Synthesis and biological evaluations of marine oxohexadecenoic acids: PPARα/γ dual agonism and anti-diabetic target gene effects

Article abstract of DOI:10.1016/j.ejmech.2018.06.034

Obesity and associated disorders such as metabolic syndrome and type 2 diabetes (T2D) have reached epidemic proportions. Several natural products have been reported as Peroxisome Proliferator-Activated Receptor (PPAR) agonists, functioning as lead compounds towards developing new anti-diabetic drugs due to adverse side effects of existing PPAR drugs. We recently isolated and identified (7E)-9-oxohexadec-7-enoic acid (1) and (10E)-9-oxohexadec-10-enoic acid (2) from the marine algae Chaetoceros karianus. Herein we report the total synthesis, pharmacological characterization, and biological evaluations of these naturally occurring oxo-fatty acids (oFAs). The syntheses of 1 and 2 afforded sufficient material for extensive biological evaluations. Both oFAs show an appreciable dose-dependent activation of PPARα and -γ with EC₅₀ values in the micromolar range, and an ability to regulate important PPAR target genes in hepatocytes and adipocytes. Moreover, both 1 and 2 are able to drive adipogenesis when evaluated in the Simpson-Golabi-Behmel syndrome (SGBS) pre-adipocyte cell model, but with lowered expression of adipocyte markers and reduced lipid accumulation compared to the drug rosiglitazone. This seems to be caused by a transient upregulation of PPARγ and C/EBPα expression. Importantly, whole transcriptome analysis shows that both compounds induce anti-diabetic gene programs in adipocytes by upregulating insulin-sensitizing adipokines and repressing pro-inflammatory cytokines.

Full text of DOI:10.1016/j.ejmech.2018.06.034

Accepted Manuscript
Synthesis and biological evaluations of marine oxohexadecenoic acids: PPARα/γ dual
agonism and anti-diabetic target gene effects
Thomas Sæther, Steinar M. Paulsen, Jørn E. Tungen, Anders Vik, Marius Aursnes,
Torgeir Holen, Trond V. Hansen, Hilde I. Nebb
PII:
S0223-5234(18)30525-7
10.1016/j.ejmech.2018.06.034
EJMECH 10504
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 4 May 2018
Revised Date: 7 June 2018
Accepted Date: 13 June 2018
Please cite this article as: T. Sæther, S.M. Paulsen, Jø.E. Tungen, A. Vik, M. Aursnes, T. Holen, T.V.
Hansen, H.I. Nebb, Synthesis and biological evaluations of marine oxohexadecenoic acids: PPARα/γ
dual agonism and anti-diabetic target gene effects, European Journal of Medicinal Chemistry (2018),
doi: 10.1016/j.ejmech.2018.06.034.
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.

ACCEPTED MANUSCRIPT
Synthesis and Biological Evaluations of Marine
Oxohexadecenoic Acids: PPARα/γ Dual Agonism
and Anti-Diabetic Target Gene Effects
1
2
3
3
3
Thomas Sæther *, Steinar M. Paulsen , Jørn E. Tungen , Anders Vik , Marius Aursnes , Torgeir
1
3
1
Holen , Trond V. Hansen and Hilde I. Nebb .
1
Department of Nutrition, Institute of Basic Medical Sciences, University of Oslo, N-0317 Oslo,
Norway.
2
MabCent-SFI, UiT The Arctic University of Norway, N-9037 Tromsø, Norway.
3
School of Pharmacy, Department of Pharmaceutical Chemistry, University of Oslo, N-0316
Oslo, Norway.
*
Corresponding Author. Telephone: + 47 22 85 15 10. E-mail: thomas.sather@medisin.uio.no.
1

ACCEPTED MANUSCRIPT
ABSTRACT (Max 200 words)
Obesity and associated disorders such as metabolic syndrome and type 2 diabetes (T2D) have
reached epidemic proportions. Several natural products have been reported as Peroxisome
Proliferator-Activated Receptor (PPAR) agonists, functioning as lead compounds towards
developing new anti-diabetic drugs due to adverse side effects of existing PPAR drugs. We
recently isolated and identified (7E)-9-oxohexadec-7-enoic acid (1) and (10E)-9-oxohexadec-10-
enoic acid (2) from the marine algae Chaetoceros karianus. Herein we report the total synthesis,
pharmacological characterization, and biological evaluations of these naturally occurring oxo-
fatty acids (oFAs). The syntheses of 1 and 2 afforded sufficient material for extensive biological
evaluations. Both oFAs show an appreciable dose-dependent activation of PPARα and -γ, with
EC values in the micromolar range, and an ability to regulate important PPAR target genes in
50
hepatocytes and adipocytes. Moreover, both 1 and 2 are able to drive adipogenesis when
evaluated in the Simpson-Golabi-Behmel syndrome (SGBS) pre-adipocyte cell model, but with
lowered expression of adipocyte markers and reduced lipid accumulation compared to the drug
rosiglitazone. This seems to be caused by a transient upregulation of PPARγ and C/EBPα
expression. Importantly, whole transcriptome analysis shows that both compounds induce anti-
diabetic gene programs in adipocytes by upregulating insulin-sensitizing adipokines and
repressing pro-inflammatory cytokines.
2

ACCEPTED MANUSCRIPT
Graphical abstract
PPARα
PPARγ
oxohexadecenoic acids
dual PPARα/γ agonists
adipocytes
inflammaꢀon↓
adiponectin↑
activation of
adipogenic
target genes
diabetes
lipogenesis ↗
β-oxidaꢀon↑
Highlights
•
•
•
•
•
We recently identified two isomeric oxohexadecenoic acids from a marine microalga.
Syntheses afforded sufficient material for extensive biological evaluations.
Both 1 and 2 are semi-potent dual PPARα/γ agonists with EC -values in the μM-range.
50
They induce anti-diabetic factors in adipocytes and β-oxidation in hepatocytes.
This makes 1 and 2 interesting molecular scaffolds for anti-diabetic drugs.
3

