Antiproliferative 3-deoxysphingomyelin analogs: Design, synthesis, biological evaluation and molecular docking of pyrrolidine-based 3-deoxysphingomyelin analogs as anticancer agents

Article abstract of DOI:10.1016/j.bioorg.2018.11.040

Sphingomyelins and glycerophospholipids are structurally related phospholipids. Nevertheless, glycerophospholipids analogs are known as antitumor agents while sphingomyelin analogs were reported as cytoprotective agents. Herein, we have addressed the development of 3-deoxysphingomyelin analogs as cytotoxic agents possessing modified sphingobases. Thus, pyrrolidine-based 3-deoxysphingomyelin analogs were synthesized and evaluated against a panel of cell lines representing four major types of cancers. Compounds 3d, 4d and 6d elicited better GI₅₀ values than the FDA approved drug miltefosine. Investigation of their impact on Akt phosphorylation as a possible mechanism for the antiproliferative activity of this class of compounds revealed that these compounds might elicit a concentration-dependent mechanism via inhibition of Akt phosphorylation at the lower concentration. Molecular docking predicted their binding modes to Akt to involve polar head binding to the Pleckstrin homology domain and hydrophobic tail extension into a hydrophobic pocket connecting the Pleckstrin homology domain and the kinase domain. As a whole, the described work suggests compounds 3d, 4d and 6d as promising pyrrolidine-based 3-deoxysphingomyelin analogs for development of novel cancer therapies.

Full text of DOI:10.1016/j.bioorg.2018.11.040

Accepted Manuscript
Anti-proliferative 3-deoxysphingomyelin analogs: Design, synthesis, biological
evaluation and molecular docking of pyrrolidine-based 3-deoxysphingomyelin
analogs as anticancer agents
Ahmed H.E. Hassan, Hye Rim Park, Yoon Mi Yoon, Hye In Kim, Sung Yeun
Yoo, Kun Won Lee, Yong Sup Lee
PII:
S0045-2068(18)30608-4
DOI:
https://doi.org/10.1016/j.bioorg.2018.11.040
YBIOO 2648
Reference:
Bioorganic Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
21 June 2018
21 November 2018
24 November 2018
Please cite this article as: A.H.E. Hassan, H. Rim Park, Y. Mi Yoon, H. In Kim, S. Yeun Yoo, K. Won Lee, Y. Sup
Lee, Anti-proliferative 3-deoxysphingomyelin analogs: Design, synthesis, biological evaluation and molecular
docking of pyrrolidine-based 3-deoxysphingomyelin analogs as anticancer agents, Bioorganic Chemistry (2018),
doi: https://doi.org/10.1016/j.bioorg.2018.11.040
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Anti-proliferative 3-deoxysphingomyelin analogs: Design, synthesis, biological
evaluation and molecular docking of pyrrolidine-based 3-deoxysphingomyelin
analogs as anticancer agents
Ahmed H.E. Hassana,b,1, Hye Rim Parkc,1, Yoon Mi Yoonc, Hye In Kimc, Sung Yeun
Yooc, Kun Won Leec, Yong Sup Leea,c,*
a Medicinal Chemistry Laboratory, Department of Pharmacy, College of Pharmacy, Kyung Hee University,
Seoul 130-701, Republic of Korea.
b Department of Medicinal Chemistry, Faculty of Pharmacy, Mansoura University, Mansoura 35516, Egypt
c Department of Life and Nanopharmaceutical Science, Kyung Hee University, Kyung Hee University, Seoul
130-701, Republic of Korea.
*Corresponding author: Yong Sup Lee (email: kyslee@khu.ac.kr)
1Both authors contributed equally.

Abstract:
Sphingomyelins and glycerophospholipids are structurally related phospholipids.
Nevertheless, glycerophospholipids analogs are known as antitumor agents while
sphingomyelin analogs were reported as cytoprotective agents. Herein, we have addressed the
development of 3-deoxysphingomyelin analogs as cytotoxic agents possessing modified
sphingobases. Thus, pyrrolidine-based 3-deoxysphingomyelin analogs were synthesized and
evaluated against a panel of cell lines representing four major types of cancers. Compounds
3d, 4d and 6d elicited better GI50 values than the FDA approved drug miltefosine.
Investigation of their impact on Akt phosphorylation as a possible mechanism for the
antiproliferative activity of this class of compounds revealed that these compounds might
elicit a concentration-dependent mechanism acting via inhibition of Akt phosphorylation at
the lower concentration and possibly via disruption of membranes at the higher
concentration. Molecular docking predicted their binding modes to Akt to involve polar head
binding to the Pleckstrin homology domain and hydrophobic tail extension into a
hydrophobic pocket connecting the Pleckstrin homology domain and the kinase domain. As a
whole, the described work suggests compounds 3d, 4d and 6d as promising pyrrolidine-based
3-deoxysphingomyelin analogs for development of novel cancer therapies.
Keywords:
3-Deoxysphingomyelin Analogs; Bioactive Lipids; Akt phosphorylation; Antiproliferative
Agents.

1. Introduction:
Despite the achieved progress in understanding cancer biology, cancer still remains a difficult
disease to treat and one of the leading causes of death all over the world [1, 2]. According to
the statistical analysis of the global burden of disease (GBD), cancers is the second among
the major cause of death [3]. Alarmingly, the lifetime risk of cancer has approached up to
53.5% for males and 47.5% for females in some societies [4]. Parenthetically, around half of
the population of these societies might develop cancer during their lifetime. The destructive
consequences of cancer go beyond the patients’ health to the socio-economical scope
influencing patients’ families, community and economy. There is an indeed need to develop
new therapeutic tools and strategies in the fight against cancer.
Glycerophospholipids are a class of lipids that possess a glycerol backbone to which fatty
acids and a polar phosphocholine head are attached. In comparison, sphingomyelins (1, Fig.1)
possess nitrogenous sphingobases (long chain bases; LCBs) instead of the glycerol backbone
of glycerophospholipids [5]. Several synthetic analogs of glycerophospholipids retaining the
glycerol backbone such as edelfosine or lacking it such as miltefosine are known as antitumor
compounds [6]. In contrast, several sphingomyelin analogs were found to elicit
cytoprotective activities via inhibition of sphingomyelinase [7, 8]. Interestingly, the literature
reports 3-O-methylated-sphingomyelin analogs as sphingomyelinase inhibitors, but
3-deoxysphingomyelin analogs in which a hydrogen atom replaces the hydroxyl group at 3-
position of the sphingobase were found to be not liable for hydrolytic cleavage by
sphingomyelinase nor possessing inhibitory activity against it [9]. This indicates that the
oxygen atom at 3-position of the sphingobase is important for sphingomyelinase substrates or
competitive inhibitors. This suggests that 3-deoxysphingomyelins might be devoid from the
cytoprotective activity arising from the reported sphingomyelinase inhibitory activity of

