Tài liệu Báo cáo khoa học: Insights into the structure of plant a-type phospholipase D Susanne Stumpe, Stephan Konig and Renate Ulbrich-Hofmann ¨ ppt

Insights into the structure of plant a-type phospholipase D
Susanne Stumpe, Stephan Ko
¨
nig and Renate Ulbrich-Hofmann
Martin-Luther University Halle-Wittenberg, Institute of Biochemistry and Biotechnology, Halle (Saale), Germany
Phospholipase D (PLD; EC 3.1.4.4), which hydrolyzes
phospholipids to phosphatidic acid and the head group
alcohol, occurs in plants, microorganisms and animals.
In plants, PLD isoenzymes play multiple regulatory
roles in diverse processes, including abscisic acid signa-
ling, programmed cell death, root hair patterning, root
growth, freezing tolerance, and other stress responses
[1]. There is growing evidence for phosphatidic acid
being an intracellular lipid messenger in plants as well
as in mammals [2,3]. Most PLDs also catalyze the
transphosphatidylation reaction in which the phos-
phatidyl moiety is transferred to a suitable alcohol. In
biotechnology, this phospholipid-transforming reaction
has been exploited on a laboratory and industrial scale
[4].
In plants, multiple PLD genes encoding isoforms
with distinct regulatory and catalytic properties have
been identified [5,6]. The most prevalent isoenzyme,
which is responsible for the common PLD activity
observed in leaves or seeds of plants, is the a-type
PLD (PLDa). The conventional PLDa does not
require phosphoinositides for activity when assayed
at millimolar levels of calcium ions. It exhibits opti-
mum activity at pH values between 5 and 6, and
nonphysiologic high calcium concentrations between
30 and 100 mm [3,7,8]. In contrast, the PLD isoen-

zymes of the b, c, d and e types from Arabidopsis
show highest activity at micromolar calcium concen-
trations, and require the presence of phosphatidy-
linositol 4,5-bisphosphate (PtdInsP
2
) [5]. The activity
Keywords
calcium binding; phospholipase D; small-
angle X-ray scattering; stability; structure
Correspondence
R. Ulbrich-Hofmann, Martin-Luther
University Halle-Wittenberg, Institute
of Biochemistry and Biotechnology,
Kurt-Mothes-Str. 3, 06120 Halle (Saale),
Germany
Fax: +49 345 5527303
Tel: +49 345 5524864
E-mail: Renate.Ulbrich-Hofmann@
biochemtech.uni-halle.de
(Received 29 December 2006, revised 23
February 2007, accepted 20 March 2007)
doi:10.1111/j.1742-4658.2007.05798.x
Phospholipases D play an important role in the regulation of cellular pro-
cesses in plants and mammals. Moreover, they are an essential tool in the
synthesis of phospholipids and phospholipid analogs. Knowledge of phos-
pholipase D structures, however, is widely restricted to sequence data. The
only known tertiary structure of a microbial phospholipase D cannot be
generalized to eukaryotic phospholipases D. In this study, the isoenzyme
form of phospholipase D from white cabbage (PLDa2), which is the most
widely used plant phospholipase D in biocatalytic applications, has been

characterized by small-angle X-ray scattering, UV-absorption, CD and
fluorescence spectroscopy to yield the first insights into its secondary and
tertiary structure. The structural model derived from small-angle X-ray
scattering measurements reveals a barrel-shaped monomer with loosely
structured tops. The far-UV CD-spectroscopic data indicate the presence
of a-helical as well as b-structural elements, with the latter being dominant.
The fluorescence and near-UV CD spectra point to tight packing of the
aromatic residues in the core of the protein. From the near-UV CD signals
and activity data as a function of the calcium ion concentration, two bind-
ing events characterized by dissociation constants in the ranges of 0.1 mm
and 10–20 mm can be confirmed. The stability of PLDa2 proved to be sub-
stantially reduced in the presence of calcium ions, with salt-induced aggre-
gation being the main reason for irreversible inactivation.
Abbreviations
PLD, phospholipase D; PLDa, a-type phospholipase D; PLDa2, isoenzyme form of phospholipase D from white cabbage; PpNp,
phosphatidyl-p-nitrophenol; PtdIns(4,5)P
2
, phosphatidylinositol 4,5-bisphosphate; SAXS, small-angle X-ray scattering.
2630 FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS
of plant PLDf is independent of calcium ions, but
the enzyme requires PtdInsP
2
to selectively hydrolyze
phosphatidylcholine. With the exception of PLDf,
which contains a Phox (PX) and a Pleckstrin (PH)
homology domain in the N-terminal region, all the
other PLD isoenzymes possess a calcium- and phos-
pholipid-binding domain (C2 domain) at the N-ter-
minus [5].
From white cabbage, the most traditional source of

PLD for biocatalytic phospholipid transformations,
two PLDs of the a type (PLD1 and PLD2) were
expressed in Escherichia coli and purified [8]. The
amino acid sequence of PLD2, called PLDa2 in the
following, is identical with that of PLD isolated from
cabbage leaves [9]. Figure 1A shows a schematic pic-
ture of the primary structure of PLDa2 with the C2
domain and four conserved regions (I–IV) specific for
the PLD superfamily, where regions II and IV are two
copies of the HxKx
4
Dx
6
GSxN (HKD) motif [10,11].
The H, K and D residues of theses motifs are abso-
lutely conserved [11]. As derived from the only crystal
structure of a PLD, the prokaryotic PLD from Strep-
tomyces sp. strain PMF, the two HKD motifs form
one catalytic site [12]. For comparison, Fig. 1B,C
shows schemes of the primary structures for a mam-
malian and a microbial PLD, demonstrating that the
PLD-specific regions I–IV are present in all members
of the PLD superfamily. However, there is low similar-
ity in the remaining parts of the molecules. Most mam-
malian PLDs contain a PX and a PH domain instead
of the C2 domain in plants, whereas microbial PLDs
lack such probably regulatory domains.
Despite the increasing amount of primary structural
data for PLDs, there is almost no information on the
secondary and tertiary structures of plant PLDs.

Although Abergel et al. reported the crystallization of
the cowpea PLD [13], no crystal structure has been
published so far. Even reports on purified plant PLDs
are extremely rare. Obviously, the biochemical and
structural characterization of plant PLD enzymes has
been hindered by their instability during and after
purification from plant tissues [14,15] or the low yields
in recombinant protein production [16], respectively.
These difficulties were recently overcome by using
Ca
2+
-mediated hydrophobic chromatography [17]
combined with recombinant PLD production in E. coli
[8]. This purification method was also successfully
applied to PLDa enzymes from soybean [18], cowpea
[19], and poppy [20,21].
In the present study, the first insights into the secon-
dary and tertiary structure of PLD a2 from cabbage
are given. The surface structure of the enzyme was
investigated by small-angle X-ray scattering (SAXS).
Further structural information was obtained by the
spectroscopic characterization of PLDa2 in the native
state. The high instability of purified plant PLDs, as
reported above, was shown to result from salt effects
that can be eliminated in low-buffered solutions at
room temperature. From spectroscopic and activity
measurements, two classes of calcium-binding events,
with dissociation constants varying by at least two
orders of magnitude, were derived.
Results

PLDa2 preparation
PLDa2 was produced according to Scha
¨
ffner et al.
[8] with slight modifications as described in Experi-
mental procedures, which improved the yield of active
enzyme. The amount of homogeneous PLDa2
obtained was 3.5–5.0 mg per liter of cell culture. The
specific activity towards the synthetic substrate phos-
phatidyl-p-nitrophenol (PpNp) was 10.2–14.7 UÆmL
)1
,
which is consistent with the activity of PLDa2 in pre-
vious preparations [8].
A
C
B
Fig. 1. Scheme of the primary structure of
PLDs. (A) PLDa2 from cabbage (P55939).
(B) PLD1 from human (Q13393). (C) PLD
from Streptomyces sp. (1V0YA). The
N-terminal C2 domain (A), the PX and PH
domains (B) and the four PLD specific con-
served regions (I–IV) are marked. Regions II
and IV represent the two HKD motifs form-
ing the active site. In (A), the positions of
the 30 tyrosine and 15 tryptophan residues
are marked by stars.
S. Stumpe et al. Structure of plant phospholipase D
FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS 2631

Structural information from SAXS
Scattering measurements with synchrotron radiation
were performed at pH 7.0 with 2 and 4.7 mgÆmL
)1
PLDa2. A typical SAXS profile of PLDa2 is shown
for 4.7 mgÆmL
)1
in Fig. 2A. From scattering intensities
I(0), a molecular mass of 97 kDa was calculated (using
BSA as standard), which is in accordance with
a monomer of PLDa2. Correspondingly, no aggre-
gates were detected at a PLDa2 concentration of
0.05–0.2 mgÆmL
)1
in size exclusion chromatography.
The molecular mass deduced from the amino acid
sequence is 91.9 kDa, whereas the molecular mass
determined by size exclusion chromatography was
85.8 kDa (data not shown). The radius of gyration R
g
(2.87 ± 0.02 nm) extracted by gnom [22] from the
scattering data was in good agreement with the R
g
value estimated by primus [23] from the Guinier plots
(2.89 ± 0.01 nm) and is in the typical range for a
spherical particle. The bell-shaped distance distribution
function illustrates the quality of the experimental data
(Fig. 2B).
Reconstruction of the overall shape of the PLDa2
from the X-ray scattering data was achieved by the

ab initio modeling program dammin [24]. The superpo-
sition of the 10 calculated models (Fig. 3) shows a
longish molecule with loosely structured apical and
basal regions. The lid-like region (Fig. 3, left) can be
construed as a separate domain, but was not explicitly
visible in all individual models. The Porod volume
(volume of the hydrated solute particle) of the low-
resolution models averages 126.7 ± 3.5 nm
3
.
Structural information from spectroscopy
UV absorption spectra (Fig. 4A) show a maximum at
280 nm, with a slight shoulder at 295 nm assigned
to the tryptophan residues. The absence of absorption
in the UV region > 310 nm, characteristic for light
scattering of high molecular mass species, provides no
A
B
Fig. 2. SAXS profile (A) and distance distribution function (B) of
PLDa2. (A) Experimental data are indicated by open circles. The
solid line shows the scattering curve fitted by the program
GNOM
[22]. The PLDa2 concentration was 4.7 mgÆmL
)1
.
Fig. 3. Ab initio low-resolution structure model of PLDa2 calculated from the SAXS pattern. The balls represent the dummy atoms used in
the simulated annealing procedure of program
DAMMIN [24] to restore the models. The model resulting from superposition of 10 individual
models is shown in three different views obtained by 90° rotation around the y-axis (middle) and the z-axis (right).
Structure of plant phospholipase D S. Stumpe et al.

2632 FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS
evidence for aggregates, and reconfirms the size exclu-
sion chromatography results.
The fluorescence spectra have an emission wave-
length maximum at 334 nm and refer to a hydrophobic
environment for most or all tryptophan side chains
(Fig. 4B). The similar shape and fluorescence emission
maxima of the spectra at both excitation wavelengths
(278 and 295 nm) indicate a complete resonance energy
transfer from the tyrosine to the tryptophan residues.
The secondary structure of PLDa2 was investigated
by far-UV CD spectroscopy (Fig. 4C). The far-UV
CD spectrum of PLDa2 exhibits: (a) a sharp maximum
at 192 nm; and (b) a wide minimum between 208
and 220 nm. The a-helix and b-strand contents of the
protein were calculated with the online programs
dichroweb [25] and k2d [26], respectively. With the
program k2d, the b-sheet content (0.45) was higher
than the a-helix content (0.08), whereas the program
dichroweb computed secondary structure contents
with 0.20 a-helix, 0.28 b-sheet, and 0.21 turn confor-
mation. Both methods agreed in indicating more
b-sheet than a-helix content.
The near-UV CD spectrum of native PLDa2 (250–
300 nm, Fig. 4D), describing the chiral environment of
the aromatic amino acid side chains, has a defined
structure that presents two sharp minima at 288 and
295 nm, and two maxima at 274–283 nm (wide) and
290 nm (sharp).
Storage stability and aggregation propensity

The instability of the purified enzyme described in the
literature [14,15] prompted us to analyze the inactiva-
tion of PLDa2 under selected conditions. Storage at
lower temperatures (0–10 °C) resulted in faster inacti-
vation of the enzyme than at room temperature
(23 °C). Under physiologic conditions (pH 7.0 and an
ionic strength of 150 mm, adjusted with sodium chlor-
ide), as well as under conditions where the enzyme is
most active (pH 5.5 and 40–100 mm calcium chloride),
PLDa2 was more rapidly inactivated than in the
absence of higher salt concentrations. In Table 1, the
rate constants of irreversible inactivation, which fol-
lowed first-order kinetics, in the absence and presence
of NaCl and CaCl
2
at pH 5.5 and 23 °C are compared.
The results show that the instability of PLDa2is
B
C
A
D
Fig. 4. Spectra of native PLDa2. (A) UV
absorbance spectrum. (B) Fluorescence
emission spectra with excitation at 278 nm
(black) and 295 nm (gray). (C) Far-UV CD
spectrum. (D) Near-UV CD spectra in the
absence (black) and the presence of
100 m
M CaCl
2

(gray). All spectra were
obtained in 50 m
M sodium acetate buffer
(pH 5.5) at 20 °C. The PLDa2 concentrations
were 253 lgÆmL
)1
(A), 25 lgÆmL
)1
(B), and
1mgÆmL
)1
(C, D), respectively.
Table 1. Influence of salts on the observed first-order rate con-
stants (k
obs
) of PLDa2 inactivation and fluorescence decrease.
PLDa2 was incubated in 50 m
M sodium acetate buffer (pH 5.5) at
23 °C. To follow the inactivation, PLDa2 (280 lgÆmL
)1
) was incuba-
ted for 5 min to 36 days, and the residual activity was measured
after dilution to 9.3 lgÆmL
)1
as described in Experimental proce-
dures. The fluorescence intensity of PLDa2 (25 lgÆmL
)1
) was
measured at an excitation wavelength of 278 nm and an emission
wavelength of 335 nm. NM, not measurable.

Additive
k
obs
(10
)7
Æs
)1
)
Inactivation
Fluorescence
decrease
– 2.5 21
10 m
M EDTA NM NM
120 m
M NaCl 6.5 31
300 m
M NaCl 21 120
40 m
M CaCl
2
6.5 33
100 m
M CaCl
2
29 120
S. Stumpe et al. Structure of plant phospholipase D
FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS 2633
caused by high ionic strength rather than by specific
ionic effects. At comparable ionic strengths of sodium

and calcium chloride, the observed kinetic constants
were nearly identical. Interestingly, however, the inacti-
vation could be further decreased by the addition of
EDTA; inactivation constants were no longer measur-
able. At room temperature, the loss of activity in
slightly buffered solutions (10 mm Pipes, pH 7.0, or
50 mm sodium acetate, pH 5.5) with 10 mm EDTA
amounted to only 17% after 5 months.
When we looked for spectroscopic alterations in
the course of inactivation, neither sodium chloride nor
calcium chloride was found to induce spectroscopic
changes in the absorption, fluorescence or far-UV CD
spectra (data not shown). Interestingly, near-UV CD
spectroscopy revealed changes in the presence of cal-
cium chloride (Fig. 4D), but not in the presence of
sodium chloride. Therefore, the spectroscopic altera-
tions could stem from specific calcium binding rather
than from nonspecific salt effects (see next paragraph).
Although no shift of the fluorescence emission max-
ima was measurable, the fluorescence intensity decre-
ased with increasing incubation time (data not shown).
Also, the decrease of the fluorescence signal followed
a first-order reaction, and allowed us to determine
rate constants (Table 1). These constants, determined
at about 10-fold lower protein concentration than
the measurements of inactivation, show the same
trends with respect to the effects of NaCl, CaCl
2
and
EDTA as the inactivation constants. As shown

by SDS ⁄ PAGE (Fig. 5), the decrease in fluorescence
intensity was caused by precipitation of PLDa2, which
is faster in the presence of salts.
Calcium binding
Whereas calcium ions did not change the absorption,
fluorescence and far-UV CD spectra of native PLDa2,
the near-UV CD signal in the range 258–292 nm was
increased in the presence of calcium chloride, with a
maximum at 280 nm (Fig. 4D). The increase in near-
UV CD intensity was specific for calcium ions, as
neither sodium nor magnesium ions showed any com-
parable effect. The increase in the near-UV CD signal
at 280 nm was dependent on the concentration of
calcium ions in a hyperbolic fashion (Fig. 6A). The
data could be fitted well using a double hyperbolic
function yielding two dissociation constants, K
D1
¼
10.24 mm and K
D2
¼ 0.123 mm. Two binding events
are also evident in a Scatchard-like plot [27], where the
relative change in the near-UV CD signal (DF) repre-
sents the amount of the Ca
2+
–PLDa2 complexes
(Fig. 6B). Calcium ions are crucial for PLDa2 activity,
and could not be replaced by other ions such as mag-
nesium ions. At concentrations of calcium ions
<1 mm, no significant activity could be detected. Up

