Tài liệu Báo cáo khoa học: Sensor of phospholipids inStreptomycesphospholipase D pdf

Sensor of phospholipids in Streptomyces phospholipase D
Yoshiko Uesugi, Jiro Arima, Masaki Iwabuchi and Tadashi Hatanaka
Research Institute for Biological Sciences (RIBS), Okayama, Japan
Phospholipase D (PLD; EC 3.1.4.4) catalyzes phos-
pholipid hydrolysis and phosphatidyl transfer
(Fig. 1A). This is a ubiquitous and important enzyme
involved in signal transduction in mammals [1,2].
Streptomyces PLDs can be categorized into two types:
one is an iron-containing enzyme, such as that from
Streptomyces chromofuscus (chromofuscus PLD) for
which tightly bound iron is necessary for its catalytic
activity [3]; and the other is a member of the PLD
superfamily whose hallmark is the possession of two
catalytic HxKxxxxD (HKD) motifs [4–6]. Because
enzymes of the latter type have a simple structure
containing two HKD motifs, they are useful as a
suitable model of mammalian PLDs.
A study of the chemical modification of PLD from
Streptomyces sp. PMF (PMFPLD) suggested that Lys,
not His, is essential for PLD activity [7]. Iwasaki et al.
[8] revealed that two HKD motifs are essential for the
activity, using the N- and C-terminal halves of Strep-
tomyces PLD. Furthermore, Leiros et al. [9] showed
that His170 in the N-terminal HKD motif of PMFPLD
acts as the initial nucleophile that attacks the phospho-
rus atom of the substrate, on the basis of the crystal
structures of PMFPLD. Previously, using two Strep-
tomyces PLDs in repeat-length independent and broad
spectrum (RIBS) in vivo DNA shuffling, we constructed
a random chimera library to investigate the recognition
of phospholipids by Streptomyces PLD. We revealed

that the N-terminal HKD motif contains the nucleo-
phile, using an inactive chimera and surface plasmon
resonance (SPR) analysis [10].
To date, the functions of the HKD motifs in cata-
lytic mechanisms have been extensively studied
[11–13]. At present, PLD-catalyzed reactions are con-
sidered to consist of two steps: first, the formation
of a covalently linked phosphatidyl enzyme interme-
diate via the His residue of the N-terminus HKD
motif; and second, the hydrolysis or transphosphati-
dylation of the intermediate by a water or alcohol
molecule (Fig. 1A).
As mentioned above, previous experimental studies
have focused on the relationship between HKD motifs
Keywords
phospholipase D; phospholipid; substrate
recognition; SPR; Streptomyces
Correspondence
T. Hatanaka, Research Institute for
Biological Sciences (RIBS), Okayama,
7549-1 Kibichuo-cho, Kaga-gun, Okayama
716-1241, Japan
Fax: +81 866 56 9454
Tel: +81 866 56 9452
E-mail:
(Received 12 January 2007, revised 14
March 2007, accepted 22 March 2007)
doi:10.1111/j.1742-4658.2007.05802.x
Recently, we identified Ala426 and Lys438 of phospholipase D from Strep-
tomyces septatus TH-2 (TH-2PLD) as important residues for activity, sta-

bility and selectivity in transphosphatidylation. These residues are located
in a C-terminal flexible loop separate from two catalytic HxKxxxxD
motifs. To study the role of these residues in substrate recognition, we eval-
uated the affinities of inactive mutants, in which these residues were substi-
tuted with Phe and His, toward several phospholipids by SPR analysis. By
substituting Ala426 and Lys438 with Phe and His, respectively, the inactive
mutant showed a much stronger interaction with phosphatidylcholine and
a weaker interaction with phosphatidylglycerol than the inactive TH-2PLD
mutant. We demonstrated that Ala426 and Lys438 of TH-2PLD play a
role in sensing the head group of phospholipids.
Abbreviations
PA, phosphatidic acid; PC, phosphatidylcholine; PG, phosphatidylglycerol; PLD, phospholipase D; POPC, 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphocholine; POPG, 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)]; POPS, 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-
L-serine];
PpNP, phosphatidyl-p-nitrophenol; RU, resonance unit; SPR, surface plasmon resonance; SUV, small unilamellar vesicle.
2672 FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS
and activity. Recently, we demonstrated that four
amino acid residues, Gly188, Asp191, Ala426 and
Lys438, of PLD from Streptomyces septatus TH-2
(TH-2PLD) are associated with PLD activity and sub-
strate recognition [10,14] (Fig. 2A). Substituting Ala426
and Lys438 with Phe and His, respectively, led to
improvements in PLD activity, thermostability, and
organic solvent tolerance and to a change in the selectiv-
ity of transphosphatidylation activity compared with
that in the original chimera [14]. This suggests that
Ala426 and Lys438 are involved in substrate recognition.
Hughes et al. [15] demonstrated that human
PLD1b interacts with fluorescence-labeled phospha-
tidylcholine (PC), whereas PLD1 does not interact

