Structural verification via convergent total synthesis of dipeptide–lipids isolated from Porphyromonas gingivalis

Article abstract of DOI:10.1016/j.tet.2016.10.010

A periodontal pathogen, Porphyromonas gingivalis, produces two serine dipeptide lipid classes that we labeled lipid 654 and lipid 430, and both contain L-serine as the terminal amino acid. The lipid 654 and lipid 430 classes are each comprised of three species with differing fatty acid substitutions, but the most abundant species demonstrate unit masses of either 654 or 430, respectively. Recently we observed that the lipid 654 can be hydrolyzed by specific lipases to lipid 430. However, a substantial percentage of the naturally occurring lipid 654 cannot be enzymatically hydrolyzed to lipid 430. The observed partial hydrolysis could be due to the presence of a mixture of stereoisomers. Testing this theory requires structural verification of our so-called 654 and 430 by total synthesis. We present herein details of the convergent synthesis of lipids 430 and 654, which confirm the proposed structure of P. gingivalis lipid 654 to be (3R and 3S)-L-serine-2. The bis(fatty acid) (3R)-L-serine-2 was prepared as well as the synthetic precursor, serine dipeptide mono-fatty acid (3R)-L-serine-1, which is the structure of lipid 430. We also synthesized the (3S)-L-serine-2 diastereomer as well as (3S)-L-serine-1. Using these synthetic standards, we confirmed that PLA2-mediated hydrolysis of lipid 654 is enantioselective in that only the (3R)-L-serine-2, but not (3S)-L-serine 2 is enzymatically hydrolyzed.

Full text of DOI:10.1016/j.tet.2016.10.010

Accepted Manuscript
Structural verification via convergent total synthesis of dipeptide–lipids isolated from
Porphyromonas gingivalis
Christopher Dietz, Theresa K. Hart, Reza Nemati, Xudong Yao, Frank C. Nichols,
Michael B. Smith
PII:
S0040-4020(16)31019-5
10.1016/j.tet.2016.10.010
TET 28151
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 3 August 2016
Revised Date: 30 September 2016
Accepted Date: 4 October 2016
Please cite this article as: Dietz C, Hart TK, Nemati R, Yao X, Nichols FC, Smith MB, Structural
verification via convergent total synthesis of dipeptide–lipids isolated from Porphyromonas gingivalis,
Tetrahedron (2016), doi: 10.1016/j.tet.2016.10.010.
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
Structural Verification Via Convergent Total Synthesis of Dipeptide–Lipids
Isolated from Porphyromonas gingivalis
Christopher Dietza, Theresa K. Harta, Reza Nematia, Xudong Yaoa, Frank C. Nichols,b and Michael B. Smitha*
Abstract
We have proven the structure of lipid 654, isolated from the dental pathogen Porphyromonas gingivalis.
The lipid was prepared by a convergent synthetic route that is significantly shorter than other published routes,
producing the targeted compound in higher yield. Further, we have shown that this dipeptide-lipid is selectivity
hydrolyzed by lipase, consistent with results from lipid 654 obtained from P. gingivalis.

ACCEPTED MANUSCRIPT
1
Structural Verification Via Convergent Total Synthesis of
Dipeptide–Lipids Isolated from Porphyromonas gingivalis
Christopher Dietza, Theresa K. Harta, Reza Nematia, Xudong Yaoa, Frank C. Nichols,b and
Michael B. Smitha*
a
Department of Chemistry, The University of Connecticut, Storrs, CT
Division of Periodontology, The University of Connecticut Health Center, Farmington, CT
b
Abstract
A periodontal pathogen, Porphyromonas gingivalis, produces two serine dipeptide lipid
classes that we labeled lipid 654 and lipid 430, and both contain -serine as the terminal amino
L
acid. The lipid 654 and lipid 430 classes are each comprised of three species with differing fatty
acid substitutions, but the most abundant species demonstrate unit masses of either 654 or 430,
respectively. Recently we observed that the lipid 654 can be hydrolyzed by specific lipases to
lipid 430. However, a substantial percentage of the naturally occurring lipid 654 cannot be
enzymatically hydrolyzed to lipid 430. The observed partial hydrolysis could be due to the
presence of a mixture of stereoisomers. Testing this theory requires structural verification of our
labeled 654 and 430 by total synthesis. We present herein details of the convergent synthesis of
lipids 430 and 654, which confirm the proposed structure of P. gingivalis lipid 654 to be (3R and
3S)-
precursor, serine dipeptide mono-fatty acid (3R)-
We also synthesized the (3S)- -serine-2 diastereomer as well as (3S)-
synthetic standards, we confirmed that PLA2-mediated hydrolysis of lipid 654 is enantioselective
in that only the (3R)- -serine-2, but not (3S)- -serine 2 is enzymatically hydrolyzed.
L
-serine-2. The bis(fatty acid) (3R)-
L
-serine-2 was prepared as well as the synthetic
-serine-1, which is the structure of lipid 430.
-serine-1. Using these
L
L
L
L
L

ACCEPTED MANUSCRIPT
2
Introduction
Inflammatory periodontal disease in adults is initiated with the accumulation of specific
bacteria in the sulcus around the teeth, followed by a chronic inflammatory reaction by the host
against the colonizing microorganisms.
The anaerobic Gram-negative organism,
Porphyromonas gingivalis, is thought to be a major periodontal pathogen1 associated with
destructive periodontal disease in adults. P. gingivalis and other phylogenetically related
organisms produce a variety of novel lipids,2 including phosphorylated dihydroceramide lipids.3
More recent work in our laboratories identified serine dipeptide lipid classes in P. gingivalis that
comprise a new class of ligands for Toll-like receptor 2 (TLR2).4 These agonists are also
produced by common oral and intestinal Bacteroidetes,5 and they are recovered in chronically
inflamed human tissues including destructive periodontal disease and atherosclerosis tissues.2
The most commonly observed serine dipeptide mono-fatty acid species is composed of a serine-
glycine dipeptide in amide linkage to a 3-OH isobranched (iso) C17:0 fatty acid. We labeled this
compound lipid 430 (see 1). The 3-OH iso-C17:0 fatty acid moiety may form a 3-OH ester
linkage to a saturated fatty acid, most frequently to iso-C15:0 (see 2).6 We labeled this isolate
lipid 654. A search of the literature indicated that this structure was originally proposed for a
serine lipid product isolated from a limited number of Flavobacteria and was termed flavolipin.7-
9
We have also observed that lipid 430 was produced by enzymatic hydrolysis of lipid 654 and
was, therefore, a logical synthetic precursor. Other fatty acids can substitute into these lipid
classes, but they are less abundant relative to branched C15:0 and 3-OH iso C17:0.6 Both of the
serine dipeptide lipids, lipid 430 and lipid 654, are recovered in lipid extracts of P. gingivalis, but
lipid 430 is recovered in low levels compared with lipid 654.

