Two-step biocatalytic route to biobased functional polyesters from ω-carboxy fatty acids and diols

Article abstract of DOI:10.1021/bm901112m

Biobased ω-carboxy fatty acid monomers 1,18-cis-9-octadecenedioic, 1,22-cis-9-docosenedioic, and 1,18-cis-9,10-epoxy-octadecanedioic acids were synthesized in high conversion yields from oleic, erucic and epoxy stearic acids by whole-cell biotransformations catalyzed by C. tropicalis ATCC20962. Maximum volumetric yields in shake-flasks were 17.3, 14.2, and 19.1 g/L after 48 h conversion for oleic acid and 72 h conversions for erucic and epoxy stearic acids, respectively. Studies in fermentor with better control of pH and glucose feeding revealed that conversion of oleic acid to 1,18-cis-9-octadecenedioic acid by C. tropicalis ATCC20962 occurred with productivities up to 0.5 g/L/h. The conversion of ω-carboxy fatty acid monomers to polyesters was then studied using immobilized Candida antarctica Lipase B (N435) as catalyst. Polycondensations with diols were performed in bulk as well as in diphenyl ether. The retension of functionality from fatty acid, to ω-carboxy fatty acid monomer and to corresponding polyesters resulted in polymers with with unsaturated and epoxidized repeat units and M_w values ranging from 25000 to 57000 g/mol. These functional groups along chains disrupted crystallization giving materials that are low melting (23-40 °C). In contrast, saturated polyesters prepared from 1,18-octadecanedioic acid and 1,8-octanediol have correspondingly higher melting transitions (88 °C). TGA results indicated that all synthesized polyesters showed high thermal stabilities. Thus, the preparation of functional monomers from C. tropicalis ω-oxidation of fatty acids provides a wide range of new monomer building blocks to construct functional polymers.

Full text of DOI:10.1021/bm901112m

Biomacromolecules 2010, 11, 259–268
259
Two-Step Biocatalytic Route to Biobased Functional Polyesters
from ω-Carboxy Fatty Acids and Diols
Yixin Yang, Wenhua Lu, Xiaoyan Zhang, Wenchun Xie, Minmin Cai, and Richard A. Gross*
NSF I/UCRC for Biocatalysis and Bioprocessing of Macromolecules, Department of Chemical and
Biological Sciences, Polytechnic Institute of NYU, Six Metrotech Center, Brooklyn, New York 11201
Received September 29, 2009; Revised Manuscript Received November 25, 2009
Biobased ω-carboxy fatty acid monomers 1,18-cis-9-octadecenedioic, 1,22-cis-9-docosenedioic, and 1,18-cis-
9,10-epoxy-octadecanedioic acids were synthesized in high conversion yields from oleic, erucic and epoxy stearic
acids by whole-cell biotransformations catalyzed by C. tropicalis ATCC20962. Maximum volumetric yields in
shake-flasks were 17.3, 14.2, and 19.1 g/L after 48 h conversion for oleic acid and 72 h conversions for erucic
and epoxy stearic acids, respectively. Studies in fermentor with better control of pH and glucose feeding revealed
that conversion of oleic acid to 1,18-cis-9-octadecenedioic acid by C. tropicalis ATCC20962 occurred with
productivities up to 0.5 g/L/h. The conversion of ω-carboxy fatty acid monomers to polyesters was then studied
using immobilized Candida antarctica Lipase B (N435) as catalyst. Polycondensations with diols were performed
in bulk as well as in diphenyl ether. The retension of functionality from fatty acid, to ω-carboxy fatty acid monomer
and to corresponding polyesters resulted in polymers with with unsaturated and epoxidized repeat units and Mw
values ranging from 25000 to 57000 g/mol. These functional groups along chains disrupted crystallization giving
materials that are low melting (23-40 °C). In contrast, saturated polyesters prepared from 1,18-octadecanedioic
acid and 1,8-octanediol have correspondingly higher melting transitions (88 °C). TGA results indicated that all
synthesized polyesters showed high thermal stabilities. Thus, the preparation of functional monomers from C.
tropicalis ω-oxidation of fatty acids provides a wide range of new monomer building blocks to construct functional
polymers.
It is well-known that many microorganisms can convert
n-alkanes and fatty acids to their corresponding R,ω-diacids,
including Candida tropicalis,4,6,7 Candida cloaca,8 Cryptococ-
cus neoforman,1 and Corynebacterium sp.9 Candida tropicalis
and similar yeasts produce R,ω-diacids through an ω-oxidation
pathway. The terminal methyl group is first hydroxylated by a
cytochrome P450 monooxygenase that is further transformed
via the action of fatty alcohol oxygenase and aldehyde dehy-
drogenase to form diacids.10 Of particular interest with regards
to this paper is the report of C. tropicalis ATCC20962, which
was engineered to block its ꢀ-oxidation pathway.4 Yi et al.11
used a mutant of C. tropicalis to convert oleic acid to 1,18-
cis-9-octadecenedioic acid. Picataggio et al.4 used engineered
C. tropicalis ATCC20962 and its P450 monooxygenase
(P450alkl) and NADPH-oxidoreductase (CPR) amplified strain
AR40 to convert methyl esters of saturated (C14 to C18) and
unsaturated (oleic, erucic) fatty acids to their corresponding R,ω-
diacids. Bioconversion of C12 to C22 n-alkanes and fatty acids
to R,ω-diacids by strain AR40 was successfully performed
without undesired modification of substrates or products via the
ꢀ-oxidation pathway. Fabritius et al.12,13 reported that a mutant
of C. tropicalis M25 converts oleic acid12 and linoleic acid13
to their corresponding 3-hydroxydicarboxylic acids. Unfortu-
nately, bioconversions using of C. tropicalis described in Yi et
al.11 and Fabritius et al.12,13 resulted in low yields and product
mixtures, including R,ω-diacids with different carbon-chain
lengths or having variable contents of hydroxylation along
chains. Furthermore, comparison of C. tropicalis ATCC20962-
catalyzed conversion rates of unsaturated acids differing in chain
lengths to their corresponding R,ω-diacids has not been studied.
Moreover, whereas the in vitro ω-hydroxylation of 9,10-
epoxystearic acid, catalyzed by cytochrome P450s in plants such
Introduction
Aliphatic R,ω-dicarboxylic acids (R,ω-diacids) are widely
used as raw materials for the manufacture of engineered plastics,
perfumes, fragrances, lubricants, and adhesives.1 The majority
of currently used R,ω-diacids are produced by chemical means
from nonrenewable petrochemical feedstocks. For examples,
adipic acid is manufactured by a two-step process involving
stoichiometricnitricacidoxidationofacyclohexanol-cyclohenanone
mixture, which is aerobically generated from cyclohexane using
a homogeneous cobalt-based catalyst.2 Dodecanedioic acid is
manufactured by a nickel-catalyzed cyclic trimerization of
butadiene, followed by hydrogenation to cyclododecane, air
oxidation to a mixture of cyclododecanone and cyclododecanol,
and, finally, nitric acid oxidation to dodecanedioic acid.3 Also,
chemical routes to R,ω-diacids can be tedious and result in
unwanted byproducts. Furthermore, chemical routes are unavail-
able to synthesize R,ω-diacids with carbon numbers greater than
13.4
Nature uses long-chain unsaturated and epoxidized dicar-
boxylic acids (mainly 9,10-epoxy octadecanedioic, 1,18-cis-9-
octadecenedioic, and 9,10-dihydroxy octadecanedioic acids), as
well as ω-hydroxyl carboxylic acids (mainly 9,10-epoxy-18-
hydroxy octadecanoic and 9,10,18-trihydroxy octadecanoic
acids) as building blocks to synthesize important plant polyesters
such as suberin and cutin.5 These and related monomers would
be useful to design and synthesize unique functional polyesters
that would also biodegrade. However, they are difficult to
synthesize by chemical methods and are currently unavailable
commercially.
* To whom correspondence should be addressed. Telephone: 718-260-
3024. Fax: 718-260-3075. E-mail: rgross@poly.edu.
10.1021/bm901112m
 2010 American Chemical Society
Published on Web 12/10/2009

260 Biomacromolecules, Vol. 11, No. 1, 2010
Yang et al.
as CYP86A1, CYP86A8, CYP94A1, and CYP94A5, has been
studied,14 corresponding conversions of epoxy-containing fatty
acids to their corresponding epoxidized R,ω-diacids by C.
tropicalis has not been reported. Resulting R,ω-diacid monomers
with an internal double bond, epoxy moiety, or other naturally
occurring functionality in fatty acids can provide important
building blocks for polymer synthesis. Also, conversions of
unsaturated and epoxy functionalized R,ω-diacids to corre-
sponding polyesters would be best performed by lipase catalysis
for reasons discussed below.
monomers 1,18-cis-9-octadecenedioic acid (ω-carboxyl OA),
1,22- cis-9-docosenedioic acid (ω-carboxyl EA), and 1,18-cis-
9,10-epoxy-octadecanedioic acid (ω-carboxyl epoxy SA). Ef-
fects of fatty acid chain length and presence of unsaturation
versus epoxy groups at C18 carbons on the C. tropicalis
ATCC20962-catalyzed conversion of these fatty acid substrates
to diacids was studied. Diacid monomers were purified and
characterized by gas chromatography (GC)-mass spectrometry
(MS) and nuclear magnetic resonance (1H NMR and 13C NMR).
Conversion of unsaturated and epoxidized ω-carboxyl fatty acids
to polyesters was performed by condensation copolymerizations
with a diol using an immobilized CALB (N435) as catalyst.
Comparison of CALB-catalyzed copolymerizations in-bulk, and
in diphenyl ether, was performed and chain growth was
monitored as a function of time. Effects of changes in ω-car-
boxyl fatty acid building blocks on thermal properties of
resulting polyesters were also investigated.
Normally, polyesters are synthesized by ester interchange
reactions or by direct esterification of diacid/diol or hydroxy-
acids.15 These reactions often require harsh reaction conditions
and metal catalysts, which are not compatible with retention of
functional group structure during polymer synthesis. An alterna-
tive approach to traditional chemical polymerization methods
is the use of cell-free, enzyme-catalyzed polycondensation
reactions.16 For example, lipase-catalyzed polymerizations of
monomers containing a double bond or an epoxy group to
synthesize polyesters have been reported by several authors.17
In one example described by Olsen and Sheares,17b trans-ꢀ-
hydromuconic acid (HMA) was copolymerized (90 °C, in-bulk,
48 h) with 1,8-octanediol using immobilized Candida antarctica
Lipase B (CALB) to give polyesters with Mn 10500 g/mol (Mw/
Mn ) 2.0) whose double bonds remained intact.17b
A natural bifunctional fatty acid of interest as a monomer
for polymer synthesis is ricinoleic acid [(Z,R)-12-hydroxy-9-
octadecenoic acid], derived from castor oil. However, esterifi-
cation of secondary hydroxyl groups by lipase-catalysis is
generally slow.18,19 Furthermore, to use ricinoleic acid as a
bifunctional monomer to prepare high molecular weight prod-
ucts, ricinoleic acid must be rigorously purified to remove
impurities that terminate chain growth.
Experimental Section
Materials. Oleic acid (technically, ∼90% pure), erucic acid (∼99%
pure), 1,3-propanediol, 1,8-octanediol, 1,16-hexadecanediol, antifoam
204, and Glucose (HK) Assay Kit were purchased from Sigma-Aldrich
(St. Louis, U.S.A.). The diols were obtained in the highest available
purity and were used as received. 1,18-Octadecanedioic acid (∼95%
pure) and thin layer chromatography (TLC) aluminum sheets were
purchased from TCI (Portland, U.S.A.) and Merck (Darmstadt,
Germany), respectively. Novozym 435 (abbreviated as N435, specified
activity 10000 PLU/g) was a gift from Novozymes (Bagsvaerd,
Denmark) and consists of CALB physically adsorbed within the
macroporous resin Lewatit VPOC 1600 (poly[methyl methacrylateco-
butyl methacrylate], supplied by Bayer). All other chemicals not listed
above were of analytical grade from Sigma-Aldrich (St. Louis, U.S.A.).
Candida tropicalis ATCC20962 was purchased from The American
Culture Collection (ATCC).
Others have pursued naturally derived functional fatty acid
monomers that, unlike ricinoleic acid, can more efficiently be
converted by enzyme-catalysis to polyesters. Veld et al.20
investigated aleuritic acid that has two secondary and one
primary (ω-position) hydroxyl groups and is derived from
ambrettolide that naturally occurs in musk abrette seed oil and
is a valuable perfume base due to its desirable odor. Aleuritic
acid was first converted to its isopropyl ester and then polym-
erized (90 °C, 550 m bar, 21 h) in a mixture of dry toluene and
dry 2,4-dimethyl-3-pentanol. Poly(aleuriteate) (Mn ) 5600
g/mol, PDI ) 3.2) was isolated in moderate yield (43%) after
precipitation. Interestingly, the polymerization was highly
selective for monomer primary hydroxyl groups with no
observable secondary hydroxyl esterification based on NMR
studies. Olsson et al.21 isolated cis-9,10-epoxy-18-hydroxyoc-
tadecanoic acid from suberin, found in the outer bark of birch.
This naturally derived monomer was converted to an epoxy-
fuctionalized polyester by N435 catalysis. Problems associated
with deriving functional fatty acid monomers from specialty
seed oils such as abrette is their low abundance. Furthermore,
fatty acids from major seed oil sources lack of ω-hydroxyl or
ω-carboxyl groups. Also, it is difficult to purify components of
suberin and cutin that consist of severe mixtures of these useful
building blocks. For these reasons, C. tropocalis ATCC20962
appears as a promising biocatalyst to provide a wide range of
bifunctional fatty acid building blocks. In this paper, we report
a new two-pot two-step route, using whole-cell bio-oxidation
followed by cell-free lipase catalysis, to convert readily renew-
able fatty acids to functional polyesters. First, fermentations of
C. tropicalis ATCC20962 were used to convert oleic, erucic
and epoxystearic acids to their corresponding diacid functional
Microorganism. Candida tropicalis ATCC20962, used in this study,
was obtained with its ꢀ-oxidation pathway blocked. This was ac-
complished by Picataggio et. al4 who disrupted the POX 4 and POX 5
genes encoding its acyl-coenzyme A oxidase.
Lipase-Mediated Epoxidation of Oleic Acid. The method used was
based on a previous literature report.22 The reaction was performed in
a 50-mL round-bottom flask containing 0.5 M oleic acid in 20 mL
toluene and 300 mg immobilized C. antarctica lipase (N435). Hydrogen
peroxide (30%, w/w) was added stepwise at 0.5 mL every hour during
the first 4 h. The reaction mixture was stirred at 600 rpm and reaction
temperature was maintained at 50 °C. After 8 h, TLC showed only a
single band indicating complete conversion of oleic acid to 9,10-epoxy
stearic acid. The reaction was terminated by removal of N435 by
filtration and solvent removal was performed under vacuum with a
rotary evaporator. The obtained product was directly used for the
biotransformation without further purification.
1
9,10-epoxy stearic acid: H NMR (8 mg in 0.7 mL CDCl3; ppm)
2.92 (2H, bs, -CH-, cis-epoxide), 2.35 (2H, t, -CH2-CO-), 1.2-1.8
(26H, bm, -CH2-), 0.89 (3H, t, -CH3).
Fermentation Procedure. C. tropicalis ATCC20962 was precultured
in 30 mL YPD medium consisting of (g/L) yeast extract, 10; peptone,
20; glucose, 20; and shaken at 250 rpm, 30 °C for 20 h in a 500 mL
flask. The cells were inoculated at 10% (v/v) to 30 mL conversion
medium consisting of (g/L) peptone, 3; yeast extract, 6; yeast nitrogen
base, 6.7; acetic acid, 3; K2HPO4, 7.2; KH2PO4 9.3; glucose, 20 in 500
mL flask. After 12 h of cultivation at 250 rpm and 30 °C, the
biotransformation phase was begun by addition of oleic acid, erucic
acid or 9,10-epoxystearic acid. The initial concentration of fatty acids
was 20 g/L and pH was adjusted to 7.5 by addition of 2 mol/L NaOH
solution. During the biotransformation, glucose (500 g/L) was fed (2.5%
per day) as cosubstrates and the pH was maintained at 7.5-8.0 by
addition of NaOH solution.

Functional Polyesters from ω-Carboxy Fatty Acids
Biomacromolecules, Vol. 11, No. 1, 2010 261
CH3], 274 (9), 259 (100) [M+
The biotransformation of oleic acid was also conducted in a 3-L
Bioflo3000 fermentor (New Brunswick Scientific Co., U.S.A.) in fed-
batch culture. The conversion medium above was used except for
addition of 0.05% antifoam 204 (Sigma) and 0.5% oleic acid. The
culture was grown at 30 °C at 900 rpm, with aeration rate of 1.5 vvm.
After 12 h fermentations (growth phase), the biotransformation phase
was started by feeding oleic acid at 2 mL/L. Glucose (500 g/L) as
cosubstrate was fed continuously at 1.2 g/L/h. During the biotransfor-
mation phase, the pH was maintained at 7.6 automatically by addition
of 4 mol/L NaOH solution and dissolved oxygen was controlled at
about 60%. To avoid excessive foaming, antifoam 204 was added as
needed. Every 12 h, a 5 mL aliquot from fermentation media was
withdrawn to determine cell growth, glucose concentration, depletion
of the fatty acid, and diacid production.
Extraction and Purification of Fermentation Products. Fermenta-
tion broths of oleic acid and erucic acid were acidified to pH 1.0 with
HCl and extracted twice with diethyl ether. To avoid epoxy ring-opening
during acidification, the fermentation broth of 9,10-epoxy stearic acid
was slowly acidified to pH 3.0 with 5 N HCl and then extracted with
diethyl ether. Diethyl ether was removed by rotoevaporation that gave
a light yellow solid for all products from oleic acid, erucic acid, and
9,10-epoxy stearic acid. The crude product (4 g), dissolved in diethyl
ether, was separated into components with a Biotage SP1 Flash
Chromatography system (Biotage Inc., A Dynax Corp. Company,
U.S.A.). Separations were performed using a Biotage SI 40+M column
(150 × 40 mm ID; 100 g KP-Sil silica; 40-63 µm particle size; flow
rate: 25-50 mL/min). The mobile phase consisted of n-hexane/diethyl
ether mixtures that were run in gradient mode with compositions ranging
from 30 to 80% diethyl ether in 10 column volumes (CVs) at 30 mL/
min flow rate. Fractions were monitored at 220 and 280 nm by UV
detection. Unreacted fatty acid substrate eluted first and then R,ω-diacid
eluted in the second peak. Fractions in the second peak showing a single
band on TLC plate were pooled and concentrated by rotoevaporation
that gave a white powder. The purity of ω-carboxyl OA, ω-carboxyl
EA, and ω-carboxyl epoxy SA used in polymerizations was >99%
(GC-MS).
[M+
-
CH3OH
-
-
(CH3)2COO(CH2)7CHO], 201 (5) [(CH3)2COO(CH2)7CHO+], 169 (4)
[CH3COO(CH2)7C+], 155 (44) [CHO(CH2)CO+], 137 (12), 109 (18),
73 (53) [TMSi+], 55 (11) [C4H7+], 31 (9) [CH3O+].
Novozym 435-Catalyzed Condensation Polymerizations
between ω-Carboxy Fatty Acids and Diols. Reactions were performed
in a parallel synthesizer (AdvantageTM 2050, Argonaut) both in-bulk
and in diphenyl ether at 90 °C. Unsaturated and epoxidized R,ω-diacids
(1.0 mmol) and diol (1.0 mmol) were transferred into reactor tubes of
the parallel synthesizer and, then, 10%-by-wt N435 (i.e., 1%-by-wt
CALB) relative to monomer was added. For solution polymerizations,
1 mL diphenyl ether was added. Control reactions without addition of
N435 were also performed in parallel. Vacuum (10 mmHg) was applied
after 2 h. To follow the progress of polymerizations, aliquots were
withdrawn at 2, 6, 12, 24, 36, and 48 h. Reactions were terminated by
addition of cooled chloroform and N435 was removed by filtration.
Filtrates were concentrated at 50 °C under vacuum to remove
chloroform and then directly (i.e., without fractionation by precipitation)
1
analyzed by gel permeation chromatography (GPC) and H and 13C
NMR. Alternatively, the concentrated filtrate solutions were added into
magnetically stirred cold methanol. Precipitated polymers were filtered,
were washed three times with cold methanol, and were dried in vacuo
overnight at 50 °C.
Polyester from ω-Carboxyl OA and 1,3-Propanediol
1
(Poly(ω-carboxyl OA-co-PD)). Yield: 80%. H NMR (10.7 mg in 0.7
mL CDCl3; ppm): 5.2-5.35 (2H, m, dCH-CH2-), 4.05 (4H, m, -(CO)-
O-CH2-), 2.23 (4H, t, -CH2-CO-), 1.8-2.0 (6H, m, dCH-CH2-, -(CO)-
O-CH2-CH2-), 1.45-1.6 (4H, m, -CH2-CH2-CO-), 1.1-1.35 (16H, m,
-CH2-). 13C NMR (76.8 mg in 0.7 mL CDCl3; ppm): 173.7 (CdO),
129.8 (dCH-), 60.8 (-CH2O-), 34.2, 29.7, 29.2, 29.1, 28.1, 27.2, 24.9.
Polyester
from
ω-Carboxyl
OA
and
1,8-Octanediol
1
(Poly(ω-carboxyl OA-co-OD)). Yield: 79%. H NMR (10 mg in 0.7
mL CDCl3; ppm): 5.3-5.4 (2H, m, dCH-CH2-), 4.05 (4H, m, -(CO)-
O-CH2-), 2.35 (4H, t, -CH2-CO-), 1.9-2.1 (4H, m, dCH-CH2-),
1.5-1.7 (8H, m, -CH2-CH2-(CO)-O-CH2-CH2-), 1.2-1.4 (24H, m,
-CH2-). 13C NMR (80 mg in 0.7 mL CDCl3; ppm): 173.9 (CdO), 129.8
(dCH-), 64.3 (-CH2O-), 34.3, 29.6, 29.2, 29.1, 28.6, 27.1, 25.8, 24.9.
Polyester from ω-Carboxyl OA and 1,16-Hexadecanediol
ω-Carboxyl OA. Yield: 95% (from crude mixture of diethyl ether
extracted products); mp 66-68 °C. 1H NMR (9.8 mg in 0.7 mL CDCl3;
ppm): 5.3-5.4 (2H, m, dCH-CH2-), 2.35 (4H, t, -CH2-CO-), 1.9-2.1
(4H, m, dCH-CH2-), 1.6 (4H, m, -CH2-CH2-CO-), 1.2-1.4 (16H, m,
-CH2-). 13C NMR (63.5 mg in 0.7 mL CDCl3; ppm): 180.7 (CdO),
129.8 (dCH-), 34.1, 29.6, 29.1, 29.0, 28.9, 27.1, 24.6. IR (KBr; cm-1):
1691 (s, CdO), 725 (m, CdC, cis-form). MS for dimethyl ester of
1
(Poly(ω-carboxyl OA-co-HD)). Yield: 89%. H NMR (8.3 mg in 0.7
mL CDCl3) (ppm): 5.3-5.4 (2H, m, dCH-CH2-), 4.05 (4H, m, -(CO)-
O-CH2-), 2.30 (4H, t, -CH2-CO-), 1.9-2.1 (4H, m, dCH-CH2-),
1.5-1.7 (8H, m, -CH2-CH2-(CO)-O-CH2-CH2-), 1.1-1.4 (40H, m,
-CH2-). 13C NMR (81.9 mg in 0.7 mL CDCl3; ppm): 173.7 (CdO),
129.7 (dCH-), 64.3 (-CH2O-), 34.3, 29.6, 29.5, 29.4, 29.2, 29.1, 29.0,
28.9, 28.6, 27.1, 25.8, 24.9.
ω-carboxyl OA, m/z (rel int, %): 340 (5) [M+], 308 (49) [M+
-
CH3OH], 276 (90) [M+ - 2CH3OH], 248 (18) [M+ - 2CH3OH -
CO], 234 (7) [M+ - 2CH3OH - CO - CH2], 165 (19) [M+ - CH3OH
- CH3OOC(CH2)6], 151 (27), 109 (37), 95 (31), 81 (100), 69 (48)
[C5H9+], 55 (81) [C4H7+], 41 (63) [C3H5+].
Polyester
from
ω-Carboxyl
EA
and
1,8-Octanediol
1
(Poly(ω-carboxyl EA-co-OD)). Yield: 79%. (a) H NMR (7.5 mg in
0.7 mL CDCl3; ppm): 5.3-5.4 (2H, m, dCH-CH2-), 4.05 (4H, m,
-(CO)-O-CH2-), 2.33 (4H, t, -CH2-CO-), 2.0 (4H, m, dCH-CH2-),
1.5-1.7 (8H, m, -CH2-CH2-(CO)-O-CH2-CH2-), 1.2-1.4 (32H, m,
-CH2-). 13C NMR (CDCl3; 75 mg in 0.7 mL ppm): 173.9, 173.8 (CdO),
129.9, 129.7 (dCH-), 64.2 (-CH2O-), 34.3, 29.7, 29.6, 29.54, 29.51,
29.4, 29.3, 29.2, 29.1, 29.06, 28.6, 27.2, 27.1, 25.8, 24.9.
ω-Carboxyl EA. Yield: 70% (from crude mixture of diethyl ether
extracted products); mp 71-73 °C. 1H NMR (10.3 mg in 0.7 mL
CDCl3; ppm): 5.3-5.4 (2H, m, dCH-CH2-), 2.33 (4H, t, -CH2-CO-),
1.9-2.1 (4H, m, dCH-CH2-), 1.6 (4H, m, -CH2-CH2-CO-), 1.2-1.4
(24H, m, -CH2-). 13C NMR (63.5 mg in 0.7 mL CDCl3; ppm): 180.4,
180.5 (CdO), 130.0, 129.7 (dCH-), 34.1, 29.7, 29.6, 29.57, 29.53,
29.5, 29.4, 29.3, 29.2, 29.1, 29.0, 27.2, 27.1, 24.6. IR (KBr; cm-1):
1691 (s, CdO), 725 (m, CdC, cis-form). MS for dimethyl ester of
Polyester from ω-Carboxyl Epoxy SA and 1,8-Octanediol
1
(Poly(ω-carboxyl epoxy SA-co-OD)). Yield: 60%. H NMR (10.0 mg
in 0.7 mL CDCl3; ppm): 4.05 (4H, m, -(CO)-O-CH2-), 2.9 (2H, bs,
-CH-, cis-epoxide), 2.29 (4H, t, -CH2-CO-), 1.2-1.7 (36H, m, -CH2-).
13C NMR (82.7 mg in 0.7 mL CDCl3; ppm): 173.8 (CdO), 64.2
(-CH2O-), 57.1 (-CH-, cis-epoxide), 34.2, 29.3, 29.1, 29.0, 28.9, 27.7,
26.5, 25.7, 24.8.
ω-carboxyl EA, m/z (rel int, %): 396 (0.7) [M+], 364 (8) [M+
-
CH3OH], 332 (13) [M+ - 2CH3OH], 304 (4) [M+ - 2CH3OH - CO],
290 (2) [M+ - 2CH3OH - CO - CH2], 165 (8), 151 (13), 109 (32),
95 (56), 81 (68), 69 (69) [C5H9+], 55 (100) [C4H7+], 41 (39) [C3H5+].
ω-Carboxyl Epoxy SA. Yield: 92% (from crude mixture of diethyl
ether extracted products); mp 85-86 °C. 1H NMR (10.0 mg in 0.7 mL
CDCl3) (ppm): 2.92 (2H, bs, -CH-, cis-epoxide), 2.35 (4H, t, -CH2-
CO-), 1.2-1.8 (24H, bm, -CH2-). 13C NMR (50.6 mg in 0.7 mL CDCl3;
ppm): 180.1 (CdO), 57.3 (-CH-, cis-epoxide), 34.0, 29.2, 29.1, 28.9,
27.6, 26.4, 24.5. MS for TMSi-ether dimethyl ester derivative of
ω-carboxyl epoxy SA, m/z (rel int, %): 429 (2) [M+ - CH3O], 413 (2)
Unlike the above copolymerizations, reaction of 1,18-octadecanedioic
acid (ω-carboxyl SA) with 1,8-octanediol was performed in a 250 mL
flask with addition of toluene using an azeotrope at 90 °C. 1,18-
Octadecanedioic acid (20 mmol) and 1,8-octanediol (20 mmol) were
transferred into a flask with 100 mL toluene and 10%-by-wt N435.
Vacuum (200-250 mmHg) was applied after 2 h. To follow the
progress of polymerizations, aliquots were withdrawn at 2, 6, 12, 24,
36, and 48 h. As above, an aliquot that was not precipitated was used

262 Biomacromolecules, Vol. 11, No. 1, 2010
Yang et al.
to determine molecular weight averages and polydispersity of products
by GPC. The final time point of this reaction was worked-up as above
to terminate the reaction and remove N435. Subsequently, the filtrate
was slowly added with stirring to methanol to precipitate the polyester.
The precipitated polymer was washed with methanol three times and
then dried under vacuum (50 °C, 10 mmHg). The resulting precipitated
product was then analyzed by 1H and 13C NMR and also used for
thermal property analysis.
Sample concentrations and injection volumes were 0.2% (w/v) and 30
µL, respectively. Tetrahydrofuran was used as eluent at a flow rate of
1 mL/min. Trisec GPC software version 3 was used for calculations.
Molecular weights were determined on the basis of conventional
calibration curve generated by narrow molecular weight polystyrene
standards obtained from Sigma.
ThermograVimetric Analysis (TGA). TGA measurements were
performed with a TA Instruments TGA2950 analyzer under nitrogen
atmosphere with about 10 mg samples at a heating rate of 10 °C/min
from 25 to 700 °C.
Differential Scanning Calorimetry (DSC). DSC scans were recorded
using a TA Instruments DSC-2920 analyzer. Polymer samples were
heated and cooled under a nitrogen flow rate of 50 mL/min and a
heating rate of 10 °C/min. Samples were first heated to 120 °C and
then quenched to -70 °C with liquid nitrogen thereby erasing their
thermal history. Data reported herein on thermal transitions were
obtained from second heating scans, recorded immediately after samples
were quenched to -70 °C by reheating from -70 to 120 °C at a heating
rate of 10 °C/min.
Polyester
from
ω-Carboxyl
SA
and
1,8-Octanediol
1
(Poly(ω-carboxyl SA-co-OD)). Yield: 86%. H NMR (7.9 mg in 0.7
mL CDCl3; ppm): 4.05 (4H, m, -(CO)-O-CH2-), 2.33 (4H, t, -CH2-
CO-), 2.0 (4H, m, dCH-CH2-), 1.5-1.7 (8H, m, -CH2-CH2-(CO)-O-
CH2-CH2-), 1.2-1.4 (32H, m, -CH2-). 13C NMR (62.5 mg in 0.7 mL
CDCl3; ppm): 173.9 (CdO), 64.2 (-CH2O-), 34.3, 29.62, 29.57, 29.5,
29.4, 29.2, 29.1, 29.0, 28.6, 25.8, 25.0.
Analytical Methods. Cell growth was determined by measuring
absorbance at 600 nm. A standard curve of absorbance versus cell dry
weight was constructed and subsequently used to convert cell growth
measurements from absorbance to cell dry weight values. Glucose was
measured by the enzymatic method using Glucose (HK) Assay Kit
(Sigma).
Thin layer chromatography analysis was done on aluminum sheets
coated with silica gel 60. Samples were spotted on the sheet using a 5
µL micro glass pipet and developed with a mobile phase system
comprising toluene and acetic acid (9:1, v/v). After developing, sheets
were sprayed with primuline solution (50 mg primuline dissolved in
800 mL acetone and 200 mL distilled water). Fatty acid derivatives
were detected after primuline spray under UV light.
Results and Discussion
The following describes a unique two-pot two-step route by
which whole-cell bio-oxidation followed by cell-free lipase
catalysis is used to convert a series of unsaturated and epoxy
fatty acid substrates to functional polyesters (Scheme 1). The
first step investigates a fermentative reaction, catalyzed by C.
tropicalis ATCC20962, to convert oleic, erucic and epoxystearic
acids to their corresponding diacid functional monomers 1,18-
cis-9-octadecenedioic acid (ω-carboxyl OA), 1,22- cis-9-docos-
enedioic acid (ω-carboxyl EA), and 1,18-cis-9,10-epoxy-
octadecanedioic acid (ω-carboxyl epoxy SA). A key question
addressed here is to what extent changes in chain length and
functionality (double bond vs epoxy groups) of fatty acids affects
their conversion to ω-carboxy diacids by C. tropicalis
ATCC20962.
Instrumental
Methods.
Chemical
Characterization
of
Fermentation Products. Gas chromatography/mass spectrometry (GC/
MS) analysis was performed at 70 eV using a ThermoFinnigan TraceGC
Ultra gas chromatograph coupled with a Trace DSQ mass spectrometer.
Purified products were esterified with BF3 in methanol (10%, w/w) at
70 °C for 20 min.23 Silylation of ω-carboxyl epoxy SA methyl ester
was performed with HMDS/TMCS/pyridine at 70 °C for 10 min.24
GC/MS experiments were performed with injector, ion source and
interface temperatures of 200, 250, and 280 °C, respectively. Samples
in hexane (1 µL) were injected in PTV split mode and run on a capillary
column (Varian CP8944 VF-5MS, 0.25 mm × 0.25 µm × 30 m). The
oven temperature was programmed at 120 °C for 1 min increasing to
260 at 20 °C/min, and then to 280 at 4.0 °C/min.
Oxidative cleavage was used to determine the position of double
bonds. Purified products were treated by acidic KMnO4 solution (2%)
at 60 °C for 20 min. After methylation of oxidized products (see above),
samples were analyzed by GC/MS.
Proton (1H) and carbon (13C) NMR spectra were recorded on a
Bruker DPX300 NMR spectrometer. The chemical shifts (ppm) reported
were referenced relative to internal tetramethylsilane (TMS, 0.00 ppm)
or to the solvent resonance at the appropriate frequency.