with fluorescence-labeled phosphatidylethanolamine.
This finding shows that the PLD–phospholipid inter-
action correlates with PLD activity, because PLD1
has no catalytic activity toward phosphatidylethanol-
amine. Recently, SPR analysis has been used to
investigate the effects of the rat PLD1 Phox homo-
logy (PX) domain on membrane binding properties
[16], and the specific association of PLD1b with its
regulator proteins, PKCa, Rac1 and ARF6 [17]. In
addition, because of the interaction of inactive PLDs
and PC retaining a covalent phosphatidyl-enzyme
intermediate determined by SPR analysis, the N-ter-
minal HKD motif was found to act as a catalytic
nucleophile [10].
In this study, to investigate the roles of Ala426
and Lys438 of TH-2PLD in substrate recognition in
more detail, we analyzed the association of inactive
mutants of TH-2PLD, in which these residues were
substituted with Phe and Ala, respectively, conco-
mitantly with the substitution of His443 of the
C-terminal HKD motif with Ala, with three phos-
pholipid substrates (Fig. 1) by SPR analysis.
Results
Preparation of inactive mutants of TH-2PLD
In a previous study, we used two homologous
Streptomyces PLDs, TH-2PLD and PLD from Strep-
tomyces sp. (PLDP), as parental enzymes by RIBS
shuffling [10]. PLDP had Phe and His corresponding
to Ala426 and Lys438 of TH-2PLD, respectively.
Thus, we substituted these residues of TH-2PLD

with Phe and His in this study. In addition, to
evaluate the effect of the residues on phospholipid
recognition by SPR analysis, we constructed inactive
mutants of TH-2PLD, in which His443 of an HKD
motif was substituted with Ala, as shown in Fig. 2B.
We then expressed the resultant genes and purified
the proteins they encode. All of the purified mutants
mostly showed a single band with the same molecu-
lar mass ( 57 kDa) as that of wild-type TH-2PLD
on SDS ⁄ PAGE (Fig. 2C). Furthermore, using west-
ern blot analysis with anti-(wild-type TH-2PLD)
serum, these mutants were found to have similar
uniformities and purities (Fig. 2D). All the mutants
had low activities toward phosphatidyl-p-nitrophenol
(PpNP) (< 0.7 lmolÆmin
)1
Æmg
)1
), whereas wild-type
TH-2PLD had high activity (59 lmolÆmin
)1
Æmg
)1
).
To confirm the folding of the inactive mutants
of TH-2PLD, their CD spectra were measured.
As shown in Fig. 3, the CD spectra showed that
the inactive mutants folded with a secondary struc-
ture similar to wild-type TH-2PLD. These results
A

B
Fig. 1. Reactions catalyzed by PLD (A) and
structure of phospholipid head groups (B).
Y. Uesugi et al. Sensor of phospholipids in PLD
FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS 2673
suggest that we successfully prepared inactive mutant
enzymes.
Association of inactive mutants with
phospholipid vesicles
To investigate the association of key C-terminal resi-
dues with the head group of phospholipid substrates,
we analyzed the binding profiles of TH-2(H443A),
TH-2-F(H443A) and TH-2-FH(H443A) for several
phospholipid vesicles with a covalent phosphatidyl-
PLD intermediate using SPR analysis. Sensorgrams
obtained by SPR analysis showed real-time biomole-
cular interaction. Overlaid sensorgrams were obtained
when TH-2(H443A), TH-2-F(H443A) and TH-2-
FH(H443A) were passed at different concentrations
over immobilized 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphocholine (POPC; Fig. 4A–C), 1-palmitoyl-2-
oleoyl-sn-glycero-3-[phospho-l-serine] (POPS; Fig. 4D–F)
or 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-rac-(1-
glycerol)] (POPG; Fig. 4G–I) vesicles. As shown in
Fig. 4A–F, TH-2-F(H443A) and TH-2-FH(H443A)
exhibited significantly higher binding abilities for
POPC and POPS vesicles than TH-2(H443A). By con-
trast, sensorgrams of TH-2-F(H443A) were similar to
those of TH-2(H443A) for POPG vesicles, and their
interactions were stronger than those of TH-2-

FH(H443A) (Fig. 4G–I). These differences in interac-
tion were not caused by the heterogeneity of the
mutants. If they were, each mutant would have shown
the same association and dissociation curves for all the
phospholipids; however, the results did not show such
A
B
CD
Fig. 2. (A) 3D structure around identified
key residues (i.e. residues 188, 191, 426
and 438 of TH-2PLD) associated with activ-
ity. The overall structure of TH-2PLD is sho-
wn using the Swiss-PDB viewer and is
based on the crystal structure of PMFPLD.
The identified key residues are indicated in
red. The N-terminal and C-terminal HKD
motifs are shown in light blue and purple,
respectively. (B) Primary structures of wild-
type TH-2PLD and its inactive mutants. The
gray box indicates the His residue of the
C-terminal HKD motif mutated to Ala. The
identified residues related to the PLD
reaction are shown in black boxes.
(C) SDS ⁄ PAGE results of purified PLDs.
Lanes 1–4 contained 2 lg of TH-2PLD, TH-
2(H443A), TH-2-F(H443A) and TH-2-FH-
(H443A), respectively. Lane M indicates
SDS ⁄ PAGE standard proteins (molecular
masses: 100 000, 80 000, 60 000, 50 000,
40 000, 30 000 and 20 000 Da). Samples

were loaded on a 10% acrylamide gel. (D)
Western blot analysis of purified PLDs using
anti-(wild-type TH-2PLD) serum. Lanes 1–3
contained 2 lg of TH-2(H443A), TH-2-
F(H443A) and TH-2-FH(H443A), respectively.
Lane M indicates prestained SDS ⁄ PAGE
standards (molecular masses: 111 000,
93 000, 53 500, 36 100 and 29 500 Da). The
samples were loaded on a 10% acrylamide
gel. The arrowhead indicates the position of
the purified PLDs.
Sensor of phospholipids in PLD Y. Uesugi et al.
2674 FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS
curves. Therefore, these results suggest that substitu-
tion of Ala426 with Phe led to considerably stronger
interactions with POPC and POPS. It should be noted
that the double mutant showed a decrease in the
strength of its interaction with POPG vesicles,
although the interactions of the mutant with POPC
and POPS vesicles remained strong.
The kinetic constant was calculated from each sensor-
gram using bia evaluation 4.1 analysis software
according to the global fitting of 1:1 binding with a mass
transfer model. Affinity constants (K
D
) for POPC vesi-
cles were 5.3 ± 0.9, 5.5 ± 1.4 and 4.8 ± 0.7 nm for
TH-2(H443A), TH-2-F(H443A) and TH-2-FH(H443A),
respectively. There was no significant difference in K
D

value between these proteins; however, the maximal
responses of these proteins differ markedly. In the case
of POPG and POPS vesicles, K
D
could not be calculated
because their response curves did not fit the evaluation
curves. In particular, sensorgrams toward POPG
showed an increase and a decrease in interactions during
the association process at a low mutant concentration
(Fig. 4G–I). The results suggest that the mutants have
more than two binding sites, with different affinities for
POPG vesicles. Unfortunately, there are no evaluation
models for determining the interaction between a ligand
and an analyte involving more than two binding sites.
Thus, the sensorgrams of mutants toward POPG could
not be analyzed appropriately.
To compare the differences in affinity for phospho-
lipids among TH-2(H443A), TH-2-F(H443A) and
TH-2-FH(H443A), the maximal responses measured
for these inactive mutant associations with POPC,
POPS and POPG vesicles are shown in Fig. 5. Injection
of 532 nm TH-2(H443A) resulted in a binding signal of
1409 resonance units (RU) for POPC vesicles, which
was similar to that of 1204 RU for POPG vesicles
(t-test; P > 0.05), whereas the association with POPS
vesicles was significantly weaker (477 RU; P < 0.001).
As shown in Fig. 4A,G, the sensorgram of TH-
2(H443A) increased more sharply in the association
phase when POPG vesicles rather than POPC vesicles
were used. In contrast, TH-2-FH(H443A) bound to

POPG vesicles slowly (Fig. 4I), and the degree of bind-
ing response for POPG vesicles was 5.1-fold lower than
that for POPC vesicles. TH-2-F(H443A) also showed a
degree of binding response for POPG vesicles 2.6-fold
lower than that for POPC vesicles. Interestingly, the
interactions of each inactive mutant with POPS vesicles
were similar and low in degree compared with those
with POPC vesicles, although each mutant exhibited
different degrees of interaction with POPG vesicles.
From these results, it is suggested that residues 426 and
438 of TH-2PLD play a role in sensing the head group
of phospholipid vesicles.
Conformational change of inactive mutants
induced by phospholipids
To analyze changes in the tertiary and secondary struc-
tures of inactive mutants induced by phospholipids, we
further measured the fluorescence and CD spectra of
inactive mutants in the absence and presence of POPC
and POPG vesicles [18]. TH-2PLD has 11 Trp residues
that contribute to its fluorescence emission spectrum.
As shown in Fig. 6A–C, the emission maxima, around
340 nm, were the same for all the inactive mutants.
With inactive mutant alone, TH-2-F(H443A) showed a
similar fluorescence emission spectrum to TH-
2(H443A), and TH-2-FH(H443A) had a higher fluores-
cence emission intensity than TH-2(H443A). For the
mutant TH-2-FH(H443A), the local environment
around Trp434 is probably changed by substituting
His for Lys438, because Lys438 is located adjacent to
Trp434 ( 4A

˚
). The fluorescence emission intensities
of TH-2-F(H443A) and TH-2-FH(H443A) increased at
340 nm with the addition of POPC vesicles, and the
degrees of increase were 0.186 and 0.083 relative to
those without POPC, respectively. Interestingly, the
fluorescence emission intensity of TH-2-FH(H443A)
with POPG vesicles decreased to that without POPG
at 0.11 degrees. However, the fluorescence emission
intensity of TH-2(H443A) did not change with or
without phospholipids. In contrast, the CD spectra of
all the inactive mutants were similar with or without
phospholipids (Fig. 6D–F). From these results, it sug-
gests that TH-2-F(H443A) and TH-2-FH(H443A) were
Fig. 3. CD spectra of PLDs. The spectrum of each PLD
(0.1 mgÆmL
)1
)in10mM potassium phosphate buffer (pH 7.0) was
measured at 25 °C.
Y. Uesugi et al. Sensor of phospholipids in PLD
FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS 2675
induced conformational changes in their tertiary struc-
tures with phospholipid introduction, although their
secondary structures remain unchanged.
Discussion
Recently, Sato et al. [19] reported that the phosphati-
dic acid (PA) contents produced by a side reaction of
several Streptomyces PLDs differ markedly during
transphosphatidylation from PC to phosphatidylgly-
cerol (PG). Among the PLDs used, TH-2PLD showed

the lowest selectivity in transphosphatidylation, and
the amount of hydrolyzed PA increased with reaction
time. This phenomenon was considered to be the result
of synthesized PG being hydrolyzed to PA during
transphosphatidylation. We speculated that TH-2PLD
recognizes synthesized PG as well as PC; therefore, the
amount of hydrolyzed PA increases. Recently, we
showed that the C-terminal flexible loop in Strepto-
myces PLD (residues 425–442) is separate from the
ABC
DEF
GHI
Fig. 4. Sensorgrams at different concentrations of inactive mutants of TH-2PLD. As substrate, POPC (A–C), POPS (D–F) and POPG (G–I)
vesicles were immobilized on an L1 sensor chip, as described in Experimental procedures. The SPR sensorgrams were obtained when
TH-2(H443A) (A,D,G), TH-2-F(H443A) (B,E,H) and TH-2-FH(H443A) (C,F,I) were passed over the phospholipid vesicles at 532, 355, 236 and
158 n
M, respectively (from top to bottom), at a flow rate of 20 lLÆmin
)1
for 5 min at 25 °C, followed by a buffer at the same flow rate
for 10 min. The affinities of the mutants to the phospholipid vesicles were determined by fitting these SPR sensorgrams using
BIA
EVALUATION
4.1 analysis software.
Sensor of phospholipids in PLD Y. Uesugi et al.
2676 FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS
two highly conserved catalytic HKD motifs, and that
it is formed at the entrance of the active-site well and
has multiple functional roles. A mutant PLD with
Ala426 and Lys438 substituted with Phe and His,
respectively, improved its selectivity in transphosphati-

dylation from PC to PG [14]. Thus, we plan to investi-
gate the relationship between the recognition of several
phospholipids, such as PC, PG and phosphatidylserine
(PS), and the residues identified by SPR analysis.
These results agree well with the present findings
that TH-2(H443A) has comparable interactions with
POPC and POPG, and that TH-2-FH(H443A) has a
much stronger interaction with POPC and a weaker
interaction with POPG than TH-2(H443A) (Fig. 5). In
addition, by determining the corresponding active
mutants in terms of their activity toward POPC vesi-
cles by the method using choline oxidase and peroxi-
dase [14], the activities of TH-2-F (3.3 lmolÆmin
)1
Æ
mg
)1
) and TH-2-FH (3.3 lmolÆmin
)1
Æmg
)1
) were
found to be twofold higher than that of TH-2PLD
(1.7 lmolÆmin
)1
Æmg
)1
) toward POPC (data not shown).
These results suggest that the activities of active
mutants correlate well with the binding data of inac-

tive mutants. These findings suggest that TH-2PLD
recognizes PC and PG similarly, and that residues
426 and 438 of TH-2PLD play an important role in
phospholipid recognition. To date, nine sequences of
Streptomyces PLDs have been determined. All Strep-
tomyces PLDs except TH-2PLD have a Phe residue
corresponding to Ala426 of TH-2PLD. By contrast, at
Lys438 of TH-2PLD, four of them have a Lys residue
and the rest have a His residue. Thus, the main cause
of the low selectivity of TH-2PLD in transphosphati-
dylation seems to be Ala426. In this study, we con-
firmed that these residues are associated with the
interaction between TH-2PLD and its substrate.
SPR analysis revealed that Streptomyces PLD inter-
acts to a much higher degree with zwitterionic phos-
pholipid vesicles (i.e. PC) than with anionic
phospholipid vesicles (i.e. PG and PS) (Fig. 5). PLDs
from mammals and poppy seedlings hydrolyze PC
most efficiently among several phospholipids [20,21].
That is, Streptomyces PLD containing HKD motifs
seems to prefer zwitterionic phospholipid vesicles, such
as POPC, to anionic phospholipid vesicles POPS and
POPG, similarly to other PLDs.
Previous SPR analysis indicated that the PLD1 PX
domain has a high phosphoinositide specificity, that is,
the K
D
value for POPC ⁄ POPE ⁄ phosphatidylinositol
3,4,5-trisphosphate (77 : 20 : 3) is 18 ± 4 nm [16].
Powner et al. [17] showed K

D
values between PLD1b
and regulator proteins, PKCa, Rac1 and ARF6 (i.e.
42 ± 1 5, 143 ± 28 and 660 ± 63 nm, respectively) [17].
In this study, K
D
values for POPC vesicles were found
to be 5.3 ± 0.9, 5.5 ± 1.4 and 4.8 ± 0.7 nm for
TH-2(H443A), TH-2-F(H443A) and TH-2-FH(H443A),
respectively. These results suggest that the affinities of
these mutants for POPC vesicles are stronger than
those of the specific association between the PLD1 PX
domain and phosphatidylinositol 3,4,5-trisphosphate,
or PLD1b and its regulators.
We speculate that the difference in the degree of
interaction with POPC among TH-2PLD mutants is a
result of the influence of two flexible loops, i.e. resi-
dues 188–203 and 425–442, of TH-2PLD on each
other. Iwasaki et al. [8] and Xie et al. [22] showed that
PLD activity is restored when the N- and C-terminal
fragments of Streptomyces PLD and PLD1 coexist,
although these fragments have only negligible activities
in isolation. From these findings, we speculate that
PLD changes its conformation markedly before and
after binding to the substrate. In fact, SPR analysis
(Figs 4,5) and fluorescence spectroscopy (Fig. 6)
showed that the inactive mutants, in which Ala426 and
Lys438 were substituted with Phe and Ala, showed ter-
tiary structural changes with phospholipid binding.
Combined with the results of SPR analysis and fluores-

cence spectroscopy, it seems that the interaction
between Streptomyces PLD and POPG differs from
that between the PLD and POPC, although this phe-
nomenon cannot be explained experimentally at pre-
sent. These two flexible loops may play a role as a
Fig. 5. Interaction of PLDs and phospholipid vesicles. The maximal
responses of TH-2(H443A), TH-2-F(H443A) and TH-2-FH(H443A)
were measured for their specific associations with phospholipids
vesicles. Each PLD (532 n
M) was injected at a flow rate of 20
lLÆmin
)1
. Each value represents the mean ± SD from three inde-
pendent experiments.
Y. Uesugi et al. Sensor of phospholipids in PLD
FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS 2677
trigger of conformational change when PLDs bind to
the substrate. Therefore, residues 426 and 438 located
in the C-terminal loop could affect the interaction of
PLDs with substrate.
TH-2(H443A) and TH-2-F(H443A) exhibited stron-
ger interactions with POPG vesicles than TH-
2-FH(H443A). Sensorgrams of these mutants showed
an increase and a decrease in the degree of interactions
during the association process at a low concentration
of mutants (Fig. 4G–I). These results suggest that
there are more than two binding sites that have differ-
ent affinities to PG vesicles in the mutants. Using com-
puter analysis with the automated docking program
autodock, Aikens et al. [23] showed that the glycerol

group of PG is bound to a region composed of
Ser453, Lys454, Asn455, Tyr457, Ser459 and Leu461
of PMFPLD. These residues correspond to Ser458,
Lys459, Asn460, Tyr462, Ser464 and Leu466, respect-
ively, in the C-terminal region of TH-2PLD. PMFPLD
and TH-2PLD are 85% homologous in primary struc-
ture. Hence, we surmise that another PG-recognizing
site is present in the C-terminal region.
Ala426 and Lys438 of TH-2PLD are located in a C-
terminal flexible loop separate from two catalytic
HKD motifs. The loop, in coordination with the N-
terminal loop, forms the entrance of the active well
comprising two HKD motifs (Fig. 2A). It is reasonable
to consider that these residues play a role in sensing
the head group of phospholipids from a geometrical
point of view. It might be possible to change the sub-
strate preference of Streptomyces PLD by substituting
these two residues with other amino acid residues.
ABC
DEF
Fig. 6. Fluorescence emission spectra of TH-2(H443A) (A),TH-2-F(H443A) (B) and TH-2-FH(H443A) (C) in the absence and presence of phosp-
holipids vesicles. The inactive mutants were excited at 290 nm and emission spectra were recorded between 300 and 380 nm. Fluores-
cence measurements were carried out at 25 °C with 1.2 l
M PLDs in 10 mM Tris ⁄ HCl (pH 7.0) containing 4 mM CaCl
2
. CD spectra of TH-
2(H443A) (D), TH-2-F(H443A) (E) and TH-2-FH(H443A) (F) in the absence and presence of phospholipids vesicles. The spectrum was meas-
ured at 25 °C with 1.7 l
M PLDs in 10 mM Tris ⁄ HCl (pH 7.0) containing 4 mM CaCl
2

. POPC or POPG of SUVs was added PLD solution at a
final concentration of 1 m
M, and incubated for over 1 h. All spectra were corrected by subtracting the spectra of the corresponding back-
ground media without PLD.
Sensor of phospholipids in PLD Y. Uesugi et al.
2678 FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS
Experimental procedures
Materials
The plasmid pETKmS2 [24] was kindly provided by
T. Yamane (Nagoya University, Japan). PpNP was pre-
pared from soybean phosphatidic acid and p-nitrophenol
according to the procedure of D’Arrigo et al. [25]. POPC,
POPS and POPG were obtained from Avanti Polar Lipids
(Alabaster, AL, USA) and used without further purification.
All the other chemicals were of the highest purity available.
Preparation of PLDs
To construct the mutant TH-2(H443A), the mutagenic gene
was amplified by PCR using the following primers: 5¢-
CCCTGCGCGCGCTCGTCGGCA-3¢ (corresponding to
the gene th-2pld, nucleotides 962–982) and 5¢-ACCAG
CTTGTGG(TG fi GC, His fi Ala)CTGCGCGTACG-3¢
[corresponding to nucleotides 1316–1340 from th-2pld,
TH-2(H443A)]. The amplified DNA fragment was
cloned, sequenced and digested with BglII and Van91I.
Next, the plasmid pUC19(TH-2) [26] was digested with
BglII and Van91I, and the product was substituted for the
corresponding region in the subcloned th-2pld. The resul-
tant mutant gene was digested with NcoI and BamHI
and ligated into the NcoI–BamHI gap of the vector
pETKmS2(TH-2(H443A)).

To prepare the mutants TH-2-F(H443A) and TH-
2-FH(H443A), a partial th-2pld was amplified by PCR using
the following primers: 5¢-ACTACGTCGACACCTCCCA
CC-3¢ (corresponding to nucleotides 575–595 from th-2pld)
and 5¢-GAAGGTG
GCTAGCTGGAGGTTG-3¢ for the
silent mutation of the NheI site (underlined) (corresponding
to nucleotides 1257–1278 from th-2pld). Then the amplified
DNA fragments were cloned, sequenced and digested with
PstI and NheI. Next, the plasmids pETKmS2(G-F) and
pETKmS2(G-FH) [14] were digested with NheI and BsiWI.
The two resulting fragments were ligated into the PstI–
BsiWI gap of the vector pETKmS2(TH-2(H443A)) to con-
struct the expression vector. The expression vectors obtained
were confirmed by DNA sequencing.
The recombinant TH-2PLD and inactive mutant enzymes
were expressed as secreted proteins with a C-terminal His6
tag, and purified with TALON metal affinity resin (Clontech,
Palo Alto, CA, USA) according to standard protocols. The
purities of proteins obtained were confirmed by SDS ⁄ PAGE
[10] and western blot analysis using an anti-(wild-type TH-
2PLD) serum.
Assay for PLD activity using PpNP
Hydrolytic activity was determined on the basis of the
hydrolytic activity of PpNP. The procedure was similar
one described previously [27]. One unit of PLD was
defined as the amount of the enzyme that releases 1 lmol
of p-nitrophenol per minute under the assay conditions.
The reactions were carried out in 1.5-mL cuvettes. The
1-mL reaction mixture consisted of 0.07–0.2 lg of purified

PLDs and 2 mm PpNP in 0.1 m sodium acetate buffer
(pH 5.5) at 37 °C.
CD spectroscopy
The folding of PLDs was confirmed by CD spectroscopy
using a J-720WI spectrometer (Jasco, Tokyo, Japan). Pro-
teins were dissolved to a final concentration of 0.1 mgÆmL
)1
in 10 mm potassium phosphate buffer (pH 7.0). Spectra
were acquired at 25 °C using a 2-mm path-length cuvette.
The spectra of PLDs that averaged 10 scans were corrected
by subtracting the spectra of the corresponding background
media without PLD.
Preparation of vesicles
An aliquot of phospholipids dissolved in chloroform was
evaporated and further dried in vacuum for at least 3 h.
The lipids were hydrated to a concentration of 10 mm in
phosphate-buffered saline for SPR analysis or in 10 mm
Tris ⁄ HCl (pH 7.0) for fluorescence and CD measurements.
The lipid suspension was vortexed vigorously, and frozen
and thawed 10 times in liquid nitrogen. To obtain small uni-
lamellar vesicles (SUVs), the suspension was passed 30 times
through polycarbonate membranes (50 nm pore diameter)
using a Lipofast extruder (Avestin, Ottawa, Canada) [28].
SPR analysis
Real-time interactions between PLD molecules and phos-
pholipids were measured using a Biacore instrument (Bia-
core 2000, Biacore AB, Uppsala, Sweden). Liposomes were
captured on the surface of a Sensor Chip L1 (Biacore AB) as
‘ligand’. The surface of the L1 sensor chip was first cleaned
with 20 mm 3-[(3-cholamidopropyl) dimethylammonio]-1-

propanesulfonic acid (CHAPS) at a flow rate of 5 lLÆmin
)1
followed by the injection of SUVs (60 lL, 0.5 mm phospho-
lipids) at a flow rate of 2 lLÆmin
)1
in buffer A (10 mm
sodium acetate, pH 5.5, 4 mm CaCl
2
). This resulted in the
deposition of  5000–7000 RU. To measure the association
of PLDs with phospholipids, a purified inactive mutant
enzyme (105–532 nm diluted in buffer A) as ‘analyte’ was
applied to the captured SUVs at a flow rate of 20 lLÆmin
)1
at 25 °C. After the binding of PLDs to phospholipids, disso-
ciation was observed at the same flow rate for 10 min. The
evaluation of the kinetic parameters of PLD binding to
phospholipids was performed by the nonlinear fitting of
binding data using bia evaluation 4.1 analysis software.
The apparent association (k
a
) and dissociation (k
d
) rate
Y. Uesugi et al. Sensor of phospholipids in PLD
FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS 2679
constants were evaluated from the differential binding curves
(sample–control) shown in Fig. 4, assuming an A + B ¼ AB
association type for protein–lipid interaction. The affinity
constant K

D
was calculated from the equation K
D
¼ k
d
⁄ k
a
.
Fluorescence spectroscopy
Fluorescence spectra were obtained with an F-4500 spectro-
fluorometer (Hitachi, Tokyo, Japan). All measurements
were carried out at 25 °C with 1.2 lm PLDs in 10 mm
Tris ⁄ HCl (pH 7.0) containing 4 mm CaCl
2
using 2-mm
path-length quartz cuvette. The excitation wavelength was
290 nm, and excitation and emission slits were 5 nm. Emis-
sion was scanned from 300 to 380 nm. PLDs were mixed
with 1 mm SUVs and incubated over 1 h. The spectra of
PLDs that averaged four scans were corrected by subtract-
ing the spectra of the corresponding background media
without PLD. The degree of change in the fluorescence
intensity was calculated as (I ) I
0
) ⁄ I
0
, where I
0
is the maxi-
mum intensity of PLD alone, and I is the maximum inten-