to 100 mm, PLDa2 activity increased with increasing
concentration of calcium ions, independent of the
counterion, chloride or acetate (Fig. 7A). At calcium
ion concentrations above 100 mm, the PLDa2 activity
dropped again, and the decrease was steeper with cal-
cium acetate than with calcium chloride. Interestingly,
the activity data in the range 0.01–100 mm CaCl
2
could be fitted in similar way as the near-UV CD data,
using a double hyperbolic function (Fig. 7B). The
resulting dissociation constants were K
D1
¼ 17.38 mm
and K
D2
¼ 0.073 mm. A modified Scatchard plot
of these data, with DA being the change in relative
activity (Fig. 7C), shows two linear parts, indicating
that at least two separate calcium-binding events affect
PLDa2 activity.
A
B
Fig. 5. Kinetics of PLDa2 precipitation followed by SDS ⁄ PAGE.
PLDa2 (280 lgÆmL
)1
) was incubated in 50 mM sodium acetate buf-
fer (pH 5.5) in the absence (A) and the presence of 300 m
M NaCl
(B) at 23 °C. After 0 min, 5 min, 30 min, 1 h, 2 h, 4 h, 8 h, 1 day,
2 days, 3 days and 4 days (lanes 1–11), samples were analyzed by

SDS ⁄ PAGE. Lane 12 shows the protein band in the precipitate
after 1 week of incubation.
B
A
Fig. 6. Calcium binding measured by near-UV CD spectroscopy. (A)
Direct plot. The solid line shows the double hyperbolic fit. (B) Modi-
fied Scatchard plot. The CD signals, measured at different concen-
trations of CaCl
2
, were taken at 280 nm and averaged over 5 min.
The PLDa2 concentration was 1 mgÆmL
)1
. DF corresponds to the
change in the CD signal [Q]
MRW
(degÆcm
2
Ædmol
)1
) at 280 nm.
Structure of plant phospholipase D S. Stumpe et al.
2634 FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS
Discussion
PLDa2 is a barrel-shaped monomer with loosely
structured tops
Despite the great number of plant PLDs identified at
the DNA or cDNA level, only a few of these enzymes
have been characterized in purified, isolated form. Cor-
respondingly, there is almost no information on the
secondary and tertiary structures of plant PLDs. Simi-

larly, structural information on other eukaryotic PLDs
is rare.
The two members of the PLD superfamily whose
crystal structures have been elucidated [28,29] are
much smaller in size, so that knowledge of their struc-
tures cannot be generalized to plant PLDs. In this art-
icle, the first information on the spatial structure of a
plant a-type PLD has been obtained on the basis of
recombinantly produced PLDa2 from cabbage. SAXS
analysis (Fig. 2) and analytic size exclusion chromato-
graphy indicate unequivocally that native PLDa2 from
cabbage is a monomeric protein. The surface structure
of PLDa2 indicates a longish molecule with loosely
structured regions at the two ends of the molecule
(Fig. 3). An assignment of these relaxed structures to
sequence regions of PLDa2 has not yet been possible.
According to the volume estimation on the basis of the
partial specific volume of a protein (0.73 mgÆg
)1
) [30],
the C2 domain comprising residues A2 to E153 should
occupy 16% of the molecular volume, and therefore
seems too large to represent the lid-like structure at
the top of the left view in Fig. 3. We speculate that the
loosely structured region on the opposite side belongs
to the C2 domain (141 amino acids) with its extended
loops, whereas the lid-like region at the top represents
another flexible part of the structure. Assuming that
the core protein containing the two essential HKD
motifs (Fig. 1) is similar in PLDa2 from cabbage (812

amino acids) and in PLD from Streptomyces sp. (506
amino acids) with the known crystal structure [29], the
C-terminal part of the plant enzyme is longer by
approximately 60 amino acids and might form this
part. A PtdInsP
2
-binding domain, which has recently
been desribed as a separate folding unit conserved in
eukaryotic PLDs and comprises about 50 amino acids
between the two HKD motifs [31], would also fit this
lid-like region.
The loosely structured parts of the molecule and the
deviation from a spherical shape may be the reason
why the experimentally determined Porod volume of
PLDa2 (126.7 ± 3.5 nm
3
) is slightly larger than the
molecular volume (111.45 nm
3
) calculated from the
partial specific volume of a protein [30]. This result is
in accordance with the smaller volume (99 nm
3
)
deduced from the radius of gyration (R
g
: 2.9 nm),
which considers the distances between scattering mas-
ses of the molecule.
Spectroscopic properties of PLDa2 reflect an

ordered tertiary structure with a high content of
b-structure
The first UV-absorption (Fig. 4A), fluorescence
(Fig. 4B) and CD spectra (Fig. 4C,D) of a plant PLD
reflect the properties of a common protein. Obviously,
tyrosine, and particularly tryptophan, residues (Fig. 1)
dominate the absorption spectra of PLDa2 from cab-
bage, with a maximum at 280 nm (Fig. 4A). The fluor-
escence spectra of PLDa2 with a maximum at 334 nm
(Fig. 4B) indicate a hydrophobic environment for most
or possibly all tryptophan residues. Therefore, trypto-
phan residues are probably mainly located in the core
of the protein. The tyrosine residues contribute to the
fluorescence spectra by fluorescence resonance energy
A
C
B
Fig. 7. Influence of the calcium concentration on PLDa2 activity. (A) Relative activity as function of calcium chloride (black) and calcium acet-
ate (gray) concentration, respectively. (B) Double hyperbolic plot of the data from (A) in the range 0–100 m
M Ca
2+
. (C) Modified Scatchard
plot of the data from (B). DA corresponds to the change in relative activity. The PLDa2 activity was measured against PpNp in sodium acet-
ate buffer (pH 5.5) at 30 °C.
S. Stumpe et al. Structure of plant phospholipase D
FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS 2635
transfer. This energy transfer seems to be very effective
in PLDa2, because no separate tyrosine fluorescence is
detectable.
The far-UV CD spectrum of PLDa2 with the maxi-

mum at 192 nm and the broad minimum at 208–
220 nm (Fig. 4C) refers to both a-helix and b-sheet
conformations. As the signal of a-helices is more
intense [32], the spectrum suggests that b-sheet
structures dominate. This assumption was confirmed
by the calculation of the a-helix and b-strand contents
of the protein with the online programs dichroweb
[25] and k2d [26], respectively.
The well-structured near-UV CD spectrum of PLDa2
(Fig. 4C) points to a tertiary structure with tight pack-
ing of the aromatic side chains in an asymmetric envi-
ronment. As all eight cysteine residues of PLDa2 should
be half-cysteines [33], no influence of disulfide bonds on
the near-UV CD spectrum must be taken into account.
The relatively low signal intensity can probably be
attributed to the compensation of positive and negative
signals of the individual aromatic residues.
PLDa2 is stable at low ionic strengths
The low stability of purified cabbage PLD was shown
to originate from a high aggregation propensity of the
enzyme under physiologic conditions. Inactivation and
precipitation of PLDa2, the latter being followed by
SDS ⁄ PAGE or decreasing fluorescence intensity, were
accelerated with growing salt concentrations (Table 1,
Fig. 5). In the comparison of the presence of NaCl
and of CaCl
2
, the ionic strength was more important
for inactivation or precipitation than the individual
ions. Hydrophobic interactions, which are known to

be favored by rising salt concentrations, are probably
the reason for this tendency. The fact that EDTA
effects a marked stabilization (Table 1), however, indi-
cates an additional destabilizing effect by calcium ions.
Destabilization of PLDa2 by calcium ions was also
indicated by the proteolytic sensitivity of PLD a2 [34].
Calcium-induced decrease of stability and biological
activity due to facilitated aggregation was also repor-
ted for a-crystallin [35].
Although the incubation with salts in the inactiva-
tion experiments was performed at higher PLDa2 con-
centrations than in the fluorescence measurements, the
observed inactivation was slower than the observed
fluorescence decrease due to protein precipitation. This
discrepancy could be due to the different experimental
conditions. Whereas the fluorescence decreases were
measured in situ, the remaining activities were deter-
mined after dilution, and therefore partial resolubiliza-
tion of precipitates is possible. We assume that the
precipitation is followed by (small) structural altera-
tions in a consecutive aggregation reaction that cause
irreversible inactivation.
It is not clear whether the destabilizing effects of
calcium ions and other salts have any physiologic rele-
vance. However, the finding that pure plant PLDs may
be more stable in the absence of calcium and other ions
at room temperature than in their presence and at lower
temperatures is of considerable practical importance.
PLDa2 binds calcium ions at two affinity levels
CD spectroscopy in the near UV-region proved to be

the only spectroscopic method able to detect specific
binding of calcium ions to PLDa2 (Fig. 4D). The sig-
nal increases in this spectrum caused by calcium ions
indicate stiffening of the PLD a2 molecule. Binding
curves obtained with this method (Fig. 6A) are similar
to the curves of PLDa2 activity as a function of cal-
cium ion concentration (Fig. 7B), showing that the
well-known effect of calcium ions on activity closely
correlates with conformational changes of the enzyme.
These results unambiguously show that activation of
PLDa2 by calcium ions is due to the binding of cal-
cium ions to the enzyme, and not (or not only) to bet-
ter structuring of the substrate aggregates.
From both the CD and the activity data, two differ-
ent calcium-binding events can be derived. The corres-
ponding dissociation constants differ by about two
orders of magnitude. Even the higher affinity (with the
dissociation constant in the range of 0.1 mm) is still
relatively low for biospecific interactions. We assume
that calcium ions are involved in enzyme activation by
guiding the protein to the membrane, mediating sub-
strate binding, or adjusting the charge distribution at
the active site. Unfortunately, the number of calcium
ions bound to the enzyme cannot be deduced from
these data. The binding event with the dissociation
constant in the range of 0.1 mm might be associated
with calcium binding to the C2 domain. For the sepa-
rately expressed C2 domain of PLDa from Arabidopsis
thaliana, one to three low-affinity calcium-binding sites
with dissociation constants in the range of 0.5 mm

were estimated from isothermal titration calorimetry
[36]. The C2 domains of other proteins showed cal-
cium-binding affinities in the range of 1–50 lm [37,38].
As the C2 domain of PLD a2 from cabbage lacks two
conserved residues that are probably responsible for
the calcium binding [8], we hypothesize that this bind-
ing event takes place at the C2 domain.
A second calcium-binding site of lower affinity at
the C2 domain (corresponding to the dissociation
constant in the range 10–20 mm) cannot be excluded. It
Structure of plant phospholipase D S. Stumpe et al.
2636 FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS
is more likely, however, that the calcium-binding event
of lower affinity is attributable to the catalytic site of
PLDa2, as found for PLDb from A. thaliana [39].
The decrease in PLD activity at concentrations
above the optimal calcium concentration (100 mm),
as demonstrated in Fig. 7A, is scarcely reported in the
literature [7]. In contrast to the activation effect of
calcium ions, inactivation at high calcium ion concen-
trations is dependent on the counterion. In contrast
to the stabilizing effects of salts according to the
Hofmeister series, the chloride salt is less destabilizing
than the acetate salt. As discussed above, the inactiva-
tion by salts is connected with aggregation and precipi-
tation of PLDa2. Obviously, the anion is specifically
involved in this process.
In summary, we conclude that calcium ions bind
specifically at two sites of PLDa2 and activate the
enzyme in two steps. As discussed above, the resulting

conformational changes destabilize the molecule. How-
ever, at higher salt concentrations, precipitation
occurs, and destabilization by calcium ions is therefore
no longer specific.
Experimental procedures
Materials
Bactotryptone and yeast extract were from Difco (Detroit,
MI, USA). All other chemicals were the purest ones com-
mercially available.
Production of PLDa2
The E. coli strain BL21 (DE3) containing plasmids pUBS520
and pld2pRSET5a [8] was grown without preculture in a
200 mL culture of 2 · YT medium (2% bactotryptone, 1%
yeast extract, and 1% NaCl, pH 7.0) with 100 lgÆmL
)1
amp-
icillin and 50 lgÆmL
)1
kanamycin at 30 °C for 15 h. At an
absorbance of 1–2 at 600 nm, the temperature was shifted to
15 °C and the cells were grown to maximum cell density (24–
48 h) without induction. Cells were harvested by centrifuga-
tion at 4 °C and 6000 g for 10 min (Avanti J-25 centrifuge,
Beckman Coulter, JA-10 rotor), and stored at ) 20 °C.
Purification of PLDa2
The pellet from a 1.2 L culture was suspended in 37 mL of
buffer (30 mm Pipes, pH 6.2, 10 mm EDTA), and the cells
were disrupted by high-pressure homogenization (APV Ho-
mogeniser; Gaulin, Lu
¨

beck, Germany). The cell debris was
separated by centrifugation at 4 °C and 48 000 g for
20 min (Avanti J-25 centrifuge, JA-30.50 rotor). Calcium
chloride was added to the supernatant (final concentration
50 mm), and after centrifugation (at 4 °C, 4800 g, 10 min,
Rotofix 32 centrifuge, Hettlich, 1620A rotor), the protein
solution was loaded onto an Octylsepharose column
(Amersham Biosciences, Piscataway, NJ, USA) equilibrated
with 30 mm Pipes (pH 6.2) containing 50 mm CaCl
2
. The
column was carefully washed with equilibration buffer, and
then the protein was eluted with 0.1 mm EDTA in 5 mm
Pipes buffer (pH 6.2). The PLDa2-containing fractions
were pooled and Tris ⁄ HCl buffer (pH 7.5) was added to a
final concentration of 20 mm. After filtration (0.2 lm;
WiCom, Heppenheim, Germany), the solution was applied
to a Source 15Q column (Amersham Biosciences). Elution
was performed using a linear gradient of 50 mL of 15–35%
1 m NaCl in Tris ⁄ HCl buffer (pH 7.5).
The fractions with highest catalytic activity and homo-
geneity in SDS ⁄ PAGE were pooled, dialyzed twice (cut-off
20 kDa; Roth, Karlsruhe, Germany) against a 100-fold vol-
ume of 10 mm Pipes buffer (pH 7.0), and stored at ) 20 °C.
Determination of protein concentration
The concentration of PLDa2 in the range between
75 lgÆmL
)1
and 1 mgÆmL
)1

was determined spectrophoto-
metrically, using the molar extinction coefficient e
280 nm
¼
123 720 m
)1
Æcm
)1
[40]. Other PLDa2 concentrations were
measured by the BCA assay (Pierce, Rockford, IL, USA)
with BSA as standard.
SDS ⁄ PAGE
SDS ⁄ PAGE was performed according to the method of
Laemmli [41]. Gels [10% (w ⁄ v) acrylamide] were stained
with Coomassie Brilliant Blue G250 and quantified by
densitometric evaluation at 595 nm (CD60; Desaga, Darm-
stadt, Germany).
Small angle X-ray solution scattering with
synchrotron radiation
Data were collected with the EMBL beamline X33 at the
DORIS storage ring, Desy, Hamburg. Measurements were
performed at a camera length of 2 m, using a Mar345
image plate detector (Marresearch, Norderstedt, Germany),
at a wavelength of 1.5 A
˚
, a temperature of 12 °C, and
PLDa2 concentrations between 2 and 4.7 mgÆmL
)1
in
10 mm Pipes buffer (pH 7.0). Calibration of the momentum

transfer axis (s) was done using collagen or tripalmitin as
standards. The experimental data were normalized to the
intensity of the incident beam and corrected for the detec-
tor response, the buffer scattering was substracted, and the
statistical errors were calculated using the program primus
[23]. To obtain the forward scattering intensity I(0) and the
radius of gyration (R
g
), the data were processed with the
program gnom [22]. The molecular mass was calculated
S. Stumpe et al. Structure of plant phospholipase D
FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS 2637
from the ratio of the forward scattering intensity of the
sample and that of the molecular mass standard BSA. The
ab initio deconvolution of the SAXS profiles to restore a
low-resolution shape of PLDa2 from the experimental data
was executed using the program dammin [24]. The final
model was computed by superposition of 10 individual
models using the program damaver [42].
Size exclusion chromatography
A Superdex 200 HR 10 ⁄ 30 FPLC column (Amersham Bio-
sciences) was equilibrated with 10 m m Pipes buffer (pH 7.0)
and 140 mm NaCl. Proteins were eluted with the same buf-
fer at a flow rate of 0.5 mLÆmin
)1
at 4 °C, and detected by
the protein absorbance at 280 nm. Aldolase (150 kDa),
ovalbumin (45 kDa), chymotrypsinogen A (25 kDa), and
cytochrome c (12.3 kDa) (Serva, Heidelberg, Germany)
were used as molecular mass standards.

Absorption and CD spectroscopy
A Jasco V-560 spectrophotometer (Jasco, Gross-Umstadt,
Germany) and a 10 mm quartz cuvette were used in UV
absorption measurements. CD spectra were obtained using
a Jasco J-810 spectropolarimeter. The protein spectra were
measured in 10 mm (near-UV) and 0.1 mm (far-UV) quartz
cuvettes by scanning at 20 nmÆmin
)1
with a resolution of
0.1 nm, a response time of 1 s, a bandwidth of 1 nm, and a
temperature of 20 °C. An average of 10 scans was recorded
and corrected by subtracting the baseline spectrum of the
buffer. The CD signal was converted to molar ellipticity
[Q]
MRW
(degÆcm
2
Ædmol
)1
) [43].
The ellipticity at 280 nm was used for monitoring cal-
cium binding. The CD signal was recorded over 5 min.
Then, calcium ions were added stepwise, and the CD signal
was recorded again. The measured signal was corrected for
the protein dilution by calcium chloride addition.
Fluorescence spectroscopy
Fluorescence experiments were performed at 20 °C with
a FluoroMax-2 spectrofluorimeter (Horiba Jobin Yvon,
Munich, Germany) using 10 · 4 mm fluorescence quartz
cuvettes. PLDa2 was excited at 278 nm (excitement of