Biotransformation of Oleic Acid, Erucic Acid, and
9,10-Epoxy Stearic Acid by C. tropicalis ATCC20962. Syn-
thesis of 9,10-epoxy stearic acid was performed by a chemo-
enzymatic method by which the peracid chemical catalysis of
epoxidation was formed by a lipase-catalyzed esterification of
oleic acid with hydrogen peroxide.25 Epoxy stearic acid was
obtained in quantitative yield and was directly used for
biotransformation after removing toluene under vacuum with a
rotary evaporator. The successful conversion of the double bond
1
to an epoxy group is evident from H NMR spectra that show
the complete disappearance of signals corresponding to oleic
Infrared spectroscopy (Avatar 360 FT-IR spectrometer) was used
to characterize products and provide information on double bond
configuration. Melting point measurements were performed using an
electrothermal melting point apparatus.
acid olefin protons a of double bond at 5.3-5.4 ppm and the
1
appearance of H NMR signals for the product at 2.90 ppm
due to methine protons A of the epoxy functionality (Figure 1).
The resulting epoxy SA was then used as a substrate for
fermentative bio-oxidation reactions.
The concentration of diacids during biotransformation was measured
by liquid chromatography/mass spectrometry (LC/MS) with purified
products as standards. The solvent delivery system was a Waters
Alliance 2795 Separation Module (Milford, MA, U.S.A.) coupled with
a Waters 2996 photodiode array detector and Waters ZQ detector with
an electron spray ionization mode. The separation was carried on a
reversed-phase column with dimensions 150 × 4.6 mm and particle
size of 5 µm. The mobile phase used for separation contained 10%
H2O, 5% acetonitrile, 5% formic acid solution (1% in water), and 80%
methanol.
Gel Permeation Chromatography (GPC). Molecular weights of
polyesters were determined by GPC using a Waters HPLC system
equipped with model 510 pump, Waters model 717 autosampler, model
410 refractive index detector, Viscotek model T-50/T-60 viscosity
detector, and 500, 103, 104, and 105 Å Ultrastyragel columns in series.
Biotransformations of oleic acid, erucic acid, and 9,10-epoxy
stearic acid, catalyzed by C. tropicalis ATCC20962, were
conducted in 500 mL shake-flasks (30 mL conversion media)
with initial fatty acid concentrations of 20 g/L. Figure 2A shows
the consumption of fatty acid carbon sources and formation of
unsaturated and epoxidized dicarboxylic acids as a function of
biotransformation time determined by LC-MS. The highest
conversion rate was obtained using oleic acid as substrate,
reaching 14.1 and 17.3 g/L (77.6 mol %) ω-carboxyl OA in 24
and 48 h, respectively. In contrast, after 72 h, the maximum
concentration of ω-carboxyl EA and ω-carboxyl epoxy SA
reached to 14.2 g/L (64.8 mol %) and 19.1 g/L (86.2 mol %),
respectively. Slower conversions of these two substrates may

Functional Polyesters from ω-Carboxy Fatty Acids
Biomacromolecules, Vol. 11, No. 1, 2010 263
Scheme 1. Lipase-Catalyzed Polycondensation of Unsaturated (a) and Epoxidized (b) Dicarboxylic Acids with Diolsa
a Dicarboxylic acids were obtained from bioconversions of oleic acid, erucic acid, and 9,10-epoxy stearic acid catalyzed by C. tropicalis ATCC20962.
boxyl EA continues to increase, indicating that extending the
fermentation time would result in still higher ω-carboxyl EA
yield.
Scale-up biotransformation of oleic acid, catalyzed by C.
tropicalis ATCC20962, was performed by fed-batch culture in
a 3 L fermentor. Previous studies by Gmunder et al.26 showed
that high glucose concentrations in fermentation media for C.
tropicalis inhibits cytochrome P450/reductase expression and
activity. Hence, after the 12 h growth phase, high glucose
concentrations in fermentation media during the biotransfor-
mation phase was avoided by slowly feeding glucose at 1.2 g/L/
h. Concurrently, oleic acid was fed at 2.0 mL/h. Resulting cell
growth, glucose consumption and production of ω-carboxyl OA
are shown in Figure 2B. Maximum ω-carboxyl OA volumetric
concentration of 31 g/L was obtained after 60 h. By conducting
ω-carboxyl OA synthesis in a fermentor instead of in-shake flask
cultures, the productivity increased from 0.36 to 0.5 g/L/h. This
is due to improved control of fermentation parameters such as
Figure 1. 1H NMR (300 MHz) spectra recorded in chloroform-d of
oleic acid (A), 9,10-epoxy stearic acid (B), and 1,18-cis-9,10-epoxy-
octadecanedioic acid (C).
pH, DO, and low level of glucose concentration during the
conversion phase that is maintained by continuous feeding in
the fermentor.
In previous studies, oleic acid bioconversion by C. tropicalis
mutants resulted in poor conversion efficiency and low yield
of ω-carboxyl OA. Yi et al.11 reported that, conversion of oleic
in part be due to their lower solubility in fermentation media.
Nevertheless, at 72 h, the conversion of erucic acid to ω-car-

264 Biomacromolecules, Vol. 11, No. 1, 2010
Yang et al.
Figure 2. Fermentation profiles for C. tropicalis ATCC20962-catalyzed conversions of oleic acid, erucic acid, and 9,10-epoxy stearic acid to
R,ω-dicarboxylic acids in shake-flask experiments (A) and (B) C. tropicalis ATCC20962-catalyzed conversion of oleic acid to ω-carboxyl oleic
acid in a 3-L fermentor.
acid using C. tropicalis mutant S76 gave ω-carboxyl OA in up
to ∼4 g/L after a 72 h fermentation in shake flask. Furthermore,
increased fermentation time resulted in decreased product yield.
This is consistent with the fact that the shorter chain-length
unsaturated diacids with 6-16 carbon atoms were also detected
in media due to degradation of ω-carboxyl OA through the
ꢀ-oxidation pathway.27 Fabritius et al.12 reported that, for
conversion of oleic acid using mutants of C. tropicalis DSM3152
and M25 in fed-batch culture, the maximum concentration of
ω-carboxyl OA reached values of 21.5 g/L at 86.5 h and 8.1
g/L at 192 h, respectively. Similarly, the decrease in ω-carboxyl
OA production was observed presumably due to metabolism
of product via the ꢀ-oxidation pathway. Indeed, 3-hydroxy-1,18-
cis-9-octadecenedioic acid was found in large amounts along
with ω-carboxyl OA in media. Picataggio et al.4 compared oleic
acid and erucic acid bioconversion using the P450 monooxy-
genase (P450alkl) and NADPH-oxidoreductase (CPR) amplified
strain AR40 derived from C. tropicalis ATCC20962. They found
that both ω-carboxyl OA and ω-carboxyl EA was obtained in
high yield and productivity. However, the effect of chain length
and functionality (double bond versus epoxy groups) of fatty
acids on their conversion to ω-carboxyl fatty acids by C.
tropicalis ATCC20962 has not been reported. Furthermore, to
our knowledge, this is the first time that ω-carboxyl epoxy SA
has been efficiently synthesized from its corresponding fatty
acid by a whole-cell biotransformation.
dimethyl ester of ω-carboxyl EA, its mass fragment spectrum
shown in Figure 3B is also in accord with that expected. The
molecular ion [M+] at m/z 396 and typical fragment ions at m/z
364 [M+ - CH3OH] and 332 [M+ - 2CH3OH], due to scission
of one and two methanol groups, respectively, were observed
+
as well as m/z 304 [M+ - 2CH3OH - CO] and m/z 69 [C5H9 ].
The mass fragment spectrum of methylated and silyated
ω-carboxyl epoxy SA is displayed in Figure 3C. As above, the
observed spectrum is consistent with that expected for the
corresponding product structure. Fragment ions at m/z 429 [M+
-
CH3O], 259 [M+
- (CH3)2COO(CH2)7CHO], 201
[(CH3)2COO(CH2)7CHO+], 155 [259 - TMS - CH3O]+, 169
[201 - CH3OH]+, 137 [169 - CH3OH]+, 109 [137 - CO]+,
73 [TMSi+], and 31 [CH3O+] were all detected.
To determine whether the double bond position remained
intact or was isomerized to another position along the fatty acid,
purified ω-carboxyl OA and ω-carboxyl EA were treated by
acidic KMnO4 solution. After methylation of oxidative products,
the samples were analyzed by GC-MS. The mass spectrum with
retention time 6.48 min in GC was identical to that of authentic
dimethyl ester of azelaic acid (see Supporting Information,
Figure S-1). Thus, as in the OA and EA used as feedstocks in
fermentations, the position of double bonds in both compounds
is between C-9 and C-10.
Structures of purified products were further confirmed by 1H
NMR, 13C NMR and IR (see deposited spectra in Supporting
Information, Figures S-2, S-3, and S-4, respectively). Indeed,
spectra obtained are consistent with that of proposed product
structures. 1H NMR spectra of ω-carboxyl OA and ω-carboxyl
EA have signals at 2.35 ppm with two overlapping triplets due
to protons for two methylene groups adjacent to carboxylic acid
moieties. Furthermore, signals at 5.35 ppm, due to olefin protons
(-CHdCH-), are observed and integration ratios are as
expected. Also, the characteristic signal corresponding to the
terminal methyl group of fatty acid starting materials OA and
Structural analysis of ω-carboxyl fatty acids was performed
by multiple methods.
