Biochem. J. (2013) 454, 491–499 (Printed in Great Britain)
491
doi:10.1042/BJ20130377
Mercury increases water permeability of a plant aquaporin through
a non-cysteine-related mechanism
Anna FRICK1 , Michael JÄRVÅ1 , Mikael EKVALL, Povilas UZDAVINYS, Maria NYBLOM and Susanna TÖRNROTH-HORSEFIELD2
Department of Chemistry and Molecular Biology, University of Gothenburg, PO Box 462, SE-405 30 Gothenburg, Sweden
Water transport across cellular membranes is mediated by a family
of membrane proteins known as AQPs (aquaporins). AQPs were
first discovered on the basis of their ability to be inhibited by
mercurial compounds, an experiment which has followed the AQP
field ever since. Although mercury inhibition is most common,
many AQPs are mercury insensitive. In plants, regulation of
AQPs is important in order to cope with environmental changes.
Plant plasma membrane AQPs are known to be gated by
phosphorylation, pH and Ca2 + . We have previously solved the
structure of the spinach AQP SoPIP2;1 (Spinacia oleracea plasma
membrane intrinsic protein 2;1) in closed and open conformations
and proposed a mechanism for how this gating can be achieved. To
study the effect of mercury on SoPIP2;1 we solved the structure
of the SoPIP2;1–mercury complex and characterized the water
transport ability using proteoliposomes. The structure revealed
mercury binding to three out of four cysteine residues. In contrast
to what is normally seen for AQPs, mercury increased the water
transport rate of SoPIP2;1, an effect which could not be attributed
to any of the cysteine residues. This indicates that other factors
might influence the effect of mercury on SoPIP2;1, one of which
could be the properties of the lipid bilayer.
INTRODUCTION
inhibition is a result of a collapse of the ar/R [10]. In cases where
Cys189 is missing other cysteine residues in the vicinity are thought
to be responsible for the inhibition [11–14]. Although most AQPs
are inhibited by mercury, many AQPs are in fact insensitive [15–
18]. These AQPs typically lack a cysteine residue at a critical
position within the water-conducting pore.
Plants depend on their ability to adapt to changes in water
supply and express a large number of AQP isoforms (35 isoforms
in the model plant Arabidopsis thaliana compared with only 13 in
humans) [19]. In addition, many plant AQPs are gated in response
to environmental stress [20]. Gating of the AQPs situated in
the plant plasma membrane (PIPs; plasma membrane intrinsic
proteins) have been shown to be mediated by phosphorylation
[21], pH [22–25] and Ca2 + [23,24] levels.
We previously solved the crystal structure of the AQP SoPIP2;1
from Spinacia oleracea in both open and closed conformations,
revealing how gating is mediated by a conformational change
of the cytoplasmic loop D. In the closed conformation, water
transport is blocked through the insertion of Leu197 into the
cytoplasmic opening of the pore, whereas in the open state, a
half-turn extension of TM helix 5 displaces loop D and thereby
Leu197 to allow passage of water (Figure 1C) [26]. This allowed
us to propose a mechanism for gating by phosphorylation and
pH in response to drought and flooding respectively. Both events
involve a divalent cation-binding site at the N-terminus, occupied
by Cd2 + in the structure, but presumed to bind Ca2 + in vivo.
Gating by phosphorylation involves Ser115 and Ser274 in loop
B and the C-terminus respectively. Phosphorylation of Ser115
prevents loop D from interacting with the Cd2 + -binding site,
leading to channel opening. In the case of Ser274 , hydrogen bonds
with residues in a neighbouring protomer stabilizes loop D in
Water flow across cellular membranes is a fundamental process
of all living cells and is mediated by a highly conserved family of
integral membrane proteins known as AQPs (aquaporins). AQPs
are tetramers with each protomer functioning as an individual
channel, consisting of six TM (transmembrane) helices and
two half-membrane-spanning helices (Figure 1A). The two halfmembrane-spanning helices meet in the middle of the membrane
where their dipole moment together with the NPA region (AsnPro-Ala, AQP signature motif) serves to efficiently prevent proton
transport (Figure 1B). Members of the AQP family can be
divided into orthodox AQPs which transport water only and
aquaglyceroporins which are permeable for other small solutes
such as glycerol or urea, with the two subgroups differing in
composition of the pore selectivity filter known as the ar/R
(aromatic/arginine region) [1].