tyrosine and tryptophan residues) or at 295 nm (excitement
of tryptophan residues only). The slit width was 5 nm for
excitation and emission. The signal was acquired with a res-
ponse time of 1 s, and the wavelength increment was 1 nm.
The fluorescence signal of the blank buffer was substracted.
Enzyme assays
The hydrolytic activity of PLDa2 was determined by meas-
uring the p-nitrophenol release from PpNp (synthesized as
described by D’Arrigo et al. [44]) at 405 nm [8]. In general,
PLDa2 was incubated in 220 lLof65mm sodium acetate
buffer (pH 5.5) containing 50 mm CaCl
2
at 30 °C. With the
addition of the substrate solution [10 mm PpNp, 5% Triton
X-100 (v ⁄ v), and 5 mm SDS], the reaction was started.
After 10 min of incubation, which is in the linear range of
the progress curve, the reaction was stopped with 60 lLof
1 m Tris ⁄ HCl buffer (pH 8.0) containing 0.1 m EDTA. One
unit (U) of PLDa2 corresponds to the release of 1 lmol
p-nitrophenolÆmin
)1
.
Acknowledgements
The authors thank Ch. Kuplens and Dr R. Scho
¨
ps for
assistance with protein purification and for preparation
of the PpNp. The financial support of the Kultus-
ministerium des Landes Sachsen-Anhalt and the
Graduiertenkolleg 1026 of the Deutsche Forschungsge-

meinschaft, Bonn (Germany), and the support of the
European Community, Research Infrastructure Action
under the FP6 ‘Structuring the European Research Area
Specific Programme’ to the EMBL Hamburg Outsta-
tion, Contract Number RII3-CT-2004-506008, are
gratefully acknowledged.
References
1 Wang X (2005) Regulatory functions of phospholipase
D and phosphatidic acid in plant growth, development,
and stress responses. Plant Physiol 139, 566–573.
2 Munnik T (2001) Phosphatidic acid: an emerging
plant lipid second messenger. Trends Plant Sci 6,
227–233.
3 McDermott M, Wakelam MJ & Morris AJ (2004) Phos-
pholipase D. Biochem Cell Biol 82, 225–253.
4 Ulbrich-Hofmann R (2000) Phospholipases used in lipid
transformations. In Enzymes in Lipid Modification
(Bornscheuer U, ed.), pp. 219–262. Wiley-VCH,
Weinheim.
5 Qin C & Wang X (2002) The Arabidopsis phospholi-
pase D family. Characterization of a calcium-indepen-
dent and phosphatidylcholine-selective PLDf1 with
distinct regulatory domains. Plant Physiol 128,
1057–1068.
6 Elia
´
s
ˇ
M, Potocky´ M, Cvrc
˘

kova
´
F&Z
ˇ
a
´
rsky´ VV (2002)
Molecular diversity of phospholipase D in angiosperms.
BMC Genomics 3.
7 Sharma S & Gupta MN (2001) Purification of
phospholipase D from Dacus carota by three-phase
partitioning and its characterization. Protein Expr
Purif 21, 310–316.
8 Scha
¨
ffner I, Ru
¨
cknagel KP, Mansfeld J & Ulbrich-
Hofmann R (2002) Genomic structure, cloning and
Structure of plant phospholipase D S. Stumpe et al.
2638 FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS
expression of two phospholipase D isoenzymes from
white cabbage. Eur J Lipid Sci Technol 104, 79–87.
9 Scho
¨
ps R, Schierhorn A, Scha
¨
ffner I, Mansfeld J &
Ulbrich-Hofmann R (2002) Identification of
phospholipase D from cabbage as N-terminally

acetylated PLD2. J Protein Chem 21, 407–411.
10 Ponting CP & Kerr ID (1996) A novel family of phos-
pholipase D homologues that includes phospholipid
synthases and putative endonucleases: identification of
duplicated repeats and potential active site residues.
Protein Sci 5, 914–922.
11 Sung TC, Roper RL, Zhang Y, Rudge SA, Temel R,
Hammond SM, Morris AJ, Moss B, Engebrecht J &
Frohman MA (1997) Mutagenesis of phospholipase D
defines a superfamily including a trans-Golgi viral pro-
tein required for poxvirus pathogenicity. EMBO J 16,
4519–4530.
12 Leiros I, McSweeney S & Hough E (2004) The reaction
mechanism of phospholipase D from Streptomyces sp.
strain PMF. Snapshots along the reaction pathway
reveal a pentacoordinate reaction intermediate and an
unexpected final product. J Mol Biol 339, 805–820.
13 Abergel C, Abousalham A, Chenivesse S, Rivie
`
re M,
Moustacas-Gardies AM & Verger R (2001) Crystalliza-
tion and preliminary crystallographic study of a recom-
binant phospholipase D from cowpea (Vigna
unguiculata L. Walp). Acta Crystallogr D Biol Crystal-
logr 57, 320–322.
14 Allgyer TT & Wells MA (1979) Phospholipase D from
savoy cabbage: purification and preliminary kinetic
characterization. Biochemistry 18, 5348–5353.
15 Abousalham A, Nari J & Noat G (1999) Phospholipase
D from plants: purification, enzymology and structural