ACCEPTED MANUSCRIPT
1
. Introduction
Obesity and associated disorders, such as metabolic syndrome and type 2 diabetes (T2D),
constitute a serious health problem. The WHO estimates that there are more than 2 billion
overweight adults worldwide, half of whom are clinically obese [1]. Over 300 million people
suffer from T2D, a number that will continue to grow due to changes in dietary patterns and a more
sedentary lifestyle. Understanding the mechanisms of metabolic control in order to prevent and
treat these disorders is therefore a top research priority. Present anti-diabetic medications include
several drug classes, such as metformin, DPP-4 inhibitors, GLP-1 analogs/agonists, SGLT-2
inhibitors, as well as drugs targeting the peroxisome proliferator-activated receptors (PPARs).
PPARs are nuclear receptors that heterodimerize with the retinoid X receptor (RXR) and
regulate target gene expression in response to lipids such as unsaturated fatty acids,
phospholipids, eicosanoids, and oxygenated fatty acids [2]. Upon ligand activation, the PPARs
undergo conformational changes that facilitate the dissociation of transcriptional co-repressors
like SMRT and NCoR [3, 4], and recruitment of transcriptional co-activators and co-activator
complexes, which include factors such as p300/CBP, SRC-1, and PGC-1α [4, 5]. Additionally,
posttranslational modifications alter the structural conformation of the receptors thereby
modifying the affinity for co-regulators that determine whether a target gene is induced or
repressed. In humans the PPAR isoforms PPARα (NR1C1) and -γ (NR1C3) are mainly
expressed in liver and adipose tissue, respectively, while PPARδ (NR1C2) is more ubiquitously
expressed [6, 7]. When PPARγ is ligand-activated it induces adipocyte growth and
differentiation by transcriptionally regulating target genes involved in lipogenesis and lipid
storage [8-10]. Moreover, activation of PPARγ maintains normal insulin sensitivity through
upregulation and secretion of adipokines such as adiponectin and leptin from adipose tissue [11].
4

ACCEPTED MANUSCRIPT
In parallel, PPARα increases lipid uptake and energy expenditure in liver by upregulating gene
targets involved in fatty acid transport, activation and oxidation [12, 13]. We recently identified
the two isomeric oxo-fatty acids (7E)-9-oxohexadec-7-enoic acid (1) and (10E)-9-oxohexadec-
1
0-enoic acid (2) from the marine algae Chaetoceros karianus, and demonstrated that they
display dual PPAR α/γ agonist activity [14]. Several synthetic PPAR drugs are already in clinical
use, like the lipid-lowering fibrates, acting as PPARα activators, and the anti-diabetic
thiazolidinediones (TZDs), targeting PPARγ. However, some of the isoform-specific PPAR
agonists like clofibrate, rosiglitazone and pioglitazone have demonstrated adverse effects such as
hepatotoxicity [15, 16], pulmonary edema [17], myocardial infarction [18], weight gain [19],
reduced bone density [20], and bladder and prostate cancer [21, 22]. A suggested way out has
been to identify natural products as candidate compounds with potency for both PPARα and -γ,
aiming at maximizing the beneficial effects, while minimizing the adverse. Several promising dual
agonists display significant improvement in both glycemic as well as dyslipidemic parameters with
no evidence of conventional side effects [23-27]. In this aspect, natural products are of interest.
Some natural occurring fatty acid PPAR-agonists are depicted below.
Fig. 1. Structure of (7E)-9-oxohexadec-7-enoic acid
(
1) and (10E)-9-oxohexadec-10-enoic acid (2),
9Z,11E,13S)-13-Hydroxyoctadeca-9,11-dienoic
acid (13S-HODE), and (9S,10E,12Z)-9-
(
Hydroxyoctadeca-10,12-dienoic acid (9S-HODE).
In this paper we report the synthesis, pharmacological characterization, and extensive
biological evaluations of the oxo-fatty acids 1 and 2.
5

ACCEPTED MANUSCRIPT
2
. Results and discussion
2
.1. Chemistry
The synthesis towards (7E)-9-oxohexadec-7-enoic acid (1) started with known methyl 7-
hydroxyheptanoate (3) [28, 29] (Scheme 1). Swern oxidation of 3 afforded 4 in 81% yield. Next
the commercially available phosphorane 5 was treated with n-BuLi and 1-bromohexane to afford
the hexyl substituted intermediate of 5 that was reacted with aldehyde 4 in an E-selective Wittig-
reaction. This gave methyl (E)-9-oxohexadec-7-enoate (6) in 53% yield from 3. The hydrolysis
of 6 to 1 proved more challenging than anticipated as several conditions failed to produce the
natural product 1 due to the formation of polymeric material. However, the acid 1 was formed
when 6 was added to a phosphate buffer solution (pH = 7.2, 37 °C) containing lipase from
Pseudomonas cepacia. The spectral data confirmed the structure of 1 with the E-configuration (J
=
15.9 Hz).
The synthetic approach for making 1 was not amendable for making (10E)-9-oxohexadec-10-
enoic acid (2). The α,β-unsaturated ketone moiety in 2 was proven to be quite labile. Hence, an
alternative synthetic route was developed. First, known 7 [30] was transferred into its Grignard
reagent that was added to a THF-solution of commercially available (E)-2-octenal. This gave the
allylic alcohol 8. Hence, alcohol 8 was converted into the bis-TBS ether using standard
conditions (TBSCl, imidazole) that was mono-deprotected using aqueous TFA. This gave 9 that
was oxidized to the aldehyde using the Dess-Martin reaction. A Lindgren-Pinnick oxidation of
the formed aldehyde followed by deprotection of the TBS-group afforded 10. Oxidation of the
2
allylic alcohol was best achieved using electrolytically activated MnO that gave the natural
product 2, albeit in 15%. The spectral data confirmed the structure of 2. All efforts to improve
the yield of 2 were fruitless. The total yield of 2 was 4% over nine steps from 8-bromooctan-1-ol.
6

ACCEPTED MANUSCRIPT
Scheme 1. Reagents and conditions:
a) (COCl) , DMSO, Et N, CH Cl
1%; (b) n-BuLi, THF, -78 °C, 1–
(
2
3
2
2
,
8
bromohexane; (c) 4, THF, 53% (two
steps); (d) Lipase from Pseudomonas
cepacia, pH = 7.2, 64%; (e) TBSCl,
imidazole, MeCN, 90%; (f) Mg, I cat.,
2
THF; (g) (E)-2-octenal, THF, 70%
(
two steps); (h) TBSCl, imidazole,
DMF, 91%; (i) TFA, H O, THF, 65%;
j) Dess-Martin periodinane, NaHCO
CH Cl ; (k) NaClO , NaH PO , MeCN,
-methyl-2-butene, H O; (l) TBAF,
2
(
3
,
2
2
2
2
4
2
2
THF, 67% (three steps); (m) MnO2,
Celite, CH Cl , 15%.
2
2
2
.2. Dose-response and toxicity
We started by evaluating oxo-fatty acid (oFA) 1 and 2 with respect to PPARα and PPARγ
agonist activity using GAL4-LBD fusion constructs and GAL4-responsive Luciferase reporters
in COS-1 cells. The oFA activity profile was compared with rosiglitazone and pirinixic acid
(positive controls) and palmitic acid (negative control). Cytotoxicity was determined
simultaneously assaying for lactate dehydrogenase (LDH) in the cell media. In addition,
7

