Effect of Headgroups on Small-Ion Permeability across Archaea-Inspired Tetraether Lipid Membranes

Article abstract of DOI:10.1002/chem.201601326

This paper examines the effects of four different polar headgroups on small-ion membrane permeability from liposomes comprised of Archaea-inspired glycerolmonoalkyl glycerol tetraether (GMGT) lipids. We found that the membrane-leakage rate across GMGT lipid membranes varied by a factor of ≤1.6 as a function of headgroup structure. However, the leakage rates of small ions across membranes comprised of commercial bilayer-forming 1-palmitoyl-2-oleoyl-sn-glycerol (PO) lipids varied by as much as 32-fold within the same series of headgroups. These results demonstrate that membrane leakage from GMGT lipids is less influenced by headgroup structure, making it possible to tailor the structure of the polar headgroups on GMGT lipids while retaining predictable leakage properties of membranes comprised of these tethered lipids. Extremophiles require robust membranes with low ion permeability to survive in harsh conditions. These Archaea organisms achieve low membrane permeability, in part, by generating lipids with unusual structural features such as tethering of their lipid tails. A systematic study probing the effect of polar headgroups in synthetic Archaea-inspired lipids showed little dependence of membrane leakage rates of small ions as a function of headgroup structure.

Full text of DOI:10.1002/chem.201601326

DOI: 10.1002/chem.201601326
Communication
&
Synthetic Lipids
Effect of Headgroups on Small-Ion Permeability across Archaea-
Inspired Tetraether Lipid Membranes
Takaoki Koyanagi,[a] Geoffray Leriche,[a] Alvin Yep,[a] David Onofrei,[b] Gregory P. Holland,[b]
Michael Mayer,[c] and Jerry Yang*[a]
Abstract: This paper examines the effects of four different
polar headgroups on small-ion membrane permeability
from liposomes comprised of Archaea-inspired glycerol-
monoalkyl glycerol tetraether (GMGT) lipids. We found
that the membrane-leakage rate across GMGT lipid mem-
branes varied by a factor of ꢀ1.6 as a function of head-
group structure. However, the leakage rates of small ions
across membranes comprised of commercial bilayer-form-
ing 1-palmitoyl-2-oleoyl-sn-glycerol (PO) lipids varied by as
much as 32-fold within the same series of headgroups.
These results demonstrate that membrane leakage from
GMGT lipids is less influenced by headgroup structure,
making it possible to tailor the structure of the polar
headgroups on GMGT lipids while retaining predictable
leakage properties of membranes comprised of these
tethered lipids.
Figure 1. Examples of lipid headgroups found in Archaea.
Figure 1 shows the structures of four common polar lipid
headgroups that are found naturally in Archaea.[5,9] The deli-
cate balance in generating a precise mixture of lipids com-
prised of these different polar headgroups is thought to be es-
sential in maintaining a viable membrane under harsh condi-
tions.[10]
Eukaryotes and prokaryotes often respond to environmental
stress by modifying the lipid composition of their mem-
branes.[1–3] Archaea, a unique classification of prokaryotes,
thrive in extreme environments, but can maintain their mem-
brane integrity under environmental stress presumably due to
the unique structural features integrated in their lipids.[4–6]
These structural features include incorporation of: 1) phytanyl
(isoprene) hydrocarbon chains, 2) tethered lipid chains, and
3) ether linkages between the glycerol backbone of the head-
group and the hydrophobic chains (Figure 1). Additionally, the
mixtures of polar lipid headgroups in Archaea membranes are
dependent on specific growth conditions.[4] For instance, under
acidic conditions, the lipid composition of Archaea will include
a high fraction of headgroups that can facilitate hydrogen
bonding between the headgroups on adjacent lipids.[7,8]
We recently reported a systematic study of the dependence
of membrane leakage on the presence of small rings incorpo-
rated into the tethered, membrane-spanning segments of syn-
thetic glycerolmonoalkyl glycerol tetraether (GMGT) lipids.[11]
We found that incorporating a single transmembrane tether
(with or without rings), phytanyl hydrophobic side chains, and
ether linkages between the glycerol backbone made it possi-
ble to generate membranes with approximately two orders of
magnitude slower rate of leakage of small ions (H+, OHÀ, Na+,
ClÀ, buffer ions) at room temperature compared to mem-
branes generated from commercial EggPC lipids. These results
demonstrate that altering the structure of the hydrophobic
portion of lipids can dramatically affect membrane leakage.