ACCEPTED MANUSCRIPT
3
We have previously observed catalytic deesterification of P. gingivalis lipid 654 to lipid
430 using commercially available preparations of phospholipase A2 (PLA2).6 PLA2 is an
important enzyme in the release of arachidonic acid from glycerol-based phospholipids and
levels of this enzyme are increased in chronically inflamed tissues. However, it was noted that a
portion of lipid 654 isolated from P. gingivalis was not hydrolyzed by PLA2. Chiral
fractionations show about 10-15% of one isoform and 85-90% of the other isoform of 654 in the
P. gingivalis 654. Recent HPLC analysis with direct infusion into the mass spectrometer suggest
that the S isoform is less than 10%. These results suggest that enzymatic hydrolysis is
stereospecific.
The lipid 654 isolated from P. gingivalis contains only -serine as determined by chiral
L
GC-MS analysis.6 The absolute configuration of the β carbon of 3-OH iso C17:0 is unknown in
the bacterial lipid 654 isolated from P. gingivalis. Our initial structural identification of 2 and of
1 was by 1H NMR and mass spectral analysis. Identification of the stereochemistry of 1 and 2,
as well as confirmation of which isoform is selectively hydrolyzed, required verification by total
synthesis for enzymatic hydrolysis studies.
This investigation prepared both (3R) and (3S)-
and (3S)- -serine-lipid 654 to ascertain if only one diastereomeric form of synthetic lipid 654
was susceptible to PLA2 enzymatic hydrolysis. If selective hydrolysis is observed, the
L
-serine-lipid 430 and thereby both (3R)
L

ACCEPTED MANUSCRIPT
4
diastereospecificity of the partial hydrolysis of P. gingivalis lipid 654 by PLA2 would be
confirmed. The biological activity associated with lipid 654 and lipid 4302,6 is sufficiently
important that synthesis is a priority for structural verification, the accumulation of synthetic
standards, and the preparation of each lipid in sufficient quantity for further biological
evaluation.
Structure 2 has been reported as an isolate from several natural products, and it has been
synthesized several times. Uchida and coworkers isolated dipeptide lipids from Flavobacterium
sp. No. 3559, and the compound labeled WB-3559 D was identified as 2.7 They reported a linear
synthesis of 2 from
D
-glucose, via a β-hydroxy aldehyde that was oxidized to the acid. Glycine
and -serine were attached sequentially in subsequent steps. In this work, WD-3559 D was
L
shown to be a platelet aggregation inhibitor, and stimulated mouse plasma euglobulin clot lysis
time.7b
Kawai and coworkers identified a serine-containing lipid along with ornithine-
containing lipids from extractable cellular lipids in Flavabacterium meningosepticum and F.
indologes, which are opportunistic pathogens.8 The serine-containing lipid was identified as 28a
and shown to be an immunoactivator that exhibited immuomodulator activity,8b and it was also
identified as a blocking agent against endotoxemia.8c Finally, Andoh and coworkers9 isolated a
dipeptide lipid from strain B-572 of Flexibacter topostinus sp. nov. and it was named topostin
D640, with structure 2, -serine, (3R). Shioiri and coworkers synthesized topostin D640 in 13
L
steps in a linear synthesis that utilized asymmetric hydrogenation of a β-keto ester
intermediate.10 Topostins are inhibitors of mammalian DNA atopoisomerase I (topo I).
Uchida and coworkers reported a synthesis of WB-3559 D7c in which (3aR,5S,6aR)-2,2-
dimethyltetrahydrofuro[2,3-d][1,3]dioxole-5-carbaldehyde11 was prepared in three steps from
D
-
glucose. Separate preparation of the triphenylphosponium ylid from 1-chloro-11-

ACCEPTED MANUSCRIPT
5
methyldodecane allowed Wittig olefination and then hydrogenation to attach the isobranched
lipid chain.
Deprotection and NaIO cleavage of the diol led to (3R)-hydroxy-15-
4
methylhexadecanal. Protection of the alcohol, oxidation of the aldehyde with Ag O and
2
deprotection gave (3R)-hydroxy-15-methylhexadecanoic acid, 5. The fatty acid side chain was
prepared from isobutyraldehyde and 10-carboxydecyltriphenylphosphorane, and coupled to 5.
Sequential coupling with protected glycine and then a protected serine gave (3R)-1 after
deprotection.7c Uchida and coworkers also prepared a derivative of 5 via a Reformatsky reaction
of 3-hydroxy-15-methylhexadecanal, added the fatty ester moiety and final coupling with a
glycine-serine dipeptide gave 1 in about 13 linear steps from the aldehyde (16 steps from -
D
glucose).7b
Scheme 1. Convergent Retrosynthetic Analysis of 1 and 2

ACCEPTED MANUSCRIPT
6
Shiozaki and coworkers prepared (3R)-5 in 7 steps from (3aR,5S,6aR)-2,2-
dimethyltetrahydrofuro[2,3-d][1,3]dioxole-5-carbaldehyde (10 steps from
D-glucose) and
coupling using both -serine and -serine gave (3R, -serine)-1 and (3R,
L
D
L
D
-serine)-1.12
Subsequently, the structure of flavolipin was revised and (3R)-5 and (3S)-5 were prepared by this
route and each was coupled with glycine and then serine. In this work, it was shown that
flavolipin was (3R,
L
-serine)-2, which had the same macrophage stimulation activity as natural
-serine) diastereomer was nearly inactive.12b
flavolipin, but the (3S,
D
Shioiri and coworkers published a 13-step synthesis of topostin D654 from decane-1,10-
diol.10 Initial conversion to 7 was followed by oxidation to 3 allowed coupling with the
magnesium enolate anion of ethyl hydrogen malonate to give 9. Asymmetric hydrogenation
using a Ru-BINAP catalyst system gave (3R)-10, which was saponified to give 5 and coupled
sequentially with a projected glycine followed by a protected serine. Deprotection gave topostin
D654, which is (3R)-2.10
The synthetic goal was to verify the structural identity of lipid 654 as 2, which would
correlate with the previously identified and synthesized flavolipin. The previous syntheses
produced hydroxy acid 5, which could be converted to serine-glycine derivative 1, which we
believed to be lipid 430. While the stereochemistry of our lipid 430 and lipid 654 were
unknown, our previous work indicated that the serine dipeptide lipids of P. gingivalis are
primarily constituted with
-serine.6 This assumption is also consistent with previous reports of
L
2 by Uchida, Shiozaki, and Shioiri. In our previous synthesis of lipids from P. gingivalis, we
prepared 7 by a LiCuCl4-Grignard coupling that was readily oxidized to aldehyde 6. This
aldehyde could be oxidized to 3 or coupled to ethyl diazoacetate to give 9.13 We therefore
envisioned a convergent strategy that utilized elements of our previous work that would take

ACCEPTED MANUSCRIPT
7
advantage of the previously published syntheses to prepare 1 and then 2 using a protected -
L
serine-glycine dipeptide. We believed that this approach would allow us to generate both
diastereomers of -serine 1 and of -serine 2, using -serine, and in a shorter linear synthesis than
L
L
L
previously reported.
Scheme 2. Synthesis of 5
Our retrosynthesis began with cleavage of the labile ester linkage from the putative
structure for lipid 654 (2), which generated the putative structure of lipid 430 (1) and 13-
methyltetradecanoic acid (3). Disconnection of the peptide linkage to the β-hydroxy fatty acid
moiety in 2 gave the
L
-serine-glycine dipeptide 4 and β-hydroxy acid derivative 5.
Disconnection of 5 led to an acetate surrogate and 13-methyltetradecanal, 6. Disconnection of 4