AQP-mediated water transport was first discovered following
the observation that mercurial compounds reversibly inhibit water
transport across the membrane of red blood cells [2]. This
allowed for the identification of the first AQP, AQP1 [3]. Since
then, mercury inhibition has been a key experiment in order
to characterize the AQP channel and to identify AQP-mediated
water transport in numerous tissues and organisms [4–7]. Mercury
preferentially binds to the thiol group of cysteine residues. In
AQP1, the cysteine residue responsible for mercury inhibition
has been identified as Cys189 within the water-conducting pore
[8]. Structural studies of AQPZ from Escherichia coli mutated to
imitate the cysteine pattern of AQP1 showed that when mercury
binds, the pore is sterically occluded [9]. Alternatively, MD
(molecular dynamics) simulations have suggested that mercury
Key words: aquaporin, mercury, plasma membrane intrinsic
protein, proteoliposome assay, water transport, X-ray crystallography.
Abbreviations used: AQP, aquaporin; ar/R, aromatic/arginine region; MD, molecular dynamics; P f , osmotic permeability coefficient; PIP, plasma
membrane intrinsic protein; So, Spinacia oleracea ; TIP, tonoplast intrinsic protein; TM, transmembrane; β-OG, octyl-β-glucanpyranoside.
1
These authors contributed equally to this work.
2
To whom correspondence should be addressed (email [email protected]).
The structural co-ordinates of the SoPIP2;1–mercury complex will appear in the PDB under accession code 4JC6.
c The Authors Journal compilation c 2013 Biochemical Society
492
Figure 1
A. Frick and others
Structure and gating mechanism of SoPIP2;1
(A) Crystal structure of SoPIP2;1 (PDB code 1Z98) showing the overall AQP fold. Loops B and E forming the half-membrane-spanning helices are coloured yellow. Water molecules in the water
conducting channel are shown as red spheres. (B) Close-up of the water conducting channel in (A) highlighting the NPA region and ar/R. (C) Stereo view of an overlay of the closed (green, PDB code
1Z98) and open (blue, PDB code 2B5F) structures of SoPIP2;1. For the closed structure, a neighbouring protomer in the tetramer is shown in lighter green. During channel closure, loop D adopts
a conformation which inserts Leu197 into the opening, whereas in the open structure loop D is rearranged and Leu197 is displaced. Residues involved in gating are shown in stick representation. A
yellow sphere represents Cd2 + bound at the N-terminus. Hydrogen bonds are indicated by dotted lines.
the closed conformation, interactions which are lost following
phosphorylation. Gating by pH involves protonation of His193
in loop D which leads to channel closure through interactions
with one of the Cd2 + ligands. In the absence of Cd2 + , hydrogen
bonds between His193 and the backbone carbonyls of Ser115
and Lys113 in loop B allows for the closed conformation to be
stabilized [27]. Since Ser115 and His193 are conserved in all PIPs
and Ser274 is conserved in the PIP2 subfamily, the proposed
mechanism is likely to be universal for these plant AQPs. The
proposed gating mechanism of PIPs allows for each gating
event to be explained individually. However, in the living plant
cell, a multitude of sometimes contradictory signals is likely
to be in effect simultaneously. Therefore, rather than a simple
on/off scenario, the gating of the channel is better viewed as an
equilibrium between the two channel states being pushed in a
certain direction by the different gating signals.
As with AQPs in general, PIPs are either inhibited by mercury
or insensitive to its effect. In contrast with several mammalian
AQPs (AQP1 [8], AQP2 [28] and AQP3 [29]) as well as
plant AQPs found in the tonoplast membrane (TIPs; tonoplast
intrinsic proteins) [4], the cysteine residues responsible for
mercury inhibition of PIPs have not yet been identified. In fact,
despite containing the same cysteine pattern, some PIPs are
inhibited by mercury, whereas others are not [30].
In order to elucidate whether the effect of mercury on PIPs could
be coupled with the gating mechanism, we set out to characterize
the interaction between mercury and SoPIP2;1, structurally as
well as functionally. In the present study, we present the 2.15 Å (1
c The Authors Journal compilation c 2013 Biochemical Society
Å = 0.1 nm) crystal structure of the SoPIP2;1–mercury complex,
revealing binding to three out of four cysteine residues. Surprisingly, water transport assays using proteoliposomes showed
that mercury activates SoPIP2;1. However, mutational studies
in which the mercury-binding cysteine residues are replaced by
serine failed to identify any of these as being involved in the
mercury activation. We therefore conclude that cysteine residues
alone cannot account for the effect of mercury on PIPs and that
other factors must be taken into consideration.
EXPERIMENTAL
Cloning, expression and purification
SoPIP2;1 mutants were created by vector PCR as described in [31]
using QuikChange® Site-Directed Mutagenesis Kit (Stratagene).
The constructs were transformed into the X33AQY1 strain of
Pichia pastoris. For wild-type SoPIP2;1, previously obtained
clones were used [32]. Overexpression and purification was
carried out as described previously [31] (for details see
Supplementary Online Data at http://www.biochemj.org/bj/
454/bj4540491add.htm).
Crystallization
A protein solution of 15 mg/ml was incubated with 5 mM
HgCl2 overnight on ice, followed by washing to remove excess
Mercury activation of SoPIP2;1
Table 1
Data collection and refinement statistics
Numbers in parentheses represent the highest resolution shell. R free is calculated from 5 % of
reflections omitted during the refinement.
Parameters
Data collection
Beamline
Wavelength
Space group
Cell dimensions
a , b , c (Å)
α, β, γ (◦ )
Resolution (Å)
R merge
I /σ I
Completeness
Redundancy
Refinement
Resolution (Å)
Number of reflections
R work /R free
Number of atoms
Protein
Cd2 +
Hg2 +
Cl −
Water
Detergent
Average B -factors
Protein
Cd2 +
Hg2 +
Cl −
Water
Detergent
RMSDs
Bond lengths (Å)
Bond angles (◦ )
Ramachandran values (%)
Favoured
Allowed
Outliers
PDB code
Values
NSLS X25
0.978
P 21 2 1 2 1
123.5, 141.8, 186.5
90, 90, 90
29.9–2.15 (2.23–2.15)
0.129 (0.639)
9.16 (1.81)
99.8 (98.9)
6.0 (5.6)
493
Data. The osmotic permeability of the vesicles with and
without the addition of 0.5 mM mercury chloride was measured
using stopped-flow spectroscopy (μSFM-20, BioLogic Science
Instruments). The Pf [osmotic permeability coefficient (cm/s)]
value was calculated according to Pf = k/[(S/V 0 ) · V w · Cout ], where
k (s − 1 ) is the fitted rate constant, (S/V 0 ) is the vesicle external
surface to internal initial volume ratio, V w is the partial molar
volume of water (18 cm3 /mol) and Cout is the external osmolality
(350 mOsm/kg) [33]. Vesicle size was measured using dynamic
light scattering (Malvern DLS). A t test was performed to evaluate
the difference in transport rate between wild-type SoPIP2;1 and its
mutants. Since the variances are assumed to be unequal a Welch’s
t test was used. Differences was assumed to be significant when
P < 0.05. For more details see Supplementary Online Data.
RESULTS AND DISCUSSION
29.9–2.15 (2.20–2.15)
167/468 (11/817)
0.195/0.226 (0.249/0.274)
15051
17
32
2
681
280
31.1
34.0
38.8
23.2
37.5
56.8
0.012
1.33
91.6
9.4
0.0
4JC6
HgCl2. After re-concentration, crystallization set-ups were made
using the hanging drop vapour diffusion method (for details see
Supplementary Online Data). Boat-shaped crystals appeared after
a few weeks and were fished and flash-frozen in liquid nitrogen
without any additional cryo-protectant.
Data collection and structure solving
Diffraction data were collected at National Synchrotron Light
Source (NSLS), Brookhaven National Laboratory, U.S.A. The
structure was solved with molecular replacement. A solution
containing two tetramers in the asymmetric unit was found using
a tetramer of the closed wild-type structure of SoPIP2;1 (PDB
code 1Z98) as a search model. The initial model was subjected to
iterative rounds of refinement and model building, resulting in a
final Rwork and Rfree of 19.5 and 22.6 % respectively. Statistics for
the data collection, refinement and model can be found in Table 1.
More details are given in the Supplementary Online Data.
Activity assays
Proteoliposomes containing the different SoPIP2;1 constructs
were prepared as described in the Supplementary Online
Overall structure of the SoPIP2;1–Hg complex
The crystal structure of SoPIP2;1–Hg (SoPIP2;1 in complex
with mercury) was solved at 2.15 Å resolution by molecular
replacement using the closed structure of SoPIP2;1 as the model
(PDB code 1Z98) (Table 1) [26]. The overall structure of
SoPIP2;1–Hg is very similar to the closed structure of SoPIP2;1;
the two structures overlay with a RMSD of 0.26 Å for 996 Cα
atoms (Figure 2A). In all eight protomers, loop D adopts the closed
conformation, efficiently blocking water transport through the
channel by inserting Leu197 into its cytoplasmic opening. Within
each water-conducting channel, seven water molecules line up in
a single file as seen previously in our closed structure of SoPIP2;1
[26].
Eleven β-OG (octyl-β-glucanpyranoside) molecules are seen
surrounding the two tetramers in the asymmetric unit, several
of which are participating in crystal contacts (Figures 2B and
2C). The two tetramers contain six and five β-OG molecules
respectively. These are unevenly distributed between protomers
with four protomers binding two molecules, three protomers binding one molecule and one protomer not binding any. All βOG molecules interact with the same region of the protein,
residues 239–241 at the extracellular side of TM helix 6, either
through hydrogen bonds with the carbonyl oxygens or via stacking
interactions with tryptophan side chains.
The central pore contains a detergent molecule
AQPs have an inherent tendency to crystallize with the fourfold axis of the tetramer coinciding with a four-fold crystal
symmetry axis, making interpretation of electron density in
this area difficult. Since the SoPIP2;1–Hg complex crystallizes
in space group P21 21 21 which lacks four-fold symmetry, this
limitation is avoided. A continuous stretch of elongated electron
density can be seen at the cytoplasmic side of the central pore
in both tetramers. Previously, lipids have been identified at this
position in the structures of human AQP5 [34] (PSF) and AQPZ
from E. coli [35] (PEE). However, a lipid molecule is too long
to fit the electron density in the SoPIP2;1–Hg structure. Instead,
the electron density is modelled as a β-OG molecule for which the
hydrocarbon tail fits perfectly in length (Figure 2D).
The hydrophobic central pore of AQPs has been suggested to
conduct gases such as CO2 , NO and NH3 [36]. However, owing to
experimental difficulties this remains a controversial topic [37].
In the case of human AQP5, the presence of a lipid originating
from the native membrane in the central pore suggests that this
AQP does not participate in CO2 transport due to steric hindrance
c The Authors Journal compilation c 2013 Biochemical Society
494
Figure 2
A. Frick and others
Overall structure of the SoPIP2;1–Hg complex
(A) Structural overlay of the SoPIP2;1–Hg complex (turquoise) and the closed structure of SoPIP2;1 (green). Water molecules from the SoPIP2;1–Hg structure are shown as red spheres. (B) Side
view of the SoPIP2;1–Hg tetramer showing positions of β-OG molecules (orange). (C) Same as in (B) seen from the extracellular side of the membrane. (D) Close-up of the β-OG molecule in the
central pore. Electron density is 2F o –F c contoured at 1.0σ .
[34]. Our observation of a detergent molecule instead of a lipid at
this position in the SoPIP2;1–Hg structure indicates that, at least
from a steric point of view, CO2 transport through the central pore
of SoPIP2;1 could be possible.