Additionally, previous computational[12] and experimental stud-
ies[13] suggest that lipid headgroups affect membrane permea-
tion of water, but have little effect on permeability of small or-
ganic molecules across membranes.[14] However, to our knowl-
edge no systematic study on the effects of headgroups on
small-ion permeability has been reported.
[a] T. Koyanagi, Dr. G. Leriche, A. Yep, Prof. J. Yang
Department of Chemistry and Biochemistry
University of California San Diego
La Jolla, CA 92093-0358 (USA)
E-mail: jerryyang@ucsd.edu
[b] D. Onofrei, Prof. G. P. Holland
Department of Chemistry and Biochemistry
San Diego State University, San Diego, CA 92182-1030 (USA)
[c] Prof. M. Mayer
Herein, we describe the synthesis of a series of Archaea-in-
spired GMGT lipid analogs containing four different head-
groups found in nature (phosphocholine, PC; phosphoethanol-
amine, PE; phosphatic acid, PA ; phosphoglycerol, PG; Figure 1)
Adolphe Merkle Institute, University of Fribourg
Chemin des Verdiers 4, 1700 Fribourg (Switzerland)
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under http://dx.doi.org/10.1002/open.201402097.
Chem. Eur. J. 2016, 22, 8074 – 8077
8074
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
shown previously with standard diacylphospholipids with PE
headgroups owing to relatively small headgroup size).[17] How-
ever, because we were able to form liposomes comprised of
1:1 mixtures between PC lipids and lipids carrying PA, PG, or
PE headgroups, we used these mixtures to evaluate the effects
of the headgroup on the relative membrane permeability
(Figure 2). Differential scanning calorimetry (DSC) measure-
Scheme 1. Synthesis of GMGT lipid derivatives.
Figure 2. Observed initial rate of pH equilibration from liposomes formed
from GMGT or PO series of lipids. Comparison of the observed initial rates of
decreased carboxyfluorescein (CF) fluorescence intensity (monitored at l(Ex/
Em)=485/517 nm) from CF-encapsulated liposomes comprised of pure
GMGTPC lipid or 1:1 mixtures of GMGTPC with GMGTPG, GMGTPE, or
GMGTPA (the inset represents a zoomed-in graph of the observed initial
rates of leakage from the GMGT lipids); and pure POPC lipid or 1:1 mixtures
of POPC with POPG, POPE, or POPA. Statistical significance was determined
using a paired Student t-test. *, **, *** indicate a p-value of <0.1, 0.01,
0.001, respectively, relative to the kobsd of the analogous PC lipid.
and investigate the effect of headgroups on small-ion mem-
brane permeability across membranes comprised of these
lipids. We prepared GMGT lipid analogs from the common diol
9 (Scheme 1), which comprised a 28-carbon aliphatic chain
and two untethered phytanyl groups attached to a glycerol
backbone through ether linkages. The phytanyl chains and
ether groups were maintained throughout this series of lipids
because these features were previously found to increase
chemical stability and to reduce small-ion permeation.[11] Brief-
ly, diol 9 was synthesized in seven steps from 1,12-dodecan-
diol, using a double Wittig reaction to form the key intermedi-
ate 6. We then generated 1,28-dibromooctacosane 7 from 6
through a hydrogenation/bromination sequence, and reacted
this dibromide with glycerol scaffold 8, followed by hydroge-
nation to afford diol 9. Finally, the syntheses of the different
GMGT analogs were carried out using derivatives of benzyldii-
sopropyl phosphoramidite as phosphorylating agents.[15,16] The
phosphite-triester products were then oxidized to the corre-
sponding phosphate-triesters, and the protecting groups were
removed by hydrogenation to afford lipids GMGTPA, GMGTPG,
and GMGTPE in good yields as mixtures of stereoisomers.
GMGTPC was generated by reaction of the diol 9 with 2-bro-
moethyl dichlorophosphate followed by nucleophilic displace-
ment of the bromine with trimethylamine, as described previ-
ously[11] (see Figures S1–S3 in the Supporting Information for
details of the syntheses of GMGT lipids).
ments show that suspensions of pure GMGTPC, GMGTPA,
GMGTPG, or GMGTPE lipids in water maintained a liquid phase
without an observed phase transition between 5 and 658C
(see Figure S5 in the Supporting Information). Additionally, we
did not observe any morphology changes by dynamic light
scattering (DLS) within the timeframe of the leakage assays
(see Figure S4 in the Supporting Information), suggesting that
liposome fusion or aggregation did not significantly contribute
to the observed rate of small-ion permeability.
