Secretory phospholipase A2 hydrolysis of phospholipid analogues is dependent on water accessibility to the active site

Article abstract of DOI:10.1021/ja067755b

A new and unnatural type of phospholipids with the head group attached to the 2-position of the glycerol backbone has been synthesized and shown to be a good substrate for secretory phospholipase A₂ (sPLA₂). To investigate the unexpe

Full text of DOI:10.1021/ja067755b

Published on Web 04/10/2007
Secretory Phospholipase A2 Hydrolysis of Phospholipid
Analogues Is Dependent on Water Accessibility to the Active
Site
Gu¨nther H. Peters,*,† Martin S. Møller,† Kent Jørgensen,§ Petra Ro¨nnholm,§
Mette Mikkelsen,§ and Thomas L. Andresen*,‡,§
Contribution from the Department of Chemistry, MEMPHYS-Center for Biomembrane Physics,
Technical UniVersity of Denmark, DK-2800 Lyngby, Denmark, Risø National Laboratory,
Technical UniVersity of Denmark, 4000 Roskilde, Denmark, and LiPlasome Pharma A/S,
Technical UniVersity of Denmark, DK-2800 Lyngby, Denmark
Received October 30, 2006; E-mail: thomas.andresen@risoe.dk
Abstract: A new and unnatural type of phospholipids with the head group attached to the 2-position of the
glycerol backbone has been synthesized and shown to be a good substrate for secretory phospholipase
A2 (sPLA2). To investigate the unexpected sPLA2 activity, we have compared three different phospholipids
by using fluorescence techniques and HPLC, namely: (R)-1,2-dipalmitoyl-glycero-3-phosphocholine
(hereafter referred to as 1R), (R)-1-O-hexadecyl-2-palmitoyl-glycero-3-phoshocholine (2R), and (S)-1-O-
hexadecyl-3-palmitoyl-glycero-2-phosphocholine (3S). Furthermore, to understand the underlying mech-
anisms for the observed differences, we have performed molecular dynamics simulations to clarify on a
structural level the substrate specificity of sPLA2 toward phospholipid analogues with their head groups in
the 2-position of the glycerol backbone. We have studied the lipids above 1R, 2R, and 3S as well as their
enantiomers 1S, 2S, and 3R. In the simulations of sPLA2-1S and sPLA2-3R, structural distortion in the
binding cleft induced by the phospholipids showed that these are not substrates for sPLA2. In the case of
the phospholipids 1R, 2R, and 3S, our simulations revealed that the difference observed experimentally in
sPLA2 activity might be caused by reduced access of water molecules to the active site. We have monitored
the number of water molecules that enter the active site region for the different sPLA2-phospholipid
complexes and found that the probability of a water molecule reaching the correct position such that
hydrolysis can occur is reduced for the unnatural lipids. The relative water count follows 1R > 2R > 3S.
This is in good agreement with experimental data that indicate the same trend for sPLA2 activity: 1R > 2R
> 3S.
Introduction
Phospholipase A2 enzymes (PLA2s) are responsible for the
remodeling of cellular phospholipids and the generation of
Cell membrane organization plays a prominent role in cell
regulation, and the earlier picture of a cell membrane being a
homogeneous mixture of lipids and proteins as suggested by
the Singer and Nicholson model is now significantly revised.1,2
A more complex picture involving membrane domains of
different compositions and regions of varying curvatures has
emerged that is believed to play important regulatory roles in
for instance membrane trafficking, protein sorting, and formation
of nanotubules.2-4 Lipid rafts, hydrophobic mismatch, mem-
brane phospholipid asymmetry, domain formation in mem-
branes, and enzymatic restructuring of the membrane are a few
examples of specific mechanisms by which membranes can take
part in signal transduction events.3,5-13
intracellular messengers.14-16 PLA2s are a diverse class of
hydrolytic enzymes that cleave the ester-linkage in the sn-2
position of aggregated glycerolphospholipids, yielding free fatty
acids and lyso-phospholipids.17-20 The PLA2 superfamily
includes the calcium-independent PLA2, the high molecular
(6) Kusumi, A.; Sako, Y. Curr. Opin. Cell Biol. 1996, 8, 566-574.
(7) Wisniewska, A.; Draus, J.; Subczynski, W. K. Cell Mol. Biol. Lett. 2003,
8, 147-159.
(8) Rieman, D.; Hansen, G. H.; Niels-Christiansen, L.-L.; Thorsen, E.;
Immerdahl, L.; Santos, A. N.; Kehlen, A.; Langner, J.; Danielsen, E. M.
Biochem. J. 2001, 354, 47-55.
(9) Caselli, A.; Mazzinghi, B.; Camici, G.; Manao, G.; Ramponi, G. Biochem.
Biophys. Res. Commun. 2002, 296, 692-697.
(10) Zhuang, L.; Lin, J.; Lu, M. L.; Solomon, K. R.; Freeman, M. R. Cancer
Res. 2002, 62, 2227-2231.
(11) Baron, G. S. EMBO J. 2002, 21, 1031-1040.
(12) Kakio, A.; Nishimoto, S.-I.; Yanagisawa, K.; Kozutsumi, Y.; Matsuzaki,
K. Biochemistry 2002, 41, 7385-7390.
† Department of Chemistry, MEMPHYS-Center for Biomembrane Phys-
ics.
(13) Simons, K.; Ikonen, E. Nature 1997, 387, 569-572.
(14) Leistad, L.; Feuerherm, A. J.; Østensen, M.; Faxvaag, A.; Johansen, B.
Clin. Chem. Lab. Med. 2004, 42, 602-610.
‡ Risø National Laboratory.
§ LiPlasome Pharma A/S.
(1) Singer, S. J.; Nicolson, G. L. Science 1972, 175, 720-731.
(2) Edidin, M. Nat. ReV. Mol. Cell Biol. 2003, 4, 414-418.
(3) Edidin, M. Curr. Opin. Struct. Biol. 1997, 7, 528-532.
(4) Brown, W. J.; Chambers, K.; Doody, A. Traffic 2003, 4, 214-221.
(5) Mouritsen, O. G. Chem. Phys. Lipids 1991, 57, 179-194.
(15) Leidy, C.; Mouritsen, O. G.; Jørgensen, K.; Peters, G. H. Biophys. J. 2004,
87, 408-418.
(16) Leidy, C.; Kaasgaard, T.; Mouritsen, O. G.; Jørgensen, K.; Peters, G. H.
Recent Res. DeVel. Biophys. 2004, 3, 163-185.
(17) Six, D. A.; Dennis, E. A. Biochim. Biochim. Acta 2000, 1488, 1-19.
9
10.1021/ja067755b CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 5451-5461
5451

A R T I C L E S
Peters et al.
weight cytosolic PLA2, and the low molecular weight secretory
PLA2 (sPLA2).17,21 Further classifications are based on their
source, subcellular location, calcium requirements, amino acid
similarity, and number of disulfide bridges.17,20 The activity of
sPLA2 depends strongly on the physical state and microstructure
of the membrane surface,15,16,22-26 and it has been recognized
that differential binding of mammalian sPLA2 to membranes
of different phospholipid compositions has physiological sig-
nificance.27,28 For instance, human tear fluid contains a high
concentration of human sPLA2-IIA as a protection against
bacterial infection.29,30 Human sPLA2-IIA does not degrade the
outer plasma membrane of a number of mammalian cells, which
are rich in zwitterionic lipids (phosphocholines; PC),31-33 but
the enzyme is able to degrade cell membranes of Gram-positive
bacteria that contain high levels of anionic lipids.23,34 The
observed increased activity in the presence of anionic lipids is
a general feature of sPLA2s23,35-37 and might be governed by
electrostatic interactions between charged residues located at
the binding surface of the enzyme and charged head groups of
the phospholipids.35,38 However, many sPLA2s, like Asp49
secretory phospholipase A2 from the snake venom of Agkistro-
don pisciVorus pisciVorus,39 can also degrade zwitterionic
lipids.23 This has been explained by the presence of one or more
aromatic residues located on the protein binding surface that
penetrate the interfacial membrane region, thereby enhancing
the binding affinity to zwitterionic membrane surfaces.40 Other
examples include the human secretory phospholipase A2 type
V and X.41 Clearly, enzyme adsorption of several sPLA2 species
are driven by the interplay between the sPLA2 binding surface
(18) White, S. P.; Scott, D. L.; Otwinowski, Z.; Gelb, M. H.; Sigler, P. B. Science
1990, 250, 1560-1563.
(19) Scott, D. L.; Otwinowski, Z.; Gelb, M. H.; Sigler, P. B. Science 1990,
250, 1563-1566.
(20) Kudo, I.; Murakami, M. Prostaglandins Other Lipid Mediators 2002, 68-
69, 3-58.
(21) Dennis, E. A. J. Biol. Chem. 1994, 269, 13057-13060.
(22) Hønger, T.; Jørgensen, K.; Biltonen, R. L.; Mouritsen, O. G. Biochemistry
1996, 35, 9003-9006.
Figure 1. View onto the binding surface (i-face) of human sPLA2-IIA
(top) and snake (Agkistrodon pisciVorus pisciVorus) venom sPLA2 (bottom).
Coloring of the surface corresponds to the properties of the residues:
positively charged (blue), negatively charged (red), polar (green), and
aromatic/hydrophobic (gray). Residues that are drawn in stick modus have
the right orientation such that they might be involved in contact with the
membrane surface.24,38,40,86 The substrate (1R) is shown in stick modus and
colored yellow.
(23) Leidy, C.; Linderoth, L.; Andresen, T. L.; Mouritsen, O. G.; Jørgensen,
K.; Peters, G. H. Biophys. J. 2006, 90, 3165-3175.
(24) Mouritsen, O. G.; Andresen, T. L.; Halperin, A.; Hansen, P. L.; Jakobsen,
A. F.; Jensen, U. B.; Jensen, M. Ø.; Jørgensen, K.; Kaasgaard, T.; Leidy,
C.; Simonsen, A. C.; Peters, G. H.; Weiss, M. J. Phys. Condens. Matter
2006, 18, 1-12.
(25) Peters, G. H.; Dahmen-Levison, U.; de Meijere, K.; Brezesinski, G.;
Toxvaerd, S.; Svendsen, A.; Kinnunen, P. K. J. Langmuir 2000, 16, 2779-
2788.
(26) Carriere, F.; Withers-Martinez, C.; van Tilberugh, H.; Roussel, A.;
Cambillau, C.; Verger, R. Biochim. Biophys. Acta 1998, 1376, 417-432.
(27) Jain, M. K.; Berg, O. G. Biochim. Biochim. Acta 1989, 1002, 127-156.
(28) Mouritsen, O. G.; Jørgensen, K. Pharm. Res. 1998, 15, 1507-1519.
(29) Huhtinen, H. T.; Gro¨nroos, J. O.; Gro¨nroos, J. M.; Uksila, J.; Gelb, M. H.;
Nevalainen, J.; Laine, V. J. O. Acta Pathol. Microbiol. Immunol. Scand.
2006, 114, 127-130.
and lipid interface properties.38,40,42,43 As shown in Figure 1,
phospholipase A2-IIA has a more positively charged binding
surface than snake venom sPLA2 and thereby prefers negatively
charged membrane surfaces.35,44
Increasing attention has been given to the human secretory
phospholipase A2 type IIA (sPLA2-IIA), a calcium-dependent,
∼14 kDa, disulfide-rich enzyme, since it has been implicated
in a large variety of physiological processes.45 sPLA2-IIA is
highly expressed in inflammatory tissues, where it hydrolyzes
phospholipids to release arachidonic acid that further down-
stream in the signaling cascade is processed by cyclooxygenases
or lipoxygenase to form proinflammatory lipid mediators known
as eicosanoids.32,46,47 sPLA2 has also been shown to be up-
(30) Qu, X. D.; Lehrer, R. I. Infect. Immun. 1998, 66, 2791-2797.
(31) Bezzine, S.; Koduri, A.; Valentin, E.; Murakami, M.; Kudo, I.; Ghomashchi,
F.; Sadilek, M.; Lambeau, G.; Gelb, M. H. J. Biol. Chem. 2000, 275, 3179-
3191.
(32) Murakami, M.; Koduri, R. S.; Enomoto, A.; Shimbara, S.; Seki, M.;
Yoshihara, K.; Singer, A.; Valentin, E.; Ghomashchi, F.; Lambeau, G.;
Gelb, M. H.; Kudo, I. J. Biol. Chem. 2001, 276, 10083-10096.
(33) Enomoto, A.; Murakami, A.; Valentin, E.; Lambeau, G.; Gelb, M. H.; Kudo,
I. J. Immunol. 2000, 165, 4007-4014.
(34) Buckland, A. G.; Heeley, E. L.; Wilton, D. C. Biochim. Biochim. Acta
2000, 1484, 195-206.
(35) Bezzine, S.; Bollinger, J. G.; Singer, A. G.; Veatch, S. L.; Keller, S. L.;
Gelb, M. H. J. Biol. Chem. 2002, 277, 48523-48534.
(36) Bayburt, T.; Yu, B.-Z.; Lin, H.-K.; Browning, J.; Jain, M. K.; Gelb, M. H.
Biochemistry 1993, 32, 573-582.
(37) Kinkaid, A. R.; Wilton, D. C. Biochem. J. 1995, 308, 507-512.
(38) Scott, D. L.; Mandel, A. M.; Sigler, P. B.; Honig, B. Biophys. J. 1994, 67,
493-504.
(42) Lin, Y.; Nielsen, R.; Murray, D.; Hubbell, W. L.; Mailer, C.; Robinson, B.
H.; Gelb, M. H. Science 1998, 279, 1925-1929.
(43) Ball, A.; Nielsen, R.; Gelb, M. H.; Robinson, B. H. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 6637-6642.
(39) Han, S. K.; Yoon, E. T.; Scott, D. L.; Sigler, P. B.; Cho, W. J. Biol. Chem.
1997, 272, 3573-3582.
(44) Canaan, S.; Nielsen, R.; Ghomashchi, F.; Robinson, B. H.; Gelb, M. H. J.
Biol. Chem. 2002, 277, 30984-30990.
(40) Gelb, M. H.; Cho, W.; Wilton, D. C. Curr. Opin. Struct. Biol. 1999, 9,
428-432.
(45) Fuentes, L.; Hernandez, M.; Nieto, M. L.; Crespo, M. S. FEBS Lett. 2002,
531, 7-11.
(41) Murakami, M.; Kudo, I. J. Biochem. 2002, 131, 285-292.
9
5452 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

PLA2 Substrate Specificity
A R T I C L E S
regulated in a variety of cancer types48,49 including prostate,50,51
colon,52,53 breast,54 and pancreatic cancer.55,56 It has been sug-
gested that the sPLA2-IIA overexpression in cancerous tissue
may be triggered by the carcinoma cells and hepatocytes and/
or inflammatory cells (e.g., macrophages) by stimulation of
inflammatory cytokines.48,57 The observation that many tumors
have elevated sPLA2 levels highlights that sPLA2 plays a central
role in tumor development and demonstrates that sPLA2 is a
potential target for therapeutic intervention.58,59 Indeed, it has
been demonstrated that elevated concentrations of sPLA2 in
cancer cells can be used as a suitable tumor-specific triggering
mechanism for liposome-based anticancer drug carriers.59-62
Liposomes comprised of natural phospholipids or anticancer
lipid-based prodrugs are hydrolyzed by sPLA2 resulting in a
controlled release of the cytotoxic drugs into the tumor
tissue.62,63 This approach results in increased circulation life
time, enhanced deposition to the malignant tissue, and protection
from the drug metabolic degradation.64 The usage of natural
lipids has a further advantage in that these systems are
biodegradable, are weakly immunogenic, possess limited in-
trinsic toxicity, and produce no antigenic or pyrogenic reac-
tions.65,66 These liposome-based drug carries can be optimized
with respect to, e.g., circulation lifetime, stability (“leakage”
of drugs), or hydrolysis by sPLA2 by changing lipid composi-
tion, modifying their surface characteristics (e.g., polymer-
grafted liposomes), or using unnatural phospholipid analogues
degradable by sPLA2.28,63,67
Figure 2. Chemical structures of the three enatiomeric pairs (six phos-
pholipids) that have been investigated by molecular dynamics simulations.
(1R) (R)-1,2-dipalmitoyl-glycero-3-phosphocholine, (1S) (S)-1,2-dipalmi-
toyl-glycero-3-phosphocholine, (2R) (R)-1-O-hexadecyl-2-palmitoyl-sn-
glycero-3-phoshocholine, (2S) (S)-1-O-hexadecyl-2-palmitoyl-sn-glycero-
3-phoshocholine,
(3S)
(S)-1-O-hexadecyl-3-palmitoyl-glycero-2-
phosphocholine, and (3R) (R)-1-O-hexadecyl-3-palmitoyl-glycero-2-
phosphocholine. The compounds with the correct stereochemistry for sPLA2
hydrolysis, 1R, 2R, and 3S have been synthesized, and liposomes constituted
of these phospholipids have been investigated experimentally.
but also are efficiently degraded by sPLA2. Although sPLA2
shows lower activity toward the unnatural phospholipid ana-
logues, it is intriguing that sPLA2 can also hydrolyze lipids with
head groups located in the 2-position. To provide further insight
on a molecular level and to explain the altered sPLA2 activity,
we report here molecular dynamics simulations of six different
sPLA2-phospholipid substrate complexes (Figure 2) and com-
pare the results with experiments. The simulations revealed that
the unnatural lipids are positioned in the sPLA2 binding pocket
in such a way that the lipids interfere with an incoming water
molecule that participates in the hydrolysis. The probability of
a water molecule reaching the correct position such that
hydrolysis can occur is reduced for the unnatural lipids. This
in part could explain the experimentally observed reduced sPLA2
activity profile toward the unnatural phospholipid analogues.
We have previously shown, in a rapid report,67 that unnatural
phospholipid analogues with the head group in the 2-position
not only form stable liposomes with intriguing phase behavior
(46) Kim, Y. J.; Kim, K. P.; Han, S. K.; Munoz, N. M.; Zhu, X.; Sano, H.;
Leff, A. R.; Cho, W. Biol. Chem. 2002, 277, 36479-36488.
(47) Glomset, J. A. Curr. Opin. Struct. Biol. 1999, 9, 425-427.
(48) Abe, T.; Sakamoto, K.; Kamohara, H.; Hirano, Y.; Kuwahara, N.; Ogawa,
M. Intern. Congress Series 2003, 1255, 351-360.
(49) Abe, T.; Sakamoto, K.; Kamohara, H.; Hirano, Y.; Kuwahara, N.; Ogawa,
M. Int. J. Cancer 1997, 74, 245-250.
(50) Sved, P.; Scott, K. F.; McLeod, D.; King, N. J. C.; Singh, J.; Tsatralis, T.;
Nikolov, B.; Boulas, J.; Nallan, L.; Gelb, M. H.; Sajinovic, M.; Graham,
G. G.; Russell, P. J.; Dong, Q. Cancer Res. 2004, 64, 6934-6940.
(51) Jiang, J.; Neubauer, B. L.; Graff, J. R.; Chedid, M.; Thomas, J. E.; Roehm,
N. W.; Zhang, S.; Eckert, G. J.; Koch, M. O.; Eble, J. N.; Cheng, L. Am.
J. Pathology 2002, 160, 667-671.
(52) Ilsley, J. N. M.; Nakanishi, M.; Flynn, C.; Belinsky, G. S.; De, Guise, S.;
Adib, J. N.; Dobrowsky, R. T.; Bonventre, J. V.; Rosenberg, D. W. Cancer
Res. 2005, 65, 2636-2643.
(53) Edhemovic, I.; Snoj, M.; Kljun, A.; Golouh, R. EJSO 2001, 27, 545-548.
(54) Yamashita, S.; Yamashita, J.; Ogawa, M. Br. J. Cancer 1994, 69, 1166-
1170.
Experimental Section
Materials and Liposome Preparation. 1,2-Dipalmitoyl-sn-glycero-
3-phosphocholine (DPPC, 1R) was purchased from Avanti Polar Lipids,
Birmingham, AL. All other lipids were synthesized. All other chemicals
were purchased from Sigma-Aldrich. One-component multilamellar
liposomes were prepared by hydrating the lipids in a Hepes buffer
solution for 1 h 10-15 °C above the main phase transition of the
multilamellar liposomes. The lipid suspension was vortexed every 15
min. Unilamellar liposomes of narrow size distribution were made by
extrusion of the multilamellar liposome suspension 10 times through
two stacked 100 nm polycarbonate filters. The Hepes buffer used was
composed of the following: 0.15 M KCl, 1 mM NaN3, 0.03 mM CaCl2,
0.01 mM EDTA, 0.01 M Hepes (pH ) 7.5).
(55) Kiyohara, H.; Egami, H.; Kurizaki, T.; Murata, K.; Ohmachi, H.; Akagi,
J.; Ohshima, S.; Yamamoto, S.; Shibata, Y.; Ogawa, M. Intern. Congress
Series 2003, 1255, 375-379.
(56) Hanada, K.; Kinoshita, E.; Itoh, M.; Hirata, M.; Kajiyama, G.; Sugiyama,
M. FEBS Lett. 1995, 373, 85-87.
(57) Yamashita, S.; Ogawa, M.; Abe, T.; Yamashita, J.; Sakamoto, K.; Niwa,
H.; Yamamura, K. Biochem. Biophys. Res. Commun. 1994, 198, 878-
884.