Mercury binds to three out of four cysteine residues
In the SoPIP2;1–Hg structure, mercury binds to three out of four
cysteine residues: Cys91 at the C-terminal end of TM helix 2, as
well as Cys127 and Cys132 , both in TM helix 3 (Figure 3A). The
occupancy for the mercury atoms varies between 0.12 and 0.6
with the vast majority having occupancy between 0.5 and 0.6. For
Cys132 , the side chain is seen to adopt two conformations, resulting
in a second minor mercury site with an average occupancy of
0.2. Although the mercury atoms bound to Cys127 and Cys132
binds between protomers in the tetramer and are exposed to the
membrane region, the mercury bound to Cys91 is buried within
the protein, close to the N-terminal end of loop D, with Thr178 as
an additional ligand (Figure 3B).
The fourth cysteine residue with no mercury bound, Cys69 , is
situated in the extracellular loop A, close to the centre of the
tetramer (Figure 3A). As in the closed structure of SoPIP2;1, this
residue is forming a disulfide bridge with the corresponding amino
acid in a neighbouring protomer, hence making it unavailable for
mercury binding. Functional studies of PIPs from Zea mays have
confirmed the participation of this residue in disulfide formation
c The Authors Journal compilation c 2013 Biochemical Society
between protomers in vivo [38]. Interestingly, in the same study,
mutation of this cysteine residue to serine in ZmPIP1;2 was
seen to be involved in mercury inhibition of a heterotetramer
of ZmPIP1;2 and ZmPIP2;5 expressed in oocytes. In contrast,
the corresponding mutant of ZmPIP2;5 had no effect on the
mercury sensitivity of either the heterotetramer or a homotetramer
consisting of ZmPIP2;5 only.
The SoPIP2;1–Hg structure reveals a new Cd2 + -binding site
As in our previously reported closed structure of SoPIP2;1, a
Cd2 + ion can be found at the N-terminus being ligated by Asp28
and Glu31 as well as two water molecules (Figure 3C). The Cd2 +
ion is co-ordinating a hydrogen-bonding network which keeps
loop D in its closed conformation, connecting loop D to the Nterminus via interactions with loop B. This network is central to
the proposed mechanism for gating by phosphorylation in which
phosphorylation of Ser115 in loop B causes a disruption of the
hydrogen-bonding network, thereby releasing loop D from its
anchor [26].
In the 2F o –F c electron density map, a strong peak (>13σ ),
interpreted as a second Cd2 + ion, is found between the C-terminus
and loop D in each protomer (Figure 3B). A corresponding peak
is also present in the anomalous density map, being of similar
magnitude (>5σ ) as the peak for the Cd2 + ion at the N-terminus
(Figure 3C). The second Cd2 + ion has an average occupancy
Mercury activation of SoPIP2;1
Figure 3
495
Mercury- and cadmium-binding sites
(A) Side view of the SoPIP2;1 tetramer showing the locations of heavy atoms. Mercury and cadmium are shown as purple and yellow spheres respectively. Ligands for the heavy atoms are shown in
stick representation with mercury ligands being labelled. (B) Close-up of mercury-binding site at Cys91 . Bonds to Cys91 and Thr178 are shown as dotted lines. Electron density is 2F o –F c contoured
at 1.0σ . (C) View from the cytoplasmic side showing the two Cd2 + -binding sites. Cd2 + ions and water molecules are shown as yellow and red spheres respectively. Bonds are indicated by dotted
lines. (D) Close-up of the new Cd2 + site at the C-terminus (Cd2) showing how it is ligated by Thr183 , Ala267 and four water molecules (red spheres). 2F o –F c electron density (blue) and anomalous
difference density (pink) are contoured at 1.0σ and 3.5σ respectively.
of 0.65–0.7 and is octahedrally co-ordinated by the carbonyl
oxygen of Ala267 in the C-terminus, the side chain of Thr183
in the beginning of loop D as well as four water molecules.
Although not modelled in the first closed structure of SoPIP2;1,
a re-examination of the anomalous density map suggests that a
second Cd2 + ion is present there as well with partial occupancy.