We evaluated the relative permeability to small ions of
membranes comprised of lipids carrying various headgroups
using a pH-equilibration assay described previously.[11,18] Briefly,
this method consists of an encapsulation of carboxyfluorescein
(CF) inside liposomes with an intraliposomal pH of 7.2. The
liposomes are then incubated in a buffered solution at pH 5.8,
and the change in fluorescence intensity of CF is monitored
over time as the pH between the inside and outside of the lip-
osomes reaches equilibrium. We chose to use CF as a fluores-
cent reporter in this assay, in part, because we previously
showed that CF does not leak from liposomes comprised of
tetraether or diacyl lipids under these experimental condi-
tions.[11] While we expect proton leakage to be the dominant
contributor to the rate of pH equilibration,[19] other ions pres-
We next assessed whether we could form liposomes com-
prised of pure GMGT lipids using a previously described proce-
dure.[11] Though pure GMGTPC, GMGTPA, and GMGTPG readily
formed stable liposomes of ꢁ120 nm diameter, (see Figure S4
in the Supporting Information), we were unable to form lipo-
somes from pure GMGTPE lipids (similar results have been
Chem. Eur. J. 2016, 22, 8074 – 8077
www.chemeurj.org
8075
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
ent in the assay medium (e.g., OHÀ, Na+, ClÀ) could also con-
tribute to the measured rate of leakage. The measured initial
rates of pH equilibration, therefore, can be considered more
generally as an estimate for membrane permeability to small
ions rather than as an estimate of permeability to protons
only. We also avoided addition of divalent cations in the buffer
to prevent negatively charged lipid polar headgroups from in-
ducing aggregation/fusion of liposomes.[20] To minimize error
in our leakage-rate measurements from photobleaching and
evaporation, we used initial rates (i.e., the first 15% change in
CF fluorescence) to evaluate the rate of pH equilibration at
room temperature (see Figure S6 in the Supporting Informa-
tion). While this CF-based assay made it possible to estimate
the relative effects of different headgroups on leakage of small
ions from liposomes, other assays may offer additional advan-
tages for estimating absolute permeability constants of specific
ions across various lipid membranes.[19,21–23]
GMGTPA (which did not affect membrane leakage compared
to GMGTPC lipids), the rate of leakage from POPE- and POPA-
containing membranes was a factor of 0.5 and 0.07 times
slower, respectively, compared to membranes formed from
pure POPC lipids. Surprisingly, liposomes comprised of a 1:1
mixture of POPC and POPA exhibited the lowest observed rate
of small-ion leakage among all lipid mixtures tested under the
experimental conditions used here. Figure 3 highlights that, in
contrast to GMGT lipids, PO lipids show a strong dependence
of headgroups on small-ion permeability.
Figure 2 shows that there was no statistically significant dif-
ference between the observed rate of leakage of small ions
from liposomes comprised of 1:1 mixtures of GMGTPC lipids
with lipids containing PE or PA headgroups compared to lipo-
somes comprised of pure GMGTPC lipids. Liposomes formed
from a 1:1 mixture of GMGTPC:GMGTPG lipids, however, exhib-
ited a 1.5-fold increased rate of leakage compared to pure
GMGTPC liposomes (the numerical values of the observed first-
order rates of membrane leakage are given in Table S1 in the
Supporting Information). Overall, the rate of small-ion leakage
from any two of the GMGT lipids differed by a factor of ꢀ1.6
as a function of headgroup.
Figure 3. Relative effects of headgroups on small-ion membrane leakage.
a) Graph of the relative variation of leakage rate from membranes comprised
of 1:1 mixtures of PC with PA, PE, or PG lipids compared to membranes
formed from pure PC lipids. Data represents the percent deviation of ob-
served initial rates of membrane leakage compared to the observed initial
rate of leakage from pure POPC or pure GMGTPC lipid membranes (zero per-
cent). b) Graph showing the relative leakage rate of membranes comprised
of 1:1 mixtures of PC with PA, PE, or PG lipids relative to the observed initial
rate of leakage from membranes comprised of pure POPC or pure GMGTPC
lipids (normalized to 1).