(58) Laye, J.; Gill, J. H. Drug DiscoVery Today 2003, 8, 710-716.
(59) Andresen, T. L.; Jensen, S. S.; Jørgensen, K. Prog. Lipid Res. 2005, 44,
68-97.
(60) Andresen, T. L.; Jensen, S. S.; Kaasgaard, T.; Jørgensen, K. Curr. Drug
DeliVery 2005, 2, 353-362.
Differential Scanning Calorimetry. Differential scanning calorim-
etry (DSC) of 5 mM multilamellar liposomes was performed using a
Microcal MC-2 (Northhampton, MA) ultrasensitive power-compensat-
ing scanning calorimeter equipped with a nanovoltmeter. The scans
were performed in the upscan mode at a scan rate of 20 °C/h. An
appropriate baseline has been subtracted from the calorimetric curves.
(61) Jørgensen, K.; Davidsen, J.; Mouritsen, O. G. FEBS Lett. 2002, 531, 23-
27.
(62) Andresen, T. L.; Davidsen, J.; Begtrup, M.; Mouritsen, O. G.; Jørgensen,
K. J. Med. Chem. 2004, 47, 1694-1703.
(63) Goyal, P.; Goyal, K.; Kumar, S. G. V.; Singh, A.; Katare, O. P.; Mishra,
D. N. Acta Pharm. 2005, 55, 1-25.
(64) Jensen, S. S.; Andresen, T. L.; Davidsen, J.; Høyrup, P.; Shnyder, S. D.;
Bibby, M. C.; Gill, J. H.; Jørgensen, K. Mol. Cancer Ther. 2004, 3, 1451-
1458.
Phospholipase A2 Activity Measurements. The employed sPLA2
(Agkistrodon pisciVoros pisciVoros) was isolated and purified from
snake venom.68 This enzyme is structurally similar to mammalian
sPLA2, indicating a similar catalytic hydrolysis mechanism for phos-
(65) Van Rooijen, N.; van Nieuwmegen, R. Immunol. Commun. 1980, 9, 243-
256.
(66) Campbell, P. I. Cytobios 1983, 37, 21-26.
(67) Andresen, T. L.; Jørgensen, K. Biochim. Biochim. Acta 2005, 1669, 1-7.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5453

A R T I C L E S
Peters et al.
pholipid substrates.69 Assay conditions for the sPLA2 lag-time measure-
ments were 0.15 mM unilamellar liposomes in a Hepes buffer (150
nM sPLA2, 0.15 M KCl, 1 mM NaN3, 0.03 mM CaCl2, 0.01 mM
EDTA, and 0.01 M Hepes (pH ) 7.5)). The catalytic reaction was
initiated by adding 8.9 µL of a 42 µM sPLA2 stock solution to 2.5 mL
of a thermostated liposome suspension equilibrated 20 min prior to
addition of the enzyme. The time elapsed between addition of sPLA2
and a sudden increase in the intrinsic fluorescence emission from sPLA2
(tryptophan emission), due to a burst in the sPLA2 activity, is defined
as the characteristic lag time of the enzyme. The emission takes place
at 340 nm after excitation at 285 nm. The total amount of hydrolyzed
lipid 1000 s after the sPLA2 burst was determined by HPLC.
HPLC Quantification. HPLC analysis was performed using a 5
µL diol column, a mobile phase composed of CHCl3/MeOH/H2O (730:
230:25, v/v), and an evaporative light scattering detector. Sample (100
µL) was withdrawn from a 0.15 mM liposome suspension, which was
rapidly mixed with 1 mL of CHCl3/methanol/acetic acid (2:4:1) in order
to quench the enzymatic reaction. The solution was washed with 1 mL
of water, and 20 µL of the organic phase were used for HPLC.
Synthesis of (R)-1-O-Hexadecyl-sn-glycero-3-phosphocholine (2R)
from (R)-O-Benzyl Glycidol (4R). Carried out as earlier described by
Andresen and co-workers.62
Synthesis of (R)-1-O-Hexadecyl-3-lyso-glycero-2-phosphocholine
(8). To a solution of 7 (253 mg, 0.442 mmol) in EtOAc/MeOH (20
mL, 1:1) under N2 at room temperature Pd/C 10% (33 mg) was added.
The reaction mixture was placed under H2 with vigorous stirring for
1.5 h, after which TLC (CH2Cl2/MeOH/H2O 65:25:4) showed no
remaining starting material. The solution was filtered through a glass
filter, which was washed successively with CH2Cl2. Concentrating in
vacuo gave 213 mg (quantitative) of 8 as a white solid, which was
dried at 0.1 mmHg for 1 h and used without further purification in the
next step. Rf ) 0.13 (CH2Cl2/MeOH/H2O 65:25:4). 1H NMR (300 MHz,
CDCl3/CD3OD 9:1): δ 4.20-4.10 (m, 3H), 3.65-3.25 (m, 8H), 3.20-
3.08 (m, 9H), 1.59 (m, 2H), 1.26 (br.s, 26H), 0.88 (t, 3H). 13C NMR
(75 MHz, CDCl3/CD3OD 9:1): δ 76.1 (d, J ) 7.4 Hz), 71.8, 71.1 (d,
J ) 5.1 Hz), 66.4 (d, J ) 6.1 Hz), 63.9 (d, J ) 4.8 Hz), 59.5 (d, J )
4.7 Hz), 54.4, 32.1, 29.9, 29.8, 29.6, 29.4, 26.3, 25.2, 22.9, 14.3.
Synthesis of (S)-1-O-Hexadecyl-3-palmitoyl-glycero-2-phospho-
choline (3S). 8 (236 mg, 0.49 mmol) was dissolved in dry CH2Cl2 (25
mL) to which Et3N (0.17 mL, 1.22 mmol), DMAP (24 mg, 0.20 mmol),
and palmitoyl chloride (0.37 mL, 1.22 mmol) were added. The yellow
reaction mixture was stirred overnight at room temperature, after which
TLC (CH2Cl2/MeOH/H2O 65:25:4) indicated that the reaction had gone
to completion. The reaction mixture was concentrated and purified by
flash chromatogaphy (CH2Cl2/MeOH/H2O 65:25:1 f CH2Cl2/MeOH/
H2O 65:25:4) to give 294 mg (83%) of 3S. Rf ) 0.28 (CH2Cl2/MeOH/
Synthesis of (S)-1-O-Hexadecyl-3-O-benzyl-glycerol (6). To a
flame-dried flask containing washed NaH (0.23 g, 9.6 mmol) under
N2 at 0 °C cetylalcohol (1.94 g, 8.02 mmol) in dry THF (20 mL) was
added. The NaH was washed by a continuous extraction of the NaH
dispersion (60% mineral oil) with dry hexane. The reaction was heated
to 80 °C for 1 h. (S)-O-Benzyl glycidol (4S) (1.05 g, 6.1 mmol) was
added followed by DMF (10 mL). The reaction was stirred overnight
at 80 °C. The reaction was cooled to room temperature and stirred 20
min after the addition of water (2 mL). Removal of solvent in vacuo
gave a brown residue, which was redissolved in ether (40 mL), washed
with brine (5 × 10 mL), and dried over Na2SO4. Purification by flash
chromatography (CH2Cl2/ether 100:1) gave 1.86 g (71%) of 6. Rf )
1
H2O 65:25:4). H NMR (300 MHz, CDCl3): δ 4.45-4.18 (m, 5H),
3.80 (m, 2H) 3.57 (m, 2H), 3.45-3.23 (m, 11H), 2.30 (t, J ) 6.7 Hz,
2H), 1.59 (m, 2H), 1.51 (m, 2H), 1.26 (br.s, 50H), 0.88 (t, 6H). 13C
NMR (75 MHz, CDCl3): δ 173.8, 76.9 (d), 71.8, 70.4 (d), 66.2 (d),
63.8 (d), 59.4 (d), 54.5, 35.5, 32.1, 29.9, 29.8, 29.6, 29.4, 26.3, 25.2,
22.9, 14.3.
Molecular Dynamics Simulations. The crystal structures of (i)
European Honeybee (Apis Mellifera)-venom phospholipase A2 com-
plexed with the transition-state analogue, 1-O-octyl-2-heptylphosphonyl-
sn-glycero-3-phosphoethanolamine (diC8(2Ph)PE), resolved to 2.0 Å19
and (ii) Human sPLA2-IIA complexed with 6-phenyl-4(R)-(7-phenyl-
heptanoylamino)-hexanoic acid resolved to 2.1 Å70 were obtained from
the Protein Data Bank71 (entry codes: 1poc and 1kqu, respectively).
The initial modeling step involved placing diC8(2Ph)PE into the binding
cleft of sPLA2-IIA, which was done by (i) deleting the inhibitor in
1kqu (keeping calcium ions and water molecules in the structure), (ii)
deleting calcium ions and water molecules in 1poc, (iii) aligning 1poc
with 1kqu, and (iii) finally deleting the bee-venom phospholipase A2
structure. The structures of the six phospholipids (Figure 1) were built
from diC8(2Ph)PE using SPARTAN version 1.0.2 (Wavefunction Inc.,
Irvine, California). Missing distance, angle, and torsion parameters were
obtained from the CHARMM27 parameter set describing similar atom
types. The structures were solvated using the program SOLVATE.72
To neutralize the systems, 18 water molecules were randomly replaced
with chloride ions. The final systems contained ∼4900 water molecules,
and the simulation cell dimensions were ∼53 × 52 × 67 Å3. For the
simulations, the molecular dynamics (MD) program NAMD73 was used
with the Charmm27 all hydrogens parameter set and with the TIP3
water model.74 Each complex was simulated at least three times (Table
1) starting from different initial conditions, which were obtained by
varying the number of steps used in energy minimization of the systems.
1
0.11 (CH2Cl2/ether 100:1). H NMR (300 MHz, CDCl3): δ 7.32 (m,
5H), 4.56 (s, 2H), 4.00 (m, 1H), 3.60-3.30 (m, 6H), 1.59 (m, 2H),
1.26 (br.s, 26H) 0.88 (t, 3H). 13C NMR (75 MHz, CDCl3): δ 138.3,
128.6, 127.9, 73.6, 72.1, 71.9, 71.7, 69.7, 32.2, 30.0, 29.9, 29.9, 29.9,
29.8, 29.7, 26.4, 23.0, 14.4.
Synthesis of (S)-1-O-Hexadecyl-3-O-benzyl-glycero-2-phospho-
choline (7). To a solution of POCl3 (79.8 µL, 0.856 mmol) in dry CH2-
Cl2 (0.5 mL) under N2 at 0 °C a solution of 6 (272.2 mg, 0.679 mmol)
and Et3N (0.13 mL, 0.910 mmol) in CH2Cl2 (2 mL) was added dropwise
over 15 min. The reaction mixture was stirred for 1 h under N2 at room
temperature, after which pyridine (0.44 mL, 5.43 mmol) and choline
tosylate (374 mg, 1.36 mmol) were added at 0 °C. After 2 h, the orange
reaction mixture was allowed to warm to room temperature and was
left stirring overnight. After 24 h of stirring, TLC (CH2Cl2/MeOH/
H2O 65:25:4) indicated that the reaction had gone to completion. Water
(0.2 mL) was added, and the reaction mixture was stirred for 1 h and
concentrated to a white solid by azeotropic distillation with toluene.
The residue was dissolved in THF/H2O 9:1 and slowly passed through
an MB-3 column (2 cm), and the solvent was removed by azeotropic
distillation with ethanol and toluene in vacuo. The crude product was
purified by column chromatography (CH2Cl2/MeOH/H2O 65:25:1) giv-
1
ing 7 in 253 mg (65%). Rf ) 0.17 (CH2Cl2/MeOH/H2O 65:25:1). H
(70) Hansford, K. A.; Reid, R. C.; Clark, C. I.; Tyndall, J. D. A.; Whitehouse,
M. W.; Guthrie, T.; McGeary, R. P.; Schafer, K.; Martin, J. L.; Fairlie, D.
P. ChemBioChem 2003, 4, 181-185.
NMR (300 MHz, CDCl3): δ 7.32 (m, 5H), 4.54 (s, 2H), 4.40 (m, 1H),
4.20 (m, 2H), 3.78-3.45 (m, 6H), 3.40 (t, J ) 6.1 Hz, 2H), 3.08 (s,
9H), 1.59 (m, 2H), 1.26 (br.s, 26H), 0.89 (t, 3H). 13C NMR (75 MHz,
CDCl3): δ 138.3, 128.6, 127.9, 73.5, 73.2 (d, J ) 7.3 Hz), 71.7, 71.6
(d, J ) 4.9 Hz), 70.9 (d, J ) 4.8 Hz), 66.3 (d, J ) 5.3 Hz), 59.2 (d,
J ) 4.8 Hz), 54.4, 32.1, 29.9, 29.8, 29.6, 29.4, 26.3, 25.2, 22.9, 14.3.
(71) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J.; Meyer, E. E.; Brice, M.
D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi, M. J. Mol. Biol.
1977, 112, 535-542.
(72) Grubmu¨ller, H. Solvate: a program to create atomic solvent models.
Electronic publication: http://www.mpibpc.gwdg.de/abteilungen/071/sol-
vate/docu.html.
(73) Kale, L.; Skeel, R.; Bhandarkar, M.; Brunner, R.; Gursoy, A.; Krawetz,
N.; Phillips, J.; Shinozaki, A.; Varadarajan, K.; Schulten, K. J. Comput.
Phys. 1999, 151, 283-312.
(74) Jorgensen, W. L.; Chandrasekhar, J.; Medura, J. D.; Impey, R. W.; Klein,
M. L. J. Chem. Phys. 1983, 79, 926-935.
(68) Maraganore, J. M.; Merutka, G.; Cho, W.; Welches, W.; Kezdy, F. J.;
Heinrikson, R. L. J. Biol. Chem. 1984, 259, 13839-13843.
(69) Vance, D. E.; Vance, J. E. Biochemistry of Lipids, Lipoproteins and
Membranes; Elsevier: Amsterdam, 1991.
9
5454 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

PLA2 Substrate Specificity
A R T I C L E S
Table 1. Summary of the Simulations Carried Outa
root-mean-square
displacement
mean
SDc
(Å)
length of
simulations
(ns)
short
nameb
number of
±
name
simulations
(R)-1,2-dipalmitoyl-glycero-3-
phosphocholine
(S)-1,2-dipalmitoyl-glycero-3-
1R
1S
2R
2S
3R
3S
6
4
3
3
3
3
each 10
each 10
each 10
each 10
each 10
each 10
1.3 ( 0.1
1.5 ( 0.1
1.4 ( 0.1
1.5 ( 0.1
1.4 ( 0.1
1.3 ( 0.1
phosphocholine
(R)-1-O-hexadecyl-2-palmitoyl-
sn-glycero-3-phoshocholine
(S)-1-O-hexadecyl-2-palmitoyl-
sn-glycero-3-phoshocholine
(R)-1-O-hexadecyl-3-palmitoyl-
glycero-2-phosphocholine
(S)-1-O-hexadecyl-3-palmitoyl-
glycero-2-phosphocholine
a The last column lists the average root-mean-square displacement of CR atoms with respect to the protein structure obtained after minimization. Averages
and standard deviations (SDs) are based on a series of simulations of a particular complex. b Abbreviation used in the text. c SD ) standard deviation
N
calculated as follows: SD )
∑
x
i)1(xi-xj)2/(N-1).
Scheme 1
The first energy minimization of the systems involved 250 steps. For
each subsequent simulation of a certain complex, the number of steps
for energy minimization was increased by 250 steps; i.e., 250, 500,
and 750 steps were used for simulations of a complex that was repeated
three times. The minimization procedure was followed by 100 ps of
heating of the systems to T ) 300 K. The simulations were carried out
for 10 ns at constant number of atoms, pressure, and temperature (NPT
ensemble). Periodic boundary conditions were applied in x, y, and
z-directions. A time step of 1 fs was used throughout all simulations.
An isotropic constant ambient pressure of 1 atm was imposed using
the Langevin piston method75 with a damping coefficient of 5 ps-1, a
piston period of 200 fs, and a decay of 500 fs. The Particle Mesh Ewald
method was used for computation of the electrostatic forces.76,77 The
grid spacing applied was approximately 1.0 Å, and a fourth order spline
was used for the interpolation. The long-range part of the electrostatic
forces was evaluated every fourth femtosecond. The van der Waals
interactions were cut off at 12 Å using a switching function starting at
10 Å. The analyses of the trajectories were performed using the
graphical program Visualization Molecular Dynamics (VMD).78
sPLA2-IIA activity, and at the same time, both lipid species
are hydrolyzed by sPLA2-IIA.23 This observation supports the
view that in the case of snake venom sPLA2 and human sPLA2-
IIA, the activity of these enzymes is determined mainly by the
physical properties of the membrane surface.79,80 Our primary
aim in the present study is to investigate by molecular dynamics
simulations the binding mode of phospholipid analogues and
to understand in molecular detail the basis for the modified
sPLA2 activity toward phospholipid analogues with head groups
located in the 2-position when compared to its natural substrate.
As summarized in Table 1, we have studied six different
phospholipids (Figure 2) in complex with human sPLA2-IIA.70
1R is the natural substrate for sPLA2. 1S has been included as
a reference, since, due to the stereospecific enzyme reaction,
this phospholipid should not be hydrolyzed by sPLA2. 2R and
2S contain an ether bond in the 1-position of the glycerol
backbone instead of the natural ester group. This replacement
was done for practical purposes, to avoid uncontrolled hydrolysis
by the enzyme when moving the head group to the 2-position.
3R and 3S are the two phospholipid enantiomers with the head
groups located at the 2-position of the glycerol backbone.
Simulation of each enzyme-phospholipid complex was repeated
at least three times to increase the statistical significance of our
Results and Discussion
We have demonstrated earlier that the presence of phospho-
glycerol (PG) in phosphocholine (PC) membranes enhances
(75) Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks, B. R. J. Chem. Phys. 1995,
103, 4613-4621.
(76) Darden, T.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98, 10089-10092.
(77) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen,
L. G. J. Chem. Phys. 1995, 103, 8577-8593.
(79) Pan, Y. H.; Yu, B.-Z.; Berg, O. G.; Jain, M. K.; Bahnson, B. J. Biochemistry
2002, 41, 14790-14800.
(80) Bahnson, B. J. Arch. Biochem. Biophys. 2005, 433, 96-106.
(78) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graph. 1996, 14, 33-38.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5455

A R T I C L E S
Peters et al.
Figure 3. (A) Differential scanning heat capacity curves obtained at a scan rate of 20 °C/h for 5 mM multilammelar liposomes composed of 1R, 2R, and
3S. (B) Characteristic sPLA2 (Agkistrodon pisciVorus pisciVorus) activity time course profiles toward unilamellar liposomes. The sPLA2 activity is monitored
by intrinsic fluorescence (solid line) from the enzyme and by 90° light scattering (dashed line) from the lipid suspension.
Figure 4. Secretory PLA2 lag times for 1R, 2R, and 3S obtained at the main phase transition temperature (gray bars). Total hydrolysis of the three phospholipids
measured by HPLC 1000 s after the burst has occurred (white bars).
results. As outlined in the Experimental Section, each simulation
for a particular sPLA2-phospholipid complex was started from
different initial conditions by varying the steps used for the
initially performed energy minimization.
It should be noted that the use of THF afforded considerably
reduced dimerization compared to using DMF alone. Phospho-
rylation using phosphorus oxychloride in DCM gave 7 in 65%
yield. Hydrogenation with Pd/C as catalyst followed by acylation
using palmitoyl chloride gave the target compound 3S in 83%
yield. One can equally well use palmitic acid and DCC as
coupling reagents under standard conditions.
Multilamellar liposomes were prepared from each lipid 1R,
2R, and 3S, and heat capacity (Cp) curves were obtained using
differential scanning calorimetry (Figure 3A). The Cp curves
show that all three lipids form liposomes with 1R, 2R, and 3S
having a main phase transition peak at 40.5, 42.6, and 40.1 °C,
respectively, reflecting the gel-to-fluid phase transition. Interest-
ingly, the pretransition that usually characterizes the phase
Synthesis and Biophysical Investigations. To compare the
results from our molecular dynamics simulations with experi-
ments, we have synthesized 2R and 3S (Figure 2) that both,
together with the commercially available 1R (DPPC), have the
right stereochemistry for sPLA2 hydrolysis. 2R was synthesized
from (R)-O-benzyl glycidol (Scheme 1) in four steps as
described in detail by our group in an earlier article.62 3S was
synthesized starting from (S)-O-benzyl glycidol (4S). The
epoxide 4S was opened regioselectively by reaction with sodium
cetyl alkoholate using THF/DMF (2:1) giving 6 in 71% yield.
9
5456 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

PLA2 Substrate Specificity
A R T I C L E S
pholipids and fatty acids are formed. Figure 4 shows the lag
time for lipids 1R, 2R, and 3S at the main phase transition,
where the enzyme is known to have the highest activity.23,62
The natural 1R (DPPC) has the shortest lag time indicating that
this is the best substrate for sPLA2. It is also seen that the
introduction of an ether bond in the 1-position (2R) results in
a longer lag time; however, the change is relatively small. A
larger change is observed when moving the phospholipid head
group to the 2-position (3S). For 3S, the lag time is increased
considerably when compared with the other two phospholipids.
Figure 4 also shows the total amount of hydrolysis 1000 s after
the accelerated hydrolysis (burst) has occurred. All three
phospholipids are hydrolyzed to a high extent after the burst
has occurred, and hence, all three lipids are good substrates for
sPLA2.