As proposed for the N-terminal Cd2 + site, the C-terminal site
most likely binds Ca2 + in vivo. The presence of two Ca2 + -binding
sites in PIPs is supported by studies on Ca2 + inhibition of the
water transport in Beta vulgaris storage roots in which a biphasic
dose–response curve was found [23]. This led to the suggestion
that PIPs contain two Ca2 + -binding sites; a higher-affinity site
in nanomolar range and a second weaker site in the micromolar
range, both situated on the cytoplasmic side of the protein. This
fits well with the two sites now identified in the SoPIP2;1–Hg
structure.
The biological relevance for the C-terminal Ca2 + site could
be to stabilize the C-terminus for gating purposes. Gating of
SoPIP2;1 by phosphorylation involves Ser274 on the C-terminus, a
residue conserved in PIP2s [21]. The closed structure of SoPIP2;1
revealed the side chain of Ser274 interacting with Pro199 and Leu200
at the end of loop D, thus stabilizing the closed conformation
[26] (Figure 1C). When adopting the open conformation, a steric
clash between Leu197 and Ser274 causes the C-terminus to be
disordered thus this interaction is lost. Since the new Cd2 + site in
the SoPIP2;1–Hg structure connects the C-terminus and loop D,
it is possible that it stabilizes the C-terminus in order to allow for
this interaction to occur. In fact, in all SoPIP2;1 structures without
Cd2 + solved to date, the C-terminus has been disordered beyond
residue 268 [27,31].
SoPIP2;1 is activated by mercury
In order to investigate the effect of mercury on SoPIP2;1, we
measured the water transport rate in a liposome swelling assay. In
this assay, the Pf value of proteoliposomes containing SoPIP2;1
was determined in the absence and presence of 0.5 mM HgCl2
using stopped-flow spectroscopy. Surprisingly, when HgCl2
was present, a higher water transport rate through SoPIP2;1 was
observed (Figure 4A). The Pf value for SoPIP2;1-containing
liposomes increased from 38.6 +
− 4.86 μm/s to 77.0 +
− 4.16 μm/s,
resulting in a total increase in water transport rate of 99.5 %
(Figure 4B). The increase in water permeability was fully
reversible with 2-mercaptoethanol. The Pf value for control
liposomes without SoPIP2;1 was 15.8 +
− 0.31 μm/s.
We have shown previously that a serine-to-glutamate mutation
designed to mimic phosphorylation of Ser188 in loop D (SoPIP2;1
S188E) displays an increased water transport rate in liposome
assays compared with wild-type SoPIP2;1 [31], suggesting that
this mutant is locked in an open state. MD simulations showed
that when this residue was phosphorylated, a conformational
change of loop D led to channel opening via interactions with
the C-terminus.
To determine whether the increased water transport rate
through SoPIP2;1 after mercury treatment is comparable with
that of SoPIP2;1 locked in an open conformation, we performed
the same assay on the SoPIP2;1 S188E mutant. Mercury had
very little effect on the water transport rate through SoPIP2;1
S188E, the Pf increased only 21.2 % from 63.8 +
− 5.01 μm/s to
77.3 +
− 1.02 μm/s (Figure 4B). In addition, wild-type SoPIP2;1
and SoPIP2;1 S188E showed very similar water transport rates
c The Authors Journal compilation c 2013 Biochemical Society
496
Figure 4
A. Frick and others
Functional effect of mercury on wild-type and mutant SoPIP2;1
(A) Typical normalized stopped-flow curves. Proteoliposomes containing wild-type SoPIP2;1 (grey) are activated by the addition of 0.5 mM HgCl2 (dark grey), whereas the slower control liposomes
(light grey) are unaffected (black). (B) The P f for untreated and mercury-treated liposomes containing wild-type SoPIP2;1 or the S188E mutant. The S188E mutant, believed to be locked in an open
conformation, shows a significantly higher transport rate (P < 0.05) than wild-type SoPIP2;1. Both are equally fast after treatment with mercury, indicating the existence of a state where the protein
is maximally open. Error bars represent the S.E.M. (C) The P f for untreated and mercury-treated liposomes containing wild-type SoPIP2;1 or cysteine-to-serine mutants. All mutants show a similar
increase in water transport rate following treatment with 0.5 mM HgCl2 as wild-type SoPIP2;1. Error bars represent the S.E.M.s.