ACCEPTED MANUSCRIPT
8
led to serine and glycine, to be prepared by coupling a serine derivative with carboxyl and
alcohol protection with a nitrogen-protected glycine. Aldehyde 6 can be prepared from 13-
methyltetradecan-1-ol (7) using our previously published coupling of the Grignard reagent
prepared from 1-bromo-2-methylpropane with 11-bromoundecan-1-ol, 8.14 We planned a
coupling reaction of ethyl diazoacetate with 6 to give a β-keto ester.13,14 Subsequent reduction of
the ketone, followed by saponification, would generate the racemic ꢀ-hydroxy acid, 5. One
approach could use the asymmetric hydrogenation procedure reported by Shioiri.10
Alternatively, a classical resolution of this β-hydroxy fatty acid would yield each enantiopure
diastereomer.
Results and Discussion
The initial steps of the synthesis focused on the fatty acid portion of the targeted lipids
and the preparation of 7, based on our published synthesis of dihydroceramides isolated from P.
gingivalis.14 The Grignard reagent prepared from 1-bromo-2-methylpropane was coupled to 8,
in the presence of a lithium tetrachlorocuprate complex with N-methylpyrrolidone as an
additive.14 A large excess of Grignard reagent was used to convert the alcohol to an alkoxide in
situ, thus avoiding the use of a protecting group. The use of excess reagent can be avoided if 11-
bromoundecanol is protected prior to the coupling reaction. In the latter case, however, there
was a decrease in overall yield for the three steps. The fact that 2-methylbromopropane is much
cheaper than the other reagents therefore led us to use the unprotected alcohol in the coupling
step. Oxidation of alcohol 7 with an oxoammonium reagent “Bobbitt’s salt”15a (4-acetamido-
2,2,6,6-tetramethylpiperidine 1-oxyl)15b,c yielded aldehyde 6 with no over-oxidation to the acid.
Upon exposure to air at ambient temperature, this aldehyde proved to be rather unstable,

ACCEPTED MANUSCRIPT
9
however, with rapid oxidation to the carboxylic acid 3. If aldehyde 6 was maintained in wet
THF solution until needed, the yield was near quantitative for the oxidation step and 91% from 8.
Racemic ꢀ–hydroxy ester 9 was prepared by the reaction of 6 with ethyl diazoacetate in
the presence of catalytic tin (II) chloride, using the protocol developed by Roskamp et al.13,14
Ester 9 was obtained in 84% yield under very mild conditions, although small amounts of
carboxylic acid 3 were also observed. Reduction of the ketone moiety in ꢀ-keto ester 9 with
sodium borohydride generated racemic ꢀ–hydroxy ester 10. Saponification of 10 using lithium
hydroxide yielded racemic ꢀ-hydroxy acid 5 in an overall yield of 62% from 8.
A traditional resolution of racemic 5 using brucine was also attempted. Racemic ꢀ-
hydroxy acid 5 was heated at reflux with brucine in different polar solvents (methanol, ethanol,
acetone, and chloroform). We obtained crystals from the reaction in chloroform, but all attempts
to separate the diastereomeric salts failed because the resolving agent co-crystallized with both
diastereomers under all conditions examined. We therefore abandoned this approach.
The best route to enantioenriched 5 was the asymmetric hydrogenation of ꢀ–keto ester 9,
as reported by Shioiri.10 Catalytic hydrogenation of the ketone moiety in 9 using ruthenium
catalyst complexed with (S)-BINAP or (R)-BINAP as chiral ligands led to ꢀ-hydroxy esters (3S)-
10 and (3R)-10 respectively. Shioiri reported the specific rotation of (3R)-10 to be [α]D23 = –
12.5° (c 1.02, CHCl3).10 Our synthetic (3R)-10 to be [α]D23 = –9.2° (c 1.02, CHCl3) and synthetic
(3S)-10 to be [α]D23 = +11.8° (c 1.02, CHCl3).
To confirm the stereoselectivity of each ester, we converted each ester to the alcohol, 5,
by LiOH saponification conditions to yield (3R)-5 in 68 % yield and (3S)-5 in 70 % yield,
respectively, from 8. The specific rotation of our synthetic β-hydroxy fatty acid (5) was

ACCEPTED MANUSCRIPT
10
23
23
determined to be [α]D +14.5° (c 1.0, CHCl3) for (3S-5) and [ꢁ]D –10.5° (c 1.0, CHCl3) for
(3R-5). There is a small variation in the specific rotation reported in the literature for (3R-5):
21
23
[α]D 10° (c 0.05, CHCl3);9 [α] –12.0° (c 1.0, CHCl3)7b,c To confirm our identification, the
D
corresponding Mosher’s ester was prepared for both (3R)-5 and (3S)-5. Based on Mosher ester
analysis, the enantiopurity of (3R)-5 was 97% ee and (3S)-5 was 98% ee. Based on analysis of
the hydroxy acids 5, and (3R)-10 was formed in at least 97 %ee and (3S)-10 was formed in at
least 98 %ee.
We also attempted to generate enantiopure or enantioenriched 5 via enzymatic resolution
of racemic 5. Sugai and Ohta reported the use of a Pseudomonas lipase with vinyl acetate for the
enzymatic transesterification of 2-hydroxytetracosanoic acid.16a,b Mori and coworkers used
lipase PS and vinyl acetate to resolve racemic 5, and used the resulting (3R)-hydroxy-15-
methylhexadecanoic acid in a synthesis of sulfobactin.16c We applied this enzymatic resolution
technique to compound 5, but we used the commercially available Amano lipase PS. Our
23
measured specific rotation data for (3R+3S)-5 obtained by the enzymatic hydrolysis was [α]D
=
+6.4° (c 1.02, CHCl ). Saponification of the acetate product gave (3R+3S)-5 and [α]D23 = –8.7°
3
(c 1.02, CHCl ). After several attempts, our best result provided (3R)-5 in about 67% ee and
3
1
(3S)-5 in about 79% ee, based on conversion to the respective Mosher’s ester followed by H
NMR analysis that showed a 84R:15S mixture for (3R)-5 and a 11R:89S mixture for (3S)-5. We
modified the exposure time of racemic 5 to the lipase, but we observed no improvement in
enantiomeric enrichment.

ACCEPTED MANUSCRIPT
11
Scheme 3. Synthesis Protected Dipeptide 13
The glycine-serine dipeptide portion of lipid 654 was prepared using standard
methodology.17 The amino group of glycine was protected as the benzyl carbamate to produce
amino acid derivative 11. Both the carboxylic acid and alcohol moieties of -serine required
L
protection to avoid undesired byproducts in the planned couplings. The reaction of
L
-serine with
thionyl chloride in methanol gave the
L
-serine methyl ester,17 which reacted with tert-
butyldimethylsilyl chloride to give the Boc-protected derivative 12. Dipeptide 13 was prepared
by coupling the amino portion of 12 to the carboxyl group of 11 using DCC, where HOBt was
used to generate the needed activated ester.
The coupling reaction of deprotected 14 and 5 proved challenging. Peptides with free
amines, especially ones containing glycine, are known to undergo cyclisation to form a
diketopiperazine derivative.18 Dipeptides substituted with an amine at one terminus and an ester
at the other can spontaneously cyclize to form a 2,5-DKP at a range of pHs.18 Deprotection of
the amino moiety by hydrogenation of dipeptide 13 using a palladium catalyst gave almost
exclusively the DKP product and little coupling to give 14. Formation of this unwanted product
could be partially suppressed with the addition of methanolic HCl to protonate the amine, but we

ACCEPTED MANUSCRIPT
12
were unable to isolate 14 free of DKP. Attempts to couple the HCl salt of 14 with ꢀ–hydroxy
acid 5 failed and coupling was only observed upon neutralization of the amine salt. The addition
of a stoichiometric amount of base led to the DKP and the yield our coupling product was only
30-40%.
Methodology developed by Shute and Rich for the coupling of dipeptides to tetrapeptides
provided a solution to the coupling problem.19
Activation of acid 5 with N-
hydroxysuccinimide7a formed the reactive ester and this crude product was added to the
hydrogenolysis product of dipeptide 13. The presence of the activated ester facilitated the
production of (3R+3S)-14.
Subsequent deprotection with TBAF followed by LiOH
saponification gave (3R+3S)-2 (75 % yield). Using an identical procedure beginning with (3R)-
5, we obtained (3R)-2 in 68 % yield), and using (3S)-5 in this sequence gave (3S)-2 in 71 %
yield. Spectroscopic evaluation demonstrated that (3R)-2 correlates with the proposed structure
of lipid 430 and the biological evaluation (see below) confirmed this conclusion.

