Spectroscopic evidence for the unusual stereochemical configuration of an endosome-specific lipid

Article abstract of DOI:10.1002/anie.201106470

At a glance: The stereochemical configuration of the diglycerophosphate backbone of the endosome-specific lipid bis(monoacylglycero)phosphate (BMP, see picture) was determined by ¹HNMR spectroscopy. Enantiomeric discrimination was facilitated b

Full text of DOI:10.1002/anie.201106470

Angewandte
Chemie
DOI: 10.1002/anie.201106470
Stereochemical Elucidation
Spectroscopic Evidence for the Unusual Stereochemical Configuration
of an Endosome-Specific Lipid**
Hui-Hui Tan, Asami Makino, Kumar Sudesh, Peter Greimel,* and Toshihide Kobayashi*
Glycerophospholipds are essential components of cell mem-
branes from archaea to bacteria and mammals. These lipids
feature either an sn-3 or sn-1 glycerophosphate (GP) back-
bone. In archaea, glycerophospholipids are exclusively phos-
phorylated at the sn-1 position. In contrast, in bacteria and
eukaryotic cells phosphorylation is restricted to the sn-3
position. It has been proposed that chiral discrimination
caused the phase separation of sn-1- and sn-3-phosphorylated
lipids, leading up to the segregation of archaea and bacteria.[1]
However, the mammalian lipid bis(monoacylglycero)phos-
phate (1, BMP) has been reported to be the only exception,
featuring an sn-1 GP backbone.[2]
Scheme 1. Strong alkaline degradation of natural BMP during bio-
chemical analysis.[2]
BMP, also known as lysobisphosphatidic acid (LBPA), is a
minor constituent of most animal tissues.[3] Despite the low
overall BMP content of less than one percent of the total
phospholipids fraction, BMP is highly enriched in specific
membrane domains of late endosomes (LEs).[4,5] In LEs
endocytosed lipids and proteins, for example, are either
recycled or destined for degradation. BMP has been proposed
to play an important role in structural and functional aspects
of LEs.[6]
BMP is a structural isomer of phosphatidylglycerol,
featuring two glycerol subunits linked by a phosphodiester
group. The stereochemical configuration of the diglycero-
phosphate (DGP) backbone of BMP was biochemically
analyzed (Scheme 1).[2] During strong alkaline hydrolysis of
the phosphate diester group the cyclic intermediates 2a and
2b were formed, leading to achiral sn-2 GP 3a as the main
product devoid of any stereochemical information. The ratio
of 3a to the mixture of 3b and 3c was determined by gas–
liquid chromatography. The ratio of 3b to 3c was determined
by selective digestion of 3c with glycerol-3-phosphate dehy-
drogenase. Owing to the harsh cleavage conditions and the
indirect detection, Brotherus et al.[2] could not exclude the
presence of up to 20% sn-2 or sn-3 phosphorylation in natural
BMP.
To date, no spectroscopic evidence of the stereochemical
configuration of the DGP backbone of BMP has been
reported. Despite the presence of two stereogenic centers in
the DGP backbone, direct NMR spectroscopic identification
of the stereochemical configuration has not been conclusive.
We envisaged the use of d-camphor ketals as NMR shift
reagents to directly determine the stereochemical configu-
ration of the BMP backbone. For comparison, we synthesized
all three DGP backbone analogues of BMP.
Previous synthetic approaches to BMP employed
phosphorus(III)-based chemistry in combination with a
protecting group at the phosphate moiety.[7] We envisaged
the utilization of classic phosphorus(V)-based chemistry to
furnish symmetrical DGPs
7 and 10 (Scheme 2 and
Scheme 3). In contrast, access to meso-DGP 17 (Scheme 5)
was pictured utilizing an H-phosphonate monoester as a key
intermediate. While PV-based chemistry allowed a short
synthetic route to 7 and 10, the PIII-based chemistry provided
the necessary control for the selective introduction of each
[*] H.-H. Tan, Dr. A. Makino, Dr. P. Greimel, Dr. T. Kobayashi
Lipid Biology Laboratory, ASI RIKEN
351-0198 (Japan)
E-mail: petergreimel@riken.jp
kobayasi@riken.jp
H.-H. Tan, Dr. K. Sudesh, Dr. P. Greimel, Dr. T. Kobayashi
School of Biological Sciences, Universiti Sains Malaysia
11800 (Malaysia)
[**] This work was supported by the Lipid Dynamics Program of RIKEN,
Naito Foundation (T.K.) and KAKENHI 22390018 (T.K.) & 21790105
(A.M.) from MEXT. H.H.T. thanks the USM Fellowship & RIKEN IPA
program for financial support. TOC micrograph by M. Murate.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201106470.
Scheme 2. Synthesis of sn-1,1’ DGP 7. TFA=trifluoroacetic acid,
pTSA=p-toluenesulfonic acid.
Angew. Chem. Int. Ed. 2012, 51, 533 –535
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
533