Simulations - Enzyme Stability. The stability of the simula-
tions of the different sPLA2-phospholipid complexes were
checked by computing the time evolution of the root mean
displacement (rmsd) of the CR atoms with respect to the protein
structure obtained after minimization. A representative time
evolution of rmsd calculated from an sPLA2-1R simulation is
provided as Supporting Information (Figure S1). Means and
standard deviations of the rmsd data calculated for the different
systems are summarized in Table 1. The relatively low rmsd
values reflect the rigid structure of sPLA2-IIA that is caused
by eight sulfur bridges. The protein structures complexed with
the different phospholipids are stable resulting in an overall rmsd
of (1.4 ( 0.1) Å, calculated from the rmsd of each simulation.
To further evaluate our simulations, we compared the time-
averaged fluctuations of the residues with the B-factor given
as part of the X-ray structure,70 and overall, good agreement
was observed between the B-factor and the simulation results
(representative results are provided as Supporting Information;
Figure S2). In some regions, larger rmsd values are seen in the
simulations than in the B-factor. Particularly, protein segments
that are located in the vicinity of the substrate (close to the head
group and the end of the acyl chains) and in two loop regions
are more flexible in the simulations than deduced from the
Figure 5. Schematic representation of the catalytic mechanism of sPLA2
and part of the hydrogen-bonding network. Atom types given in parentheses
refer to Protein Data Bank nomenclature.
behavior of natural phospholipids,15,16 such as DPPC (1R, seen
at 31.9 °C), is abolished for the unnatural phospholipid 3S with
the head group in the 2-position. This may be due to the
symmetry in the lipid that is introduced by moving the head
group from the 3- to the 2-position.67 Furthermore, the ther-
mograms revealed that the introduction of an ether bond in the
1-position does not change the phase behavior drastically (Figure
3A).
Unilamelar liposomes for sPLA2 activity measurements were
prepared by extrusion through 100 nm polycarbonate filters. A
typical sPLA2 (Agkistrodon pisciVorus pisciVorus) fluorescence
activity profile can be seen in Figure 3B. After addition of
sPLA2, there is a lag time period that reflects a relatively low
sPLA2 activity, which suddenly changes in a burst of activity
as a consequence of accelerated hydrolysis. The fluorescence
is measured as intrinsic fluorescence from the enzyme. Simul-
taneously with the sudden increase in fluorescence, a decrease
in the 90° static light scattering is observed indicating changes
in suspension morphology as non-bilayer forming lyso-phos-
Figure 6. Time evolution of selected distances involving the cofactor calcium. Atom types correspond to the Protein Data Bank nomenclature. The distances
were extracted from an sPLA2-1R complex simulation and are representative for all simulations.
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5457

A R T I C L E S
Peters et al.
Figure 7. Time evolution of selected distances that involve Asp91 or His47. S and atom types refer to substrate and Protein Data Bank nomenclature,
respectively. The distances were extracted from an sPLA2-1R simulation and are representative for all simulations.
Figure 8. Time evolution of selected distances that involve the substrate (S). See Figure 7 caption for more details.
B-factor. The latter observation is not surprising, since these
loop regions are solvent exposed and may be affected by crystal
packing.81 We also calculated the rmsd variation between
simulations of a particular complex and between all simulations.
We determined an average structure for each simulation using
the last 5 ns of the trajectory and calculated consequently the
rmsd between the average structures. The rmsd within a series
of simulations of a particular complex varies between 0.48 and
0.65 Å yielding (0.57 ( 0.07) Å. The rmsd calculated between
average structures of different sPLA2-phospholipid complexes
varies between 0.60 and 0.98 Å resulting in (0.72 ( 0.09) Å.
Both average rmsd values are the same within the statistical
uncertainties indicating that all simulations converge to similar
structures.
Simulations - Key Distances. Further structural information
was extracted from the simulations by calculating distances
between selected atoms (residues). We also calculated the
corresponding interaction energies (data not shown), and as
expected, there is a direct correlation between calculated
interaction energies and monitored distances.82 In the following,
we therefore only report distances. The distances were chosen
accordingly to their importance in the calcium-dependent
enzymatic reaction.83 X-ray crystallography revealed that the
catalytic device of sPLA2 is essentially characterized by an Asp-
His dyad, a calcium-binding site, and a water molecule, which
(82) Peters, G. H.; Iversen, L. F.; Andersen, H. S.; Møller, N. P. H.; Olsen, O.
H. Biochemistry 2004, 43, 18-28.
(83) Scott, D. L.; White, S.; Otwinowski, Z.; Yuan, W.; Gelb, M. H.; Sigler, P.
B. Science 1990, 250, 1541-1546.
(81) Peters, G. H.; Bywater, R.P. J. Mol. Recog. 2002, 15 393-403.
9
5458 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

PLA2 Substrate Specificity
A R T I C L E S
distances that involve the substrate (hereafter referred to as S)
(S(O22)-Gly29(HN), S(O22)-Ca2+, and S(O4)-Ca2+; Figure
8). Atom types indicated in parentheses refer to the Protein Data
Bank nomenclature, which will be used throughout. The time
evolutions presented in Figures 6 to 8 are representative for all
simulations.
Overall, coordinations of Ca2+ by Asp48 and Gly29 are rather
stable (Figure 6). The increase of the Ca2+-Asp48(OD1)
distance taking place in concert with the change in Ca2+-Gly29-
(O) and occurring over a relatively short time interval was also
observed in some of the other simulations. In all cases, the
fluctuations occurred only over a short time period and at
different time intervals, and we conclude that there was no
systematic pattern in these fluctuations. As discussed above,
Asp91 forms hydrogen bonds with Tyr51, Tyr66, and His47
(Figure 7). The Asp91(OD2)-Tyr66(OH) distance is stable
throughout the simulations, whereas distances between Asp91
and Tyr51 or His47 show larger fluctuations than those observed
for Asp91-Tyr66. The distance change observed for His47-
Asp91 between ∼1.8 ns and ∼6.1 ns corresponds to a
reorientation of the histidine (180 degrees flip of the ring
system), and hence, a similar change is observed for His47-
(ND1)-S(C21). After ∼6.1 ns, the His residue flips back in
the orientation, where His47 can participate in hydrolysis. The
flip-flop of His47 was only observed in a few simulations and
encountered for shorter time periods than that shown in Figure
7. Substrate-Ca2+ distances are stable throughout the simula-
tions (Figure 8), and Ca2+ binds tightly to the phosphate oxygen.
Noticeable, the S(O22)-Gly29(HN) distance is stable but shows
relatively large fluctuations.
Figure 9. Average distances of His47(ND1)-S(C21) and S(O22)-Ca2+
extracted from the simulations of the different sPLA2-phospholipid
complexes. Averages and standard deviations (SDs) were calculated from
a series of simulations of a particular complex, where only the last 5 ns of
each simulation were used. The dashed-dotted and the dotted lines indicate
the upper and lower errors of His47(ND1)-S(C21) and S(O22)-Ca2+
respectively extracted from the sPLA2-1R simulations. Note, 1R is the
natural substrate for sPLA2.
acts as the attacking nucleophile.83-85 Figure 5 displays a
schematic representation of the catalytic mechanism indicating
that the calcium ion is coordinated by two carboxylate oxygen
atoms of Asp48, by a backbone oxygen of Gly29, and two
oxygens of the substrate.83,84 The latter two coordinations
involve a phosphate oxygen and the sn-2 carbonyl oxygen and
serve as stabilization of the substrate. His47, which abstracts a
proton from the incoming water molecule, is stabilized by Asp91
that additionally forms hydrogen bonds with Tyr51 and Tyr66.