Molecular weights of resulting ω-carboxyl OA, ω-carboxyl
EA and ω-carboxyl epoxy SA, determined by LC-MS, are 312,
368, and 328, respectively, in agreement with those calculated
for each product. Methyl ester and silylated (only for ω-carboxyl
epoxy SA) derivatives of purified ω-carboxyl products were
analyzed by GC-MS. Indeed, for derivatized ω-carboxyl OA,
ω-carboxyl EA and ω-carboxyl epoxy SA, single peaks with
retention times of 10.54, 13.97, and 13.17 min, respectively,
were detected. The mass fragment spectrum for each compound
is listed in the Experimental Section and displayed in Figure 3.
The mass fragment spectrum for ω-carboxyl OA dimethyl ester
(Figure 3A) shows a molecular ion [M+] at m/z 340 and
1
EA at 0.88 ppm is not observed. As above, comparison of H
NMR spectra of 9,10-epoxy SA and ω-carboxyl epoxy SA are
consistent with formation of the ω-oxidized product (Figure 1).
Unique to these spectra is the signal at 2.90 ppm corresponding
to methine protons A and the 13C signal at 57.3 ppm corre-
sponding to carbons of cis-epoxy groups.23 Observation of epoxy
proton signals in the product shows this group remains intact
during the biotransformation. IR spectra of ω-carboxyl OA and
ω-carboxyl EA show the characteristic vibrational absorption
band at 725 cm-1 located in the fingerprint region corresponding
to the presence of cis-configuration double bonds.11
fragment ions at m/z 308 [M+ - CH3OH] and 276 [M+
-
2CH3OH] due to scission of one and two methanol groups,
respectively. Fragment ions at m/z 248 [M+ - 2CH3OH - CO],
m/z 234 [248 - CH2]+, and m/z 69 [C5H9+] were also detected.
The mass fragment spectrum was in agreement with that
reported previously by Yi et al.,11 obtained by GC-MS analysis
of ∆9-cis-1,18-octadecenedioic acid dimethyl ester. For the

Functional Polyesters from ω-Carboxy Fatty Acids
Biomacromolecules, Vol. 11, No. 1, 2010 265
Figure 3. Mass spectral analysis of oleic acid, erucic acid, and 9,10-epoxy stearic acid metabolites produced by C. tropicalis ATCC20962. The
mass spectra were obtained from GC peaks appearing at the retention times of A) 10.54 min (1,18- cis-9-octadecenedioic acid dimethyl ester
derivative; B), 13.97 min (1,22- cis-9-docosenedioic acid dimethyl ester derivative; C), and 13.17 min (1,18-cis-9, 10-epoxy-octadecanedioic
acid, TMS ether dimethyl ester derivative.
Novozym
435-Catalyzed
Copolymerization
of
values of poly(ω-carboxyl OA-co-OD) reached 44000 g/mol
and 2.1, respectively. A decrease in Mw/Mn was observed from
about 2.5 to 2.1 at 12-36 and 48 h, respectively. This decrease
in polydispersity is explained by that, when reaction times are
prolonged and high viscosity reaction media are formed, more
rapid diffusion of lower molecular weight chains to propagating
chain ends depletes the population of low-molecular weight
species.28 When this reaction was performed in diphenyl ether
at 90 °C, poly(ω-carboxyl OA-co-OD) reached an Mw of 57000
g/mol in 36 h with Mw/Mn of 2.0. The attainment of higher
molecular weights in solution is explained by higher diffusivity
of reactants as well as easier removal of water from reaction
media. However, given the (i) difficulty of removing diphenyl
ether from reaction products (discussed further below), (ii)
overall benefits in process efficiency by running bulk polymer-
izations, and (iii) ability to form high molecular weight
polyesters without solvent, the focus of polymerization studies
below is on use of bulk polymerization conditions.
Unsaturated and Epoxidized r,ω-Dicarboxylic Acids with
Diols. After successful preparation by whole-cell bio-oxidations
of ω-carboxyl OA, ω-carboxyl EA, and ω-carboxyl epoxy SA,
the second step of this newly developed two-pot, two-step route
to unsaturated and epoxidized fatty acid-based polyesters was
undertaken. That is, cell-free immobilized lipase catalysis was
used to copolymerize R,ω-diacids with a series of diols. Of
interest was to what extent do changes in structures of fatty
acid derived R,ω-diacids and R,ω-alkylene diols affect lipase-
catalyzed conversions of monomers to polyesters.
N435, consisting of CALB immobilized on a macroporous
polyacrylate resin, was used as the lipase catalyst for polycon-
densations of ω-carboxy fatty acids and diols. Previous work
by Mahapatro et al.16g demonstrated that, N435-catalyzed
copolycondensations in-bulk, between adipic acid and 1,8-
octanediol (OD), could be conducted at temperatures up to at
least 90 °C without negatively impacting the outcome of
reactions. Hence, to (i) melt unsaturated and epoxy diacids, (ii)
attain monophasic liquid media for bulk copolymerizations, and
(iii) minimize diffusion limitations under such conditions, all
polymerizations conducted herein were performed at 90 °C.
The time course of changes in weight average molecular
weight (Mw) and polydispersity (Mw/Mn) for N435-catalyzed in-
bulk polycondensations of 1,8-octanediol with ω-carboxy OA
to prepare poly(ω-carboxyl OA-co-OD) are displayed in Figure
4. Poly(ω-carboxyl OA-co-OD) Mw increased from about 15000
to 37000 g/mol from 2 to 24 h. Beyond 24 h, further increases
in molecular weight occurred slowly. By 48 h, Mw and Mw/Mn
While using the same ω-carboxyl diacid, the effect of varying
diol chain length on unsaturated polyester molecular weight was
investigated. Figure 4A shows that increasing the diol chain
length from 1,8-octanediol to 1,16-hexadecanediol (HD) gave
a similar trend in Mw increase as a function of time. The most
notable difference in the molecular weight behavior during
poly(ω-carboxyl OA-co-OD) and poly(ω-carboxyl OA-co-HD)
synthesis is that, for extended reaction times, Mw/Mn decreased
for the former but increased for the later. Increase in Mw/Mn is
indicative of an increase in events of transesterification that
appears to be more prevalent when using HD as comonomer.

266 Biomacromolecules, Vol. 11, No. 1, 2010
Yang et al.
Figure 4. Molecular weight (Mw (A) and polydispersity index (Mw/Mn (B) as a function of reaction time for N435-catalyzed polymerization of
ω-carboxyl OA with 1,3-propanediol, 1,8-octanediol, and 1,16-hexadecandediol. Reactions were performed in bulk at 90 °C.
Figure 5. Molecular weight (Mw, A) and polydispersity index (Mw/Mn, B) as a function of reaction time for N435-catalyzed polymerization of
1,8-octanediol with ω-carboxyl OA, ω-carboxyl EA, and ω-carboxyl epoxy SA. Reactions were performed in bulk at 90 °C.
The rapid chain growth found herein for N435-catalyzed
polymerizations of ω-carboxyl OA with longer chain diols OD
and HD was similarly observed for N435-catalyzed bulk
polymerizations of 12- and 16-carbon ω-hydroxyacids that,
performed at ambient pressure, gave polyesters in 4 h with Mn
up to 23000 g/mol (nonfractionated).16g By using 1,3-pro-
panediol (PD) in place of OD and HD as comonomer for
copolymerizations with ω-carboxyl OA, Mw increased more
slowly with time. While Mw for poly(ω-carboxyl OA-co-OD)
increased from 16000 to 27000 and 35000 g/mol at 6, 12, and
24 h, respectively, Mw for poly(ω-carboxyOA-co-PD) increased
from 11000 to 19000 and 23000 g/mol at 6, 12, and 24 h,
respectively. However, similarly to poly(ω-carboxyl OA-co-
OD), Mw/Mn of poly(ω-carboxyl OA-co-PD) decreased with
extended reaction time due, as described above, to higher
diffusivity of low molecular weight fragments to condense,
thereby narrowing the polydispersity. The slow reactivity of PD
relative to OD and HD is interesting, given that PD is much
closer in structure to glycerol, the acyl acceptor associated with
lipase activity in nature, than is OD or HD. Indeed, the relatively
low reactivity of shorter chain R,ω-alkylene glycols during
N435-catalyzed polymerizations was observed by others. Uyama
et al.29 reported that reactions between short chain diols (1,2-
ethane diol and 1,4-butanediol) and succinic anhydride gave
only oligomeric products, whereas polymerizations between
glutaric anhydride and 1,14-dihydroxytetradecane at 60 °C gave
polyesters with Mn 7600 g/mol and Mw/Mn of 2.1. The difficulty
in obtaining polyesters by N435-catalyzed polymerization using
diol monomers with four or fewer carbons and nonactivated
diacid comonomers was discussed by Azim et al.30 Similarly,
Mahapatro et al.16f reported that, for bulk N435-catalyzed
copolymerizations of adipic acid with 1,4-butanediol, 1,6-
hexanediol, and 1,8-octanediol at 70 °C, chain growth occurred
most rapidly with 1,8-octanediol and slowest with 1,4-butane-
diol.
