Journal of the American Chemical Society
cysteine-modified lysolipids would react by NCL with
palmitoyl-CoA 1 generated in situ by cgFAS I. As a test, we
available palmitoyl-CoA, forming phospholipid 3 (Figure S4B).
Briefly, we treated lysophospholipid 2 (1 mM) with palmitoyl-
CoA (1 mM) in 10 mM phosphate (Na HPO /NaH PO )
2
4
2
4
buffer, pH 7.4 containing 10 mM TCEP at 37 °C.
Phospholipid formation was followed using HPLC-MS
combined with evaporative light-scattering detection (ELSD)
(
determined that 820 μM of phospholipid 3 was formed after
3
h, corresponding to a yield of 82%.
In previous work, we have observed that amphiphilic species
are preferentially acylated by amphiphilic reactants, likely
13,16,23
promoted by coassembly in micelles or membranes.
To
better understand the role self-assembly plays in the formation
of the phospholipid product, we investigated the reactivity of
lysophospholipid 2 with nonamphiphilic small-chain thioesters.
Malonyl- and acetyl-CoA were selected as reactive thioester
partners with 2. Although both substrates contain a reactive
thioester moiety that can react by NCL with cysteine-modified
lysophospholipid 2, the absence of a long-chain hydrophobic
tail precludes assembly into structures such as micelles.
Therefore, we anticipated a difference in their reactivity with
Figure 3. Characterization of phospholipid 3 vesicular structures. (A)
Phase-contrast microscopy image of membrane-bound vesicles
resulting from the self-assembly of 3. Scale bar denotes 5 μm. (B)
Fluorescence microscopy image of vesicles formed by hydration of a
thin film of 3. Membranes were stained with 0.1 mol % BODIPY-FL
DHPE. Scale bar denotes 5 μm. (C) TEM image of negatively stained
vesicles of 3. Scale bar denotes 100 nm. (D) Fluorescence microscope
image demonstrating the encapsulation of HPTS in vesicles of 3. Scale
bar denotes 5 μm.
2
in comparison to the previously tested palmitoyl-CoA 1. As
expected, when we attempted to react 2 with malonyl- or
were unable to detect product formation (Figure S5).
To determine the ability of noncanonical phospholipid 3 to
form membrane-bound vesicles, microscopy studies were
performed. Neither palmitoyl-CoA 1 nor lysophospholipid 2
formed membranes in aqueous solution. Phospholipid 3
readily formed membrane-bound assemblies when hydrated
3. After 30 min of reaction, small vesicular structures were
(Figure 3). Lipid vesicles were initially identified by phase-
contrast (Figure 3A) and fluorescence microscopy using the
S6A). Under these conditions, vesicles of 1−10 μm diameter
were observed after hydration and tumbling of 3 in phosphate
buffer, pH 7.4 at 37 °C for 1 h. Transmission electron
microscopy (TEM) also corroborated the formation of
vesicular structures (Figure 3C). The encapsulation ability of
the phospholipid vesicles was demonstrated by hydrating a
thin lipid film of 3 in the presence of 8-hydroxypyrene-1,3,6-
trisulfonic acid (HPTS), a highly polar fluorescent dye,
followed by removal of excess dye by spin-filtration and vesicle
We next investigated the one-pot chemoenzymatic for-
mation of membranes in the presence of biologically relevant
24
cell membrane components, including cholesterol, ionic
small molecules such as guanidine hydrochloride
25−27
28
(GuHCl),
and short-chain alkanols such as decanol.
Natural cell membranes are heterogeneous bilayers composed
of multiple phospholipids, as well as other lipid species such as
29,30
cholesterol.
Since the lipid profile of our vesicles is
homogeneous, we wanted to explore the effect of incorporating
biologically relevant additives into our system. Therefore, we
added cholesterol (400 μM), GuHCl (400 μM) and 1-decanol
(400 μM) to the one-pot in situ chemoenzymatic reaction
forming phospholipid 3. It has been suggested that
stoichiometric addition of cholesterol, GuHCl, and decanol
leads to the curvature stabilization and fusion of fatty acid
Having characterized the individual enzymatic and chemical
reactions, we next explored combining enzymatic palmitoyl-
CoA 1 synthesis with chemical phospholipid 3 synthesis in a
one-pot reaction (Figure 4). Briefly, we added lysophospho-
lipid 2 (400 μM) to 10 mM phosphate (Na HPO /NaH PO )
27
vesicles. We expected similar interactions of such additives
with our phospholipid vesicles. We observed that the additives
did not perturb the formation of phospholipid 3 membranes
and led to the formation of larger, more stable vesicles. Vesicles
were stable over 48 h at 37 °C, as observed by fluorescence
microscopy using BODIPY-FL DHPE (Figure 4D).
2
4
2
4
buffer, pH 7.4 containing cgFAS I (1 μM), acetyl-CoA (1
mM), malonyl-CoA (1 mM), and NADPH (10 mM) along
with TCEP (10 mM) at 37 °C. Phospholipid formation was
followed using HPLC-MS-ELSD measurements. Optimization
of the reaction conditions enabled rapid coupling between
cgFAS I generated 1 and 2. The one-pot reaction afforded the
corresponding phospholipid 3 as the prominent product within
In summary, we have developed a chemoenzymatic route to
synthesize noncanonical phospholipids from water-soluble
precursors. Given our approach, there should be flexibility to
diversify the lipid species generated in the reaction. Even
though we utilized cgFAS I to selectively produce palmitoyl-
3
0 min (Figure 4A). All of lysophospholipid 2 was consumed
in less than 4 h (Figure 4 B) to afford 367 μM of phospholipid
8
535
J. Am. Chem. Soc. 2021, 143, 8533−8537