.
Angewandte
Communications
glycerol enantiomer to furnish 17. Moreover, we omitted the
introduction of a protecting group at the phosphate moiety to
further reduce the number of synthetic steps.
In light of these results, we adopted an alternative
approach (Scheme 5) to 17. In contrast to isopropylidene-
protected 8, diphenyl phosphite treatment of benzyl-pro-
The symmetrical sn-1,1’ DGP (7) was prepared from 2,3-
O-isopropylidene-sn-glycerol (4, Scheme 2). Treatment of
alcohol 4 with phosphoryl chloride gave the desired phos-
phate diester 5. Transketalization of this intermediate in the
presence of d-camphor dimethyl ketal (6) afforded d-
camphor bisketal 7. Even a large excess of 6 and prolonged
reaction times of up to seven days resulted in very low yields.
The yield was significantly increased by utilizing a two-step,
one-pot sequence. First, the isopropylidene protecting group
was removed with neat TFA, and the residue was treated with
6 to form desired sn-1,1’ DGP (7). The symmetrical sn-3,3’
DGP (10) was prepared by the same route (Scheme 3),
starting from 1,2-O-isopropylidene-sn-glycerol (8).
Scheme 5. Synthesis of sn-3,1’ DGP 17. PivCl=pivaloyl chloride.
tected 13 gave H-phosphonate 14 in good yield. Contrary to
isopropylidene-protected 11, benzyl-protected 14 exhibited
considerable stability against hydrolysis, easily tolerating
chromatographic purification. Subsequently, 15 was prepared
by a one-pot condensation–oxidation procedure.[8] After the
pivaloyl chloride mediated activation of 14, iodine-mediated
oxidation of the mixed phosphonate diester intermediate was
performed. The isopropylidene protection was removed by
neat TFA giving 16. After hydrogenolysis, the crude inter-
mediate was treated with 6 to give desired meso d-camphor
bisketal 17.
Scheme 3. Synthesis of sn-3,3’ DGP 10.
Our initial route (Scheme 4) towards the remaining meso
analogue 17 required the direct H-phosphonylation of 8.
Treatment of 8 with diphenyl phosphite afforded only
undesired phenyl phosphite and phosphorous acid. In the
case of phosphorous trichloride, only small amounts of 11
were isolated. In contrast, 8 was readily converted to the
phosphordiamidite 12. Careful hydrolysis of 12 gave crude 11.
Further purification of 11 failed, because of its high suscept-
ibility to hydrolysis. Unfortunately, the presence of small
amounts of alcohol 8 in crude 11 rendered this approach
unsuitable to access enantiomerically pure meso-DGP 17.
1
The H NMR spectra of bisketals 7 and 10 revealed the
presence of a pair of diastereomers, owing to the nonstereo-
specific formation of the chiral camphor ketals. In contrast,
the asymmetric nature of bisketal 17 gave rise to four
diastereomers in total. The most prominent shift differences
between bisketals 7, 10, and 17 are those observed from d =
4.50 to 3.70 ppm (Figure 1A–C). In general, the sn-2/2’
protons exhibit the strongest low-field shift, in the range of
d = 4.50 to 4.27 ppm. The signals of the glycerol CH2 protons
adjacent to the phosphate moiety are centered around d =
4.25 and 4.15 ppm. All signals of terminal protons of the DGP
backbones emerge between d = 3.98 and 3.70 ppm.
Despite the considerable similarity of the 1H NMR
spectra of bisketals 7, 10, and 17, each spectrum exhibits a
distinct pattern. In short, a signal at d = 3.79 ppm and the
absence of any signal between d = 4.50 and 4.33 ppm is typical
for sn-1,1’ DGP 7 (Figure 1A). The characteristic signals for
meso-DGP 17 are centered at d = 3.82 and 4.41 ppm (Fig-
ure 1C). In contrast, sn-3,3’ DGP 10 exhibits signals between
d = 4.50 and 4.33 ppm, whereas the area around d = 3.82 ppm
Scheme 4. Synthesis of H-phosphonate 11. HMDSO=hexamethyldisi-
loxane.
534
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 533 –535