analysis. Current Topics Plant Biol 1, 123–132.
16 Wang X, Xu L & Zheng L (1994) Cloning and expres-
sion of phosphatidylcholine-hydrolyzing phospholipase
D from Ricinus communis L. J Biol Chem 269, 20312–
20317.
17 Lambrecht R & Ulbrich-Hofmann R (1992) A facile
purification procedure of phospholipase D from cab-
bage and its characterization. Biol Chem Hoppe Seyler
373, 81–88.
18 Abousalham A, Teissere M, Gardies AM, Verger R &
Noat G (1995) Phospholipase D from soybean (Glycine
max L.) suspension-cultured cells: purification, struc-
tural and enzymatic properties. Plant Cell Physiol 36,
989–996.
19 El Maarouf H, Carrie
`
re F, Rivie
`
re M & Abousalham A
(2000) Functional expression in insect cells, one-step
purification and characterization of a recombinant phos-
pholipase D from cowpea (Vigna unguiculata L. Walp).
Protein Eng 13 , 811–817.
20 Lerchner A, Mansfeld J, Scha
¨
ffner I, Scho
¨
ps R, Beer
HK & Ulbrich-Hofmann R (2005) Two highly homolo-
gous phospholipase D isoenzymes from Papaver

somniferum
L. with different transphosphatidylation
potential. Biochim Biophys Acta 1737, 94–101.
21 Oblozinsky M, Schoeps R, Ulbrich-Hofmann R &
Bezakova L (2003) Two uncommon phospholipase D
isoenzymes from poppy seedlings (Papaver somniferum
L.). Biochim Biophys Acta 1631, 153–159.
22 Svergun DI (1992) Determination of the regularization
parameter in indirect-transform methods using percep-
tual criteria. J Appl Crystallogr 25, 495–503.
23 Konarev PV, Volkov VV, Sokolova AV, Koch MHJ &
Svergun DI (2003) PRIMUS: a Windows PC-based sys-
tem for small-angle scattering data analysis. J Appl
Crystallogr 36, 1277–1282.
24 Svergun DI (1999) Restoring low resolution structure of
biological macromolecules from solution scattering
using simulated annealing. Biophys J 76, 2879–2886.
25 Whitmore L & Wallace BA (2004) DICHROWEB, an
online server for protein secondary structure analyses
from circular dichroism spectroscopic data. Nucleic
Acids Res 32, W668–W673.
26 Andrade MA, Chacon P, Merelo JJ & Moran F (1993)
Evaluation of secondary structure of proteins from
UV circular dichroism spectra using an unsupervised
learning neural network. Protein Eng 6, 383–390.
27 Fersht A (1999) Structure and Mechanism in Protein
Science: a Guide to Enzyme Catalysis and Protein
Folding, p. 208. W.H. Freeman, New York.
28 Stuckey JA & Dixon JE (1999) Crystal structure of a
phospholipase D family member. Nat Struct Biol 6 ,

278–284.
29 Leiros I, Secundo F, Zambonelli C, Servi S & Hough E
(2000) The first crystal structure of a phospholipase D.
Structure Fold Des 8, 655–667.
30 Harding SE (1997) Hydrodynamic properties of
proteins. In Protein Structure: a Practical Approach
(Creighton TE, ed.), pp. 219–251. Oxford University
Press, New York, NY.
31 Zheng L, Shan J, Krishnamoorthi R & Wang X (2002)
Activation of plant phospholipase Db by phosphati-
dylinositol 4,5-bisphosphate: characterization of
binding site and mode of action. Biochemistry 41,
4546–4553.
32 Johnson WC Jr (1990) Protein secondary structure and
circular dichroism: a practical guide. Proteins 7,
205–214.
33 Hwang IS, Park SJ, Roh T, Choi M & Kim HJ (2001)
Investigation of sulfhydryl groups in cabbage phospholi-
pase D by combination of derivatization methods and
matrix-assisted laser desorption ⁄ ionization time-of-flight
mass spectrometry. Rapid Commun Mass Spectrom 15,
110–115.
34 Younus H, Scho
¨
ps R, Lerchner A, Ru
¨
cknagel KP,
Schierhorn A, Saleemuddin M & Ulbrich-Hofmann R
(2003) Proteolytic sensitivity of a recombinant phospho-
lipase D from cabbage: identification of loop regions

S. Stumpe et al. Structure of plant phospholipase D
FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS 2639
and conformational changes. J Protein Chem 22,
499–508.
35 del Valle LJ, Escribano C, Pe
´
rez JJ & Garriga P (2002)
Calcium-induced decrease of the thermal stability and
chaperone activity of alpha-crystallin. Biochim Biophys
Acta 1601, 100–109.
36 Zheng L, Krishnamoorthi R, Zolkiewski M & Wang X
(2000) Distinct Ca
2+
binding properties of novel C2
domains of plant phospholipase Da and b. J Biol Chem
275, 19700–19706.
37 Bittova L, Sumandea M & Cho W (1999) A structure–
function study of the C2 domain of cytosolic phospholi-
pase A2. Identification of essential calcium ligands and
hydrophobic membrane binding residues. J Biol Chem
274, 9665–9672.
38 Nalefski EA, Wisner MA, Chen JZ, Sprang SR, Fukuda
M, Mikoshiba K & Falke JJ (2001) C2 domains from
different Ca
2+
signaling pathways display functional and
mechanistic diversity. Biochemistry 40, 3089–3100.
39 Pappan K, Zheng L, Krishnamoorthi R & Wang X
(2004) Evidence for and characterization of Ca
2+

bind-
ing to the catalytic region of Arabidopsis thaliana phos-
pholipase Db. J Biol Chem 279, 47833–47839.
40 Gill SC & von Hippel PH (1989) Calculation of protein
extinction coefficients from amino acid sequence data.
Anal Biochem 182, 319–326.
41 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
42 Volkov VV & Svergun DI (2003) Uniqueness of
ab initio shape determination in small-angle scattering.
J Appl Crystallogr 36, 860–864.
43 Kelly SM, Jess TJ & Price NC (2005) How to study
proteins by circular dichroism. Biochim Biophys Acta
1751, 119–139.
44 D’Arrigo P, Piergianni V, Scarcelli D & Servi S (1995)
A spectrophotometric assay for phospholipase D. Anal
Chimica Acta 304, 249–254.
Structure of plant phospholipase D S. Stumpe et al.
2640 FEBS Journal 274 (2007) 2630–2640 ª 2007 The Authors Journal compilation ª 2007 FEBS