ACCEPTED MANUSCRIPT
13
Scheme 4. Synthesis of 1 and 2 from 5 and 13
The final synthetic step was esterification of the alcohol moiety in (3R+3S)-14 with fatty
acid 3. A DCC mediated coupling gave product (3R+3S)-15 in > 90% yield. The sensitivity of
the ester linkage in the lipid 654 derivatives led to deprotection of 14 and 16 using two different
methods. Lipid derivative (3R+3S)-14 was deprotected in a straightforward manner in two steps.
First, the silyl ether was removed using tetrabutylammonium fluoride to liberate the hydroxyl
derivative of the serine moiety. After purification, this product was subjected to saponification
with lithium hydroxide to yield (3R+3S)-1 in 87% yield over the two steps.
The desilylation of lipid derivative (3R+3S)-16 was done in the same manner as before
but a different hydrolysis of the methyl ester was used because saponification of the fatty acid

ACCEPTED MANUSCRIPT
14
occurred under basic conditions. After treatment with TBAF, reaction with potassium
trimethylsilanolate led to (3R+3S)-2. This S 2 type of deprotection was successful, but the yield
N
was lower over the two steps, giving 66% of (3R+3S)-
L-serine-2. In an identical manner, (3R)-
16 led to (3R)-L-serine-2 (65% overall yield) and (3S)-16 led to (3S)-2 (78% overall yield).
Spectroscopic analysis of (3R)-
L
-serine-2 showed that it correlates structurally with lipid 654.
ͦͧ
The specific rotation of 2 from various literature syntheses is quite varied: [α]
= +19.75° (c
͸
ͦͧ
͸
ͦͧ
͸
ͦͧ
0.8, CHCl ),7a [α]
= +19.5° (c 0.3, CHCl ),7c [α]
= +17.1° (c 2.0, CHCl ),9 [α]
= +12.8
͸
3
3
3
ͦͧ
and +13.3° (c 0.13; 0.14, CHCl ; prepared from different precursors,12a [α]
= +18.9° (c 0.39,
3
͸
ͦͧ
CHCl ).12b and [α]
= +18.4° (c 0.4, CHCl ).10 The specific rotation of our synthetic (3R)-
-
L
3
3
͸
ͦͧ
serine-2 was [α]
= +12.85° (c 1.02, CHCl ) and the specific rotation of our (3S)- -serine-2
L
3
͸
ͦͧ
was [α]
= +6.3° (c 1.02, CHCl ). Our specific rotation data was consistent with some
3
͸
literature reports but showed a discrepancy relative to others. However, the enantiopurity of 10
and 5 is clear based on our Mosher’s ester analysis, and there is no evidence of racemization in
the subsequent conversion to 1 and to 2.
Enzymatic Hydrolysis of Synthetic Lipid 654 ((2R)- -serine-2)
L
HPLC purified synthetic (3R)-
L
-serine-2 and (3S)-
-serine-26,4 were subjected to
L
enzymatic hydrolysis using a commercially available PLA2 preparation. Approximately 250 ng
of each synthetic lipid was sonicated in Tris buffer (10 mM, pH 7.5, 1 mL) containing 150 mM
NaCl and 2.5 mM CaCl . Control samples received no enzyme. Approximately 100 U of
2
porcine pancreatic PLA2 (Sigma) was added and the samples were stirred for 5 days at 25°C. At

ACCEPTED MANUSCRIPT
15
5 days, the samples were acidified with glacial acetic acid and extracted with CHCl (1 mL x 3
3
extractions).
The pooled extracts were dried and suspended in HPLC solvent
(hexane:isopropanol:water, 6:8;0.75, v/v/v). The samples were evaluated using MRM-MS as
previously described6,4 and the ratio of lipid 430 to lipid 654 was determined for both control
and enzyme treated samples. The results in Figure 1 show that the (3R) diastereomer of
synthetic lipid 654, (3R)-
L
-serine-2, is hydrolyzed by PLA2 but the (3S)- -serine-2 is not. Of
L
note, we recently reported that the lipid 430 recovered following porcine pancreatic PLA2
hydrolysis of P. gingivalis lipid 654 is a potent inhibitor of osteoblast function both in vivo and
in vitro.6
Figure 1. Hydrolysis of (3R)-
L
-serine-2 and (3S)- -serine -2 by porcine pancreatic (PP,
L
secretory) PLA2. Each lipid preparation was aliquoted (250 ng/sample) and the samples were
dried. Samples were processed as described in the text and each trial was run in triplicate.
Multiple Reaction Monitoring (MRM) was performed using a Sciex QTrap 4000. The 654.3-
381.3 or 430.3- 382.3 m/z transitions were used to calculate the lipid 430/lipid 654 ratios.
Hydrolysis of (3R)- -serine-2 (R syn 654) was significantly different (denoted by *) from all
L
other trials by one factor ANOVA with Scheffe contrasts comparisons. No other pairwise
comparisons were significantly different.
Conclusions

ACCEPTED MANUSCRIPT
16
We have prepared (3R)-
-serine 1 by a convergent synthetic route that is significantly shorter than other published
routes, producing the targeted compounds in higher yield. Based on sour total synthesis and
biological comparison with natural bacterial lipid we have demonstrated that lipid 654 is (3R)-
serine-2 and lipid 430 is (3R)- -serine-1. In summary, the convergent synthesis described herein
confirms the proposed structures of lipid 654 produced from P. gingivalis. Our convergent
synthesis gave lipid 430, (3R)- -serine-1, in 8 linear steps and 24.7% overall yield, using the
protected dipeptide 13 that was prepared separately. We similarly prepared (3S)- -serine-1 in 8
linear steps and 26.8% overall yield. Lipid 654, (3R)- -serine-2, was prepared in 9 linear steps
and 34% overall yield. We similarly prepared (3S)-L-serine-2 in 9-linear steps and 31.1% overall
yield. Using the same approaches, (3R+3S)- -serine-1 was prepare in 8 linear steps and 49.9%
overall yield and (3R+3S)- -serine-2 was prepared in 9 linear steps ad 46.3% overall yield.
L
-serine-2 and (3R)-
L
-serine 1, as well as (3S)-
L-serine-2 and
(3S)-
L
L
-
L
L
L
L
L
L
Demonstration that there is a difference in enzymatic reactivity with the different
diastereomers allows us to use the (R) and (S) synthetic preparations of lipid 654 to evaluate
biological properties of enantioenriched or enantiopure lipids to engage the innate immune
system, including the engagement of Toll-like receptor 2. Furthermore, it is important to
determine the levels of (R) and (S) lipid 654 present in other members of the Bacteroidetes
phylum as well as the levels of these isoforms in diseased tissues where lipid 654 is known to
accumulate. These studies have begun, and the results will be the subject of a future
communication.
Acknowledgements