ACCEPTED MANUSCRIPT
cytotoxicity was assessed using standard XTT assays. The dose-response curves are shown in
Fig. 2 and summarized with corresponding EC -values in Table 1.
5
0
Fig. 2. The synthetic oxo-fatty acids (7E)-9-oxohexadec-7-enoic acid (1) and (10E)-9-oxohexadec-10-enoic acid
2) activate PPARα and -γ and show low cytotoxicity. Dose-response of 1 (7E)-9-OHE and 2 (10E)-9-OHE,
(
compared to pirinixic acid, rosiglitazone and palmitic acid on (A) PPARα activity and (B) PPARγ activity. The
assays were run with Gal4-DBD-NR-LBD chimeric constructs in COS-1 cells, using LBD from human PPARα and
PPARγ, respectively. (C) Corresponding cytotoxicity data were obtained by measuring lactate dehydrogenase
(
LDH) in media or (D) by running XTT assays. The results are shown as mean±SEM. The data represent three
biological replicates run in quadruplicates. RLU: relative light units.
Table 1. PPAR EC50 values and corresponding cytotoxicity of the oxo-fatty acids.
PPARα agonism
EC50 (µM)
PPARγ agonism
EC50 (µM)
Viability at EC50
LDH (%)
Viability at EC50
XTT (%)
Compound
1
20
26
10
8.5
n.d.
0.2
n.d.
>95
>95
>95
>95
>95
>90
>90
>95
>95
>90
2
Pirinixic acid
Rosiglitazone
Palmitic acid
39
n.d.
35
8

ACCEPTED MANUSCRIPT
As can be seen both oFAs show an appreciable PPAR agonist dose-response, with EC50
-
values in the micromolar range (Fig. 2; Table 1). At the same time the cell toxicity, as measured
by lactate dehydrogenase leakage (LDH) and reduced metabolic NAD(P)H flux (XTT), is none
to moderate in the same concentration range (10-30 µM). With respect to PPARα agonism, both
1
and 2 displayed EC -values and agonist activity comparable to pirinixic acid (WY-1464).
50
However, when assaying 1 and 2 in the PPARγ activity assay, their EC -values were
5
0
approximately 50 fold higher and their PPARγ agonist activities about 50% lower of that
observed with rosiglitazone (BRL 49653; Fig. 2).
Fig. 3. The oxo-fatty acids 1 and 2 display PPARα/γ
dual specificity. The assays were run with Gal4-DBD-
NR-LBD chimeric constructs in COS-1 cells, using the
ligand binding domain (LBD) from human nuclear
receptors. The doses used were 0, 25 and 50 µM of 1
(7E)-9-OHE and 2 (10E)-9-OHE, respectively. The
results are shown as mean±SEM. The data represent
three biological replicates run in quadruplicates. RLU:
relative light units. *p<0.05, **p<0.05, ***p<0.001
2
.3. Selectivity
We next addressed the selectivity of the two oxo-fatty acids with respect to nuclear receptor
agonist activity. To this end we transfected COS1 cells with GAL4-LBD fusions of human
PPARα, -δ, -γ, LXRα, -β, and RXRα, and treated them with increasing concentrations of the
oFAs. Both natural products were able to activate PPARα and -γ in the range of 25 to 50 µM,
9

ACCEPTED MANUSCRIPT
while neither of them activated the LXRs or RXRα (Fig. 3). A small but significant PPARδ
agonist activity was observed with 50 µM of 2.
To be able to study the agonist activity of 1 and 2 in a more relevant context we expressed
full-length human PPARα and –γ together with RXRα in COS-1 cells and assayed their
transactivity on human CPT1A- and PLIN1-driven luciferase reporters in the presence or absence
of the oxo-fatty acids. As can be seen in Fig. 4 both 1 and 2 were able to activate the PPARα-
target gene promoter CPT1A, while only 1 activated the promoter of PLIN1, a bona fide PPARγ
target gene. This activation was dependent of functional PPAR response elements (PPREs).
Fig. 4. The oxo-fatty acids 1 and 2 activate human
natural promoters in a PPAR/PPRE-depend manner.
COS-1 cells were transfected with full-length PPARα, the
heterodimerization partner RXRα and a CPT1A-driven
reporter, with or without a functional PPAR recognition
element (PPRE) or PPARγ, RXRα and a PLIN1-driven
reporter, with (black bars) or without (white bars) a
functional PPRE. The cells were then treated with 1 (7E)-9-
OHE or 2 (10E)-9-OHE for 18 hours. The results are shown
as mean±SEM. The data represent three biological
replicates run in quadruplicates. RLU: relative light units.
***p<0.001
2
.4. Activation of endogenous PPAR target genes
As both oxo-fatty acids displayed PPAR agonism in vitro, we asked whether 1 and 2 were
able to activate fully chromatinized PPAR target genes in the human hepatocellular carcinoma
10

ACCEPTED MANUSCRIPT
cell line Huh7. Huh7 cells express both PPARα and –γ [31] (Ct: 26.3 and 24.6, respectively at
baseline), but is often used in combination with PPARα agonist to assess regulation of liver-
specific PPARα target genes. While both 1 and 2 were able to induce CPT1A and ANGPTL4
expression, only 1 activated ACSL3 (Fig. 5). PLIN1 was only upregulated significantly by 2 but
to very low levels. We were surprised to find that ACOX1, a classical PPARα target gene did not
respond to any of the treatments. Several of the assayed target genes displayed a weak but
significant downregulation (Fig. 5). Whether this was due to cytotoxic effects were not studied
further.
Fig. 5. The oxo-fatty acids 1 and 2
activate endogenous target genes in
human hepatocarcinoma cells. Huh7
cells were stimulated with 50 µM of 1
(7E)-9-OHE, 2 (10E)-9-OHE or pirinixic
acid (PIRI) for 24 hours. Gene expression
was analysed by qPCR using specific
SYBR green primers. The results are
shown as mean±SEM. The data represent
3
biological replicates run in duplicates.
*p<0.05, **p<0.01, ***p<0.001
2
.5. Adipocyte differentiation
Turning our focus to adipocyte specific target gene activation, we took advantage of the
human Simpson-Golabi-Behmel syndrome (SGBS) pre-adipocyte cell line [32], and
11

ACCEPTED MANUSCRIPT
differentiated them into adipocytes using an 8 days differentiation protocol as described in the
experimental section. On day 8 we either stimulated the cells with 1 or 2, or rosiglitazone. As can
be seen in Fig. 6 both 1 and 2 induce the expression of CPT1A, while only 2 was able to
upregulate ANGPTL4 expression. The oFAs 1 and 2 were not able to induce classical adipocyte
markers, such as PLIN1, FABP4 and ADIPOQ. Even rosiglitazone failed to induce these genes
any further. The reason for this is probably that differentiated SGBS cells already have a high
expression of these genes. As an exception to this, UCP1 was induced 10 times compared to
baseline by rosiglitazone.
Fig. 6. The oxo-fatty acids 1 and 2
activate endogenous target genes in
human adipocytes. SGBS cells were
differentiated following standard protocol.
On day eight the cells were stimulated with
either 25 µM 1 (7E)-9-OHE, 25 µM 2
(10E)-9-OHE, or 2 µM rosiglitazone
(ROSI) for 24 hours. Gene expression was
analysed by qPCR using specific SYBR
green primers. The results are shown as
mean±SEM. The data represent three
biological replicates run in duplicates.
*p<0.05, **p<0.01, ***p<0.001
To address a potential role for 1 and 2 in adipocyte differentiation we compared the
adipogenic potential of both compounds to rosiglitazone, by exchanging the classical PPARγ
agonist with 1, 2 or DMSO the first 4 days of the differentiation protocol. On day 0, 4, 8 and 12
12

ACCEPTED MANUSCRIPT
cells were either harvested for RNA isolation, to monitor the expression of target genes, or
stained with Oil-Red-O to follow the inclusion of neutral lipids in lipid droplets.
Fig. 7. The oxo-fatty acids 1 and 2 drive adipocyte
differentiation, but to
a lesser extent than
rosiglitazone SGBS cells were differentiated in
Quickdiff and 3FC medium for 4 and 8 days,
respectively, following standard protocol. The
quickdiff medium was supplemented with either 25
µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE, or 2 µM
rosiglitazone. DMSO (0.025%) and BSA (4.2 µM)
were used as negative control. Adipocyte
differentiation was monitored by staining of neutral
lipids using Oil-Red-O. The images are representative
for three biological replicates run in duplicates.
Interestingly both 1 and 2 increased the number of cells with adipocyte-like, multilocular
morphology, as well as the total volume of lipid droplets (no. of droplets × size). As expected,
rosiglitazone was much more potent in inducing adipogenesis. Interestingly, the droplets formed
during treatment with 1 seemed to be slightly bigger than the ones seen with rosiglitazone and 2
(
Fig. 7, day 12). To evaluate the gene regulatory events underlying the morphological changes
we assayed 20 genes relating to adipogenesis (Fig. 8), lipid storage and metabolism (Fig. 9),
13

ACCEPTED MANUSCRIPT
Fig. 8. The oxo-fatty acids 1 and 2 induce an adipogenic gene program in SGBS cells. SGBS cells were
differentiated in Quickdiff and 3FC medium for 4 and 8 days, respectively, following standard protocol. The
quickdiff medium was supplemented with either 25 µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE, or 2 µM
rosiglitazone. DMSO or DMSO:BSA were used as control. Gene expression was analysed by qPCR using specific
SYBR green primers with RNA from cells harvested on day 0, 4, 8, and 12 (D0-D12). The results are shown as
mean±SEM. The data represent minimum 3 biological replicates run in duplicates. *p<0.05, **p<0.01, ***p<0.001
14

ACCEPTED MANUSCRIPT
Fig. 9. The oxo-fatty acids 1 and 2 induce gene involved in lipid metabolism and storage in SGBS cells. SGBS
cells were differentiated in Quickdiff and 3FC medium for 4 and 8 days, respectively, following standard protocol. The
quickdiff medium was supplemented with either 25 µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE, or 2 µM rosiglitazone.
DMSO or DMSO:BSA were used as control. Gene expression was analysed by qPCR using specific SYBR green
primers with RNA from cells harvested on day 0, 4, 8, and 12 (D0-D12). The results are shown as mean±SEM. The
data represent minimum three biological replicates run in duplicates. *p<0.05, **p<0.01, ***p<0.001
15

ACCEPTED MANUSCRIPT
Fig. 10. The oxo-fatty acids 1 and 2 induce adipokines and browning factors. SGBS cells were differentiated in
Quickdiff and 3FC medium for 4 and 8 days, respectively, following standard protocol. The quickdiff medium was
supplemented with either 25 µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE, or 2 µM rosiglitazone. DMSO or
DMSO:BSA were used as control. Gene expression was analysed by qPCR using specific SYBR green primers with
RNA from cells harvested on day 0, 4, 8, and 12 (D0-D12). The results are shown as mean±SEM. The data represent
minimum 3 biological replicates run in duplicates. *p<0.05, **p<0.01, ***p<0.001
and adipokine signalling and browning (Fig. 10) using quantitative PCR. Both 1 and 2 were able
to induce the classical adipogenic factors, PPARG, CEBPA and CEBPB (Fig. 8). The onset of the
adipogenic gene program is also reflected in the upregulation of PLIN1, FABP4, CD36 and
SCD1 which play important roles in fatty acid metabolism, transport and storage. This was seen
16

ACCEPTED MANUSCRIPT
with both oFAs (Fig. 9). However, the expression of most of these genes seem to be 5 to 10-fold
lower on day 8-12 after stimulation with 25 µM 1 and 2 than with rosiglitazone. This might be
related to what seems to be a transient induction of several important adipogenic driver genes
when treating the cells with 1 and 2 (Fig. 8). Here gene expression peaked at day 4, while the
rosiglitazone-treated cells are able to support a stable or increased expression of e.g. PPARG and
CEBPA for the whole duration of the experiment. Part of the same pattern can be seen in Fig. 10
with the adipokines and thermogenic factors. Still, it should be noted that both 1 and 2
significantly upregulate the expression of adiponectin (ADIPOQ) and the mitochondrial brown
fat uncoupling protein 1 (UCP1) over the 12 day differentiation period (Fig. 10).
2
.6. Adipocyte transcriptomics
To get a comprehensive view of the differential, transcriptional changes occurring during
adipocyte differentiation, we sequenced RNA from SGBS cells differentiated for 8 days in
medium supplemented with oxo-fatty acid 1, 2 or rosiglitazone. Data from the treated cells were
compared to SGBS cells at day 0. As can be seen from the Venn diagrams in Fig. 11A there was
a significant overlap between the genes regulated by 1 and 2, with 82.2% among the top 500
upregulated and 85.8% overlap among the top 500 downregulated genes. The overlap between
rosiglitazone and the oxo-fatty acids was 42.0% and 58.4% for the same set of upregulated and
downregulated genes, respectively. Pathway enrichment analysis made based on the KEGG
collection displayed a significant overrepresentation of upregulated genes involved in e.g. PPAR
signalling (hsa03320), fatty acid biosynthesis (hsa00061), and steroid biosynthesis (hsa00100)
for 1 and 2, as well as for rosiglitazone (Fig. 11B). Interestingly, biosynthesis of unsaturated
17

ACCEPTED MANUSCRIPT
UPREGULATED
ROSI
DOWNREGULATED
A
180
290
173
249
2
08
2
38
ROSI
B
C
(
7E)
(10E)
(7E)
(10E)
KEGG Pathway Pathway ID 9-OHE 9-OHE
ROSI
9-OHE 9-OHE
ROSI
Pathway ID KEGG Pathway
Fatty acid elongation hsa00062
Biosynthesis of unsaturated fatty acids hsa01040
Steroid biosynthesis hsa00100
hsa04512 ECM-receptor interaction
hsa04668 TNF signaling pathway
hsa04933 AGE-RAGE signaling in diabetic complications
hsa04510 Focal adhesion
Fatty acid biosynthesis hsa00061
PPAR signaling pathway hsa03320
hsa04350 TGF-beta signaling pathway
hsa04611 Platelet activation
Fatty acid degradation hsa00071
Propanoate metabolism hsa00640
hsa04015 Rap1 signaling pathway
Terpenoid backbone biosynthesis hsa00900
Valine, leucine and isoleucine degradation hsa00280
Pyruvate metabolism hsa00620
hsa04610 Complement and coagulation cascades
hsa04974 Protein digestion and absorption
hsa04520 Adherens junction
Butanoate metabolism hsa00650
hsa04810 Regulation of actin cytoskeleton
hsa04550 Signaling pathways reg. pluripotency of stem cells
hsa04390 Hippo signaling pathway
Citrate cycle (TCA cycle) hsa00020
beta-Alanine metabolism hsa00410
Glyoxylate and dicarboxylate metabolism hsa00630
Tyrosine metabolism hsa00350
hsa04151 PI3K-Akt signaling pathway
hsa04670 Leukocyte transendothelial migration
hsa04640 Hematopoietic cell lineage
hsa04621 NOD-like receptor signaling pathway
hsa04068 FoxO signaling pathway
Regulation of lipolysis in adipocytes hsa04923
Glycerolipid metabolism hsa00561
Fat digestion and absorption hsa04975
AMPK signaling pathway hsa04152
Glycolysis / Gluconeogenesis hsa00010
Peroxisome hsa04146
hsa04380 Osteoclast differentiation
hsa04060 Cytokine-cytokine receptor interaction
hsa04022 cGMP-PKG signaling pathway
hsa04014 Ras signaling pathway
hsa04062 Chemokine signaling pathway
hsa00330 Arginine and proline metabolism
hsa00010 Glycolysis / Gluconeogenesis
hsa04066 HIF-1 signaling pathway
Adipocytokine signaling pathway hsa04920
Glycerophospholipid metabolism hsa00564
Insulin resistance hsa04931
Retinol metabolism hsa00830
Insulin signaling pathway hsa04910
Complement and coagulation cascades hsa04610
Longevity regulating pathway hsa04211
Phospholipase D signaling pathway hsa04072
Transcriptional misregulation in cancer hsa05202
Pathways in cancer hsa05200
hsa04010 MAPK signaling pathway
hsa04064 NF-kappa B signaling pathway
hsa04145 Phagosome
hsa04919 Thyroid hormone signaling pathway
%
overlap with gene set
% overlap with gene set
20
0
50
0
D
18

ACCEPTED MANUSCRIPT
Fig. 11. Adipocyte transcriptomics. RNA from SGBS cells differentiated for 8 days in medium supplemented with
either 25 µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE or 2 µM rosiglitazone (all treatments n=2-4), were sequenced
on a Illumina NextSeq500, using strand-specific TruSeq RNA Sample Preparation, generating 2 × 75 bp paired-end
reads. The reads were aligned, counted and normalized as Reads Per Kilobase of transcript per Million mapped
reads (RPKM). (A) Venn diagrams of the top 500 up- or downregulated genes from the three treatments. (B) KEGG
pathway enrichment analysis of the top 500 upregulated and (C) downregulated genes, relative to day 0, displayed in
falling order based on percent overlap with the KEGG gene sets. Minimum overlap with input list: 3. p-value cutoff:
p<0.01. (D) Principal Component Analysis (PCA) of the 8274 DEGs in our dataset. Principal component 1 and 2
explains 55.6% and 23.6% of the total variance, respectively. Prediction ellipses represent the area for which a new
observation from the same group will fall inside with a probability of 0.95.
fatty acids and fatty acid elongation (hsa00062 and hsa01041) seem to be overrepresented only
in the rosiglitazone gene set. When analysing the downregulated genes, pathways such as extra
cellular matrix-receptor interaction (hsa04512), TNF signalling (hsa04668), and AGE-RAGE
signalling in diabetic complications (hsa04933) stood out (Fig. 11C). The principal component
analysis made, based on the 8274 differentially expressed genes (DEGs) in our dataset showed
that both oxo-fatty acid and rosiglitazone-treated cells differ significantly from the day 0 cells
(
Fig. 11D). Interestingly, the transcriptomes from oFAI cells cluster with the oFAII cells, and
separate from the rosiglitazone transcriptomes when the total variance is explained using these
principal components. Taking a closer look at the genes in the top tier enriched pathways (Fig.
1
2) the pattern from the differentiation experiments again becomes evident (Fig. 8-10). Even
though both oFAs and rosiglitazone drive the expression of most genes in the sets in the same
direction, the potency of and are lower, leading to a gradual decrease in expression and an
1
2
earlier turning point from up- to downregulation (PPAR signalling) or vice versa (ECM-receptor
interaction). The same is also seen for genes involved in fatty acid biosynthesis, adipokine
19

ACCEPTED MANUSCRIPT
signalling, biosynthesis of unsaturated fatty acids, and AGE-RAGE signalling in diabetic
complications (Suppl. Fig. 1). In the pathway analysis shown in Fig. 12 and Suppl. Fig. 1, all
genes in the KEGG gene set were included, independent of expression level. To get a better
ECM-RECEPTOR
INTERACTION
understanding of the difference between the
rosiglitazone treated cells and the cells treated with
PPAR SIGNALLING
(7E)
(10E)
(7E)
(10E)
ROSI 9-OHE 9-OHE
ROSI 9-OHE 9-OHE
FABP4
LPL
TNC
RELN
PCK1
THBS1
ITGB8
ITGA11
VTN
1 and 2, we plotted the top 500 expressed genes in
ADIPOQ
PLIN1
AQP7
ITGB3
THBS2
COL1A1
ITGA2
FN1
the rosiglitazone transcriptome in falling order and
compared them to the same genes in the oFA
transcriptomes (Fig. 13). Most of the genes express
approximately 10-fold lower in the cells treated
PLIN4
SORBS1
UCP1
PLIN5
RXRG
SPP1
CD36
ITGA4
COL1A2
TNXB
SCD
ACSL1
FABP7
NR1H3
ANGPTL4
HMGCS2
ACADL
FABP3
FABP5
CPT1B
PPARG
ME1
ITGA5
COMP
ITGA10
LAMA3
SDC1
SDC4
with
1 and 2, resulting in the veil-like clustering
LAMA1
HMMR
THBS3
CD44
below the rosiglitazone line. Interestingly, ten
outliers are easily distinguishable for both
treatments (Fig. 13; red circles).
ITGA3
ITGAV
LAMC2
ITGB1
COL6A3
COL6A6
ITGB5
GP1BB
ITGA1
CD47
DBI
ACSL6
PLTP
ACSBG1
ACADM
CPT2
PLIN2
SLC27A1
ACSL5
ACAA1
ACOX1
FADS2
ACOX2
EHHADH
RXRA
LAMC1
LAMA5
COL9A2
AGRN
LAMB2
COL6A2
ITGA8
ITGB4
LAMB4
COL6A1
LAMA4
LAMB1
HSPG2
VWF
Fig. 12. The oxo-fatty acids
1
and
2
induce
a
transcriptional PPAR signalling
program
and
SLC27A5
SCP2
ACSL3
SLC27A4
ACSL4
CYP27A1
PDPK1
PPARA
UBC
downregulate genes involved in extra cellular matrix-
receptor interaction. RNA sequencing data (RPKM) from
SGBS cells differentiated for 8 days in medium supplemented
with either 25 µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE or 2
µM rosiglitazone (ROSI) (n=2-4) were averaged and expressed
as log2 fold-induction over unstimulated cells (T0). Gene sets
were grouped based on the Kyoto Encyclopedia of Genes and
Genomes (KEGG).
COL4A2
COL4A5
LAMA2
THBS4
DAG1
RXRB
GK
SCD5
COL4A1
LAMB3
ITGB7
SV2A
ACOX3
OLR1
PPARD
SLC27A6
PCK2
ITGA9
COL4A6
TNN
ILK
ITGA6
ITGA7
COL9A3
CD36
CPT1A
CPT1C
MMP1
Log2 (fold-induction)
-15
Log2 (fold-induction)
1
5
15
-15
20