sphingomyelin analogs. Furthermore, their structural similarity with glycerophospholipids
might suggest 3-deoxysphingomyelins analogs as potential candidates for development of
cytotoxic agents. To the best of our knowledge, the literature reports no effort to develop
antitumor 3-deoxysphingomyelins analogs despite this apparent structural similarity.
Recently, a novel cytotoxic pyrrolidine-derived sphingomimetic base (2) was designed via
conformational restriction [10]. Considering the pro-apoptotic activity of this pyrrolidine-
based derivative as well as the literature reports indicating that 3-deoxysphingomyelins are
devoid from the sphingomyelinase inhibitory activity responsible for the cytoprotective
activity, in addition to the structural similarity with antitumor alkylphospholipids analogs, we
have anticipated that pyrrolidine-based 3-deoxysphingomyelin analogs might elicit a
promising antiproliferative activity against cancer cells. Accordingly, novel pyrrolidine-based
3-deoxysphingomyelin analogs belonging to four series were designed, synthesized,
biologically evaluated and mechanistically studied. As shown in Fig. 1, the design involved
conformational restriction between the 2-amino group and the C-5 carbon atom of the LCB
moiety, excision of the alkyl chain of the LCB moiety beyond the introduced pyrrolidine ring,
as well as, the hydroxyl group at C-3 carbon atom (highlighted in red color) while retaining
the long chain FAs (highlighted in green color) to afford the novel series 3 of pyrrolidine-
based 3-deoxysphingomyelin analogs. Series 3 was designed to retain a chiral configuration
similar of that of natural sphingomyelins. However, the effect of stereochemistry inversion of
the pyrrolidine ring on the biological activity was planned to be assessed after synthesis of
series 4, which represent the opposite enantiomeric form of series 3. Displacement of the N-
acylamide group of series 3 by one carbon atom would produce the novel series 5 while
displacement of the N-acylamide group of series 4 by one carbon atom and direct attachment
of the phosphocholine to 2-position of the pyrrolidine ring would provide series 6. According

to the proposed design, the impact of incorporation of various saturated and unsaturated fatty
acid moieties on the biological activity would be explored. Herein, we would like to report
our results.
CH3
O
P
O
P
N
N
O
H3C
O
O
O
O
14
O
O
17
O
H3C
Edelfosine
Miltefosine
R1
O
H
N
OH
OH
O
N
HN
O
P
O
O
R2
C
13H27
OH
1
2
Sphingomyelins
O
O
N
R1
R1
O
O
P
N
N
N
O
O
O
P
O
O
O
3
4
O
N
O
P
O
O
N
O
O
O
R1
O
O
P
N
N
O
R1
5
6
Figure 1. Reported antitumor alkylphospholipids and the design of pyrrolidine-based 3-deoxysphingomyelin
analogs 3, 4, 5 and 6
2. Results and discussion
2.1. Chemistry
Accessing the desired molecules in few synthetic steps is one of the aims of good synthetic
plan. As concise synthesis would help to approach this ideal goal through step economy, the
targeted compounds were synthesized in two pots over three synthetic steps starting from
commercially available chiral pool of the starting materials D-prolinol (7), L-prolinol (10),

(S)-3-pyrrolidinemethanol (12) and (R)-3-pyrrolidinol (14). The required acid chlorides (8a–
f) were obtained from palimitic acid, palmitoleic acid, oleic acid, erucic acid, linoleic acid,
and α-linolenic acid respectively following a procedure similar to the reported literature
method [11]. In the first pot, the starting pyrrolidine derivatives 7, 10, 12 and 14 were reacted
with acid chlorides 8a–f in the presence of triethylamine as a base and dimethylformamide as
a solvent to afford the corresponding amides 9a–d, 11a–f, 13a–f and 15a–f respectively. In
the second pot, the intermediate amide derivative 9a–d, 11a–f, 13a–f and 15a–f were first
reacted with 2-choloro-1,3,2-dioxaphospholane-2-oxide at 0 ºC to afford the corresponding
cyclic phospholane esters followed by opening of the ring via heating under pressure with
trimethylamine to yield the targeted designed compounds.

O
R
8a~8d
Cl
O
O
H
N
R
R
O
P
a
b
OH
N
N
N
O
OH
O
O
7
9a
9d
9b
9c
3a (
3d (
3b (
3c (
89%), 87%),
(84%),
(76%)
(86%),
(78%),
61%),
79%)
O
N
O
8a~8f
H
N
R
R
O
P
a
b
OH
N
N
O
OH
O
O
10
11a
11d
11b
11e
11c
11f
4a
4d
4b
4e
4c
(54%), (50%),
4f
(86%), (65%)
(83%),
(82%),
(89%),
(89%),
(72%),
(78%)
(61%),
(79%),
8a~8f
O
N
O
O
R
b
OH
a
O
O
P
OH
N
N
HN
O
R
12
13a
13d
13b
13e
13c
13f
5a
5d
5b
5e
5c
(60%), (85%),
(90%),
(57%),
(90%),
(62%),
(91%)
(58%)
(62%),
(69%),
5f
(87%), (63%)
O
P
N
O
OH
O
8a~8f
OH
O
R
O
O
a
b
N
N
HN
R
14
15a
15d
5b
15e
5c
6a
6d
6b
6e
6c
(68%), (78%),
6f
(73%), (74%)
(74%),
(66%),
(82%)
(72%,)
(79%)
(62%),
(73%),
15f
(71%),
d, R =
a, R =
b, R =
c, R =
14
7
11
7
e, R =
f, R =
5
7
4
7
7
7
Scheme 1. Reagents and conditions: a) NEt3, DMF, rt; b) i) 2-chloro-2-oxo-1,3,2-dioxaphospholane, NEt3,
benzene, 0 ºC to rt, ii) NMe3, CH3CN, 60 ºC
2.2. Biological evaluation
2.2.1. Inhibition of Cell line proliferation:
The antiproliferative activity of the synthesized novel 3-deoxysphingomyelin analogs was
evaluated using MTT assay against a panel of four human cancer cells representing breast
cancer, non-small-cell lung cancer, liver cancer, and skin cancer. Based on structural
similarity of the tested compounds with alkylphospholipids, Akt phosphorylation inhibition
was anticipated to be the mechanism of action of these compounds. Therefore, miltefosine (a
member of alkylphospholipids that has been approved by FDA for treatment of cutaneous

metastatic breast cancer) was used as a reference standard for activity comparisons as it was
reported to elicit cytotoxic activity through inhibition of Akt phosphorylation, as well as,
other mechanisms [12]. Except for MCF-7 cell line (breast cancer), miltefosine was found to
elicit a limited inhibitory activity on the growth of all other used cell lines showing a growth
inhibition not exceeding 36.6% at 100 μM concentrations.
Breast cancer is the most common cancer among females and it is ranked the second in
regard to the costs associated with premature-mortality [13-15]. In addition, miltefosine; the
reference drug used in the conducted biological assay is a FDA approved drug for metastatic
breast cancer [12]. Therefore, the impact of the designed 3-deoxysphingomyelin analogs on
the proliferation of MCF-7 cell line as a representative for breast cancers was investigated.
As displayed in Table 1 and illustrated in Fig. 2, miltefosine elicited an average
antiproliferative activity against the employed MCF-7 cell line (69.44% inhibition of growth
at 100 μM concentration and GI50 value equals to 28.4 μM). Among all evaluated
compounds, the saturated palimitic acid and the monounsaturated erucic acid derivatives
were the most active compounds whereas the polyunsaturated linoleic and α-linolenic acids
derivatives were the least active. The activities were lower for derivatives of oleic acid; a
monounsaturated fatty acids shorter than erucic acid, and even more decreased for the more
shorter monounsaturated FA; palmitoleic acid. 3-Deoxysphingomyelin analogs possessing
pyrrolidine core of type 3 and its enantiomeric antipode 4 were more potent than the
corresponding 3-deoxysphingomyelin analogs belonging to types 5 and 6. This might indicate
more activity of 3-deoxysphingomyelin analogs possessing 2-substitutedpyrrolidinyl core
than those possessing 3-substitutedpyrrolidinyl core aginst MCF-7 cells. It was noted that
compounds of core type 3 and the corresponding enantiomeric antipode compounds 4 showed
almost identical activity. This might indicate that stereochemistry of the pyrrolidine ring has