ACCEPTED MANUSCRIPT
17
The work was supported, in part, by NIH Grant DE 021055. We thank the UCONN
Mass Spectra facility, under the direction of Dr. You-Jun Fu, for their efforts.
Experimental Section
All glassware was oven-dried, and all reactions were performed under a nitrogen
atmosphere. All chemicals were purchased from the Sigma-Aldrich Chemical Co., and used
without further purification unless otherwise noted. Tetrahydrofuran (THF) was distilled from
sodium benzophenone ketyl, methylene chloride (dichloromethane, DCM) was distilled from
calcium hydride, and dimethylformamide (DMF) was distilled in vacuo from calcium hydride.
Ethyl acetate (EtOAc), methanol (MeOH), and diethyl ether were used as obtained from the
vendor. Bobbitt’s salt15 was provided as a gift from Dr. James Bobbitt (Department of
Chemistry, The University of Connecticut)), but it is commercially available from Sigma-
Aldrich as 4-acetamido-2,2,6,6-tetramethylpiperidine 1-oxyl. Thin-layer chromatography was
done on Sorbent Technologies aluminum-backed TLC plates with fluorescent indicator and 0.2
mm silica gel layer thickness, and p-anisaldehyde or phosphomolybdic acid were used as
developing agents. Column chromatography was done using 60 Å porosity, 32-63 µm silica gel.
13
1
13
1H and C NMR were collected on a Bruker Avance 300 (300.13 MHz H, 75.48 MHz C),
Bruker DRX-400 (400.144 MHz 1H, 100.65 MHz 13C) or a Bruker Avance 500 (500.13 MHz 1H,
125.65 MHz 13C). Chemical shifts are reported in ppm downfield from tetramethylsilane (TMS)
in the following format chemical shift, multiplicity (s=singlet, d=doublet, t=triplet, q=quartet,
m=multiplet). Coupling constants are reported in Hz. Mass spectrometry data was collected on a
HP 5870B GC/MSD mass spectrometer with an HP-1 column, and high resolution MS and

ACCEPTED MANUSCRIPT
18
MS/MS spectra of synthetic 2 and 1 were obtained by direct infusion of the target lipids into a
QTOF mass spectrometer, QSTAR Elite from Sciex (Foster City, CA) or directly by using
AccuTOF™ DART, JEOL (Peabody, MA). Multiple reaction monitoring (MRM) transitions
were selected based the MS/MS spectra of 2 and 1. Lipid samples were injected into a triple
quadrupole mass spectrometer, 4000 QTrap from Sciex (Foster City, CA). IR spectra were taken
on FT/IR-410/C031560585 JASCO and Nexus 670 FT-IR E.S.P, neat, unless otherwise stated.
All melting points to an upper limit of 270 °C were obtained using a Uni-melt capillary melting
point apparatus or Digimelt MPA160. Perfect. For products described as waxy solid or semi-
solids, melting points could not be obtained.
Methyl
L
-serinate
Thionyl chloride (2.75 mL, 37.8 mmol) was added, dropwise, to a suspension of
L-serine
(4.015 g, 38.2 mmol) in 50 mL of dry methanol. The solution was heated at reflux with stirring
for 5 h and then cooled to ambient temperature. The solvent was removed in vacuo and the
crude white solid was redissolved in a minimal amount of methanol. This solution was treated
with an equal volume of toluene and the flask was swirled periodically to induce crystallization.
Filtration yielded methyl
-serinate as a crystalline solid (3.9675 g, 33.3 mmol, 87.2%).20 MP,
L
1
161-162 °C. H NMR (400 MHz, Methanol-d4) δ 4.14 (t, J=3.9 Hz, 1H), 4.01 (dd, J=11.8, 4.5
Hz, 1H), 3.93 (dd, J=11.8, 3.1 Hz, 1H), 3.85 (s, 3H). 13C NMR (101 MHz, Methanol-d4) δ 169.4,
60.7, 56.1, 53.71.
Methyl O-(tert-butyldiphenylsilyl)- -serinate 12
L

ACCEPTED MANUSCRIPT
19
A stirring solution of methyl -serinate (2.003 g, 16.8 mmol) in 15 mL of DMF was
L
treated with imidazole (3.435 g, 50.4 mmol).17 (tert-Butyl)diphenylsilyl chloride (4.80 mL, 18.5
mmol) was added, dropwise, and the reaction was stirred overnight at ambient temperature. At
this time, the solvent was removed in vacuo and the resulting oil was dissolved in 100 mL of
diethyl ether. This solution was washed with water (5 X 400 mL), 10 % LiCl (2 X 100), and
brine (50 mL). Drying with MgSO4, filtering, and concentration under reduced pressure yielded
methyl O-(tert-butyldiphenylsilyl)-
1H NMR (400 MHz, CDCl
L
-serinate, 12, as a yellow oil (5.331 g, 14.9 mmol, 88.7%).
)
δ 7.64 (td, J = 7.7, 1.6 Hz, 4H), 7.50-7.35 (m, 6H), 3.98 (dd, J =
3
9.9, 4.3 Hz, 1H), 3.89 (dd, J = 9.9, 3.8 Hz, 1H), 3.71 (s, 3H), 3.58 (t, J = 4.1 Hz, 1H), 1.04 (s,
13
9H). C NMR (101 MHz, CDCl
) δ 135.6, 133.1, 133.0, 129.9, 127.8, 127.7, 66.1, 56.5, 52.1,
3
26.8, 19.3. HRMS (AccuTOF): [MH]+ Calc’d for C20H28NO3Si m/z 358.1838. Found: m/z
358.1819.
((Benzyloxy)carbonyl)glycine [N-Cbz-Glycine], 11
Glycine (3.000 g, 40.0 mmol) and sodium bicarbonate (6.710 g, 79.9 mmol) were
dissolved in 100 mL of H O at ambient temperature and this solution was cooled to 0 °C using
2
an ice bath. Benzyl chloroformate (8.56 mL, 59.9 mmol) in 25 mL of dioxane was added,
dropwise. This solution was slowly warmed to ambient temperature and stirred overnight. The
reaction mixture was transferred to a separatory funnel and washed with EtOAc (2 X 50 mL).
The aqueous layer was acidified with 10 % HCl to a pH=1 (to pH paper) and extracted with
EtOAc (3 X 50 mL). The later organic layers were dried, filtered, and solvents evaporated at
reduced pressure to yield ((benzyloxy)carbonyl)glycine, 11,21 as a white solid (6.982 g, 33.4
1
mmol, 83.5 %), MP, 112-114 °C. H NMR (400 MHz, Methanol-d4) δ 7.40–7.24 (m, 5H), 5.10