ACCEPTED MANUSCRIPT
A
B
1
00000
100000
Rosiglitazone
Rosiglitazone
(10E)-9-OHE
(
7E)-9-OHE
1
0000
10000
1000
100
10
COL3A1
COL1A2
COL3A1
COL1A2
COL1A1
FSTL1
COL1A1
FSTL1
FN1
FN1
COL6A3
MMP2
COL6A3
MMP2
1
000
1
00
1
0
KRT79
PDK4
PDK4
PDE3B
PDE3B
KRT79
1
1
Gene no.
Gene no.
Fig. 13. The fatty acids 1 and 2 slow down fibroblast-to-adipocyte transition. Top 500 expressed genes in SGBS
cells (day 8) differentiated in medium supplemented with rosiglitazone at was sorted in falling order based on RNA
sequencing data (black line). The corresponding expression data from SGBS cells differentiated in medium
supplemented with 1 (7E)-9-OHE (A) and 2 (10E)-9-OHE (B) was superimposed on the same graph (grey circles).
Outliers are indicated with red circles and gene symbol.
A closer look at Table 2, listing these genes, shows that six out of seven overexpressed genes
relate to extra cellular matrix. This probably reflects the more fibroblast-like properties of these
cells, and indicates a restricted or delayed adipogenigenesis. The reduced levels of PDK4 and
PDE3B point in the same direction (Table 2). That being said, lower levels of phosphodiesterase
3
might be beneficial as this would lower the anti-lipolytic effect of insulin and enable full
induction of UCP1 expression and lipolysis in brown adipocytes in response to adrenergic
stimulus [33, 34].
Our data clearly show that while being able to activate both PPAR
α and PPARγ, oxo-fatty
acids and have a reduced adipogenic effect compared to rosiglitazone. Even if we corrected
1
2
21

ACCEPTED MANUSCRIPT
Table 2. Outliers in the Top 500 expressed gene set
a
Symbol
COL3A1
COL1A2
FN1
Gene name
Relative G.E.
Collagen type III alpha 1 chain
Collagen type I alpha 2 chain
Fibronectin 1
+
+
+
+
+
+
+
-
COL1A1
FSTL1
Collagen type I alpha 1 chain
Follistatin like 1
MMP2
COL6A3
KRT79
PDK4
Matrix metallopeptidase 2
Collagen type VI alpha 3 chain
Keratin 79
Pyruvate dehydrogenase kinase 4
Phosphodiesterase 3B
-
PDE3B
-
a
Relative to rosiglitazone stimulated SGBS cells.
for their lower potency by adding twice their EC -values in the stimulation- and differentiation
5
0
cocktails, their reduced efficacy, especially with respect to PPARγ activity (Fig. 2), restricted
gene activation and delayed adipogenigenesis (Fig. 7-10). This may at first glance be interpreted
as a disadvantage. However, as many of the known side effects of the thiazolidinediones (TZDs)
have been attributed to their potency and efficacy as classical PPAR
in fact be an advantage. The strong adipogenic potential of the TZDs evidently contributes to the
weight gain associated with the use of this of this type of anti-diabetic drugs [35, 36]. While
and are able to drive adipogenesis, their ability to activate target genes and support lipid
γ-specific agonists, this may
1
2
storage in adipocytes is several fold lower than that of rosiglitazone (Fig. 7-9). Whether this is all
due to reduced lipid synthesis or also reduced synthesis of lipid droplet coating factors, like the
PLIN family (Fig. 14), is difficult to know.
The continuous increase in expression for many of the PPARγ target genes observed with
TZD-stimulated adipogenesis (Fig. 9 and 10) has been reported in numerous papers, and is
22

ACCEPTED MANUSCRIPT
proposed to be caused by these agonists’ ability to induce the biosynthesis of endogenous PPAR
γ
ligands [37-39]. As unsaturated and elongated carboxylic acids is the common chemical denominator
Fig. 14. Regulation of the lipid droplet-associated protein Perilipin family in SGBS cells. Comparison of gene
expression (RPKM) of PLIN1-5 from SGBS cells differentiated for 8 days in medium supplemented with either 25 µM
1
(7E)-9-OHE, 25 µM 2 (10E)-9-OHE, or 2 µM rosiglitazone. Gene expression at day 0 is included for comparison.
The results are shown as mean±SD, and represent 2-4 biological replicates. *p<0.05, **p<0.01.
of the many proposed endogenous PPAR ligands, e.g. PUFAs, prostaglandins and other oxidized
eicosanoids, it is intriguing that two of the metabolic pathways necessary to produce such
molecules (hsa00062: biosynthesis of unsaturated fatty acids; hsa01041fatty acid elongation) are
underrepresented in the top tire DEGs from the oxo-fatty acid-stimulated cells (Fig. 11). The
oxo-fatty acids, which structurally resemble the linoleic acid metabolites 9
HODE [40], and even more so the corresponding oxo-fatty acids 9-oxo-ODE and 13-oxo-ODE
41], are still not able to support a PPAR-driven, self-reinforcing adipogenesis. Whether this is
due to and not being able to bind covalently to PPAR as other -unsaturated ketones [41,
2] or if it a stability issue, should be the focus of future structure-activity-relationship studies.
The reduced PPAR agonistic activity of the oFAs is still strong enough to induce a robust
expression of leptin (LEP) and the insulin-sensitizing adipokine adiponectin (ADIPOQ) in the
S-HODE and 13S-
[
1
2
γ
α,β
4
γ
23