no significant impact on the elicited biological activity. The most potent
3-deoxysphingomyelin analog was erucic acid derivative 3d exhibiting GI50 value equals to
15.7 µM. Relative to miltefosine whose GI50 values was 28.4 μM, this represent 1.8 folds
increase in activity.
Table 1. In vitro cytotoxicity of the synthesized 3-deoxysphingomyelin analogs against selected human cancer
cells.
MCF-7a
% Inhibitione GI50
A-549b
% Inhibitione
HepG2c
A-431d
Comp
f
% Inhibitione % Inhibitione
3a
86.9
62.2
76.6
93.0
87.5
63.1
72.9
91.0
38.6
30.1
74.6
67.3
71.5
80.5
10.8
25.4
80.9
63.5
69.4
92.9
24.4
22.6
69.4
29.6
44.6
4.7
22.3
13.3
10.1
56.9
30.9
12.6
11.1
38.5
17.1
17.5
51.3
26.0
39.6
46.6
24.3
24.6
11.1
27.8
10.9
72.8
21.5
21.2
31.3
55.6
19.6
27.3
62.6
41.5
17.3
24.5
42.7
7.3
3b
ND
ND
15.7
29.6
ND
ND
24.8
ND
ND
42.2
ND
51.3
36.2
ND
ND
51.1
ND
ND
23.5
ND
ND
28.4
3c
10.1
56.9
36.0
8.3
3d
4a
4b
4c
12.9
49.0
15.7
12.9
71.6
16.8
25.3
49.9
21.6
18.7
42.8
21.6
11.9
58.7
11.5
11.5
36.6
4d
4e
4f
2.4
5a
74.8
32.6
55.1
54.3
NT
5b
5c
5d
5e
5f
3.6
6a
82.5
20.9
33.8
67.3
0.3
6b
6c
6d
6e
6f
6.1
Miltefosine
35.4
a Human breast carcinoma cells.
b Human lung cancer cells.
c Human hepatocellular carcinoma cells.
d Human Epidermoid carcinoma cells.
e % Inhibition at 100 μM.
f GI50 was defined as the concentration resulting in 50% inhibition of cell proliferation.

100
90
80
70
60
50
40
30
20
10
0
3a 3b 3c 3d 4a 4b 4c 4d 4e 4f 5a 5b 5c 5d 5e 5f 6a 6b 6c 6d 6e 6f Ref.
Palmitoyl Palmitoleoyl oleoyl Erucoyl Linoleoyl α-Linolenoyl Miltefosine
Figure 2. Percentages inhibition of MCF-7 cell line proliferation after treatment with synthesized
3-deoxysphingomyelin analogs and miltefosine at 100 μM.
Out of all cancers, lung cancer causes the highest mortality [3, 13-15]. Among lung cancer
cases, 85% are non-small lung cancers. Consequently, we investigated the efficacy of the
prepared 3-deoxysphingomyelin analogs against A-549 cell line, a non-small-cell lung
adenocarcinoma. The reference standard miltefosine elicited a limited inhibition of the
growth of the employed A-549 cell line by 36.6% at 100 μM dose. As indicated in Table 1
and outlined in Fig. 3, only derivatives of palmitic and the monounsaturated erucic acids
elicited antiproliferative activity comparable to or higher than the reference miltefosine
whereas the shorter monounsaturated acids; palmitoleic acid and oleoic acid, and the
polyunsaturated linoleic and α-linolenic acids derivatives were of low activity. In addition,
compounds possessing type 3 pyrrolidine core and the corresponding enantiomeric antipode
compounds having type 4 pyrrolidine core showed almost similar potency confirming that
stereochemistry of the pyrrolidine ring has no significant impact on the elicited biological
activity. The most active compound (5a); a palmitic derivative of type 5 (S)-3-
pyrrolidinemethanol core elicited inhibition of the growth of A-549 cell line by 71.6%.

Whereas, erucic acid derivatives of (R)-3-pyrrolidinol (6d) and D-prolinol (3d) came in the
second rank of activity by elicited percent of inhibition of growth by 58.7 and 56.9 %
respectively. This indicates that the optimum activity against A-549 was related to the FA
type more than the type of the pyrrolidine ring.
80
70
60
50
40
30
20
10
0
3a 3b 3c 3d 4a 4b 4c 4d 4e 4f 5a 5b 5c 5d 5e 5f 6a 6b 6c 6d 6e 6f Ref.
Palmitoyl
Palmitoleoyl
oleoyl
Erucoyl
Linoleoyl
α-Linolenoyl
Miltefosine
Figure 3. Percentages inhibition of A-549 cell line proliferation after treatment with synthesized
3-deoxysphingomyelin analogs and miltefosine at 100 μM.
Among primary liver cancers in adult patients, hepatocellular carcinoma comes to the fore. In
addition, it is the third leading cause for mortality among cancer deaths [16]. Surgically
unmanageable hepatocellular carcinoma patients show 6-20 months average survival time
and 5-years survival less than 5%. Accordingly, we investigated the antiproliferative activity
of the designed 3-deoxysphingomyelin analogs against HepG2 cell line as a model for
hepatic carcinoma. The employed HepG2 cell line, miltefosine elicited a low efficacy at 100
μM concentration measured as 31.3% inhibition of growth (Fig. 4). Among evaluated
3-deoxysphingomyelin analogs, the most actives were erucic acid derivatives of (R)-3-
pyrrolidinol (6d) and D-prolinol (3d) which elicited growth inhibition percentages of 72.8%

and 56.9% respectively. The only palmitic acid derivative that elicited more than 50% growth
inhibition was 5a, which was a derivative of (S)-3-pyrrolidinemethanol. All derivatives of
other FAs were less active than the reference miltefosine.
80
70
60
50
40
30
20
10
0
3a 3b 3c 3d 4a 4b 4c 4d 4e 4f 5a 5b 5c 5d 5e 5f 6a 6b 6c 6d 6e 6f Ref.
Palmitoyl
Palmitoleoyl
oleoyl
Erucoyl
Linoleoyl
α-Linolenoyl
Miltefosine
Figure 4. Percentages inhibition of HepG2 cell line proliferation after treatment with synthesized
3-deoxysphingomyelin analogs and miltefosine at 100 μM.
During the last decades, skin cancer mortality rates did not improve despite the progress in
cancer’s biology research and the development of novel antitumor agents [17]. In A-431 cell
line (a cell line model of epidermal squamous cells carcinoma), it was found that evolvement
of constitutively active Akt results in acquired resistance to the EGFR inhibitor erlotininb
[18]. As prepared 3-deoxysphingomyelin analogs possess structural similarity with
alkylphospholipids, which are known to influence Akt phosphorylation, we investigated the
antiproliferative activity of the prepared 3-deoxysphingomyelin analogs against A-431 cell
line. The elicited pattern of activity by the prepared 3-deoxysphingomyelin analogs against
A-431 cell line was similar to the pattern of activity elicited against other evaluated cell lines.
Thus, derivatives of palmitic acid and erucic acid elicited higher activity than the

corresponding derivatives of other FAs. Meanwhile, compounds possessing type 3
pyrrolidine core and the corresponding enantiomeric antipode compounds having type 4
pyrrolidine core showed almost similar potency. This confirms that the stereochemistry
inversion of the pyrrolidine ring has no significant impact on the elicited biological activity.
While miltefosine elicited only 35.4% growth inhibition the most active compounds which
were palmitic acid derivatives 6a and 5a produced 82.5% and 74.8% growth inhibition
respectively (Fig. 5). The highest activity of palmitic acid and erucic acid derivatives were for
those derivatives with (R)-3-pyrrolidinol core rather than other pyrrolidine cores. With
exception of oleic acid derivative 5c, all FAs derivatives other than palmitic acid and erucic
acid showed growth inhibition percent lower than 50%. Collectively, the data of the activity
of the synthesized 3-deoxysphingomyelin analogs against the employed cell lines indicate
that all of the designed pyrrolidine cores can provide antiproliferative 3-deoxysphingomyelin
analogs with significant activity when it is substituted the appropriate fatty acid. In addition,
the stereochemistry configuration of the pyrrolidine ring has minimal impact on the elicited
biological activity whereas the antiproliferative activity is highly impacted by the type of FA.
It was found that palmitic and erucic acids derivative yield the most effective antiproliferative
3-deoxysphingomyelin analogs. Similar to miltefosine, the activity of most of the synthesized
3-deoxysphingomyelin analogs were more pronounced against MCF-7 cells than other cell
lines.