ACCEPTED MANUSCRIPT
20
13
(s, 2H), 3.84 (s, 2H). C NMR (101 MHz, Methanol-d4) δ 173.6, 159.1, 138.1, 129.4, 129.0,
128.8, 67.7, 43.1.
Methyl N-(((benzyloxy)carbonyl)glycyl)-O-(tert-butyldiphenylsilyl)- -serinate, 13
L
A solution of ((benzyloxy)carbonyl)glycine (1.748 g, 8.36 mmol) in 10 mL of DMF was
cooled to 0°C using an ice bath. 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI, 1.603
g, 8.36 mmol), HOBt (1.280 g, 8.36 mmol) and then diisopropylethylamine (DIPEA, 1.46 mL,
8.36 mmol) were added in that order. The solution was stirred for 15 min before the addition of
methyl O-(tert-butyldiphenylsilyl)- -serinate, 12 (2.988 g, 8.36 mmol). The reaction was
L
warmed to ambient temperature and stirred for 48 h. At that time, the DMF was removed in
vacuo and the resulting oil was dissolved in 100 mL of diethyl ether. The ether was washed with
water (5 X 400 mL), saturated NaHCO3 (3 X 100) mL, and brine (50 mL). The organic layer
was dried, filtered, and solvents evaporated at reduced pressure. Purification by column
chromatography (1
butyldiphenylsilyl)-
MHz, CDCl
%
MeOH:DCM) yielded (((benzyloxy)carbonyl)glycyl)-O-(tert-
1
L
-serinate, 13, as a clear oil (3.390 g, 6.18 mmol, 73.9%). H NMR (400
)
δ 7.59 (t, J = 7.6 Hz, 4H), 7.49–7.34 (m, 6H), 6.77 (d, J = 6.9 Hz, 1H), 5.36 (s,
3
1H), 5.12 (q, J= 12.3, 7.6 Hz, 2H) 4.68 (dt, J = 8.3, 2.9 Hz, 1H), 4.13 (dd, J = 10.3, 2.8 Hz, 1H),
3.93–3.75 (m, 3H), 3.74 (s, 3H), 1.04 (s, 9H). 13C NMR (101 MHz, CDCl
)
δ 170.5, 168.6,
3
156.4, 136.1, 135.5, 132.8, 132.6, 130.1, 128.6, 128.2, 128.1, 127.9, 127.8, 67.3, 64.2, 54.1, 52.5,
44.4, 26.7, 19.3. HRMS (AccuTOF): [MH]+ Calc’d for C30H37N2O6Si m/z 549.2421. Found: m/z
549.2423.
13-Methyltetradecan-1-ol, 7

ACCEPTED MANUSCRIPT
21
Magnesium turnings (2.90 g, 0.119 mol) were added to a 500 mL round bottom flask
with a stirbar and flame-dried under a nitrogen atmosphere. After cooling, the Mg turnings were
suspended in dry THF (50 mL) and 1-bromo-2-methylpropane (8.00 mL, 0.0736 mol) was
rapidly added with vigorously stirring. Upon initiation, dry THF (50 mL) was added and the
flask was immersed in an ice bath to modulate the exothermic reaction. The reaction was stirred
until the flask cooled to ambient temperature and the temperature was lowered to -78 °C using
dry ice acetone bath. 11-Bromoundecan-1-ol (8, 4.005 g, 0.0172 mol) dissolved in THF (20 mL)
and LiCl (1.53 g, 0.036 mol) and copper (II) chloride (2.54 g, 0.019 mol), dissolved in THF (30
mL), were added. The mixture was slowly warmed to ambient temperature and stirred
overnight. The reaction was quenched with satd NH Cl (100 mL) before being transferred to a
4
separatory funnel with diethyl ether (200 mL) and H O (100 mL). The layers were separated and
2
the organic layer was washed with satd NaHCO (50 mL X 3) and brine (50 mL). The organic
3
phase was dried with MgSO , gravity filtered, and solvents removed in vacuo. The crude product
4
was purified by column chromatography on silica gel (10% EtOAc:hexane) to yield 13-
1
methyltetradecan-1-ol, 7,10,22 as a white solid (3.793 g, 0.017 mol, 96.7%), MP, 26-27 °C. H
NMR (400 MHz, CDCl
) δ 3.64 (t, 2H), 1.63–1.44 (m, 3H), 1.27 (m, 18H), 1.17 (m, 3H), 0.86
3
(d, J = 6.6 Hz, 6H); 13C NMR (101 MHz, CDCl
28.1, 27.6, 25.9, 22.8.
) δ 63.3, 39.2, 33.0, 30.1, 29.9, 29.8, 29.8, 29.6,
3
13-Methyltetradecanal, 6
4-(Acetyamino)-2,2,6,6-tetramethyl-1-oxo-piperidinium tetrafluoroborate15 (Bobbit’s
salt, 5.285 g, 0.0176 mol) was added to 13-ethyltetradecanol (7, 3.498 g, 0.0153 mol) dissolved
in 250 mL of dry DCM. An equal mass of silica was slowly added to the stirring mixture. The

ACCEPTED MANUSCRIPT
22
bright yellow solution stirred at ambient temperature until the color faded, giving a solution with
a pale yellow tint. The solution was filtered to remove the silica, and the filtrate was
concentrated in vacuo. The crude product was purified by column chromatography (5%
EtOAc:hexanes) to yield 13-methyltetradecanal, 6,23 as a clear solid, (3.319 g, 0.0147 mol,
1
95.8%), MP, ~25 °C. H NMR(400 MHz, CDCl
) δ 9.76 (t, J = 1.9 Hz, 1H), 2.41 (td, J = 7.4,
3
1.9 Hz, 2H), 1.63 (quin, J = 7.2 Hz, 2H), 1.51 (sep, 1H), 1.36–1.22 (m, 16H), 1.15 (m, 2H), 0.86
(d, J = 6.6 Hz, 6H) 13C NMR (101 MHz, CDCl
)
δ 13C NMR (101 MHz, CDCl ) δ 203.0, 44.0,
3
3
39.1, 30.0, 29.8, 29.7, 29.6, 29.5, 29.4, 29.2, 28.0, 27.5, 22.7, 22.2.
13-Methyltetradecanoic acid, 3
13-Methyltetradecanal (6, 1.251 g, 5.53 mmol) was dissolved in 50 mL of THF, and 5
mL of H O and 0.5 mL of 1M HCl were added. The reaction mixture was stirred for 72 h at
2
ambient temperature. The resulting solution was concentrated under reduced pressure, the
residue dissolved in 15 mL of H2O, and extracted with DCM (15 mL X 3). The resulting
solution was filtered through a small plug of silica gel and rinsed with DCM before being
concentrated in vacuo to yield 13-methyltetradecanoic acid10,24 (3) as a white solid (1.307 g, 5.39
mmol, 97.6%), MP, 48-49 °C. 1H NMR (400 MHz, CDCl
) δ 2.35 (t, J = 7.5 Hz, 2H), 1.63
3
(quin, J = 7.4 Hz, 2H), 1.51 (sep, J = 6.6 Hz, 1H), 1.39–1.18 (m, 16H), 1.15 (q, J = 6.7 Hz, 2H),
0.86 (d, J = 6.6 Hz, 6H). 13C NMR (101 MHz, CDCl
)
δ 13C NMR (101 MHz, CDCl ) δ 179.6,
3
3
39.1, 34.0, 30.0, 29.8, 29.7, 29.7, 29.5, 29.3, 29.1, 28.0, 27.5, 24.8, 22.7.
Ethyl 15-methyl-3-oxohexadecanoate, 9

