Synthesis of polyfunctional fatty amines from sophorolipid-derived 17-hydroxy oleic acid

Article abstract of DOI:10.1007/s11746-006-1248-1

A series of mono- and diamines and one triamine have been prepared using methyl 17-hydroxy oleate as the common starting material. The 1 7-hydroxy oleate is an abundant bioderived material obtained from acid alcoholysis of sophorolipids, which in turn are

Full text of DOI:10.1007/s11746-006-1248-1

Synthesis of Polyfunctional Fatty Amines
from Sophorolipid-Derived 17-Hydroxy Oleic Acid
Jonathan A. Zerkowski* and Daniel K.Y. Solaiman
USDA, ARS, Eastern Regional Research Center, Wyndmoor, Pennsylvania 19038
ABSTRACT: A series of mono- and diamines and one triamine tional groups (carboxyl, olefin, and hydroxy) could be ad-
have been prepared using methyl 17-hydroxy oleate as the com-
mon starting material. The 17-hydroxy oleate is an abundant
bioderived material obtained from acid alcoholysis of
sophorolipids, which in turn are produced by fermentation of
agricultural by-products. Incorporation of the amino unit(s) can
be selectively performed at either end of the chain or at its mid-
dle. The chief synthetic reactions used are allylic bromination,
Curtius rearrangement, and the Mitsunobu reaction. These fatty
amines also possess functionality such as hydroxy groups, car-
boxylic acids, and C–C double bonds. The amines are isolated in
protected form using a variety of protecting groups, the identities
dressed individually or sequentially for purposes of further de-
rivatization. The roughly equidistant spacing of the three func-
tional groups along the alkyl chain presents a unique opportu-
nity; ricinoleic acid, for example, is not subject to ready
derivatization at its alkane terminus. The class of derivatives
we have chosen to investigate initially are amines. A great deal
of work has appeared on oxygenation of FA, but fewer studies
have been devoted to amination. We report here some prelimi-
nary synthetic manipulations that convert 17-hydroxy oleic
acid derived from sophorolipids into several classes of poly-
of which can be selected on the basis of the intended use of the functional fatty amines (Fig. 2). For ease of purification and
amine. These novel compounds will be of interest in the prepara-
tion of highly functionalized polymers and surfactants, among
other areas.
characterization, the amines are generated as derivatives with
common, readily removable protecting groups attached (10),
but to reduce wordiness these are termed “amines” in the rest
of this paper.
Paper no. J11281 in JAOCS 83, 621–628 (June 2006).
The rationale behind our focus on amines is their use in a
wide range of important technological situations. Since the
compounds depicted in Figure 2 are novel and have not them-
selves been used in such situations, some justification may be
in order. We expect that the most likely applications of these
multifunctional fatty amines will be in polymer chemistry,
where they will be useful for preparing new types of
polyamides or polyureas, and in surfactant chemistry, where
the ability to position two charges at opposite ends of an alkyl
KEY WORDS: Biorenewables, fatty amines, hydroxy fatty
acids, sophorolipids.
The conversion of bulk agricultural products and co-products,
such as vegetable oils, tallow, restaurant greases, bioglycerol,
and soy molasses, into value-added materials is an area of bur-
geoning scientific and commercial interest. Proposed uses for
such value-added materials include surfactants, gelators, drug-
delivery adjuvants, and building blocks for polymers (1,2). Mi-
crobial, enzymatic, and chemical routes have been used, and a
wide range of molecular structures may result (3). One promis-
ing class of compounds is glycolipids, particularly
sophorolipids, which can be produced abundantly from renew-
able feedstocks (4–6). The crude sophorolipids as obtained
from a fermentation run consist of a variety of lactonic and free
acid forms in various states of acetylation (hence the pluralized
name). Although the sophorolipid molecule offers much po-
tential for derivatization in its own right (7,8), it can also be
subdivided into the carbohydrate and FA components. The lat-
ter, which are hydroxylated predominantly at the ω-1 position,
are attractive raw materials for many purposes. A good com-
parison is afforded by the hydroxy-containing ricinoleic acid,
which has been the subject of many synthetic modifications
(1,9).
As Figure 1 shows, this starting material is a trifunctional-
ized hydrocarbon. We hypothesized that each of its three func-
FIG. 1. The 17-hydroxy oleic acid that is the precursor for new com-
pounds discussed in this work is obtained from acid alcoholysis of
sophorolipids, which are predominantly acetylated macrolactones.
Molecule 1 can be derivatized at its carboxyl, olefin, or hydroxy group.
*To whom correspondence should be addressed at Fats, Oils, and Animal
Coproducts Research Unit, USDA, ARS, ERRC, 600 East Mermaid Ln.,
Wyndmoor, PA 19038. E-mail: jzerkowski@errc.ars.usda.gov
Copyright © 2006 by AOCS Press
621
JAOCS, Vol. 83, no. 7 (2006)

622
J.A. ZERKOWSKI AND D.K.Y. SOLAIMAN
FIG. 2. Synthetic transformations used to introduce amine units into 17-hydroxy oleate. Reagents and conditions: a: LiOH, MeOH/H2O. b:
diphenylphosphoryl azide (DPPA) in t-BuOH, NEt3, ∆. c: N-bromosuccinimide in CCl4, ∆. d: NaN3, Bu4NBr, dimethylformamide/toluene, ∆. e: (1)
triphenylphosphine (PPh3), ∆. (2) H2O, ∆. f: di-t-butyl dicarbonate (Boc2O)/NEt3. g: DPPA, di-isopropylazodicarboxylate (DIAD), PPh3. h. phthal-
imide, PPh3, DIAD. i: N-(9-fluorenylmethoxycarbonyloxy)-succinimide (Fmoc-OSu), NEt3. j: o-nitrobenzenesulfonyl (Nosyl)-chloride, NEt3. k:
DPPA in BnOH, NEt3, ∆. l:o-NO2C6H4SO2NH-propyl, DIAD, PPH3. PG, protecting group. For a discussion of the regioisomerism of the allylic
substituents in 5, 11, 12, and 15 (indicated in the figure), see the text.