Figure 5 shows the time course of Mw and Mw/Mn for N435-
catalyzed polymerization of 1,8-octanediol with ω-carboxyl EA
(C22 unsaturated diacid), ω-carboxyl epoxy SA (C18 epoxy
diacid), and ω-carboxyl OA, respectively. Poly(ω-carboxyl EA-
co-OD) Mw rapidly increased from 12000 to 21000 and 25000
g/mol in 2, 6, and 12 h, respectively. Similarly, Mw of poly(ω-
carboxyl epoxy SA-co-OD) rapidly increased from 19000 to
28000 and 32000 g/mol in 2, 6, and 12 h, respectively. Typical
of diffusion limited reactions, further extending the reaction
times beyond 12 h resulted in slow increases in Mw so that, by
48 h, copolymers of ω-carboxyl EA and ω-carboxyl epoxy SA
reached Mw values of 29000 and 39000 g/mol, respectively.
Increase in the diacid chain length by 4-methylene units for
EA relative to OA and epoxy SA did not result in enhanced
Mw increase (see Figure 5A). Variations in Mw/Mn between these
three copolymerizations are difficult to explain and may be due
to experimental error.
For all of the above-described polymerizations control
experiments were performed without addition of N435. GPC
analysis of reaction products under control conditions showed
that no polymers (Mw < 1000 g/mol) were formed.
To ensure that conversion of ω-carboxyl fatty acids to
polyesters occurred without disturbing double bond and epoxy
moieties, 1H and 13C NMR spectra of corresponding polyesters
were recorded. Chemical shifts of polyesters are listed in the
Experimental Section and 1H NMR spectra of poly(ω-carboxyl
OA-co-OD), poly(ω-carboxyl EA-co-OD), and poly (ω-carboxyl
epoxy SA-co-OD) are deposited in Supporting Information
(Figures S-5 and S-6). For copolyesters with unsaturated and

Functional Polyesters from ω-Carboxy Fatty Acids
Biomacromolecules, Vol. 11, No. 1, 2010 267
Table 1. Thermal Properties of Saturated, Unsaturated, and Epoxidized Polyesters Catalyzed by Novozym 435a
b
b
c
d
d
polyester
Mw (g/mol)
Mw/Mn
Td (°C)
Tm (°C)
∆Hm (J/g)
poly(ω-carboxyl OA-co-OD)
poly(ω-carboxyl OA-co-PD)
poly(ω-carboxyl OA-co-HD)
poly(ω-carboxyl EA-co-OD)
poly(ω-carboxyl epoxy SA-co-OD)
poly(ω-carboxyl SA-co-OD)
44000
25000
45000
29000
39000
76000
2.1
2.1
2.4
2.1
3.1
2.0
388
384
375
385
381
383
23/36
16/22
35/57
35/40
33
11.7/21.1
1.4/14.4
8.3/83.4
30.6/1.1
52.9
77/88
14.8/90.3
a Polycondensations of unsaturated and epoxidized diacids with diols were performed in bulk containing 1.0 mmol unsaturated diacid and 1.0 mmol
1,8-octanediol at 90 °C for 48 h. Polycondensation of 1,18-octadecanedioic acid with 1,8-octanediol was performed in toluene in a round-bottom flask.
b Data from GPC were measured using THF as eluent. Samples without precipitation for unsaturated and epoxidized polyesters were used for measurements.
c Data from TGA under nitrogen atmosphere at a heating rate of 10 °C/min from 25 to 700 °C. d Data from DSC from the second heating scan run at 10
°C/min.
1
epoxy moieties, H NMR signals at 5.35 and 2.90 ppm due to
vinyl and epoxy protons were found and integration intensities
are consistent with these groups remaining intact. Low intensity
signals at 3.65 ppm correspond to protons of -CH2-OH end
groups. Retention of double bond and epoxy group structure
during conversion of monomers to polyesters is an important
advantage of lipase-catalyzed condensation reactions. In contrast,
chemical-catalyzed condensation methods conducted at high
temperatures (>150 °C) with acid, base, or organo-metallic
reagents cause the formation of undesired side products by
reactions at unsaturated and epoxy groups.17d
with the longer chain length of ω-carboxyl EA relative to
ω-carboxyl OA. An increase in the alkylene diol chain length
of ω-carboxyl OA copolymers from 3 to 8 and 16 carbons
results in corresponding increases in Tm values (Table 1). Large
differences in melting transitions for poly(ω-carboxyl OA-co-
HD) and poly(ω-carboxyl SA-co-OD) are intriguing. Poly(ω-
carboxyl OA-co-HD) has only one point of cis-unsaturation that
provides a kink along a repeat unit consisting of a C18
unsaturated diacid and C16 saturated diol. That one kink, along
a 34-carbon AA-BB repeat unit structure, results in a predomi-
nant Tm transition at 57 °C. In contrast, poly(ω-carboxyl SA-
co-OD), with an AA-BB repeat unit structure consisting of only
26 carbons, but that is fully saturated, has a predominant melting
transition that is 31 °C higher (88 °C). This attests to the
powerful affects of cis-unsaturation that nature uses on a regular
basis to manipulate the fluidity and, therefore, biological
properties of lipid bilayers in living systems.32 Comparison of
thermal transitions for poly(ω-carboxyl OA-co-OD) and poly(ω-
carboxyl epoxy SA-co-OD) shows the relative affects of cis-
unsaturation and epoxidation on disruption of crystalline order
in nonbranched aliphatic polyesters containing ω-carboxy fatty
acid repeat units. These two polyesters, with cis-unsaturation
and epoxy moieties, have similar predominant melting transi-
tions at 36 and 33 °C, respectively. However, ∆Hm of the latter
is much higher (21 vs 53 J/g), while the former also has a
substantial lower melting transition at 23 °C. Hence, it is
concluded that cis-unsaturation relative to epoxidation has a
greater affect on disrupting crystalline organization of non-
branched long-chain aliphatic polyesters.
Thermal
Properties
Analysis
of
Saturated,
Unsaturated, and Epoxy Functionalized Polyesters. Thermal
properties of polyesters were analyzed by TGA and DSC. For
comparison with unsaturated and epoxy functionalized polyes-
ters described above, it was deemed useful to prepare the
corresponding nonfunctionalized polyester from 1,18-octade-
canedioic (ω-carboxy stearic acid, ω-carboxy SA). To that end,
N435-catalyzed synthesis of poly(ω-carboxyl SA-co-OD) was
undertaken. Due to the high melting point of ω-carboxyl SA
(122-123 °C),31 bulk polycondensations at 90 °C with OD were
unsuccessful. From previous work in our laboratory where
polymers of similar structure were synthesized by N435 catalysis
in diphenyl ether, difficulties were encountered when trying to
remove residual diphenyl ether from products. To circumvent
these problems, poly(ω-carboxyl SA-co-OD) synthesis was
conducted in toluene under vacuum so that the reaction was
maintained at 90 °C and water was removed azeotropically (see
Experimental Section). At 48 h, GPC analysis showed that a
high molecular weight polyester had formed with Mw 76000
1
g/mol and Mw/Mn 2.0. H NMR analysis of this product was
Conclusion
consistent with that expected (see Supporting Information,
Figure S-5). Subsequently, poly(ω-carboxyl SA-co-OD) was
isolated by precipitation, solvent was removed under vacuum
at 50 °C, and the resulting product was used for thermal analysis.
Table 1 lists results from DSC and TGA analyses of saturated,
unsaturated and epoxidized polyesters and their corresponding
molecular weight averages. Comparison of TGA scans shows
they are all alike in appearance and have temperatures at which
10% of weight loss was observed (Td) that are similar in value
ranging from 375 to 388 °C. Hence, the presence of epoxy and
unsaturated values as well as the use of diols differing in
structure led to polyesters of similarly high thermal stability.
Differential scanning calorimetry (DSC) traces of unsaturated
poly(ω-carboxyl OA-co-OD) and poly(ω-carboxyl EA-co-OD)
both show two peak melting transitions (Tm values) at 23/36
and 35/40 °C, respectively. Although the 40 °C transition has
a low ∆Hm, poly(ω-carboxyl EA-co-OD) lacks the melting
transition at 23 °C and the Tm at 35 °C has a ∆Hm of 30 J/g.
Higher melting transitions and ∆Hm values of poly(ω-carboxyl
EA-co-OD) relative to poly(ω-carboxyl OA-co-OD) is consistent
This work reported a unique two-pot, two-step route, using
whole-cell bio-oxidation followed by cell-free lipase catalysis
to convert readily renewable fatty acids to functional polyesters.
Whole-cell biotransformation was found to be an effective way
to synthesize ω-carboxyl fatty acids with internal functional
groups from renewable feedstocks. ω-Carboxyl OA, ω-carboxyl
EA, and ω-carboxyl epoxy SA were produced using fermenta-
tions of C. tropicalis ATCC20962 in good yield and with intact
double bond or epoxy groups. This biochemical route to
oxidation of terminal fatty acid methyl groups to carboxyl
moieties fills a gap in the expansive repertoire of organic
chemical methods.
Immobilized CALB (N435) is an efficient catalyst for the
polycondensation of unsaturated and epoxidized ω-carboxyl
fatty acids with diols to give double bond- or epoxy-function-
alized polyesters with high molecular weight. The effect of
changes in ω-carboxyl fatty acid building blocks and length of
diols on thermal properties of the resulting polyesters was
studied. Linear unsaturated and epoxidized polyesters had high

268 Biomacromolecules, Vol. 11, No. 1, 2010
Yang et al.
(12) Fabritius, D.; Schafer, H. J.; Steinbuchel, A. Appl. Microbiol.
Biotechnol. 1996, 45, 342–348.
(13) Fabritius, D.; Schafer, H. J.; Steinbuchel, A. Appl. Microbiol.
Biotechnol. 1997, 48, 83–87.
(14) Kandel, S.; Sauveplane, V.; Olry, A.; Diss, L.; Benveniste, I.; Pinot,
F. Phytochem. ReV. 2006, 5, 359–372.
(15) Odian, G. Principles of Polymerization; John Wiley & Sons: New
York, 2004; p 94.
(16) (a) Magolin, A. L.; Creene, J. Y.; Klibanov, A. M. Tetrahedron. Lett.
1987, 28, 1607–1612. (b) Chaudhary, A. K.; Lopez, J.; Beckman, E. J.;
Russell, A. J. Biotechnol. Prog. 1997, 13, 318–325. (c) Linko, Y. Y.;
Wang, Z. L.; Seppala, J. Enzymol. Microb. Technol. 1995, 17, 506–
511. (d) Linko, Y. Y.; Lamsa, M.; Wu, X.; Uosukanum, E.; Seppala,
J.; Linko, P. J. Biotechnol. 1998, 66, 41–50. (e) Uyama, H.; Inada,
K.; Kobayashi, S. Polym. J. 2000, 32, 440–443. (f) Mahapatro, A.;
Kalra, B.; Kumar, A.; Gross, R. A. Biomacromolecules 2003, 4, 544–
551. (g) Mahapatro, A.; Kumar, A.; Kalra, B.; Gross, R. A. Macro-
molecules 2004, 37, 35–40. (h) Ebata, H.; Toshima, K.; Matsumura,
S. Macromol. Biosci. 2007, 7, 798–803.
(17) (a) Wallace, J. S.; Morrow, C. J. J. Polym. Sci., Polym. Chem. Ed.
1989, 27, 2553. (b) Olson, D. A.; Sheares, V. V. Macromolecules
2006, 39, 2808–2814. (c) Uyama, H.; Kuwabara, M.; Tsujimoto, T.;
Kobayashi, S. Biomacromolecules 2003, 4, 211–215. (d) Warwel, S.;
Demes, C.; Steinke, G. J. Polym. Sci., Part A: Polym. Chem. 2001,
39, 1601–1609.
molecular weights of 25000-57000 g/mol with low melting
points (23-40 °C), whereas saturated polyesters from ω-car-
boxyl SA with 1,8-octanediol had a higher melting transition
(77-88 °C). Increasing the chain length of diols resulted in
higher molecular weights and melting point of unsaturated
polyesters from ω-carboxyl OA. The results of this work pave
the way for a new family of fatty acid derived polyesters from
a wide-range of ω-carboxyl fatty acid monomers. Unlike
polyesters from saturated diacids, polyesters with epoxy and
unsaturated moieties as described herein can be modified,
derivatized, or cross-linked at these sites, which will prove
valuable in developing curable coatings as well as for decoration
of these polymers with bioactive moieties for medical applications.
Acknowledgment. We thank the NSF and Industrial members
of the NSF-Industry/University Cooperative Research Center
(NSF I/UCRC) for Biocatalysis and Bioprocessing of Macro-
molecules at Polytechnic Institute of NYU for their financial
support, intellectual input, and encouragement during the course
of this research.
(18) Kelly, A. R.; Hayes, D. G. J. Appl. Polym. Sci. 2006, 101, 1646–
Supporting Information Available. Supplementary data of
mass spectrum of oxidative products from ω-carboxyl OA and
1656.
(19) Ebata, H.; Toshima, K.; Matsumura, S. Macromol. Biosci. 2007, 7,
798–803.
(20) Veld, M. A. J.; Palmans, A. R. A.; Meijer, E. W. J. Polym. Sci., Part
A: Polym. Chem. 2007, 45, 5968–5978.
(21) Olsson, A.; Lindstrom, M.; Lversen, T. Biomacromolecules 2007, 8,
757–760.
(22) Warwel, S.; Ruesch gen klaas, M. J. Mol. Catal. B: Enzym. 1995, 1,
1
ω-carboxyl EA, H and 13C NMR spectra of ω-carboxyl OA,
ω-carboxyl EA, and ω-carboxyl epoxy SA, IR spectra of
1
ω-carboxyl OA and ω-carboxyl EA, and H NMR spectra of
polyesters. This material is available free of charge via the
Internet at http://pubs.acs.org.
29–35.
(23) Dawidowicz, E. A.; Thompson, T. E. J. Lipid. Res. 1971, 12, 636.
(24) Blau, K.; Halket, J. Handbook of DeriVatiVes for Chromatography,
2nd ed.; John Wiley & Sons: New York, 1993.
(25) Cecilia, O.; Samuel, C.; Dietlind, A.; Bo, M.; Rajni, H. Eur. J. Lipid
Sci. Technol. 2005, 107, 864–870.
(26) Gmunder, F. K.; Kappeli, O.; Fiechter, A. Appl. Microbiol. Biotechnol.
1981, 12, 135–142.
(27) Yi, Z. H.; Rehm, H. J. Appl. Microbiol. Biotechnol. 1988, 28, 520–
526.
References and Notes
(1) Chan, E. C.; Kuo, J. Enzyme Microb. Technol. 1997, 20, 585–589.
(2) Sato, K.; Aoki, M.; Noyori, R. Science 1998, 281, 1646–1647.
(3) Debuan, F.; Ha¨nssle, P. J. Synth. Lubr. 1985, 1, 254–267.
(4) Picataggio, S.; Rohrer, T.; Deanda, K.; Lanning, D.; Reynolds, R;
Mielenz, J.; Eirich, L. D. Bio/technology (N.Y.) 1992, 10, 894–898.
(5) Kolattukudy, P. E. Science 1980, 208, 990–1000.
(6) Yi, Z. H.; Rehm, H. J. Eur. J. Appl. Microbiol. Biotechnol. 1982, 15,
175–179.
(28) Jiang, Z. Z.; Liu, C.; Xie, W. H.; Gross, R. A. Macromolecules 2007,
(7) Hill, F. F.; Venn, I.; Lukas, K. L. Appl. Microbiol. Biotechnol. 1986,
40, 7934–7943.
24, 168–174.
(29) Uyama, H.; Wada, S.; Fukui, T.; Kobayashi, S. Biochem. Eng. J. 2003,
16, 145–152.
(30) Azim, H.; Dekhterman, A.; Jiang, Z. Z.; Gross, R. A. Biomacromol-
ecules 2006, 7, 3093–3097.
(31) Tulloch, A. P.; Spencer, J. F. T.; Gorin, P. A. J. Can. J. Chem. 1962,
40, 1326–1338.
(32) Rawicz, W.; Olbrich, K. C.; McIntosh, T.; Needham, D.; Evans, E.
Biophys. J. 2000, 79, 328–339.
(8) Green, K. D.; Turner, M. K.; Woodley, J. M. Enzyme Microb. Technol.
2000, 27, 205–211.
(9) Broadway, N. M.; Dickinson, M.; Ratledge, C. J. Gen. Microbiol.
1993, 139, 1337–1349.
(10) Eschenfeldt, W. H.; Yeyan, Z. H.; Samaha, H.; Stols, L.; Eirich, D. C.;
Ronald, W.; Donnelly, M. I. Appl. EnViron. Microbiol. 2003, 69, 5992–
5999.
(11) Yi, Z. H.; Rehm, H. J. Appl. Microbiol. Biotechnol. 1989, 30, 327–
331.
BM901112M