ACCEPTED MANUSCRIPT
23
Tin (II) chloride dehydrate (0.417 g, 1.85 mmol) was suspended in 15 mL of dry DCM.
Ethyl diazoacetate (2.60 g, 19.4 mmol) was added, dropwise via syringe, and the reaction was
stirred for 15 min. 13-Methylpentadecanal (6, 4.186 g, 18.5 mmol) was dissolved in 10 mL of
DCM and added to the reaction, dropwise, via cannula. The mixture was stirred at ambient
temperature for about 6 h, until nitrogen evolution ceased. The reaction mixture was diluted
with 100 mL of DCM and washed with a satd brine solution (2 X 50 mL). The aqueous layers
were combined and extracted with DCM (2 X 25 mL). All organic layers were combined, dried
with MgSO4, filtered, and solvents evaporated at reduced pressure. The resulting crude product
was purified via column chromatography (5% EtOAc:hexanes) to yield ethyl 15-methyl-3-
1
oxohexadecanoate,10,25 9, as a clear solid (4.871 g, 0.0156 mol, 84.3%), MP, ~25 °C. H NMR
(400 MHz, CDCl ) δ 4.20 (q, J = 7.1 Hz, 2H), 3.42 (s, 2H), 2.53 (t, J = 7.3 Hz, 2H), 1.60 (quin, J
3
= 7.2 Hz, 2H), 1.50 (sep, J = 6.5 Hz, 1H), 1.30-1.20 (m, 20H), 1.15 (q, J = 7.3, 6.8 Hz, 2H), 0.86
13
(d, J = 6.6 Hz, 6H). C NMR (101 MHz, CDCl ) δ 203.0, 167.3, 61.4, 49.4, 43.1, 39.1, 30.0,
3
29.8, 29.7, 29.7, 29.5, 29.4, 29.1, 28.0, 27.5, 23.5, 22.7, 14.2.
Ethyl 3-hydroxy-15-methylhexadecanoate, 10
A solution of ethyl 15-methyl-3-oxohexadecanoate (9, 3.114 g, 0.001 mol) in 2.5 mL of
ethanol and 22.5 mL of THF was cooled to 0 °C on an ice bath. Sodium borohydride (0.301 g,
0.008 mol) was added and the resulting slurry was stirred vigorously for 1 h. The reaction was
quenched by the slow addition of 50 mL of a 10 % aq citric acid solution. Subsequent
neutralization with satd K CO solution was followed by extraction with EtOAc (3 X 50 mL).
2
3
The organic layers were washed with 25 mL of satd brine solution, dried with MgSO4, filtered,
and the solvents were removed at reduced pressure. The crude product was purified by column

ACCEPTED MANUSCRIPT
24
chromatography (10% EtOAc:hexanes) to yield ethyl 3-hydroxy-15-methylhexadecanoate,10 10,
1
as a clear oil (2.674 g, 0.0085 mol, 85.3 %). H NMR (400 MHz, CDCl
)
δ 4.17 (q, J = 7.1 Hz,
3
2H), 4.06–3.93 (m, 1H), 2.91 (d, J = 3.9 Hz, 1H), 2.50 (dd, J = 16.4, 3.1 Hz, 1H), 2.39 (dd, J =
13
16.4, 9.0 Hz, 1H), 1.60–1.20 (m, 23H), 1.15 (q, J = 6.6 Hz, 2H), 0.86 (d, J = 6.6 Hz, 6H). C
NMR (101 MHz, CDCl ) δ 173.2, 68.1, 60.7, 41.4, 39.1, 36.6, 30.0, 29.8, 29.7, 29.6, 29.6, 28.0,
3
27.5, 25.5, 22.7, 14.2.
Ethyl (3R)-hydroxy-15-methylhexadecanoate, (3R)-10
In a nitrogen atmosphere glove box, (R)-BINAP (45 mg, 0.072 mmol) and bis(2-
methylallyl)(1,5-cyclooctadiene)ruthenium(II) (20 mg, 0.063 mmol) were dissolved in 9 mL of
freshly distilled acetone. Methanolic HBr (0.22 mL, 0.73 M, 0.16 mmol) was added, and the
solution was stirred for 1 h. The reaction was concentrated to dryness in vacuo before being
redissolved in 15 mL of ethanol and ꢀ-keto ester 9 (0.962 g, 3.08 mmol) was added. The system
was flushed with hydrogen and heated at reflux for 48 h. The resulting solution was filtered and
the solvents removed at reduced pressure. Subsequent purification via column chromatography
23
(10% EtOAc:hexanes) yielded ethyl (3R)-hydroxy-15-methylhexadecanoate, (3R)-1010 [ꢁ]D
–
1
9.2° (c 1.0, CHCl3), as a clear solid (0.901 g, 2.86 mol, 93.0 %), MP, ~25 °C. H NMR (400
MHz, CDCl ) δ 4.17 (q, J = 7.2 Hz, 2H), 4.05–3.94 (m, 1H), 2.91 (d, J = 4.0 Hz, 1H), 2.50 (dd, J
3
= 16.4, 3.1 Hz, 1H), 2.39 (dd, J = 16.4, 9.0 Hz, 1H), 1.60–1.20 (m, 23H), 1.15 (q, J = 6.6 Hz,
2H), 0.86 (d, J = 6.6 Hz, 6H). 13C NMR (101 MHz, CDCl ) δ 173.2, 68.1, 60.7, 41.4, 39.1, 36.6,
3
1
30.0, 29.8, 29.7, 29.6, 29.6, 28.0, 27.5, 25.5, 22.7, 14.2. A H NMR analysis that showed a
97R:03S mixture for (3R)-5.

ACCEPTED MANUSCRIPT
25
Ethyl (3S)-hydroxy-15-methylhexadecanoate, (3S)-10
In a nitrogen atmosphere glove box, (S)-BINAP (45 mg, 0.072 mmol) and bis(2-
methylallyl)(1,5-cyclooctadiene)ruthenium(II) (20 mg, 0.063 mmol) were dissolved in 9 mL of
freshly distilled acetone. Methanolic HBr (0.22 mL, 0.73 M, 0.16 mmol) was added, and the
solution stirred for 1 h. The reaction was concentrated to dryness in vacuo before being
redissolved in 15 mL of ethanol and
was flushed with hydrogen gas and heated at reflux for 48 h. The resulting solution was filtered
and the solvents removed at reduced pressure. Subsequent purification via column
ꢀ-keto ester 9 (0.954 g, 3.05 mmol) was added. The system
chromatography (10% EtOAc:hexanes) yielded ethyl (3S)-hydroxy-15-methylhexadecanoate,
(3S)-10 as a clear solid (0.917 g, 2.92 mol, 95.5 %), MP, ~25 °C. [ꢁ]D23 +11.8° (c 1.0, CHCl3),
1H NMR (400 MHz, CDCl
) δ 4.17 (q, J = 7.1 Hz, 2H), 4.06–3.93 (m, 1H), 2.91 (d, J = 4.0 Hz,
3
1H), 2.50 (dd, J = 16.4, 3.1 Hz, 1H), 2.39 (dd, J = 16.4, 9.0 Hz, 1H), 1.55–1.39 (m, 3H), 1.38–
13
1.21 (m, 20H), 1.15 (q, J = 7.1, 6.6 Hz, 2H), 0.86 (d, J = 6.6 Hz, 6H). C NMR (101 MHz,
CDCl ) δ 173.2, 68.1, 60.7, 41.4, 39.1, 36.6, 30.0, 29.8, 29.7, 29.6, 29.6, 28.0, 27.5, 25.5, 22.7,
3
14.2. A 1H NMR analysis that showed a 02R:98S mixture for (3S)-5.
Enzymatic Resolution
Lipase PS (.275 g) was added to a stirred solution of (±)-5 (5.50 g, 1.92 mmol) and 2,6-
di-tert-butyl-4-methylphenol (5 mg) in vinyl acetate (8.8 mL), and stirred at 52-58 °C for 168 h.
The resulting mixture was cooled to ambient temperature and filtered through Celite and
concentrated under reduced pressure. Column chromatography (MeOH: DCM), yielded (3R)-5
ͦͧ
(128 mg) as a white solid; [α]
+6.4° (c 1.02, CHCl ). After saponification of the acetate
3
͸