chain could be harnessed to make bola-amphiphiles (11). Bio-
logical applications are also compelling. α-Amino oleic acid
has been used as an anticalcification agent in prosthetic heart
valves (12). Isomeric amino oleic acids of the sort prepared in
this paper will shed light on structure/function relationships in
that application, as well as perhaps finding use as coatings in
other biomaterials-type devices. Long-chain fatty amines also
constitute the core of ceramide analogs, which are under inves-
tigation as inducers of apoptosis (13). Fluorescent derivatives
of amino FA have been used as probes to study enzymes in-
volved in lipid regulation and biosynthesis (14,15). Finally,
there are a number of newer technologies where fatty amines
are being used: as matrices for nucleation of inorganic
nanocrystals (16); as surface-bound monolayers for anchoring
cells (17) or immobilizing enzymes (18); as cross-linking
agents for composite materials (19); and as anticorrosion or an-
tioxidation units in lubricant formulations (20). In regard to this
last area, some of the molecules shown in Figure 2 could be
converted into amino versions of estolides (21). In any of these
fields, the opportunity to incorporate some extra functionality
of choice (e.g., a fluorescent or metal-chelating unit) or to alter
the hydrophobicity profile by, for example, hydroxylating the
olefin, would open up rich new avenues of exploration.
EXPERIMENTAL PROCEDURES
General. Sophorolipids were prepared as previously reported,
using oleic acid and glucose as carbon sources (5). Chemicals
and reagents were obtained commercially from Sigma-Aldrich
(St. Louis, MO) and Lancaster Synthesis/Alfa Aesar (Ward
Hill, MA). All solvents and reagents were used as received. Sil-
ica gel used for column chromatography was obtained from
Fisher Scientific (Fairlawn, NJ). NMR spectra were recorded
on a Varian Associates Gemini 200 MHz instrument. LC-MS
data were recorded on a Waters/Micromass ZMD instrument
with atmospheric pressure CI (APCI), using an elution gradi-
ent of 20:80 water/methanol to 100% acetonitrile over 30 min
(or minor variations on those conditions) on a 2.1 × 150 mm
Waters Symmetry C18 3.5 µ column. See Table 1 for spectro-
scopic data.
Methyl 17-hydroxy oleate (1). Crude sophorolipids (20 g)
were dissolved in 1% H2SO4/MeOH (250 mL) and heated to
60–65°C for 16–20 h. Methanol was removed on the rotary
evaporator, and the resulting brown syrup was diluted with 100
mL of water and extracted with ethyl ether (3 × 100 mL). The
organic fractions were condensed to give 8.5 g material. Yields
can be calculated only approximately, since the starting mater-
JAOCS, Vol. 83, no. 7 (2006)

POLYFUNCTIONAL FATTY AMINES
623
TABLE 1
Spectra of Products
Compound
13C NMRa
1H NMR
Mass spectrab
3
23.6, 25.9, 26.9, 27.3, 28.5, 29.2, 29.3,
29.6, 29.7, 30.2, 39.5, 40.7, 68.2, 79.1,
129.9, 130.0, 156.1
1.20 (d, 3H, 7.5 Hz), 1.24–1.48 (br s, 16 H),
1.48–1.56 (br s, 11 H), 1.56–1.74 (m, 2 H),
1.90–2.16 (m, 4H), 3.10 (q, 2H, 7.3 Hz),
3.72–3.90 (m, 1H), 4.55 (br s, 1H),
370, 314 (loss of t-Bu), 270
(loss of Boc), 252
(loss of Boc + H2O)
5.24–5.45 (m, 2H)
4
23.6, 25.9, 27.0, 27.3 (m), 28.5, 29.2,
29.3, 29.6, 29.7, 30.3, 39.5, 41.1, 43.6
(br), 68.2, 80.3, 129.9, 130.0, 154.8, 157.8 1.93–2.17 (m, 4H), 3.24 (q, 2H, 7.2 Hz),
3.30–3.41 (m, 4H), 3.41–3.57 (m, 4H),
1.20 (d, 3H, 7.5 Hz), 1.24–1.43 (br s, 16 H),
1.43–1.65 (br s, 11 H), 1.65–1.79 (m, 2 H),
482, 426 (loss of t-Bu), 382
(loss of Boc), 278, 270 (loss of
Boc-piperazine-CO), 252
3.71–3.90 (m, 1H), 4.48 (t, 1H, 7.0 Hz),
5.26–5.47 (m, 2H)
5
6
7
8
23.6, 25.0, 25.8 (m), 28.5, 29.2 (m), 29.5
(m), 32.3 (m), 34.2, 35.8 (m), 39.4 (m),
51.6, 52.5, 68.2, 79.2, 130.8, 130.9,
131.1, 131.3, 155.5, 174.4
1.21 (d, 3H, 7.4 Hz), 1.24–1.48 (br s, 16 H),
1.48–1.56 (br s, 11 H), 1.56–1.78 (m, 2 H),
1.93–2.14 (m, 2H), 2.32 (t, 2H, 7.2 Hz), 3.68 (s,
3H), 3.74–3.89 (m, 1H), 3.91–4.11 (m, 1H), 4.40
(br s, 1H), 5.22–5.37 (m, 1H), 5.46–5.65 (m, 1H)
1.12 (d, 3H, 7.5 Hz), 1.20–1.43 (br s, 16 H),
1.43–1.52 (br s, 9 H), 1.52–1.78 (m, 4H),
1.92–2.15 (m, 4H), 2.33 (t, 2H, 7.3 Hz),
3.53–3.80 (m, 1H), 3.66 (s, 3H), 4.34 (br s, 1H),
5.26–5.44 (m, 2H)
428, 383 (loss of CH3CHOH),
372 (loss of t-Bu), 354 (loss of
t-Bu and H2O), 328 (loss of
Boc), 311, 293, 278
21.4, 25.0, 26.1, 27.3 (m), 28.5, 29.2
(m), 29.3, 29.5, 29.8, 34.2, 37.5, 46.7,
51.6, 79.0, 129.9, 130.0, 155.5, 174.4
412, 356 (loss of t-Bu), 312
(loss of Boc)
18.8, 25.0, 26.9, 27.2, 29.2 (m), 29.7,
33.8, 34.2 (m), 47.6, 51.5, 123.1, 129.9
(m), 132.1, 133.9, 168.7, 174.4
1.14–1.43 (br s, 16 H), 1.47 (d, 3H, 7.5 Hz),
1.55–1.80 (m, 4 H), 1.91–2.16 (m, 4H), 2.31
(t, 2H, 7.2 Hz), 3.67 (s, 3H), 4.24–4.44 (m, 1H),
5.26–5.45 (m, 2H), 7.66–7.77 (m, 2H),
442
7.77–7.92 (m, 2H)
21.9, 25.6, 26.9, 27.2, 27.3, 28.5, 29.2
(m), 29.3, 29.7 (m), 30.2, 37.5, 40.7,
51.3, 79.1, 125.4, 129.8, 130.0, 130.7,
132.9, 133.4, 135.3, 147.9, 156.1
1.12 (d, 3H, 7.4 Hz), 1.15–1.41 (m, 16 H),
1.41–1.76 (br s, 13 H), 1.89–2.15 (m, 4H), 3.13
(q, 2H, 7.6 Hz), 3.42–3.63 (m, 1H), 4.54 (br s,
1H), 5.16 (br d, 1H), 5.25–5.44 (m, 2H), 7.70–7.82
(m, 2H), 7.83–7.97 (m, 1H), 8.14–8.21 (m, 1H)
1.13 (d, 3H, 7.4 Hz), 1.22–1.44 (br s, 16 H),
1.44–1.68 (br s, 22 H), 1.94–2.14 (m, 4H), 3.14
(q, 2H, 7.2 Hz), 3.53–3.75 (m, 1H), 4.33 (br s,
1H), 4.56 (br s, 1H), 5.31–5.44 (m, 2H)
554, 498 (loss of t-Bu),
454 (loss of Boc)
10
11
21.4, 26.1, 26.9, 27.3, 28.5, 29.3 (m),
29.5, 29.6, 29.7, 29.8, 30.1, 37.4, 41.0,
46.9, 79.3, 129.9, 130.0, 151.1, 151.3
469, 413 (loss of t-Bu), 369
(loss of Boc), 357 (loss of 2
t-Bu), 313, 295, 269
loss of 2 Boc), 252
607, 507 (loss of Boc), 440,
385 (loss of Fmoc), 267
23.6, 25.8 (m), 26.8 (m), 28.5, 29.2 (m),
29.4, 30.1, 32.3, 35.6, 39.3 (m), 40.7,
47.4, 53.0, 66.5, 68.2, 79.1, 120.1,
125.2, 127.1, 127.7, 130.5, 131.8 (m),
141.4, 144.1, 155.9, 156.1
1.19 (d, 3H, 7.4 Hz), 1.23–1.40 (m, 16 H),
1.40–1.58 (s, 13H), 1.95–2.08 (m, 2H), 3.10
(t, 2H, 7.3 Hz), 3.71–3.86 (m, 1H), 4.24 (m, 1H),
4.42 (t, 1H, 8.0 Hz), 4.53 (br s, 1H), 4.59 (d, 2H,
7.9 Hz), 4.73 (br s, 1H), 5.22–5.40 (m, 1H),
5.44–5.68 (m, 1H), 7.30–7.53 (m, 4H), 7.58–7.68
(m, 2H), 7.78 (d, 2H, 6.5 Hz)
1.10–1.38 (br s, 16 H), 1.38–1.53 (br s, 11 H),
1.46 (d, 3H, 7.6 Hz), 1.53–1.69 (m, 2 H),
1.90–2.16 (m, 2H), 2.30 (t, 2H, 7.2 Hz), 3.66 (s,
3H), 3.87–4.14 (m, 1H), 4.23–4.45 (m, 1H), 4.41
(br s, 1H), 5.16–5.36 (m, 1H), 5.42–5.62 (m, 1H),
7.64–7.76 (m, 2H), 7.76–7.90 (m, 2H)
1.13 (d, 3H, 7.5 Hz), 1.18–1.39 (br s, 16 H),
1.39–1.54 (br s, 11 H), 1.54–1.73 (m, 2 H),
1.88–2.15 (m, 2H), 1.98 (s, 3H), 2.32 (t, 2H,
7.4 Hz), 3.67 (s, 3H), 3.85–4.11 (m, 2H), 4.42 (br s,
1H), 5.17–5.35 (m, 1H), 5.38 (br s, 1H),
5.45–5.65 (m, 1H)
12
13
18.8, 25.0 (m), 25.7, 25.8, 26.7, 26.8,
28.5, 28.8, 29.0 (m), 29.1, 29.3 (m), 32.3
(m), 33.8, 34.2, 35.7, 47.5, 51.5, 52.4,
79.2, 123.1, 130.8, 130.9, 131.0, 131.2,
132.1, 133.9, 155.4, 168.7, 174.4
557, 501 (loss of t-Bu), 457
(loss of Boc), 440, 408
21.1, 23.6, 25.0, 25.8 (m), 26.0 (m),
28.5, 29.0, 29.2 (m), 29.4, 32.2 (m),
34.2, 35.7, 37.0, 45.4, 51.6, 52.4, 79.2,
130.8, 131.0, 131.1, 131.3, 155.5,
169.4, 174.4
469, 413 (loss of t-Bu),
369 (loss of Boc), 352, 320
14
15
23.6, 25.8, 26.7, 28.5, 29.2 (m), 29.4,
29.8, 30.0, 30.4, 32.3, 35. 7 (m), 39.4,
41.2, 50.0, 52.4, 66.7, 68.2, 79.2, 128.2,
128.4, 128.6, 130.9 (m), 131.2 (m),
136.8, 155.5, 156.5
11.6, 19.5, 21.5, 25.1, 25.8, 26.5 (m),
28.5, 29.3 (m), 29.8, 30.0, 32.2 (m), 35.7
(m), 41.2, 45.4, 52.4, 54.4, 66.7, 79.2,
124.0, 124.2, 128.2 (m), 128.6, 130.8,
131.0, 131.2, 131.5, 131.6, 133.3,
134.3, 136.8, 148.1, 155.5, 156.5
1.20 (d, 3H, 7.4 Hz), 1.23–1.42 (br s, 14 H),
1.41–1.79 (br s, 15 H), 1.94–2.18 (m, 2H), 3.20
(q, 2H, 7.6 Hz), 3.71–3.86 (m, 1H), 3.93–4.10
(m, 1H), 4.42 (br s, 1H), 4.83 (br s, 1H), 5.10 (s, 2H),
5.21–5.39 (m, 1H), 5.45–5.65 (m, 1H), 7.38 (m, 5H)
0.91 (t, 3H, 7.4 Hz), 1.13 (d, 3H, 7.5 Hz),
1.18–1.42 (br s, 16 H), 1.42–1.77 (br s, 15 H),
1.92–2.12 (m, 2H), 3.11–3.35 (m, 4H), 3.70–3.83
(m, 1H), 3.85–4.06 (m, 1H), 4.52 (br s, 1H), 4.77
(br s, 1H), 5.11 (s, 2H), 5.23–5.36 (m, 1H),
5.46–5.65 (m, 1H), 7.39 (br s, 5H), 7.65–7.78
(m, 3H), 7.97–8.10 (m, 1H)
519, 419 (loss of Boc), 358,
311, 294
745, 645 (loss of Boc), 628,
595, 581, 538, 520, 494, 464
aIn 13C NMR spectra, “m” indicates overlapped peaks that are poorly resolved, with δ∆ < 0.1 ppm.
bThe first mass listed is in all cases MH+. BOC, t-butyl carbonate; Fmoc, 9-fluorenylmethoxycarbonyl.
JAOCS, Vol. 83, no. 7 (2006)

624
J.A. ZERKOWSKI AND D.K.Y. SOLAIMAN
ial is heterogeneous, but in assuming it to be pure di-acetyl
oleic sophorolactone (M.W. = 688) this amount of product
would represent a yield of 93%. Compound 1 was isolated out
of the mixture of hydroxy methyl esters (see Results and Dis-
cussion) by column chromatography on silica gel with 2:1 hex-
ane/ethyl acetate.
17-Hydroxy oleic acid (2). Compound 1 (1.25 g, 4 mmol)
was dissolved in 50 mL 1:1 MeOH/H2O, and LiOH hydrate
(430 mg, 10.2 mmol) was added. The mixture was heated to
40–45°C overnight. Methanol was removed on the rotary evap-
orator, 10% wt/vol citric acid solution (100 mL) was added to
acidify the product, and the mixture was extracted twice with
50 mL ethyl ether. Evaporation of the solvent yielded an off-
white solid, 1.18 g (98%).
Synthesis of terminal amine (3). To a solution of 17-hydroxy
oleic acid 2 (810 mg, 2.7 mmol) in 10 mL t-BuOH heated to
50–55°C under N2 was added triethylamine (417 mL, 3 mmol)
and a solution of diphenylphosphoryl azide (DPPA; 825 mg, 3
mmol) in 2 mL THF. Heating was increased to afford reflux,
and the mixture was stirred for 24 h. It was then cooled and
concentrated on the rotary evaporator, then applied to a silica
gel column. Elution with 2:1 hexane/ethyl acetate afforded
molecule 3 (408 mg, 41%). The 16-hydroxy heptadecyl isocy-
anate 3a was also recovered (295 mg, 37%) and was reacted
SCHEME 1
with N-butyloxycarbonyl-piperazine in THF and NEt3 at 50°C and compound 5 (656 mg, 48%) was isolated by column chro-
overnight to obtain 4 (see Scheme 1).
matography on silica gel (2:1 hexane/ethyl acetate).
Synthesis of mid-chain amine (5). A solution of 1 (1 g, 3.2
Synthesis of 17-amino-oleic acid derivatives (6 and 7). To a
mmol), N-bromosuccinimide (NBS; 571 mg, 3.2 mmol), and magnetically stirred solution of 1 (1.56 g, 5 mmol), DPPA (1.38
benzoyl peroxide (30 mg) in 30 mL CCl4 was heated to reflux g, 5 mmol), and PPh3 (1.31 g, 5 mmol) in 50 mL THF under
under N2. After 90 min, TLC indicated that no 1 remained, and N2 a solution of di-isopropylazodicarboxylate (DIAD, 1.01 g,
all of the suspended solid floated. The reaction flask was cooled 5 mmol) in 3 mL THF was added dropwise over the course of
to room temperature and the contents were filtered. Solvent was 5 min. Stirring was continued overnight for 24 h. Method A
removed on the rotary evaporator and was further concentrated was then applied, followed by addition of Boc2O (1.09 g, 5
twice out of THF to remove all traces of chlorinated solvents. mmol). Silica gel chromatography in 4:1 hexane/ethyl acetate
The resulting colorless syrup was dissolved in 50 mL 1:1 di- yielded 6 (622 mg, 45% based on recovered starting material)
methylformamide/toluene, and n-Bu4NBr (50 mg) and NaN3 as well as starting material 1 (510 mg).
(520 mg, 8 mmol) were added. The solution was heated to re-
In a variant of this reaction, 1 (1.48 g, 4.7 mmol), phthalim-
flux (of toluene) under N2. Progress of the reaction was moni- ide (691 mg, 4.7 mmol), and PPh3 (1.24 g, 4.7 mmol) were dis-
tored by TLC to observe disappearance of the intermediate solved in 20 mL THF in a round-bottomed flask submerged in
bromo compound; heating was continued for 24 h. The mix- an ice-water bath, and a solution of DIAD (950 mg, 4.7 mmol)
ture was cooled to room temperature and poured into 300 mL in 3 mL THF was added. The solution was stirred under N2 and
H2O, then extracted with toluene (3 × 50 mL). The organic allowed to warm to room temperature overnight. Solvent was
layer was dried over MgSO4 and filtered.
removed on the rotary evaporator, and the material was puri-
A sequence of steps, which was used as a general procedure fied by column chromatography to yield 1.45 g of compound 7
in several reactions (see below), was then performed (Method (3.3 mmol, 70%).