To examine how these results compared to the effect of
headgroups on standard bilayer-forming diacylphospholipids,
we also examined leakage of small ions from 1-palmitoyL-2-
oleoyL-sn-glycerol (PO) lipids. Because POPA and POPE have
phase-transition temperatures near room temperature, we
generated liposomes by mixing POPC lipids 1:1 with POPG,
POPE, or POPA and compared their relative rate of leakage to
liposomes comprised of pure POPC lipids (i.e., the analogous
procedure we used to measure relative leakage from the
GMGT lipids with the same four headgroups). Again, DSC
measurements showed that liposomes comprised of 1:1 mix-
tures of POPC with POPA, POPE, or POPG maintained a liquid
phase at room temperature and did not exhibit a phase transi-
tion between 5 and 658C (see Figure S5 in the Supporting In-
formation), and POPC has a known phase transition at
À28C.[24] DLS measurements also confirmed that lipid fusion or
aggregation is not expected to significantly contribute to the
observed rate of membrane leakage (see Figure S4 in the Sup-
porting Information).
The reduced permeation of small ions observed for lipo-
somes containing POPC mixed with POPA or POPE (compared
to pure POPC liposomes) could arise from increased intermo-
lecular hydrogen bonding between headgroups of neighbor-
ing lipids. Such intermolecular hydrogen bonding may lead to
exclusion of water molecules near the membrane surface and
increased membrane packing, as suggested through X-ray dif-
fraction,[25] FT-IR,[26] and computation[27] studies. For the GMGT
lipids, we previously showed that leakage of small ions from
tethered lipids was significantly reduced compared to bilayer-
forming lipids,[11] presumably as a result of favorable lipid pack-
ing of hydrocarbon chains in neighboring lipids within the
membrane. Such inherently tight membrane packing in GMGT
lipids, thus, may not be as influenced by membrane surface ef-
fects induced by the presence of PE or PA headgroups. On the
other hand, for lipids with PG headgroups, the presence of
multiple hydroxyl groups may lead to an increase in the
number of water molecules in between lipid headgroups near
the membrane surface, which could cause a decrease in lipid
packing and an increase in membrane leakage.[13] We expect
such an effect on leakage by PG headgroups would be more
pronounced in PO lipids (which presumably are inherently
more loosely packed) compared to GMGT lipids. The results
from permeability experiments (Figure 2) support such a hy-
pothesis.
In the case of the PO series of lipids, we found significant ef-
fects of headgroups on the observed rates of small-ion mem-
brane permeation (Figure 2). Similar to the series of GMGT
lipids, the PG headgroup increased leakage compared to the
PC headgroup. However, in the PO series, the PG headgroup
had a much larger effect on membrane leakage (increased by
a factor of 2.5) compared to liposomal membranes from lipids
with PC headgroups only (see Table S1 in the Supporting Infor-
mation). Furthermore, in contrast to the case with GMGTPE or
Chem. Eur. J. 2016, 22, 8074 – 8077
www.chemeurj.org
8076
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[5] Y. Koga, H. Morii, Microbiol. Mol. Biol. Rev. 2007, 71, 97–120.
We have, thus, presented a systematic study of the effects of
lipid headgroups on the leakage properties of membranes
comprised of GMGT or PO lipids. Liposomes containing PO
lipids exhibited a strong dependence on headgroups for mem-
brane leakage. Such effects on small-ion permeability from
headgroups may limit the utility of PO (and possibly other bi-
layer-forming) lipids in applications and studies where main-
taining a consistent membrane permeability is important
across various headgroups. On the other hand, membrane
leakage from liposomes comprised of synthetic GMGT lipids
was not strongly influenced by headgroups. These findings
suggest that GMGT lipids may offer greater flexibility for tailor-
ing the functionality presented in the lipid headgroups, with-
out significantly compromising membrane permeability. Such
versatility in the design of lipid headgroups may open up op-
portunities to use GMGT lipids in a range of applications in-
cluding the incorporation of receptor-targeting molecules on
lipids for development of liposomal drug-delivery systems[28] or
the incorporation of ligands as well as charged headgroups to
attract specific analytes or binding partners to the membrane
surface for biophysical studies.[29]
[6] D. L. Valentine, Nat. Rev. Microbiol. 2007, 5, 316–323.