ACCEPTED MANUSCRIPT
26
ͦͧ
͸
1
product, (3S)-5 (177 mg) was purified to give a white solid; [α]
–8.7° (c 1.02, CHCl ). H
3
NMR spectra were identical with those of (±)-5.16c Conversion to the corresponding Mosher’s
1
ester as described above was followed by H NMR analysis that showed a 84R:15S mixture for
(3R)-5 and a 11R:89S mixture for (3S)-5.
Ethyl (R)-15-methyl-3-(((S)-3,3,3-trifluoro-2-methoxy-2-phenylpropanoyl) oxy)-
hexadecanoate
Under a nitrogen atmosphere, (S)-(–)-α-Methoxy-α-(trifluoromethyl)phenylacetic acid
(Mosher’s acid,26 23 mg, 0.098 mmol) was dissolved in 2 mL of dry DCM and cooled to 0 °C in
an ice bath. Subsequent addition of DCC (24 mg, 0.12 mol) was followed by stirring for 15 min,
at which time ethyl (3R)-hydroxy-15-methylhexadecanoate, (3R)-10 (29 mg, 0.092 mmol) was
added, along with a catalytic amount of DMAP. The reaction was equilibrated to ambient
temperature slowly and stirred overnight. Evaporation of solvents in vacuo was followed by
addition of 15 mL of EtOAc and washing with H2O 3 X 10 mL and 10 mL of satd brine. The
resulting organic layer was dried over MgSO4, filtered, and the solvents removed at reduced
pressure. Subsequent purification via column chromatography (1-10% EtOAc:hexanes) yielded
ethyl (R)-15-methyl-3-(((S)-3,3,3-trifluoro-2-methoxy-2-phenylpropanoyl)oxy)hexadecanoate as
1
a clear oil (38 mg, 0.072 mmol, 78 %). H NMR (400 MHz, CDCl
)
δ 7.55-7.50 (m, 2H), 7.42-
3
7.37 (m, 3H), 5.52-5.45 (p, J = 6.4 Hz, 1H), 4.06 (q, J = 7.1 Hz, 2H), 3.53 (s, 3H), 2.64 (dd, J =
16.0, 7.9 Hz, 1H), 2.55 (dd, J= 16.0, 5.0 Hz, 1H) 1.80-1.60 (m, 2H), 1.55-1.47 (sep, J = 6.6 Hz,
1H), 1.36-1.12 (m, 23H), 0.86 (d, J = 6.6 Hz, 6H).
Ethyl (S)-15-methyl-3-(((S)-3,3,3-trifluoro-2-methoxy-2-phenylpropanoyl)oxy)-
hexadecanoate

ACCEPTED MANUSCRIPT
27
Under a nitrogen atmosphere, (S)-(–)-α-Methoxy-α-(trifluoromethyl)phenylacetic acid
(Mosher’s acid,26 23 mg, 0.098 mmol) was dissolved in 2 mL of dry DCM and cooled to 0 °C in
an ice bath. Subsequent addition of DCC (23 mg, 0.11 mmol) was followed by stirring for 15
min, at which time ethyl (3S)-hydroxy-15-methylhexadecanoate, (3S)-10 ((28 mg, 0.089 mmol)
was added, along with a catalytic amount of DMAP. The reaction was equilibrated to ambient
temperature slowly and stirred overnight. Evaporation of solvents in vacuo was followed by
addition of 15 mL of EtOAc and washing with H2O 3 X 10 mL and 10 mL of satd brine. The
resulting organic layer was dried over MgSO4, filtered, and the solvents removed at reduced
pressure. Subsequent purification via column chromatography (1-10% EtOAc:hexanes) yielded
ethyl (R)-15-methyl-3-(((S)-3,3,3-trifluoro-2-methoxy-2-phenylpropanoyl)oxy)hexadecanoate as
1
a clear oil (45 mg, 0.084 mmol, 94 %). H NMR (400 MHz, CDCl
)
δ 7.50-7.55 (m, 2H), 7.37-
3
7.42 (m, 3H), 5.48 (quin, J = 6.3 Hz, 1H), 4.12 (q, J = 7.1 Hz, 2H), 3.55 (s, 3H), 2.69 (dd, J =
16.0, 8.0 Hz, 1H), 2.59 (dd, J= 16.0, 4.8 Hz, 1H), 1.70-1.57 (m, 2H), 1.52 (sep, J = 6.6 Hz, 1H),
1.36-1.12 (m, 23H), 0.86 (d, J = 6.6 Hz, 6H).
3-Hydroxy-15-methylhexadecanoic acid, (3R+3S)-5
Hydroxy ester 10 (2.674 g, 0.0085 mol) was dissolved in 8 mL of MeOH and 16 mL of
THF and 1 M aqueous LiOH (9.40 mL, 0.0094 mol) were added. The mixture was stirred
overnight at ambient temperature. The reaction was then acidified with 1 M HCl and extracted
with EtOAc (3 X 50 mL). The organic layers were washed with 25 mL of brine, dried with
MgSO4, filtered, and the solvents were removed at reduced pressure. The crude solid was
recrystallized from hexanes to yield 3-hydroxy-15-methylhexadecanoic acid,10 (3R+3S)-5, as a
1
white solid (2.257 g, 0.0079 mol, 92.7 %), MP, 54-56 °C. H NMR (400 MHz, CDCl
) δ 4.09–
3

ACCEPTED MANUSCRIPT
28
3.98 (m, 1H), 2.58 (dd, J = 16.6, 3.2 Hz, 1H), 2.48 (dd, J = 16.6, 8.9 Hz, 1H), 1.59–1.43 (m, 3H),
13
1.35-1.20 (m, 18H), 1.15 (q, J = 6.7 Hz, 2H), 0.86 (d, J = 6.6 Hz, 6H). C NMR (101 MHz,
CDCl ) δ 177.4, 68.1, 41.2, 39.1, 36.6, 30.0, 29.8, 29.7, 29.6, 29.6, 29.5, 28.0, 27.5, 25.5, 22.7.
3
3(R)-Hydroxy-15-methylhexadecanoic acid, (3R)-5
Hydroxy ester (3R)-10 (0.818 g, 0.0026 mol) was dissolved in 8 mL of methanol and 16
mL of THF and 1 M aqueous LiOH (2.9 mL, 0.0029 mol) were added. The mixture was stirred
overnight at ambient temperature. The reaction was then acidified with 1 M HCl and extracted
with EtOAc (3 X 50 mL). The organic layers were washed with 25 mL of brine, dried with
MgSO4, filtered, and the solvents were removed at reduced pressure. The crude solid was
recrystallized from hexanes to yield (3R)-3-hydroxy-15-methylhexadecanoic acid,10 (3R)-5, as a
1
white solid (0.708 g, 0.0024 mol, 92 %), [ꢁ]D23 –10.5° (c 1.0, CHCl3), MP, 54-55 °C. H NMR
(400 MHz, CDCl ) δ 4.09–3.98 (m, 1H), 2.58 (dd, J = 16.6, 3.2 Hz, 1H), 2.48 (dd, J = 16.6, 8.9
3
Hz, 1H), 1.59–1.43 (m, 4H), 1.35-1.20 (m, 18H), 1.15 (q, J = 6.7 Hz, 2H), 0.86 (d, J = 6.6 Hz,
13
6H). C NMR (101 MHz, CDCl ) δ 177.8, 68.1, 41.1, 39.1, 36.6, 30.0, 29.8, 29.7, 29.6, 29.5,
3
28.0, 27.5, 25.5, 22.7.
3(S)-Hydroxy-15-methylhexadecanoic acid, (3S)-5
Hydroxy ester (3S)-10 (0.815 g, 0.0026 mol) was dissolved in 8 mL of methanol and 16
mL of THF and 1 M aqueous LiOH (2.9 mL, 0.0029 mol) were added. The mixture was stirred
overnight at ambient temperature. The reaction was then acidified with 1 M HCl and extracted
with EtOAc (3 X 50 mL). The organic layers were washed with 25 mL of brine, dried with
MgSO4, filtered, and the solvents were removed at reduced pressure. The crude solid was