ACCEPTED MANUSCRIPT
adipocyte model (Fig. 10). At the same time their PPAR
catabolism by inducing genes involved in -oxidation, like CPT1A and ACSL3 in hepatocytes
Fig. 5) and CPT1A mature adipocytes (Fig. 6). This provides the oFAs with additional
beneficial effects, known to be inherent properties of other PPAR dual agonists [23, 25, 27,
3]. Certain adipokines contribute to the development of obesity-associated insulin resistance,
α agonistic activity triggers fatty acid
β
(
α/γ
4
and studies on the link between obesity and insulin resistance have shown that there is an
extensive crosstalk between adipocytes and immune cells in the adipose tissue (AT) [44, 45]. Pro-
Fig. 15. The oxo-fatty acids 1 and 2 downregulate pro-inflammatory cytokines in SGBS cells. Comparison of
gene expression (RPKM) of a selection of pro-inflammatory cytokines from SGBS cells differentiated for 8 days in
medium supplemented with either 25 µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE, or 2 µM rosiglitazone. Gene
expression at day 0 is included for comparison. The results are shown as mean±SD, and represent 2-4 biological
replicates. **p<0.01.
inflammatory cytokines like IL-6, IL-1
β and TNFα are mostly produced and secreted by
macrophages infiltrating the AT. However as part of the crosstalk, macrophage-derived factors
potently stimulate adipocytes to produce pro-inflammatory cytokines [46-48]. On the other side,
activation of PPARγ by TZDs is shown to modulate this inflammatory-like state by suppressing
cytokine production [49, 50]. Thus, it is agreeable that TNF signaling (hsa04668) is one of the
24

ACCEPTED MANUSCRIPT
top three overrepresented pathways among the differentially downregulated genes (Fig. 11). As
can be seen in Fig. 15, both oFAs and rosiglitazone are repressing cytokine production.
Interestingly, IL-1
β
seems to be an exception: only
1
and
2
are able to repress the IL1B gene
is a major pro-
expression, while rosiglitazone seems to have no effect (Fig. 15). Since IL-1
β
inflammatory cytokine involved in AT metainflammation, and known to reduce adipose insulin
sensitivity by downregulating IRS1 and GLUT4 expression [51, 52], it is tempting to suggest that
transcriptional regulation of this interleukin 1B is another potential advantage of the oxo-fatty
acids. Indeed IRS2 expression increases with both
1 and 2, while there was a tendency towards
increased expression of IRS1 and SLC2A4 (GLUT4) (Fig16).
Fig. 16. Antidiabetic effects of oxo-fatty acids 1 and 2 in SGBS cells. Gene expression (RPKM) of insulin receptor
substrates IRS1 and IRS2 and the glucose transporters SLC2A1 and SLC2A4 were measured as proxies for increased
insulin sensitivity and glucose uptake in SGBS cells differentiated for 8 days in medium supplemented with either
2
5 µM 1 (7E)-9-OHE, 25 µM 2 (10E)-9-OHE, or 2 µM rosiglitazone. Gene expression at day 0 is included for
comparison. The results are shown as mean±SD, and represent 2-4 biological replicates. **p<0.01
The adverse effects reported with both PPARα-activating fibrates and PPARγ-activating
thiazolidinedione, such as hepatotoxicity, pulmonary edema, myocardial infarction, weight gain,
reduced bone density, and bladder and prostate cancer [15-22] were not addressed further in this
25

ACCEPTED MANUSCRIPT
paper. To proceed in this matter we would have to initiate in vivo studies in mice, which given
the current in vitro cytotoxicity profile of
1 and 2 was not alternative. Certain of the reported side
effects, like cancers of the bladder and prostate, is difficult to detect in animal models. Moreover,
clinical trials are often too short to detect drug-related tumorigenesis. Therefore, the carcinogenic
potential of most compounds is evaluated based on the extrapolation of minimum 2-year studies
in rats or mice. Before testing the oFAsin long term in vivo experiments, less cytotoxic analogues
would have to be synthesised and characterized.
3
. Conclusion
Through the synthesis, pharmacological characterization, and biological evaluations of the oxo-
fatty acids (7 )-9-oxohexadec-7-enoic acid ( ) and (10 )-9-oxohexadec-10-enoic acid ( ), we
have shown that both compounds are semi-potent dual PPAR agonists. While being derived
E
1
E
2
α/γ
from a marine alga, both oFAs regulate important PPAR target genes in human hepatocytes and
adipocytes. Transcriptome analyses show that both compounds activate anti-diabetic gene
programs in adipocytes by upregulating insulin-sensitizing adipokines and repressing pro-
inflammatory cytokines. Simultaneously, they do not accumulate lipids in the adipocytes to the
same extent as the classical thiazolidinediones. Together these features make the
oxohexadecenoic acids interesting molecular scaffolds for designing new ligands with improved
stability and a better toxicity profile to test in animal models relevant for T2D and metabolic
syndrome.
26

ACCEPTED MANUSCRIPT
4
4
4
. Experimental section
.1. Chemistry
.1.1. Methyl 7-hydroxyheptanoate
Methyl 7-hydroxyheptanoate (
The physical and spectral data of
3
) was synthesized using literature protocols [28, 29, 53, 54].
3
were in full agreement with those reported in the literature.
4
.1.2. Methyl (E)-9-oxohexadec-7-enoate (6)
Phosphorane (190 mg, 0.60 mmol, 1.18 equiv.) in THF (4 mL) was added
.6 M in hexane, 1.30 equiv.) at -78 °C and stirred for 20 min. 1-Bromohexane (83 mg, 0.51
5
n-BuLi (0.41 mL,
1
mmol, 1.00 equiv.) dissolved in CH Cl (0.4 mL) was added dropwise and the solution was
2
2
allowed to warm up to room temperature overnight. The reaction mixture was diluted with H O
2
(10 mL) and extracted with CH Cl (3 × 10 mL). The combined organic layers were dried
2 2
(
Na SO ), before being concentrated in vacuo. The crude product was dissolved in CH Cl (1.5
2
4
2
2
mL) and aldehyde
4 (80 mg, 0.51 mmol, 1.00 equiv.) in CH Cl (1.0 mL) was added. The
2 2
solution was stirred overnight before being being concentrated in vacuo. The crude product was
purified by column chromatography on silica (hexanes/EtOAc 9:1) to afford the title compound
6
as a colourless oil. Yield: 75 mg (53%). TLC (hexanes/EtOAc 9:1, KMnO stain): Rf = 0.12;
4
1
H NMR (300 MHz, CDCl3)
.53 (t, = 7.4, 2H), 2.33 (t,
δ
6.82 (dt,
J
= 15.9, 6.9 Hz, 1H), 6.10 (dt,
J = 15.9, 1.5 Hz, 1H),
2
1
2
1
3
J
J
= 7.4, 2H), 2.22 (td, J
= 7.5, 1.5 Hz, 2H), 1.72 – 1.57 (m, 4H),
13
.56 – 1.42 (m, 2H), 1.42 – 1.25 (m, 10H), 0.90 (t,
J = 6.6, 3H); C NMR (101 MHz, CDCl ) δ
3
01.0, 174.2, 146.8, 130.6, 51.6, 40.4, 34.0, 32.3, 31.9, 29.4, 29.3, 28.8, 27.9, 24.8, 24.5, 22.8,
+
+
4.2. HRMS (TOF ES ) Exact mass calculated for C17
H
30NaO
3
[M+Na] : 305.2092, found:
05.2102.
27

ACCEPTED MANUSCRIPT
4
.1.3. (E)-9-Oxohexadec-7-enoic acid (1)
The methyl ester (20 mg, 0.07 mmol) was dissolved in phosphate buffer (pH = 7.2, 0.7 mL)
6
and heated to 37 °C. Lipase from Pseudomonas cepacia (10 mg) was added and the reaction was
stirred at ambient temperature for 24 hours. The mixture was and extracted with CH Cl (3 × 10
2
2
mL). The combined organic layers were dried (Na SO ), before being concentrated in vacuo. The
2
4
crude product was purified by column chromatography on silica (hexanes/EtOAc 97.5:2.5, then
9
0:10) to afford the title compound
1
as a white solid. Yield: 12 mg (64%). TLC
1
(
hexanes/EtOAc 7:3, KMnO stain):
4
R
f
= 0.12; H NMR (400 MHz, CDCl
3
)
δ
6.80 (dt,
= 7.4 Hz, 2H), 2.22
= 7.0, 1.5 Hz, 2H), 1.70 – 1.55 (m, 4H), 1.53 – 1.45 (m, 2H), 1.44 – 1.34 (m, 2H), 1.34 –
J = 15.9,
6
.9 Hz, 1H), 6.09 (d, J = 15.9 Hz, 1H), 2.52 (t, J = 7.5 Hz, 2H), 2.36 (t, J
(qd,
J
1
3
1
4
.22 (m, 8H), 0.87 (t,
J
= 4.8, 3H); C NMR (101 MHz, CDCl )
δ
201.1, 178.6, 146.8, 130.6,
3
+
0.4, 33.8, 32.3, 31.9, 29.4, 29.3, 28.7, 27.9, 24.5, 24.5, 22.8, 14.2. HRMS (TOF ES ) Exact
+
mass calculated for C H NaO [M+Na] : 291.1936, found: 291.1940.
1
6
28
3
4
.1.4. ((8-Bromooctyl)oxy)(tert-butyl)dimethylsilane (7)
The compound was prepared essentially as previously reported [30]. 8-Bromooctan-1-ol
6.00 mmol, 1.25 g) was dissolved in dry MeCN (12.0 mL) and added TBSCl (6.3 mmol, 0.94 g)
7
(
and then imidazole (9.00 mmol, 0.630 g). The mixture was stirred under argon for 1 hour. The
mixture was filtered, evaporated, and purified by flash chromatography on silica gel
1
(EtOAc:heptane, 49:1) to give the known compound
7
as a colourless oil. Yield 1.74 g (90%). H
1
NMR data was in agreement with literature [30]. H NMR (400 MHz, Chloroform-
d)
δ
3.60 (t,
J
=
6.6 Hz, 2H), 3.41 (t,
J
= 6.9 Hz, 2H), 1.85 (quint.,
J
= 6.9 Hz, 2H), 1.55 – 1.39 (m, 4H), 1.30
(m, 6H), 0.89 (s, 9H), 0.04 (s, 6H).
28

ACCEPTED MANUSCRIPT
4
.1.5. (E)-16-((Tert-butyldimethylsilyl)oxy)hexadec-6-en-8-ol (8)
Magnesium turnings (6.40 mmol, 0.154 g, 2.00 equiv.) and a magnet were placed in a flame
dried round bottomed flask that was flushed with argon. A small amount of iodine (ca. 6 mg)
was added. A solution of the bromide (4.16 mmol, 1.34 g, 1.30 equiv.) in dry THF (12.0 mL)
7
was prepared, and about 2 mL of this solution was quickly added to the magnesium turnings.
After a few minutes of vigorous stirring, the reaction mixture went from clear brown to
colourless. The remaining THF-solution of bromide
one hour the solution was cloudy and the remaining magnesium was black. The mixture was
cooled on an ice bath and added a solution of ( )-2-octenal (3.20 mmol, 0.403 g, 1.00 equiv.) in
7 was then added over five minutes. After
E
dry THF (4.00 mL) over five minutes. The cooling bath was removed and the mixture stirred for
one hour. After one hour, the mixture was cooled on an ice bath and quenched with sat. aq.
NH Cl (40 mL). The mixture was extracted with EtOAc (3 × 20 mL), dried (MgSO ), evaporated
4
4
and purified by flash chromatography on silica gel (EtOAc:heptane, 15:85) to give the desired
product
8 as a colourless oil. Yield: 0.832 g (70%). Rf =0.35 (EtOAc:heptane 3:7, KMnO -stain).
4
1
H NMR (400 MHz, Chloroform-
q, = 6.7 Hz, 1H), 3.59 (t, = 6.6 Hz, 2H), 2.07 – 1.95 (m, 2H), 1.59 – 1.22 (m, 21H), 0.89 (s,
H), 0.88 (m, 3H), ( 0.04 (s, 6H). C NMR (101 MHz, CDCl )
d) δ 5.62 (m, 1H), 5.44 (broad dd, J = 15.4, 7.2, Hz, 1H), 4.02
(
J
J
1
3
9
3
δ
133.2, 132.4, 73.4, 63.5, 37.5,
3
3.0, 32.3, 31.5, 29.7, 29.7, 29.5, 29.0, 26.1 (3xC), 25.9, 25.6, 22.7, 18.5, 14.2, -5.1. Exact mass
+
calculated for C H O SiNa [M+Na] : 393.3159, found 393.3159.
2
2
46
2
4
.1.6. (E)-9-((Tert-butyldimethylsilyl)oxy)hexadec-10-en-1-ol (9)
Step 1. Synthesis of ( )-5-(Hept-1-en-1-yl)-2,2,3,3,15,15,16,16-octamethyl-4,14-dioxa-3,15-
disilaheptadecane. The alcohol (2.18 mmol, 0.805 g, 1.00 equiv.) was dissolved in dry DMF
E
8
29