A): Triphenylphosphine (PPh3, 838 mg, 3.2 mmol) was added
Synthesis of proximal/distal diamines (8 and 10). An analo-
to the toluene solution, and it was heated to 50–55°C. Stirring gous procedure to the one for synthesizing 6 was followed,
under N2 proceeded for another 24 h, following which 1 mL of using 3 (715 mg, 1.94 mmol) as starting material. Instead of
water and 100 mL THF were added. Stirring at the same tem- protection with Boc2O, o-nitrobenzenesulfonyl chloride (663
perature continued for another 24 h.
mg, 3 mmol) and NEt3 (420 µL, 3 mmol) were added after per-
At the end of this sequence, a reagent for introducing the forming Method A. Column chromatography in 3%
protecting group of choice onto the newly formed amino group MeOH/CHCl3 afforded 8 (266 mg, 36% allowing for recov-
was added, which for this example was di-t-butyl dicarbonate ered starting material) and 3 (224 mg).
(Boc2O, 700 mg, 3.2 mmol), along with 3 mL NEt3. This mix-
To prepare molecule 10, molecule 6 (625 mg, 1.5 mmol) and
ture was stirred under N2 at room temperature for 72 h. At the LiOH (64 mg, 1.5 mmol) were dissolved in 10 mL 2:1
end of this time, solvent was removed on the rotary evaporator, MeOH/H2O and heated to 45°C overnight. MeOH was removed
JAOCS, Vol. 83, no. 7 (2006)

POLYFUNCTIONAL FATTY AMINES
625
on the rotary evaporator, and the aqueous portion was acidified by reacting o-nitrophenylsulfonyl chloride with an excess of
with citric acid and then washed with ether (2 × 50 mL). The propylamine) were added, followed by the dropwise addition
ether extracts were dried over Na2SO4 to afford a colorless oil, of DIAD (81 mg, 0.4 mmol) in 2 mL THF over 3 min. The so-
which was purified by column chromatography on silica gel with lution was stirred at room temperature under nitrogen for 18 h,
9:1 CHCl3/MeOH. Molecule 9, the carboxylic acid, was ob- then another portion of PPh3, DIAD, and the sulfonamide (half
tained (250 mg, 0.63 mmol) as well as starting material (305 the previous amounts) was added. Stirring was continued for
mg). Molecule 9 was dissolved in 10 mL t-BuOH and 2 mL another 6 h, solvent was removed on the rotary evaporator, and
THF, heated to 55°C, and then NEt3 (104 µL, 0.75 mmol) and the crude product was purified by column chromatography on
DPPA (204 mg, 0.74 mmol) were added. The heat was increased silica gel, first with 2:1 hexane/ethyl acetate and then with 9:1
to reflux the solution for 18 h. Workup proceeded as for 3 to give chloroform/ethyl acetate, to give 15 (107 mg, 74% based on re-
10 (107 mg, 36%).
covered 14).
Synthesis of proximal/mid diamine (11). Molecule 3 (335
mg, 0.91 mmol) was treated with NBS (160 mg, 0.9 mmol) and
benzoyl peroxide (20 mg) as for 5 above, then reacted with
RESULTS AND DISCUSSION
NaN3 and processed with Method A. At the end of this se- Preparation of starting material. Methyl 17-hydroxy oleate (1)
quence, N-(9-fluorenylmethoxycarbonyloxy)-succinimide was readily obtained by acid alcoholysis of sophorolipids that
(Fmoc-OSu, 300 mg, 0.92 mmol) and NEt3 (0.5 mL) were were prepared according to literature procedures using oleic
added, and the mixture was stirred overnight. Solvent was re- acid as the feedstock. Since one of the long-term goals of this
moved on the rotary evaporator, and the residue was chro- work is to convert agricultural by-products to useful materials,
matographed on silica gel with 5% MeOH/CHCl3 to afford 11 it might seem counterintuitive to begin by using a purified
(217 mg, 39%).
chemical as feedstock rather than a chemically complex waste.
Synthesis of mid/distal diamine (12). Molecule 5 (200 mg, At this early stage of the project, however, it is important to
0.47 mmol) was treated with phthalimide (76 mg, 0.52 mmol), validate the kinds of synthetic conversions that can be accom-
DIAD (104 mg, 0.52 mmol), and PPh3 (136 mg, 0.52 mmol) plished without introducing ambiguities due to chemical het-
as for molecule 7 above to afford 12 (193 mg, 74%).
erogeneity. Even using pure feedstock, the hydroxy FAME ob-
To remove the phthalimido protecting group, 12 (130 mg, tained consisted of a mixture of products, predominantly 1 with
0.23 mmol) was dissolved in 8 mL 2-propanol/H2O (6:1, smaller amounts of the corresponding stearate (approximately
vol/vol), and NaBH4 was added in four portions of 20–25 mg 15%), the 18-hydroxy oleate (5%), and palmitates (5%, pre-
each over the course of 20 min. After another 90 min, another sumably 15- and 16-hydroxy although we did not investigate
30 mg was added, and the reaction was stirred for another 90 those compounds). Compound 1 could be purified away from
min. At the end of this time, glacial acetic acid was added (cau- these minor analogs by column chromatography. Yields were
tiously—evolution of gas) to an apparent pH of 5 (approx. 1 very good, better than 90%, but it is important to note that by
mL was added), then the mixture was heated to 75°C for 2 h. It discarding the disaccharide portion, 10 g of sophorolipid start-
was cooled to room temperature and stirred overnight. The so- ing material can theoretically yield only 4.5 g of 1. For steps
lution was evaporated to dryness on the rotary evaporator, and where a free carboxylic acid rather than the ester was required,
pyridine (5 mL) and acetic anhydride (0.5 mL) were added. hydrolysis with LiOH in aqueous methanol gave excellent
The reaction was stirred at room temperature for 24 h, then yields.
concentrated under vacuum to a syrup and chromatographed
Synthesis of monoamines. To introduce an amino function-
on silica gel with 1:1 hexane/ethyl acetate to remove a by-prod- ality at the end of the molecule proximal to the carboxylic acid,
uct, then 100% ethyl acetate to afford 13 (66 mg, 62%; see several synthetic strategies can be envisioned. α- or β- substi-
Scheme 1).