[7] H. Shimada, N. Nemoto, Y. Shida, T. Oshima, A. Yamagishi, J. Bacteriol.
2008, 190, 5404–5411.
[8] P. L. G. Chong, Chem. Phys. Lipids 2010, 163, 253–265.
[9] A. Pearson, A. E. Ingalls, Annu. Rev. Earth Planet. Sci. 2013, 41, 359–384.
[10] K. Yamauchi, M. Kinoshita, Prog. Polym. Sci. 1993, 18, 763–804.
[11] T. Koyanagi, G. Leriche, D. Onofrei, G. P. Holland, M. Mayer, J. Yang,
Angew. Chem. Int. Ed. 2016, 55, 1890–1893; Angew. Chem. 2016, 128,
1922–1925.
[12] J. C. Mathai, S. Tristram-Nagle, J. F. Nagle, M. L. Zeidel, J. Gen. Physiol.
2008, 131, 69–76.
[13] M. Jansen, A. Blume, Biophys. J. 1995, 68, 997–1008.
[14] M. J. Spooner, P. A. Gale, Chem. Commun. 2015, 51, 4883–4886.
[15] M. Iwashita, K. Makide, T. Nonomura, Y. Misumi, Y. Otani, M. Ishida, R. Ta-
guchi, M. Tsujimoto, J. Aoki, H. Arai, J. Med. Chem. 2009, 52, 5837–5863.
[16] W. Huang, W. Sun, Z. Song, Y. Yu, X. Chen, Q. Zhang, Org. Biomol. Chem.
2012, 10, 5197–5201.
[17] M. Sipai Altaf Bhai, Y. Vandana, Y. Mamatha, V. V. Prasanth, J. Pharm. Sci.
Innov. 2012, 1, 13–21.
[18] K. Arakawa, T. Eguchi, K. Kakinuma, Bull. Chem. Soc. Jpn. 2001, 74, 347–
356.
[19] S. Paula, G. Volkov, N. Van Hoek, T. H. Haines, D. W. Deamer, Biophys. J.
1996, 70, 339–348.
[20] N. Düzgünes¸, J. Wilschut, R. Fraley, D. Papahadjopoulos, Biochim. Bio-
phys. Acta Biomembr. 1981, 642, 182–195.
[21] N. Sakai, S. Matile, Chirality 2003, 15, 766–771.
[22] A. V. Jentzsch, D. Emery, J. Mareda, S. K. Nayak, P. Metrangolo, G. Res-
nati, N. Sakai, S. Matile, Nat. Commun. 2012, 3, 905.
[23] Z. Huang, F. C. Szoka, J. Am. Chem. Soc. 2008, 130, 15702–15712.
[24] P. J. Davis, B. D. Fleming, K. P. Coolbear, K. M. Keough, Biochemistry
1981, 20, 3633–3636.
[25] J. M. Boggs, Biochim. Biophys. Acta 1987, 906, 353–404.
[26] W. Hübner, A. Blume, Chem. Phys. Lipids 1998, 96, 99–123.
[27] K. Murzyn, T. Róg, M. Pasenkiewicz-Gierula, Biophys. J. 2005, 88, 1091–
1103.
Acknowledgements
This research was supported by the Air Force Office of Scientif-
ic Research (FA9550-12-1-0435, FA9550-16-1-0151, and FA9550-
14-1-0014).
[28] T. M. Allen, P. R. Cullis, Adv. Drug Delivery Rev. 2013, 65, 36–48.
[29] E. C. Yusko, J. M. Johnson, S. Majd, P. Prangkio, R. C. Rollings, J. Li, J.
Yang, M. Mayer, Nat. Nanotechnol. 2011, 6, 253–260.
Keywords: archaea
synthetic lipid
· biomimetic · lipids · membranes ·
[1] S. Jain, A. Caforio, A. J. M. Driessen, Front. Microbiol. 2014, 5, 641.
[2] A. Gliozzi, A. Relini, P. L.-G. Chong, J. Membr. Sci. 2002, 206, 131–147.
[3] A.-S. Andersson, J. Biol. Chem. 1996, 271, 6801–6809.
[4] P. L.-G. Chong, U. Ayesa, V. P. Daswani, E. C. Hur, Archaea 2012, 2012, 1–
11.
Received: March 19, 2016
Published online on May 3, 2016
Chem. Eur. J. 2016, 22, 8074 – 8077
www.chemeurj.org
8077
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim