Kains, G. Bontempi, C. Decaestecker, W.
Berger, F. Lefranc, R. Kiss, Neoplasia 2009, 11, 485–496.
[15] S. Rorive, N. Belot, C. Decaestecker, F. Lefranc, L. Gordower,
S. Micik, C. A. Maurage, H. Kaltner, M. M. Ruchoux, A. Dan-
guy, H. J. Gabius, I. Salmon, R. Kiss, I. Camby, GLIA 2001,
33, 241–255.
[3] E. Cieckiewicz, L. Angenot, T. Gras, R. Kiss, M. Frederich,
Phytomedicine 2012, 19, 278–283.
[4] L. E. Xodo, V. Rapozzi, M. Zacchigna, S. Drioli, S. Zorzet,
Curr. Med. Chem. 2012, 19, 799–807.
[5] R. Y. Yang, G. A. Rabinovich, F. T. Liu, Expert Rev. Mol. Med.
2008, 10, e17.
[6] G. A. Rabinovich, Br. J. Cancer 2005, 92, 1188–1192.
[7] F. Lefranc, V. Mathieu, R. Kiss, World J. Biol. Chem. 2011, 2,
193–201.
[16] A. Satelli, P. S. Rao, P. K. Gupta, P. R. Lockman, K. S. Sriv-
enugopal, U. S. Rao, Oncol. Rep. 2008, 19, 587–594.
[17] J. Xue, Y. Pan, Z. Guo, Tetrahedron Lett. 2002, 43, 1599–1602.
[18] M. Frederich, A. Marcowycz, E. Cieckiewicz, V. Megalizzi, L.
Angenot, R. Kiss, Planta Med. 2009, 75, 1634–1637.
Received: March 25, 2015
Published Online: August 7, 2015
6074
www.eurjoc.org
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2015, 6061–6074