tution through enolates or conjugate addition will be worth in-
Synthesis of triamine (15). Molecule 5 (450 mg, 1.05 mmol) vestigating in the longer term, but since those routes require
was hydrolyzed with LiOH and worked up according to the several steps and possibly harsher conditions, we initially chose
procedure above for converting 6 to 10. The resulting car- to convert the carboxylic acid functionality of 1 directly into
boxylic acid (375 mg, 0.91 g) was then dissolved in 5 mL THF; an amine. For this we used the Curtius rearrangement as im-
5 mL benzyl alcohol and NEt3 (280 µL, 2 mmol) were added, plemented with DPPA. This reagent has attained popularity as
and the solution was heated to 55°C, at which point DPPA (275 a convenient way to introduce azide groups, and when used at
mg, 1 mmol) was added and heating was increased to reflux the high temperature in an alcohol gives the carbamate-protected
THF. Heating was continued for 18 h, then the reaction mix- amine (22). Molecule 3 was obtained in acceptable yields of
ture was cooled to room temperature and the THF was removed 30–40%. Note that the FA chain has been shortened by one car-
on the rotary evaporator. Ether (100 mL) was added, and the bon owing to the rearrangement mechanism. A minor by-prod-
solution was extracted with H2O (5 × 100 mL). Diamine 14 uct (approximately 10%) was the t-butyl ester of 17-hydroxy
was purified by chromatography on silica gel.
oleic acid, which could be recycled or used in other syntheses
To add the distal amino unit, 14 (140 mg, 0.27 mmol) was if desired. A more significant by-product, 3a, was also ob-
dissolved in 10 mL THF, and PPh3 (105 mg, 0.4 mmol) and N- served, the amount of which (30–40%) could not be lessened
propyl-o-nitrobenzenesulfonamide (99 mg, 0.4 mmol, prepared by varying reaction conditions such as temperature, base, or
JAOCS, Vol. 83, no. 7 (2006)

626
J.A. ZERKOWSKI AND D.K.Y. SOLAIMAN
co-solvent (although see the synthesis of 14, which follows, for as above. This route has the advantage of avoiding use of the
a way to circumvent it). We were surprised to discover that this hazardous sodium azide. Alternatively, 1 could be subjected to
material was the isocyanate, as could be demonstrated by con- the Mitsunobu reaction with phthalimide as the nitrogen car-
verting it to the desired carbamate 3 by further heating with t- rier. This reaction is convenient to perform and proceeds in
butanol and NEt3, or to the urea 4 by reaction with mono-Boc good yield. Both of these reactions should proceed with inver-
piperazine. This isocyanate appears to be less prone to hydrol- sion of configuration at C17, although we have not investigated
ysis than shorter-chain analogs: A sample of it that had been chirality in any of the molecules reported in this work.
stored in ethyl acetate solution survived with little or no appar-
Synthesis of diamines. By combining the methodologies
ent decomposition at room temperature for several months, and used to generate monoamines from 1, three positional isomers
samples of it were regularly chromatographed on silica gel. of diamines can be obtained. Using the terminology described
Equally significant was the observation that the distal hydroxy in Figure 1, we refer to these as proximal/mid, proximal/distal,
group seemingly does not need to be protected in this route: Al- and mid/distal. Not all combinations of preparative steps are
though we did not explicitly search for them, neither the cyclic equally apt, however. For example, use of allylic bromination
urethane nor polymeric self-condensation products were ob- to derivatize the chain adjacent to the olefin is also likely to
served.
modify benzylic protecting groups, so if this combination of
To append an amine unit at the midsection of 1, we used al- procedures and structural elements is to be used, the benzylic
lylic substitution. Again, other methods should ultimately be group would need to be introduced after the bromine, not vice
investigated (e.g., azide opening of an epoxide) but for now we versa. We have explored several synthetic pathways for the
preferred to retain the C–C double bond as a site that could be preparation of diamines.
further modified. The allylic bromide was readily prepared
To make the proximal/distal isomer, we examined two
with NBS and was then converted to the azide with NaN3 (23). routes. First was to use 3 in a Mitsunobu-type reaction with
We sought to avoid isolation and purification of this potentially DPPA, analogous to the preparation of 6, presented earlier. In
dangerous intermediate, so it was used in a one-pot procedure this case the final step was protection of the newly installed
(24) by reaction with PPh3 (the Staudinger reaction), then water amine with the o-nitrobenzenesulfonyl, or Nosyl, group. This
to hydrolyze the iminophosphorane, and finally a reagent for diamine therefore possesses orthogonally protected amino
introducing a protecting group to the newly generated free groups at opposite ends of the carbon chain: The Nosyl group
amine. Allylic reaction of an oleic acid derivative should, of can be removed with nucleophilic reagents (28), and the Boc
course, give two products, one substituted at the 8-position and group with acid. As expected, the Boc group of 3 was unaf-
one at the 11-position. We observed two poorly resolved peaks fected by the conditions needed to append this second amine.
in the HPLC trace of molecule 5, both of which have the same Second was to proceed in the “opposite” direction: The methyl
mass. Conditions to separate these isomers from each other ester 6 was hydrolyzed to the free acid and subjected to Curtius
using column chromatography on silica were sought but not rearrangement conditions with DPPA in t-BuOH. Again, the
found. Carbon NMR shows four olefinic peaks. A further com- NHBoc unit (at C17) was unaltered by these conditions. As in
plication is that diastereomeric mixtures should also be pro- the preparation of 3, a significant amount of a by-product was
duced, since allylic substitution generates a new chiral center. observed, which by analogy we assume to be the isocyanate.