90
80
70
60
50
40
30
20
10
0
Palmitoyl
Palmitoleoyl
oleoyl
Erucoyl
Linoleoyl
α-Linolenoyl
Miltefosine
Figure 5. Percentages inhibition of A-431 cell line proliferation after treatment with synthesized
3-deoxysphingomyelin analogs and miltefosine at 100 μM.
2.2.2. Inhibition of Akt phosphorylation assay
As mentioned in the introduction, antitumor alkylphospholipids, which are structurally
similar to the prepared 3-deoxysphingomyelin analogs, inhibit Akt phosphorylation, which
can induce cell death. Accordingly, we anticipated that inhibition of Akt phosphorylation
might be the mechanism of action of the designed antiproliferative 3-deoxysphingomyelin
analogs, at least in part. Therefore, cellular assay of the Akt phosphorylation inhibitory
activity of the prepared 3-deoxysphingomyelin analogs, as well as, miltefosine as a reference
compound was performed using MCF-7 cell line to check for the validity of the assumption
that Akt phosphorylation inhibition is the mechanism of action of the designed
antiproliferative
3-deoxysphingomyelin
analogs.
Among
the
synthesized
3-deoxysphingomyelin analogs, three compounds; 4a, 5a and 6a were selected. These
selected compounds have variable pyrrolidine rings, but with identical alkyl chains. The GI50
value of compound 4a was almost similar to the reference standard miltefosine, while those
for 5a and 6a were higher. However, at a high dose of 100 µM, all of the three compounds

elicited higher antiprolifertive activity against MCF-7 cell line than that of the standard
miltefosine. In the course of the experiment, MCF-7 cells pretreated with compounds to be
evaluated were stimulated with insulin to induce Akt phosphorylation followed by cell lysis
and determination of the amount of phosphorylated Akt. As Table 2 shows, the standard
reference miltefosine elicited 100% inhibition of Akt phosphorylation at 50 µM concentration
with IC50 value of 20.0 µM. On the other hand, compound 4a elicited incomplete inhibition
of Akt phosphorylation measured at 94.2% inhibition of at the same 50 µM concentration.
However, compound 4a exhibited IC50 value for inhibition of Akt phosphorylation close to
that of the reference miltefosine. In fact, both of compound 4a and miltefosine also possessed
close GI50 values for inhibition of growth of MCF-7 cell line despite the higher percent
inhibition of growth elicited by compound 4a at the higher concentrations. This might
indicate that the evaluated 3-deoxysphingomyelin analogs elicits a concentration dependent
mechanism of action in which the effective mechanism of action inducing the cell
cytotoxicity might differ depending on the concentration. It seems that these compounds at
the high concentrations might be unselective membrane-disrupting detergents. This
conclusion is supported by the assay results of compounds 5a and 6a as the results show that
these compounds are less efficient inhibitors of Akt phosphorylation (Table 2). Yet, these
compounds produces higher percent for inhibition of MCF-7 cells’ growth than miltefosine at
the higher concentrations despite their higher GI50 values relative to the reference miltefosine.
To confirm the relation between the mechanism of action and the concentration, the
correlation constants were calculated. As shown in Table 2, the statistical analysis of the
correlation of the GI50 of compounds 4a, 5a, 6a and miltefosine; to the IC50 value of Akt
phosphorylation inhibition elicited by these compounds showed high correlation value.
However, the correlation between the percent inhibitions of growth elicited by the same set of
compounds at the higher 100 µM concentration and the IC50 value of Akt phosphorylation

was very low. This might confirm that the mechanisms contributing to the elicited
cytotoxicity at this high concentration might arise from unselective membrane-disruption by
these compounds. In summary, these class of compounds might possess concentration-
dependent mechanism for their antiproliferative activity where the inhibition of Akt
phosphorylation might the mechanism of action for these compounds at the lower
concentrations, while other mechanism (possibly membrane-disruption) might possibly
contribute to the elicited cytotoxicity at higher concentration.
Table 2. Akt phosphorylation inhibitory activities of selected 3-deoxysphingomyelin analogs and their in silico
estimated energetic terms for the best Akt-docked pose.
MCF-7 Growth
% Inhibitiona
GI50
29.6
Akt Phosphorylation
Interaction
Energy
(kcal/mol)
Full Energy Estimated ΔG
Comp
(kcal/mol)
(kcal/mol)
% Inhibitionc
IC50
b
d
4a
5a
87.5
74.6
80.9
69.4
94.2
93.6
73.3
100
21.1
26.4
32.5
20.0
-64.27
-58.34
-58.72
-69.27
-2479.07
-2490.84
-2476.56
-2480.24
-9.43
-9.02
-8.07
-8.43
42.2
51.1
28.4
6a
Miltefosine
Correl. with Akt
0.1704
0.9936
-0.9298
1.0000
0.8419
NCf
NCf
Phosphophorylatione
a % Inhibition at 100 μM.
b GI50 was defined as the concentration resulting in 50% inhibition of cell proliferation.
c % Inhibition of AKT phosphorylation at 50 µM in MCF-7.
d IC50 was defined as the concentration resulting in 50% inhibition of AKT phosphorylation in MCF-7.
e The calculated correlation constant with IC50 values of Akt phophorylation; f Value was not calculated.
2.2.3. In silico simulation study
Akt is a protein kinase that incorporates a Pleckstrin homology domain (PH domain) which is
responsible for binding lipids. The structure of PH domain consists of two antiparallel β-
sheets and a terminal α-helix. Despite the fact that the PH domain exists across several
kinases, the variation of loops’ length and sequence would accounts for the selectivity for
binding different lipids. Besides, a single tryptophan in the α-helix is the only conserved
amino acid among the whole sequence of the PH domain.