sity in the presence of phospholipids [29].
Statistical analysis
All statistical evaluations were performed using an unpaired
Student’s t test. All data are presented as mean ± SD of at
least three determinations.
Acknowledgements
We thank Ms M. Taniai (Hayashibara Biochemical
Laboratories) for technical advice on SPR analysis. This
research was financially supported by the Sasakawa Sci-
entific Research Grant from The Japan Science Society.
References
1 McDermott M, Wakelam MJO & Morris AJ (2004)
Phospholipase D. Biochem Cell Biol 82, 225–253.
2 Jenkins GM & Frohman MA (2005) Phospholipase
D: a lipid centric review. Cell Mol Life Sci 62, 2305–
2316.
3 Yang H & Roberts MF (2002) Cloning, overexpression,
and characterization of a bacterial Ca
2+
-dependent
phospholipase D. Protein Sci 11, 2958–2968.
4 Hammond SM, Altshuller YM, Sung TC, Rudge SA,
Rose K, Engebrecht J, Morris AJ & Frohman MA
(1995) Human ADP-ribosylation factor-activated phos-
phatidylcholine-specific phospholipase D defines a new
and highly conserved gene family. J Biol Chem 270,
29640–29643.
5 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.
6 Koonin EV (1996) A duplicated catalytic motif in a new
superfamily of phosphohydrolases and phospholipid
synthases that includes poxvirus envelope proteins.
Trends Biochem Sci 21, 242–243.
7 Secundo F, Carrea G, D’Arrigo P & Servi S (1996)
Evidence for an essential lysyl residue in phospholipase
D from Streptomyces sp. by modification with diethyl
pyrocarbonate and pyridoxal 5-phosphate. Biochemistry
35, 9631–9636.
8 Iwasaki Y, Horiike S, Matsushima K & Yamane T
(1999) Location of the catalytic nucleophile of phospho-
lipase D of Streptomyces antibioticus in the C-terminal
half domain. Eur J Biochem 264, 577–581.
9 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.
10 Uesugi Y, Mori K, Arima J, Iwabuchi M & Hatanaka
T (2005) Recognition of phospholipids in Streptomyces
phospholipase D. J Biol Chem 280, 26143–26151.
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 Stuckey JA & Dixon JE (1999) Crystal structure of a
phospholipase D family member. Nat Struct Biol 6,
278–284.
13 Gottlin EB, Rudolph AE, Zhao Y, Matthews HR &
Dixon JE (1998) Catalytic mechanism of the phospholi-
pase D superfamily proceeds via a covalent phosphohis-
tidine intermediate. Proc Natl Acad Sci USA 95,
9202–9207.
14 Uesugi Y, Arima J, Iwabuchi M & Hatanaka T (2007)
C-terminal loop of Streptomyces phospholipase D has
multiple functional roles. Protein Sci 16, 197–207.
15 Hughes WE, Larijani B & Parker PJ (2002) Detecting
protein–phospholipid interactions. J Biol Chem 277,
22974–22979.
16 Stahelin RV, Ananthanarayanan B, Blatner NR, Singh
S, Bruzik KS, Murray D & Cho W (2004) Mechanism
of membrane binding of the phospholipase D1 PX
domain. J Biol Chem 279, 54918–54926.
17 Powner DJ, Hodgkin MN & Wakelam MJO (2002)
Antigen-stimulated activation of phospholipase D1b by
Rac1, ARF6, and PKCa in RBL-2H3 cells. Mol Biol
Cell 13, 1252–1262.
18 Qin S, Pande AH, Nemec KN & Tatulian SA (2004)
The N-terminal a-helix of pancreatic phospholipase A
2
Sensor of phospholipids in PLD Y. Uesugi et al.
2680 FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS
determines productive-mode orientation of the enzyme
at the membrane surface. J Mol Biol 344, 71–89.
19 Sato R, Itabashi Y, Hatanaka T & Kuksis A (2004)

Asymmetric in vitro synthesis of diastereomeric phos-
phatidylglycerols from phosphatidylcholine and gly-
cerol by bacterial phospholipase D. Lipids 39, 1013–
1018.
20 Horwitz J & Davis L (1993) The substrate specificity of
brain microsomal phospholipase D. Biochem J 295,
793–798.
21 Oblozinsky M, Ulbrich-Hofmann R & Bezakova L
(2005) Head group specificity of phospholipase D isoen-
zymes from poppy seedling (Papaver somniferum L.).
Biotechnol Lett 295, 793–798.
22 Xie Z, Ho WT & Exton JH (1998) Association of
N- and C-terminal domains of phospholipase D is
required for catalytic activity. J Biol Chem 273,
34679–34682.
23 Aikens CL, Laederach A & Reilly PJ (2004) Visualizing
complexes of phospholipids with Streptomyces phospho-
lipase D by automated docking. Proteins 57, 27–35.
24 Mishima N, Mizumoto K, Iwasaki Y, Nakano H &
Yamane T (1997) Insertion of stabilizing loci in vectors
of T7 RNA polymerase-mediated Escherichia coli
expression systems: a case study on the plasmids invol-
ving foreign phospholipase D gene. Biotechnol Prog 13,
864–868.
25 D’Arrigo P, Piergianni V, Scarcelli D & Servi S (1995)
A spectrophotometric assay for phospholipase D. Anal
Chim Acta 304, 249–254.
26 Mori K, Mukaihara T, Uesugi Y, Iwabuchi M & Hata-
naka T (2005) Repeat-length-independent broad-spec-
trum shuffling, a novel method of generating a random

chimera library in vivo. Appl Environ Microbiol 71,
754–760.
27 Hatanaka T, Negishi T, Kubota-Akizawa M &
Hagishita T (2002) Purification, characterization, clon-
ing and sequencing of phospholipase D from Strepto-
myces septatus TH-2. Enzyme Microb Technol 31,
233–241.
28 MacDonald RC, MacDonald RI, Menco BP, Takeshita
K, Subbarao NK & Hu LR (1991) Small-volume extru-
sion apparatus for preparation of large, unilamellar
vesicles. Biochim Biophys Acta 1061, 297–303.
29 Feng J, Wehbi H & Roberts MF (2002) Role of trypto-
phan residues in interfacial binding of phosphatidylino-
sitol-specific phospholipase C. J Biol Chem 277,
19867–19875.
Y. Uesugi et al. Sensor of phospholipids in PLD
FEBS Journal 274 (2007) 2672–2681 ª 2007 The Authors Journal compilation ª 2007 FEBS 2681