Although the existence of molecule 5 (or any of the other mid-
The proximal/mid isomer was prepared by subjecting the
chain-substituted products) as an inseparable isomeric mixture Boc-protected 3 to NBS substitution followed by the standard
might be problematic for some applications, such as medicinal sequence of reduction (Method A), then protecting with the
chemistry, we believe that this situation should not be a draw- Fmoc group. Again, the two protecting groups are orthogonal,
back in polymeric, materials, or surfactant applications. That responding to acid and base, while the unprotected hydroxy
epoxidized plant oils, for example, exist as “a vast mixture of group remains open to further chemistry. The final diamine is
isomers” (25) has not prevented them from finding many uses the mid/distal isomer, synthesis of which began with allylic-
and being produced in hundreds of thousands of tons a year. substituted 5, which was then phthalimidated using Mitsunobu
This sequence of reactions (formation of an azide, reduction conditions to give 12. Repeatedly, this phthalimidation reac-
with PPh3, hydrolysis with water, and protection) is one we tion worked in high yield and conveniently. The only foresee-
used regularly in the synthesis of these fatty amines. It is far able drawback was that the phthalimide protecting group has a
from optimal, since monitoring of the progress of the reactions reputation for being difficult to remove, so we wished to test a
is difficult. The intermediate iminophosphorane, for example, reported improvement for deprotection (29). A one-pot reduc-
is probably hydrolytically unstable and incompatible with sil- tion with NaBH4 and acid treatment followed by acetyl cap-
ica gel chromatography. Nonetheless, we have continued to use ping worked well to give 13, without altering the methyl ester,
these steps since the likely alternative of reduction of the azide the Boc group, or the olefin. This route looks promising as a
over palladium would also reduce the olefin.
way to introduce an acyl unit that would be more interesting
To replace the 17-hydroxy group with an amino unit, we than CH3CO at the distal nitrogen.
used two variants of the Mitsunobu reaction (26). First, the
Some comment on the choice of protecting groups (PG) in
azide unit was directly appended to the oleate with DPPA (27), these molecules is in order. For these initial experiments, we
followed by the same sequence of PPh3, water, and protection were simply motivated by discovering what works. In the longer
JAOCS, Vol. 83, no. 7 (2006)

POLYFUNCTIONAL FATTY AMINES
627
der, New Syntheses with Oils and Fats as Renewable Raw Ma-
terials for the Chemical Industry, Angew. Chem. Int. Ed.
39:2206–2224 (2000).
term, the pattern of PG adopted will also be determined by the
intended use of the compound. For example, to use a di-
aminocarboxylic acid such as 12 as the diamine component in
the synthesis of a polyamide (in conjunction with a dicarboxylic
co-monomer), the PG on the two amino units should be remov-
able in one step by the same reagent, and the PG on the car-
boxylic acid should be orthogonal. By contrast, to use 12 in a
head-to-tail fashion to prepare a homopolyamide, the carboxylic
acid and one amine should be protected with groups that are re-
moved simultaneously, while the second amine remains orthog-
onal. The steps we have used so far to introduce amines into 1
provide five classes of protecting groups: Boc (removable with
acid), benzyloxycarbonyl (H2), Fmoc (base), phthalimide
(NaBH4/acid), and Nosyl (nucleophile, e.g. thiophenol).
Triamine synthesis. To combine the above steps into the syn-
thesis of a triply substituted derivative, we began with the al-
lylic amine 5. Again, this molecule is a likely starting point,
since the free-radical bromination used to produce it is incom-
patible with several protecting groups. First, the methyl ester
was successfully hydrolyzed. The DPPA-mediated Curtius re-
action was then performed in benzyl alcohol to give benzyloxy-
carbonyl-protected diamine 14, which is a second example of
the proximal/mid class of diamines. It is worth noting that none
of the isocyanate by-product was observed, as it was when t-
BuOH was the solvent during the synthesis of 3; the primary
benzyl alcohol is, as would be expected, much more successful
at attacking the OCN unit than the tertiary alcohol. Removal of
the excess benzyl alcohol was tedious but could be accom-
plished by repeated extraction of an ether solution of the prod-
uct with water. The final step was to replace the distal hydroxy
group with an amino unit, and for this we chose Mitsunobu
conditions, which are mild. The nitrobenzenesulfonamido unit
has several advantages for this step: not only is the NH acidic
enough for use in the Mitsunobu protocol, but also the prepara-
tion of alkyl- or aryl-substituted nitrobenzenesulfonamides is
easy, starting from the sulfonyl chloride and a primary amine.
In this way, a Nosyl-protected secondary amine, where the
added N-substituent could in principle bear a wide variety of
groups, can be introduced. We used the propyl group for ease
and because its proton NMR signature is characteristic but un-
complicated. Triamine 15 was produced in acceptable yield.
This molecule, along with the other polyfunctional fatty
amines, amino alcohols, and amino acids described in this
work, demonstrates the range of value-added compounds that
can be obtained in a few synthetic steps from an abundant
bioderived starting material.
2. Lichtenthaler, F.W., and S. Peters, Carbohydrates as Green Raw
Materials for the Chemical Industry, C.R. Chimie 7:65–90
(2004).
3. Hayes, D.G., Enzyme-Catalyzed Modification of Oilseed Mate-
rials to Produce Eco-Friendly Products, J. Am. Oil Chem. Soc.
81:1077–1103 (2004).
4. Kitamoto, D., J. Isoda, and T. Nakahara, Functions and Poten-
tial Applications of Glycolipid Biosurfactants—From Energy-
Saving Materials to Gene Delivery Carriers, J. Biosci. Bioeng.
94:187–201 (2002).
5. Solaiman, D.K.Y., R.D. Ashby, A. Nuñez, and T.A. Foglia, Pro-
duction of Sophorolipids by Candida bombicola Grown on Soy
Molasses as Substrate, Biotech. Lett. 26:1241–1245 (2004).
6. Guilmanov, V., A. Ballistreri, G. Impallomeni, and R.A. Gross,
Oxygen Transfer Rate and Sophorose Lipid Production by Can-
dida bombicola, Biotechnol. Bioeng. 77:489–494 (2002).
7. Singh, S.K., A.P. Felse, A. Nunez, T.A. Foglia, and R.A. Gross,
Regioselective Enzyme-Catalyzed Synthesis of Sophorolipid
Esters, Amides, and Multifunctional Monomers, J. Org. Chem.
68:5466–5477 (2003).
8. Bisht, K.S., W. Gao, and R.A. Gross, Glycolipids from Candida
bombicola: Polymerization of a 6-O-Acryloyl Sophorolipid De-
rivative, Macromolecules 33:6208–6210 (2000).