Angewandte
Chemie
In conclusion, we prepared the d-camphor bisketal
derivative of natural BMP to identify the stereochemical
configuration of its DGP backbone by 1H NMR spectroscopy.
As reference materials we synthesized the sn-1,1’, sn-3,3’, and
1
sn-3,1’ DGP analogues. Comparison of the H NMR spectra
revealed that natural BMP features the unusual sn-1,1’ DGP
backbone. In contrast to previous biochemical analysis,[2] the
presence of other DGP backbone configurations in natural
BMP was below the minimum detection limit. BMP is
enriched in the lumenal side of LEs,[9] where lipids are
exposed to hydrolysis catalyzed by degrading enzymes. The
sn-1,1’ DGP backbone of BMP is advantageous, because
phospholipases preferentially hydrolyze sn-3 phospholipids.
Besides in mammalian cells, the occurrence of BMP has only
been reported in alkalophilic bacteria[10] and in the amoeba
Dictyostelium discoideum.[11] Bacterial BMP has been
reported to exhibit an sn-3,1’ DGP backbone,[10] whereas
the stereochemical configuration of amoeban BMP has not
been reported. This observation suggests that sn-1 phospho-
lipids, besides in archaea, are restricted to the endocytic
organelles of eukaryotic cells. An exciting possibility is that
BMP in eukaryotic cells originates from endocytosed archaea.
Figure 1. 1H NMR spectra of the DGP backbone region of d-camphor
bisketals in [D6]benzene A) from sn-1,1’ DGP 7, B) from sn-3,3’ DGP
10, C) from sn-3,1’ DGP 17, and D) from natural BMP 19. The dashed
lines highlight signals at d=4.42, 3.82, and 3.79 ppm.
as well as around d = 3.79 ppm is devoid of any signals
(Figure 1B).
Received: September 13, 2011
Published online: December 1, 2011
Next, we isolated and purified natural BMP from baby
hamster kidney (BHK) cells as described.[5] Mass spectro-
metric analysis revealed that natural BMP from BHK cells
predominantly features oleic acid residues (data not shown).
To determine the stereochemical configuration of the DGP
backbone of isolated BMP (1 mg), we first cleaved the fatty
acid residues by mild alkaline methanolysis (Scheme 6). After
neutralization by ion-exchange column, the fatty acid methyl
esters were removed by hexane extraction. Subsequently, the
lyophilized intermediate 18 was converted into its d-camphor
bisketal derivative 19 by treatment with an excess of 6 under
mild acidic conditions.
Keywords: configuration determination · NMR spectroscopy ·
.
phospholipids · structure elucidation
[1] a) G. Wꢀchtershꢀuser, Mol. Microbiol. 2003, 47, 13 – 22; b) J.
Peretꢁ, P. Lꢁpez-Garcꢂa, D. Moreira, Trends Biochem. Sci. 2004,
29, 469 – 477; c) Y. Koga, J. Mol. Evol. 2011, 72, 274 – 282.
[2] J. Brotherus, O. Renkonen, J. Herrmann, W. Fischer, Chem.
Phys. Lipids 1974, 13, 178 – 182.
[3] a) F. Hullin-Matsuda, C. Luquain-Costaza, J. Bouvier, I. Delton-
Vandenbroucke, Prostaglandins Leukotrienes Essent. Fatty Acids
2009, 81, 313 – 324; b) E. Orsꢁ, M. Grandl, G. Schmitz, Chem.
Phys. Lipids 2011, 164, 479 – 487.
[4] T. Kobayashi, E. Stang, K. S. Fang, P. de Moerloose, R. G.
Parton, J. Gruenberg, Nature 1998, 392, 193 – 197.
[5] T. Kobayashi et al., J. Biol Chem. 2002, 277, 32157-32164, see the
Supporting Information.
[6] a) H. Matsuo et al., Science 2004, 303, 531-534, see the Support-
ing Information; b) T. Hayakawa et al., Biophys. J. 2007, 92,
L13-16, see the Supporting Information; c) I. Delton-Vanden-
broucke, J. Bouvier, A. Makino, N. Besson, J.-F. Pageaux, M.
Lagarde, T. Kobayashi, J. Lipid Res. 2007, 48, 543-552;d) I.
Le Blanc et al., Nat. Cell Biol. 2005, 7, 653-664, see the
Supporting Information.
Finally, 19 (0.4 mg) was subjected to NMR spectroscopic
analysis (Figure 1D). One of the most prominent aspects of
1
the H NMR spectrum of 19 is the apparent lack of signals
between d = 4.50 and 4.33 ppm. Furthermore, only a weak
shoulder is present at 3.82, originating from a signal around
3.79 ppm. This signal pattern is in good agreement with the
typical features of sn-1,1’ DGP 7 (Figure 1A). Interestingly,
the presence of other stereoisomers is well below the
1
minimum detection limit of H NMR spectroscopy.
[7] a) J. Chevallier, N. Sakai, F. Robert, T. Kobayashi, J. Gruenberg,
S. Matile, Org. Lett. 2000, 2, 1859 – 1861; b) G. Jiang, Y. Xu, T.
Falguieres, J. Gruenberg, G. D. Prestwich, Org. Lett. 2005, 7,
3837 – 3840; c) G. Jiang, Y. Xu, G. D. Prestwich, J. Org. Chem.
2006, 71, 934 – 939.
[8] P. Greimel, M. Lapeyre, Y. Nagatsuka, Y. Hirabayashi, Y. Ito,
Bioorg. Med. Chem. 2008, 16, 7210 – 7217.
[9] T. Kobayashi, K. Startchev, A. J. Whitney, J. Gruenberg, Biol.
Chem. 2001, 382, 483 – 485.
[10] M. Nishihara, H. Morii, Y. Koga, J. Biochem. 1982, 92, 1469 –
1479.
[11] J. M. Rodriguez-Paris, K. V. Nolta, T. L. Steck, J. Biol. Chem.
1993, 268, 9110 – 9116.
Scheme 6. Conversion of natural BMP (1) to its d-camphor bisketal
derivative 19.
Angew. Chem. Int. Ed. 2012, 51, 533 –535
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
535