To understand how the tested 3-deoxysphingomyelin analogs might interacts with Akt, a
molecular modeling study was conducted using the reported crystal structure of Akt. As
mentioned in the introduction, there is a structural similarity between the tested
3-deoxysphingomyelin analogs and alkylphospholipids. In addition, miltefosine which is a
synthetic member of alkylphospholipids was used as a reference standard in the conducted
biological evaluations. Therefore, the 3-deoxysphingomyelin analogs 4a, 5a, 6a, as well as,
miltefosine were docked blindly to the PH domain of the crystal structure of Akt (PDB code:
3O96) after appropriate preparation employing the SwissDock algorithm that implements
31
CHARMM force field [19, 20]. A previous report utilized high resolution field-cycling P
NMR along with docking study showed that miltefosine binds the PH domain in a site other
than the site that binds phosphoinositides, but located in the vicinity of it [21]. In addition, it
showed that the positively charged head is directed towards Tyr18 and Glu17 and interacting
with Glu17. Considering the whole structure of Akt, this site is adjacent to a hydrophobic
site, which is the binding site of some allosteric inhibitors such as shown in Fig. 6. This
hydrophobic site connects the PH domain with the kinase domain and, thus, maintains Akt in
a closed configuration [22]. This might offer a possibility for the hydrophobic tails of
phospholipidic inhibitors of Akt to extend into this site after the ionic head interacts with the
crucial Glu17 [21]. In lieu of the above-mentioned reports, analysis was performed for the
clustered poses obtained from the in silico simulation considering the binding sites,
interactions and their calculated energy terms. In the course of the analysis, only clusters
showing interaction with Glu17 were considered as this interaction is crucial for binding of
phosphoinositides and also for triggering inhibition of Akt by miltefosine and other
alkylphospholipids [21]. As shown in Table 2, the calculated interaction energies for the best-
selected poses of compounds 4a, 5a, 6a, and miltefosine showed excellent correlation with

the experimentally determined IC50 values for inhibition of Akt phosphorylation (calculated
correlation value equals to 0.8419).
N
O
N
N
N
N
NH
N
H
Figure 6. Structure of the reported cocrystallized allosteric small molecules inhibitors (PDB code: 3O96).
As illustrated in Fig. 7, the positively charged trimethylammonium moiety of miltefosine’s
head fits in a pocket within the PH domain eliciting polar electrostatic and hydrogen bonding
interactions with Glu17 as previously reported. Also, the negatively charged phosphate
moiety of the head of miltefosine interacts via polar electrostatic and hydrogen bonding
interactions with Arg273 as well as hydrogen bonding interaction with Asn54. Meanwhile,
the hydrophobic alkyl chain, as anticipated, extends into the hydrophobic pocket occupied by
the hydrophobic small molecule inhibitor. The alkyl chain partially overlaps with the
imidazoquinoxaline moiety of the co-crystallized small molecule inhibitor and elicits
hydrophobic interaction with Trp80. As shown in Table 2, the in silico calculation estimated
the interaction between miltefosine and Akt to be stabilized by -69.27 kcal/mol and the
ligand-enzyme complex to be stabilized by -2480.24 kcal/mol while the estimated free energy
to be -8.43 kcal/mol.
As shown in Fig. 7, the predicted binding modes for compounds 4a and 5a were comparable
to that of miltefosine, however, with different energetic terms. Thus, both of the polar heads
of compounds 4a and 5a elicited polar electrostatic and hydrogen bonding interactions via the
trimethylammonium moiety with Glu17, polar electrostatic and hydrogen bonding
interactions between the phosphate moiety with Arg273, as well as, hydrogen bonding

interaction with Asn54. In addition, Tyr326 is involved in two hydrogen bonding interactions
with both of the positively charged and the negatively charged moieties of the polar heads of
compounds 4a and 5a. Both of the alkyl chains of compounds 4a and 5a overlap with the
imidazoquinoxaline moiety of the co-crystallized small molecule inhibitor and elicit
hydrophobic interaction with Trp80. While the substituents’ position and the stereochemistry
are different between compounds 4a and 5a, this did not influence the binding of the polar
head nor the orientation of the hydrophobic tail. This might explains why the elicited
biological activity is more influenced by the alkyl chain type than by the pyrrolidine ring.
Despite binding modes similarity, the calculated energetic terms were different between
compounds 4a and 5a. For compound 4a, the interaction energy was significantly higher than
compounds 5a, yet, lower than that for miltefosine. However, the estimated free energy was
higher than miltefosine and the ligand-enzyme complex energy was very close to that of
miltefosine (Table 2). This might explain the close IC50 values for inhibition of Akt
phosphorylation of compound 4a and miltefosine. However in case of compound 5a, the
estimated interaction energy and the estimated free energy were further decreased than that
for 4a. This might explain the lowered potency of compound 5a relative to that of compound
4a and miltefosine with regard to inhibition of Akt phosphorylation. Although the polar head
of docked compound 6a elicits similar interactions to compounds 4a, 5a and miltefosine, the
alkyl chain tail is out of the plane of those of compounds 4a, 5a and miltefosine and,
accordingly, not overlapping with imidazoquinoxaline moiety of the co-crystallized small
molecule inhibitor. Consequently, the in silico calculations predicted that compound 6a
possesses lower interaction energy, less ligand-enzyme complex energy and free energy than
those for compounds 4a, 5a and miltefosine (Table 2). This would explain the lower IC50
value for inhibition of Akt phosphorylation of compound 6a.

Figure 7. Overlapping Akt-Docked poses of compounds 4a, 5a 6a and Miltefosine with the co-crystallized
small molecule inhibitor (PDB code: 3O96). Blue color is Miltefosine, orange color is 4a, pink color is 5a, gray
color is 6a while yellow color for the co-crystallized small molecule inhibitor.
3. Conclusion
In this study, twenty two novel pyrrolidine-based 3-deoxysphingomyelin analogs belonging
to four series were designed, synthesized, evaluated for their antiproliferative activity, and
were mechanistically studied employing three highly active compounds out of the screened
set of compounds. The mechanistic study was done using cellular assay of inhibition of Akt
phosphorylation, in addition to molecular docking study to get insights into the factors that
underlie the measured Akt phosphorylation inhibition. The designed 3-deoxysphingomyelin
analogs elicit antiproliferative activity in analogy to the structurally similar
alkylphospholipids but in contrast to the reported cytoprotective activity of previously
developed sphingomyelin analogs. The results showed that the type of the fatty acid moiety
in the structure of these compounds has more influence on the biological activity than the
variation of the pyrrolidine core. Among evaluated fatty acid, pyrrolidine-based
3-deoxysphingomyelin possessing palmitic or erucic acid were the most active. Similar to the
FDA approved reference drug miltefosine, the prepared 3-deoxysphingomyelin analogs were
more active against breast cancer cell line than other cell line. Some of palmitic or erucic acid

derivatives were more active than miltefosine against MCF-7 cell line, as well as, other cell
lines. Particularly, 3-deoxysphingomyelin analogs 3d, 4d and 6d were the most potent with
cellular GI50 ranges of 15.7~24.8 μM against breast cancer which is lower than that of
miltefosine which was 28.4 μM. Cellular assay of Akt-phosphorylation inhibition for the
studied compounds revealed that these class of compounds elicits a concentration dependent
mechanism of action where Akt phosphorylation might be the prevalent mechanism of action,
whereas, at high concentrations other mechanisms of action (possibly membrane-disruption
because of a detergent-like property for these compounds) might contribute to the elicited
cytotoxicity.
4. Experimental
4.1. Chemistry:
Unless otherwise specified, all reactions and manipulations were performed under nitrogen
atmosphere using standard Schlenk techniques. Solvent were purified and dried prior to use.
For TLC, precoated silica gel (E. Merck Kiesegel 60F254, layer thickness 0.25 mm) was
used. For column chromatography Merck Kiesegel 60 Art 9385 (230-400 mesh) was used.
NMR spectra were recorded on Bruker AC 400 spectrometer or Agilent 500 spectrometer
1
operating at 400 MHz/100 MHz or 500 MHz/125 MHz for H NMR and 13C NMR
respectively. Chemical shifts (δ) are reported in ppm, downfield from internal TMS standard.
For HRMS data, Jeol accuTOF (JMS-T100TD) with DART ion source (Ionsense, Tokyo,
Japan) was employed.
4.1.1. General procedure for the preparation of acid chlorides 8a–f [11].
To a stirred solution of the fatty acid (1 eq., 0.78 mmol) and a drop of DMF in dry
dichloromethane (5 ml) was added dropwise oxalyl chloride (3 eq., 297 mg, 2.34 mmol) at

0ºC. The reaction mixture was warmed slowly up to rt and stirred for further 3 h. Solvent and
excess oxalyl chloride were removed in vacuo and the resulting acid chloride was used next
step without purification.
4.1.2. General procedure for the preparation of intermediate compounds 9a–d, 11a–f, 13a–f
and 15a–f.
To a solution of the appropriate pyrrolidine derivative (1.17 mmol) and triethylamine (237
mg, 2.34 mmol) in dry DMF (4 ml) was added slowly a solution of the appropriate acid
chloride (0.78 mmol) in dry DMF (1 ml) at rt. After stirring for 5 h, the reaction mixture was
diluted with EtOAc (100 ml) and washed by water (100 ml×3). The combined water layer
was washed again with EtOAc. The combined organic layer was washed by brine, dried over
anhydrous MgSO4 and concentrated. The residue was purified by flash column
chromatography (EtOAc/n-hexane = 1:5)
(R)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)hexadecan-1-one (9a):
Compound 9a was obtained from 7 and palmitoyl chloride (8a) according to the general
procedure. Yield 84%. 1H NMR (CDCl3) δ 4.24–4.19 (m, 1H), 3.65 (dd, 1H, J1 = 11.2 Hz, J2
= 2.5 Hz), 3.59–3.43 (m, 3H), 2.30 (t, 2H, J = 7.8 Hz), 2.08–1.99 (m, 1H), 1.97–1.91 (m,
1H), 1.89–1.81 (m, 1H), 1.68–1.54 (m, 3H), 1.37–1.20 (m, 24H), 0.88 (t, 3H, J = 7.0 Hz).
(R,Z)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)hexadecan-9-en-1-one (9b):
Compound 9b was obtained from 7 and palmitoleoyl chloride (8b) according to general
procedure. Yield 86%. 1H NMR (CDCl3) δ 5.38–5.30 (m, 2H), 5.18 (brs, 1H), 4.25–4.19 (m,
1H), 3.66 (d, 1H, J = 11.5 Hz), 3.58–3.42 (m, 3H), 2.30 (t, 2H, J = 7.8 Hz), 2.09–1.97 (m,
5H), 1.96–1.80 (m, 2H), 1.66–1.53 (m, 3H), 1.38–1.22 (m, 16H), 0.88 (t, 3H, J = 7.1 Hz).

(R,Z)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)octadecan-9-en-1-one (9c):
Compound 9c was obtained from 7 and oleoyl chloride (8c) according to general procedure.
1
Yield 78%. H NMR (CDCl3) δ 5.39–5.30 (m, 2H), 5.22 (dd, 1H, J1 = 7.6 Hz, J2 = 1.8 Hz),
4.26–4.19 (m, 1H), 3.69–3.64 (m, 1H), 3.59–3.42 (m, 3H), 2.30 (t, 2H, J = 7.4 Hz), 2.09–1.97
(m, 5H), 1.96–1.83 (m, 2H), 1.65 (t, 2H, J = 6.7 Hz), 1.62–1.53 (m, 1H), 1.30 (d, 20H, J =
20.4 Hz), 0.88 (t, 3H, J = 6.6 Hz).
(R,Z)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)docos-13-en-1-one (9d):
Compound 9d was obtained from 7 and erucoyl chloride (8d) according to general procedure.
1
Yield 76%. H NMR (CDCl3) δ 5.39–5.31 (m, 2H), 5.23 (d, 1H, J = 6.7 Hz), 4.25–4.19 (m,
1H), 3.69–3.64 (m, 1H), 3.59–3.43 (m, 3H), 2.30 (t, 2H, J = 7.5 Hz), 2.08–1.97 (m, 5H),
1.96–1.83 (m, 2H), 1.65 (t, 2H, J = 7.3 Hz), 1.62–1.53 (m, 1H), 1.37–1.23 (m, 28H), 0.88 (t,
3H, J = 6.5 Hz).
(S)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)hexadecan-1-one (11a):
Compound 11a was obtained from 10 and palmitoyl chloride (8a) according to the general
procedure. Yield 83%. 1H NMR (CDCl3) δ 5.21 (brs, 1H), 4.24–4.18(m, 1H), 3.65 (d, 1H, J =
11.1 Hz), 3.59–3.43 (m, 3H), 2.30 (t, 2H, J = 7.5 Hz), 2.08–1.99 (m, 1H), 1.98–1.91 (m, 1H),
1.89–1.81 (m, 1H), 1.68–1.55 (m, 3H), 1.35–1.20 (m, 24H), 0.88 (t, 3H, J = 7.0 Hz).
(S,Z)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)hexadecan-9-en-1-one (11b):
Compound 11b was obtained from 10 and palmitoleoyl chloride (8b) according to general
procedure. Yield 89%. 1H NMR (CDCl3) δ 5.38–5.30 (m, 2H), 5.18 (dd, 1H, J1 = 7.8 Hz, J2 =
2.1 Hz), 4.26–4.19 (m, 1H), 3.66 (ddd, 1H, J1 = 11.04 Hz, J2 = 7.7 Hz, J3 = 2.5 Hz), 3.59–

3.42 (m, 3H), 2.30 (t, 2H, J = 7.4 Hz), 2.09–1.97 (m, 5H), 1.96–1.80 (m, 2H), 1.67–1.53 (m,
3H), 1.38–1.25 (m, 16H), 0.88 (t, 3H, J = 7.1 Hz).
(S,Z)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)octadecan-9-en-1-one (11c):
Compound 11c was obtained from 10 and oleoyl chloride (8c) according to general procedure.
1
Yield 72%. H NMR (CDCl3) δ 5.39–5.30 (m, 2H), 5.17 (dd, 1H, J1 = 7.6 Hz, J2 = 2.0 Hz),
4.26–4.19 (m, 1H), 3.69–3.64 (m, 1H), 3.59–3.43 (m, 3H), 2.30 (t, 2H, J = 7.4 Hz), 2.09–1.97
(m, 5H), 1.96–1.83 (m, 2H), 1.65 (t, 2H, J = 6.9 Hz), 1.61–1.53 (m, 1H), 1.29 (d, 20H, J =
20.5 Hz), 0.88 (t, 3H, J = 6.6 Hz).
(S,Z)-1-(2-(Hydroxymethyl)pyrrolidin-1-yl)docos-13-en-1-one (11d):
Compound 11d was obtained from 10 and erucoyl chloride (8d) according to general
procedure. Yield 82%. 1H NMR (CDCl3) δ 5.39–5.31 (m, 2H), 5.18 (dd, 1H, J1 = 7.8 Hz, J2 =
2.2 Hz), 4.26–4.19 (m, 1H), 3.69–3.64 (m, 1H), 3.59–3.43 (m, 3H), 2.30 (t, 2H, J = 7.4 Hz),
2.08–1.98 (m, 5H), 1.97–1.80 (m, 2H), 1.68–1.53 (m, 3H), 1.35–1.22 (m, 28H), 0.88 (t, 3H, J
= 6.3 Hz).
(9Z,12Z)-1-(S)-2-(Hydroxymethyl)pyrrolidin-1-yl)octadeca-9,12-dien-1-one (11e):
Compound 11e was obtained from 10 and linoleoyl chloride (8e) according to general
procedure. Yield 89%. 1H NMR (CDCl3) δ 5.37–5.31 (m, 4H), 5.28 (dd, 1H, J1 = 7.7 Hz, J2 =
2.2 Hz), 4.25–4.19 (m, 1H), 3.69–3.64 (m, 1H), 3.60–3.44 (m, 3H), 2.77 (t, 2H, J = 6.4 Hz),
2.31 (t, 2H, J = 7.8 Hz), 2.07–2.00 (m, 5H), 1.98–1.83 (m, 2H), 1.68–1.54 (m, 3H), 1.32–1.22
(m, 14H), 0.89 (t, 3H, J = 7.0 Hz).
(9Z,12Z,15Z)-1-((R)-2-(Hydroxymethyl)pyrrolidin-1-yl)octadeca-9,12,15-trien-1-one (11f):

Compound 11f was obtained from 10 and α-linolenoyl chloride (8f) according to general
procedure. Yield 78%. 1H NMR (CDCl3) δ 5.41–5.33 (m, 6H), 5.32–5.27 (m, 1H), 4.25–4.19
(m, 1H), 3.69–3.64 (m, 1H), 3.60–3.43 (m, 3H), 2.84–2.78 (m, 4H), 2.30 (t, 2H, J = 7.8 Hz),
2.07–2.00 (m, 5H), 1.98–1.82 (m, 2H), 1.68–1.54 (m, 3H), 1.38–1.23 (m, 8H), 0.89 (t, 3H, J
= 6.7 Hz).
(S)-1-(3-(Hydroxymethyl)pyrrolidin-1-yl)hexadecan-1-one (13a):
Compound 13a was obtained from 12 and palmitoyl chloride (8a) according to general
procedure. Yield 90%. 1H NMR (CDCl3) δ 3.69–3.50 (m, 4H), 3.46–3.36 (m, 1H), 3.30–3.20
(m, 1H), 2.45 (doublet of septets, 1H, J1 = 41.7 Hz, J2 = 7.0 Hz), 2.24 (quartet, 2H, J = 8.4
Hz), 2.11–1.92 (m, 1H), 1.86–1.57 (m, 3H), 1.40–1.09 (m, 24H), 0.88 (t, 3H, J = 7.0 Hz).
(S,Z)-1-(3-(Hydroxymethyl)pyrrolidin-1-yl)hexadec-9-en-1-one (13b):
Compound 13b was obtained from 12 and palmitoleoyl chloride (8b) according to general
procedure. Yield 90%. 1H NMR (CDCl3) δ 5.38–5.30 (m, 2H), 3.69–3.50 (m, 4H), 3.46–3.37
(m, 1H), 3.30–3.20 (m, 1H), 2.45 (doublet of septets, 1H, J1 = 41.4 Hz, J2 = 7.0 Hz), 2.25
(quartet, 2H, J = 8.3 Hz), 2.11–1.93 (m, 5H), 1.85–1.58 (m, 3H), 1.37–1.24 (m, 16H), 0.88 (t,
3H, J = 7.0 Hz).
(S,Z)-1-(3-(Hydroxymethyl)pyrrolidin-1-yl)octadec-9-en-1-one (13c):
Compound 13c was obtained from 12 and oleoyl chloride (8c) according to general procedure.
1
Yield 91%. H NMR (CDCl3) δ5.38–5.30 (m, 2H), 3.76 (brs, 1H), 3.69–3.50 (m, 4H), 3.46–
3.37 (m, 1H), 3.30–3.19 (m, 1H), 2.45 (doublet of septets, 1H, J1 = 41.5 Hz, J2 = 7.0 Hz),
2.24 (quartet, 2H, J = 8.2 Hz), 2.11–1.92 (m, 5H), 1.85–1.57 (m, 3H), 1.29 (d, 20H, J = 14.9
Hz), 0.88 (t, 3H, J = 7.0 Hz).

(S,Z)-1-(3-(Hydroxymethyl)pyrrolidin-1-yl)docos-13-en-1-one (13d):
Compound 13d was obtained from 12 and erucoyl chloride (8d) according to general
procedure. Yield 57%. 1H NMR (CDCl3) δ 5.38–5.31 (m, 2H), 3.71–3.50 (m, 4H), 3.46–3.38
(m, 1H), 3.30–3.21 (m, 1H), 2.55–2.36 (m, 2H), 2.24 (quartet, 2H, J = 8.0 Hz), 2.11–1.93 (m,
5H), 1.86–1.59 (m, 3H), 1.38–1.22 (m, 28H), 0.88 (t, 3H, J = 7.1 Hz).
(9Z,12Z)-1-((S)-3-(Hydroxymethyl)pyrrolidin-1-yl)octadeca-9,12-dien-1-one (13e):
Compound 13e was obtained from 12 and linoleoyl chloride (8e) according to general
procedure. Yield 62%. 1H NMR (CDCl3) δ 5.42–5.28 (m, 4H), 3.71–3.50 (m, 4H), 3.47–3.36
(m, 1H), 3.30–3.21 (m, 1H), 2.77 (t, 2H, J = 6.4 Hz), 2.61 (brs, 1H), 2.50–2.37 (m, 1H), 2.25
(quartet, 2H, J = 6.5 Hz), 2.12–1.93 (m, 5H), 1.85–1.58 (m, 3H), 1.39–1.24 (m, 14H), 0.89 (t,
3H, J = 6.9 Hz).
(9Z,12Z,15Z)-1-((S)-3-(Hydroxymethyl)pyrrolidin-1-yl) octadeca-9,12,15-trien-1-one (13f):
Compound 13f was obtained from 12 and α-linolenoyl chloride (8f) according to general
procedure. Yield 58%. 1H NMR (CDCl3) δ 5.43–5.28 (m, 6H), 3.69–3.50 (m, 4H), 3.46–3.37
(m, 1H), 3.30–3.20 (m, 1H), 2.80 (t, 4H, J = 5.6Hz), 2.45 (doublet of septets, 1H, J1 = 41.4
Hz, J2 = 7.0 Hz), 2.25 (quartet, 2H, J = 8.3 Hz), 2.11–1.92 (m, 5H), 1.85–1.58 (m, 3H), 1.38–
1.25 (m, 8H), 0.98 (t, 3H, J = 7.6 Hz).
(R)-1-(3-Hydroxypyrrolidin-1-yl)hexadecan-1-one (15a):
Compound 15a was obtained from 14 and palmitoyl chloride (8a) according to general
1
procedure. Yield 74%. H NMR (CDCl3) δ 4.50 (d, 1H, J = 18.9 Hz), 3.67–3.41 (m, 4H),