9. Baumann, H., M. Buhler, H. Fochem, F. Hirsinger, H. Zoe-
belein, and J. Falbe, Natural Fats and Oils—Renewable Raw
Materials for the Chemical Industry, Angew. Chem. Int. Ed.
Engl. 27:41–62 (1988).
10. Greene, T.W., and P.G.M. Wuts, Protective Groups in Organic
Synthesis, 2nd edn., John Wiley & Sons, New York, 1991.
11. Chevalier, Y., New Surfactants: New Chemical Functions and
Molecular Architectures, Curr. Opin. Colloid Interface Sci.
7:3–11 (2002).
12. Chen, W., F.J. Schoen, and R.J. Levy, Mechanism of Efficacy
of 2-Amino Oleic Acid for Inhibition of Calcification of Glu-
taraldehyde-Pretreated Porcine Bioprosthetic Heart Valves, Cir-
culation 90:323–329 (1994).
13. Dagan, A., C. Wang, E. Fibach, and S. Gatt, Synthetic, Non-
Natural Sphingolipid Analogs Inhibit the Biosynthesis of Cellu-
lar Sphingolipids, Elevate Ceramide and Induce Apoptotic Cell
Death, Biochim. Biophys. Acta 1633:161–169 (2003).
14. Muller, G., H. Jordan, C. Jung, H. Kleine, and S. Petry, Analy-
sis of Lipolysis in Adipocytes Using a Fluorescent Fatty Acid
Derivative, Biochimie 85:1245–1256 (2003).
15. Gehring, L., D. Haase, K. Habben, C. Kerkhoff, H.H. Meyer,
and V. Kaever, Synthesis of an Unsaturated Fatty Acid Ana-
logue (18-(4′-azido-2′-hydroxybenzoylamino)-oleic acid) and
Its Interaction with Lysophosphatidylcholine: Acyl-CoA-O-
acyltransferase, J. Lipid Res. 39:1118–1126 (1998).
16. Shankar, S.S., H. Joshi, R. Pasricha, N.R. Pavaskar, A.B. Man-
dale, and M. Sastry, A Low-Temperature, Soft Chemistry
Method for the Synthesis of Zirconia Nanoparticles in Ther-
mally Evaporated Fatty Amine Thin Films, J. Colloid Interface
Sci. 269:126–130 (2004).
17. Gole, A., V. Dixit, N. Lala, S.R. Sainkar, A. Pant, and M. Sas-
try, Patterned Assembly of Yarrowia lipolytica Yeast Cells onto
Thermally Evaporated Octadecylamine Films, Colloids Surf. B:
Biointerfaces 25:363–368 (2002).
18. Gole, A., S. Vyas, S.R. Sainkar, A. Lachke, and M. Sastry, En-
hanced Temperature and pH Stability of Fatty Amine-Endoglu-
canase Composites: Fabrication, Substrate Protection, and Bio-
logical Activity, Langmuir 17:5964–5970 (2001).
ACKNOWLEDGMENTS
We gratefully acknowledge the technical assistance of Bun-Hong
Lai (chromatography and fermentations), Marshall Reed (microbiol-
ogy), and Kerby Jones (MS and HPLC).
REFERENCES
19. Xu, J., Z. Liu, S.Z. Erhan, and C.J. Carriere, Cross-linkers Con-
trol the Viscoelastic Properties of Soybean Oil-Based Biomate-
rials, J. Am. Oil Chem. Soc. 81:813–816 (2004).
1. Biermann, U., W. Friedt, S. Lang, W. Luhs, G. Machmuller, J.O.
Metzger, M. Rusch gen. Klaas, H.J. Schafer, and M.P. Schnei-
JAOCS, Vol. 83, no. 7 (2006)

628
J.A. ZERKOWSKI AND D.K.Y. SOLAIMAN
20. Ochoa, N., F. Moran, N. Pebere, and B. Tribollet, Influence of 26. Mitsunobu, O., The Use of Diethyl Azodicarboxylate and Triph-
Flow on the Corrosion Inhibition of Carbon Steel by Fatty
Amines in Association with Phosphonocarboxylic Acid Salts,
Corrosion Sci. 47: 593–604 (2005).
enylphosphine in Synthesis and Transformation of Natural Prod-
ucts, Synthesis:1–28 (1981).
27. Lal, B., B.N. Pramanik, M.S. Manhas, and A.K. Bose,
Diphenylphosphorylazide, a Novel Reagent for the Stereospe-
cific Synthesis of Azides from Alcohols, Tetrahedron Lett.
18:1977–1980 (1977).
21. Cermak, S., and T. Isbell, Estolides—The Next Biobased Func-
tional Fluid, inform 15: 515–517 (2004).
22. Shioiri, T., K. Ninomiya, and S. Yamada, Diphenylphosphory-
lazide. A New Convenient Reagent for a Modified Curtius Re- 28. Fukuyama, T., C.-K. Jow, and M. Cheung, 2- and 4-Nitroben-
action and for the Peptide Synthesis, J. Am. Chem. Soc.
94:6203–6205 (1972).
23. Koziara, A., K. Osowska-Pacewicka, S. Zawadzki, and A.
zenesulfonamides: Exceptionally Versatile Means for Prepara-
tion of Secondary Amines and Protection of Amines, Ibid.
36:6373–6374 (1995).
Zwierzak, One-Pot Transformation of Alkyl Bromides into Pri- 29. Dasgupta, F., and P.J. Garegg, Reductive Dephthalimidation: A
mary Amines via the Staudinger Reaction, Synthesis:202–204
(1985).
Mild and Efficient Method for the N-Phthalimido Deprotection
During Oligosaccharide Synthesis, J. Carbohydr. Chem,
7:701–708 (1988).
24. Vaultier, M., N. Knouzi, and R. Carrie, Reduction d’Azides en
Amines Primaires par une Methode Generale Utilisant la Reac-
tion de Staudinger, Tetrahedron Lett. 24:763–764 (1983).
25. Rusch gen. Klaas, M., and S. Warwel, Complete and Partial
Epoxidation of Plant Oils by Lipase-Catalyzed Perhydrolysis,
Ind. Crops Prod. 9:125–132 (1999).
[Received December 2, 2005; accepted April 8, 2006]
JAOCS, Vol. 83, no. 7 (2006)