The oxyanion hole that stabilizes the transition state after
nucleophilic attack is formed by the backbone HN group of
Gly29 assisted by the charge of the calcium ion. In view of
these interactions, we have monitored distances between Ca2+
and residues/atoms in sPLA2 (Asp48(OD1), Asp48(OD2), and
Gly29(O); Figure 6), distances that involve Asp91 or His47
(Asp91(OD1)-Tyr51(OH), Asp91(OD1)-Tyr66(OH), His47-
(HE2)-Asp91(OD1), and His47(ND1)-S(C21); Figure 7), and
Visualization of the time evolution of these distances is an
important tool to study the stability of these interactions but
provides no information of statistical significance. We have
therefore calculated averages over time and configurational
space (i.e., simulations) using the last 5 ns of each simulation.
Before calculating the averages, all distances were checked if
Figure 10. Trajectories of two water molecules (A and B) monitored in an sPLA2-1R simulation. Distances are calculated with respect to S (C21) (top)
and H47 (bottom).
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5459

A R T I C L E S
Peters et al.
Figure 5). If water molecules cannot reach this position, it might
explain the lack of sPLA2 activity.
Simulations - Nucleophilic Water Molecule. A prerequisite
for catalysis is that a water molecule can come in close vicinity
to His47(ND1) and S(C21). We therefore have monitored the
movement of water molecules, and trajectories of two water
molecules taken from a sPLA2-1R simulation are shown in
Figure 10. In order to quantify the positioning of the potential
catalytic water molecules, we have measured the probability of
water molecules being at a certain distance from both His47-
(ND1) and S(C21), hereafter referred to H-S. We counted the
number of water molecules that are within a certain distance
from H-S. The diffusion of water molecules is governed mainly
by thermal motion, and hence, water molecules randomly diffuse
into the catalytic site.82 Since this is a random event, relatively
large differences are observed between a series of simulations
of the same complex. We have therefore arbitrarily normalized
each result from a particular complex simulation with the water
count determined at 6.0 Å from H-S of the simulation. Average
relative water counts for each series are provided in the
Supporting Information (Table S2). Comparing 1R with 1S
(Figure 11A) reveals that the lack of sPLA2 activity toward 1S
is related to positioning of water molecules. Within the time
frame of the simulations of 1S, no water molecule could be
detected within 4 Å of H-S, and the relative water count within
4.5 Å is much less than that found for 1R (Figure 11A). Also
clear differences are seen for 1R, 2R, and 3S (Figure 11B).
Independent of the distance analyzed, the relative water count
follows the pattern 1R > 2R > 3S. Within 4 Å of H-S, the
relative water count decreases by ∼30% from 1R to 2R and an
additional ∼25% from 2R to 3S implying that the sPLA2 activity
toward these phospholipids follows the same trend: 1R > 2R
> 3S. This observation is in good agreement with the
experimental activity data.62,67
Figure 11. (A) Average relative water counts extracted from the 1R and
1S simulations. (B) Average relative water counts extracted from the 1R,
2R, and 3S simulations. Before calculating the averages, the data extracted
from each simulation were arbitrarily normalized by the water count at 6
Å of the simulation.
Conclusions
We have shown that unnatural phospholipid analogues where
the head group (normally located at the sn-3 position) was placed
at the 2-position are relatively good substrates for sPLA2.67 We
therefore decided to carry out molecular dynamics simulations
to understand on a molecular level why sPLA2 can tolerate
phospholipids with their head group in the 2-position and to
explain the different sPLA2 activity toward natural and unnatural
phospholipid analogues. In the simulations of sPLA2-2S and
sPLA2-3R, the distance between His47 and the carbonyl carbon
(C21; Figure 5) of the substrate is larger than that observed in
the simulations of sPLA2-1R (natural substrate). The phos-
pholipids cannot arrange in the right geometry in the binding
cleft due to the additional space and are therefore not hydrolyzed
by sPLA2. Although the phospholipids 1R, 2R, and 3S align
perfectly in the binding cleft, these lipids are hydrolyzed by
sPLA2 with different efficiencies. Our simulations revealed that
the difference observed in sPLA2 activity is caused by less
efficient access of water molecules to the active site. During
the hydrolysis, a water molecule that acts as the nucleophile in
the reaction is positioned between His47 and C21 (Figure 5).
We have monitored the number of water molecules that enter
the region between His and C21 for the different sPLA2-
phospholipid complexes and found that the probability of a water
molecule reaching the correct position such that hydrolysis can
occur is reduced for the unnatural phospholipids. The relative
they were stable throughout the simulations. Overall, the
distances are very stable, and differences in average distance
between the different complexes are within the statistical
uncertainties (a summary is provided as Supporting Information;
Table S1). Relatively large variations are observed for His47-
(ND1)-S(C21) and S(O22)-Ca2+ distances in the different
complexes (Figure 9). Particularly, 2S and 3R have on average
a larger His47(ND1)-S(C21) distance than that in the complex
with the natural substrate (1R). This could explain the lack of
sPLA2 activity toward these phospholipids. The phospholipids
are not perfectly aligned in the binding cleft of sPLA2 to allow
for hydrolysis to occur. Surprisingly, the distances monitored
in sPLA2-1S are (within the statistical uncertainties) the same
as those in sPLA2-1R (Figure 9, Table S1). The reasoning that
1S is not a substrate for sPLA2 (enzymatic hydrolysis is a
stereospecific reaction) must depend on other factors. As
discussed above, the enzymatic reaction requires that a water
molecule acting as a nucleophile is positioned correctly relative
to His 47(ND1) and the carbonyl carbon of the substrate (C21,
(84) Janssen, M. J. W.; van de Wiel, W. A. E. C.; Beiboer, S. H. W.; van
Kampen, M. D.; Verheij, H. M.; Slotboom, A. J.; Egmond, M. R. Protein
Eng. 1999, 12, 497-503.
(85) Sekar, K.; Yu, B.-Z.; Rogers, J.; Lutton, J.; Liu, X.; Chen, X.; Tsai, M.-
D.; Jain, M. K.; Sundaralingam, M. Biochemistry 1997, 36, 3104-3114.
(86) Zhou, F.; Schulten, K. Proteins 1996, 25, 12-27.
9
5460 J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007

PLA2 Substrate Specificity
A R T I C L E S
water count follows 1R > 2R > 3S, which is in excellent
Supporting Information Available: Root-mean displacement
as function of simulation time; time-averaged fluctuations of
residues, averaged distances, and averaged relative water count.
This material is available free of charge via the Internet at
http://pubs.acs.org.
agreement with experimentally determined activity data.62,67
Acknowledgment. Financial support from the Danish Na-
tional Research foundation via a grant to the MEMPHYS-Center
for Biomembrane Physics is acknowledged. Simulations were
performed at the Danish Center for Scientific Computing at the
University of Southern Denmark.
JA067755B
9
J. AM. CHEM. SOC. VOL. 129, NO. 17, 2007 5461