2.87 (brs, 1H), 2.28–2.20 (m, 2H), 2.06–1.90 (m, 2H), 1.63 (quintet, 2H, J = 7.3 Hz), 1.36–
1.20 (m, 24H), 0.88 (t, 3H, J = 7.2 Hz).
(R,Z)-1-(3-Hydroxypyrrolidin-1-yl)hexadec-9-en-1-one (15b):
Compound 15b was obtained from 14 and palmitoleoyl chloride (8b) according to general
1
procedure. Yield 82%. H NMR (CDCl3) δ 5.38–5.30 (m, 2H), 4.49 (d, 1H, J = 21.3 Hz),
3.67–3.42 (m, 4H), 3.26 (brs, 1H), 2.28–2.21 (m, 2H), 2.06–1.93 (m, 6H), 1.63 (t, 2H, J = 7.0
Hz), 1.38–1.23 (m, 16H), 0.88 (t, 3H, J = 6.8 Hz).
(R,Z)-1-(3-Hydroxypyrrolidin-1-yl)octadec-9-en-1-one (15c):
Compound 15c was obtained from 14 and oleoyl chloride (8c) according to general procedure.
1
Yield 72%. H NMR (CDCl3) δ 5.39–5.29 (m, 2H), 4.88 (brs, 1H), 4.52 (m, 1H), 3.69–3.39
(m, 4H), 2.30–2.17 (m, 2H), 2.08–1.88 (m, 6H), 1.72–1.59 (m, 2H), 1.37–1.22 (m, 20H), 0.88
(t, 3H, J = 7.1 Hz).
(R,Z)-1-(3-Hydroxypyrrolidin-1-yl)docos-13-en-1-one (15d):
Compound 15d was obtained from 14 and erucoyl chloride (8d) according to general
1
procedure. Yield 66%. H NMR (CDCl3) δ 5.38–5.31 (m, 2H), 4.48 (d, 1H, J = 19.4 Hz),
3.67–3.41 (m, 5H), 2.28–2.20 (m, 2H), 2.05–1.92 (m, 6H), 1.62 (t, 2H, J = 7.2 Hz), 1.37–1.22
(m, 28H), 0.88 (t, 3H, J = 7.0 Hz).
(9Z,12Z)-1-((R)-3-Hydroxypyrrolidin-1-yl)octadeca-9,12-dien-1-one (15e):
Compound 15e was obtained from 14 and linoleoyl acid chloride (8e) according to general
procedure. Yield 71%. 1H NMR (CDCl3) δ 5.42–5.30 (m, 4H), 5.26 (brs, 1H), 4.26–4.19 (m,
1H), 3.66 (d, 1H, J = 11.1 Hz), 3.59–3.43 (m, 3H), 2.77 (t, 2H, J = 6.3 Hz), 2.30 (t, 2H, J =

7.8 Hz), 2.09–2.01 (m, 5H), 1.97–1.81 (m, 2H), 1.65 (t, 2H, J = 6.8 Hz), 1.39–1.24 (m, 14H),
0.89 (t, 3H, J = 7.1 Hz).
(9Z,12Z,15Z)-1-((R)-3-Hydroxypyrrolidin-1-yl)octadeca-9,12,15-trien-1-one (15f):
Compound 15f was obtained from 14 and α-linolenoyl chloride (8f) according to general
1
procedure. Yield 79%. H NMR (CDCl3) δ 5.42–5.27 (m, 6H), 4.47 (d, 1H, J = 21.5 Hz),
4.11 (brs, 1H), 3.67–3.42 (m, 4H), 2.80 (t, 4H, J = 5.6 Hz), 2.28–2.21 (m, 2H), 2.11–1.99 (m,
5H), 1.97–1.92 (m, 1H), 1.62 (t, 2H, J = 7.1 Hz), 1.32 (brs, 8H), 0.97 (t, 3H, J = 7.5 Hz).
4.1.3. General procedure for the preparation of compounds 3a–d, 4a–f, 5a–f, and 6a–f.
To a stirred solution of the appropriate pyrrolidine-amide intermediate (9a–d, 11a–f, 13a–f,
or 15a–f, 0.35 mmol) and triethylamine (147 mg, 1.05 mmol) in benzene (10 mL) was added
2-chloro-2-oxo-1,3,2-dioxaphospholane (156 mg, 0.35 mmol) at 0ºC and warmed up to rt.
After stirring for 5 h, the precipitated triethylamine hydrochloride was filtered off and washed
with benzene. The combined filtrate was concentrated, diluted with CH3CN (10 mL) and
transferred into pressure bottle. The mixture was cooled to -78ºC and treated with
trimethylamine (2 mL) at the same temperature under nitrogen. The reactor was closed and
heated at 65ºC for 15 h. After cooling to rt, the reactor was opened. The solvent and excess
TMA were evaporated and the residue was purified by silica gel column chromatography
(CHCl3/MeOH/H2O = 6:25:4).
(R)-(1-Palmitoylpyrrolidin-2-yl)methyl 2-(trimethylammonio)ethyl phosphate (3a):
1
Compound 3a was obtained from 9a according to general procedure. Yield 61%. H NMR
(400 MHz, CDCl3) δ 4.38 (brs, 2H), 4.28 (brs, 2H), 4.17–4.10 (m, 1H), 4.00–3.93 (m, 1H),
3.84–3.73 (m, 2H), 3.58–3.49 (m, 1H), 3.47–3.43 (m, 1H), 3.36 (s, 9H), 2.19 (t, 2H, J = 7.6

Hz), 2.08–1.98 (m, 2H), 1.93–1.82 (m, 2H), 1.62-1.50 (m, 2H), 1.25 (brs, 24H), 0.88 (t, 3H, J
= 6.8 Hz). 13C NMR (100 MHz, CDCl3) δ 172.2, 66.2, 64.1, 59.2, 57.0, 54.2 (3C), 47.2, 34.9,
31.9, 29.7 (5C), 29.6, 29.6, 29.5, 29.3, 27.2, 25.6, 24.8, 23.7, 22.7, 14.1; HRMS calcd for
C26H54N2O5P [M+H]+ m/z 505.3765, found m/z 505.3815.
(R,Z)-(1-Hexadec-9-enoylpyrrolidin-2-yl)methyl 2-(trimethylammonio)ethyl phosphate (3b):
1
Compound 3b was obtained from 9b according to general procedure. Yield 89%. H NMR
(400 MHz, CDCl3) δ 5.38–5.32 (m, 2H), 4.38–4.25 (m, 2H), 4.18–4.11 (m, 1H), 4.00–3.94
(m, 1H), 3.92–3.75 (m, 2H), 3.58–3.52 (m, 1H), 3.48–3.42 (m, 1H), 3.40 (s, 9H), 2.20 (t, 2H,
J = 7.6 Hz), 2.07–1.98 (m, 6H), 1.93–1.83 (m, 2H), 1.59–1.52 (m, 2H), 1.29 (brs, 18H), 0.88
(t, 3H, J = 6.8 Hz). 13C NMR (100 MHz, CDCl3) δ 172.2, 129.9, 129.7, 66.2, 64.1, 59.2, 57.1,
54.3 (3C), 47.2, 34.9, 31.8, 29.8 (2C), 29.7 (2C), 29.5, 29.4, 29.2, 28.9, 27.2, 24.8, 23.7, 22.6,
14.1; HRMS calcd for C26H52N2O5P [M+H]+ m/z 503.3608, found m/z 503.3649.
(R,Z)-(1-Oleoylpyrrolidin-2-yl)methyl 2-(trimethylammonio)ethyl phosphate (3c):
1
Compound 3c was obtained from 9c according to general procedure. Yield 87%. H NMR
(400 MHz, CDCl3) δ 5.38–5.33 (m, 2H), 4.35–4.20 (m, 2H), 4.18–4.10 (m, 1H), 4.01–3.95
(m, 1H), 3.90–3.76 (m, 2H), 3.59-3.49 (m, 1H), 3.40 (s, 9H), 3.38–3.30 (m, 1H), 2.20 (t, 2H,
J = 7.6 Hz), 2.05–1.98 (m, 6H), 1.93–1.81 (m, 2H), 1.62–1.50 (m, 2H), 1.28 (d, 20H, J = 11.0
Hz), 0.88 (t, 3H, J = 7.2 Hz). 13C NMR (100 MHz, CDCl3) δ 172.2, 129.9, 129.7, 66.2, 64.1,
59.2, 57.1, 54.2 (3C), 47.2, 34.9, 31.8, 29.7 (2C), 29.6, 29.5, 29.4, 29.4, 29.3 (2C), 29.2, 27.3,
27.2, 24.8, 23.7, 22.6, 14.1; HRMS calcd for C28H56N2O5P [M+H]+ m/z 531.3921, found m/z
531.3983.
(R,Z)-(1-Docos-13-enoylpyrrolidin-2-yl)methyl 2-(trimethylammonio)ethyl phosphate (3d):