in the presence of mercury, although factors such as varying
protein levels in the two reconstitutions need to be taken
into account. Taken together, this suggests that once in the
open conformation, the addition of mercury does not cause an
increase in water transport, indicating that the activating effect
of mercury is coupled with the open or closed state of the
channel.
An activating effect of mercury on AQPs has only been reported
previously for rat AQP6 [39]. AQP6 is found in intracellular
vesicles in renal epithelia and has low intrinsic water permeability.
However, when expressed in oocytes, treatment with mercury
resulted in a 10-fold increase in water transport. This was
accompanied by anion conductance which could also be induced
by acidic pH in the absence of mercury. Site-directed mutagenesis
has identified Cys155 and Cys190 as being responsible for the
mercury effect in AQP6 [40]. It was suggested that when mercury
binds to these two cysteine residues, AQP6 is trapped in a
conformational state which allows water and ion transport.
Interestingly, Cys190 corresponds to Cys189 in AQP1, the residue
known to be involved in mercury inhibition in this and other
AQPs, suggesting that the inhibitory effect of mercury on AQPs
is not merely a question of whether a cysteine residue is present
within the water-conducting pore or not.
c The Authors Journal compilation c 2013 Biochemical Society
Mercury activation of SoPIP2;1 cysteine-to-serine mutants does
not significantly differ from wild-type
To investigate which of the cysteine residues in SoPIP2;1 is
responsible for mercury activation, we made a series of mutants
in which the cysteine residues were replaced with serine. In
addition to making the single mutants C69S, C91S, C127S and
C132S, we also made the double mutant C127S/C132S as well as
the triple mutant C91S/C127S/C132S. The effect of mercury on
water transport through these mutants was assayed in liposomes
as described above.
In contrast with what could be expected, all SoPIP2;1 mutants
displayed a similar response to mercury treatment as wild-type
SoPIP2;1 (P > 0.05), with the Pf value increasing 59–164 %
(Figure 4C). One single mutant, C132S, stands out as having
a lower degree of activation than the others with only 59 %
increase in Pf value, however, the double mutant C127S/C132S
as well as the triple mutant C91S/C127S/C132S shows no such
effect. We therefore conclude that neither C132S nor any of the
other cysteine residues investigated in the present study are likely
to be involved in mercury activation of SoPIP2;1. This is in
striking contrast with AQP6 where the C155A/C190A double
mutant displayed negligible mercury activation, identifying these
residues as responsible for mercury activation of AQP6 [40].
Mercury activation of SoPIP2;1
Cysteine residues and mercury response in plant AQPs
From analysis of the literature, a complex picture of the relationship between cysteine residues and mercury response in plant
AQPs emerges (for a sequence alignment, see Supplementary Figure S1 at http://www.biochemj.org/bj/454/bj4540491add.htm).
Site-directed mutagenesis identified the cysteine residue
responsible for mercury inhibition of two TIPs from A. thaliana as
being located in TM helix 3, the same helix as Cys127 and Cys132
in SoPIP2;1 [4]. The sequence alignment reveals that although
there is no exact match, this cysteine is only two residues away
from Cys132 . However, in a study of eight PIPs from Populus
trichocarpa, three variants containing the same cysteine residues
as SoPIP2;1, as well as the pin-pointed cysteine from the A.
thaliana TIPs displayed varying mercury sensitivity with two
being inhibited and one being insensitive [30]. Of the remaining
five PIPs included in the study, all were inhibited by mercury
despite having the exact same pattern of cysteine residues as
SoPIP2;1. Interestingly, a slight increase in water transport after
mercury treatment was seen for a cysteine-lacking AQP from
the legume Medicago truncatula (MtAQP1) when expressed in
oocytes, although this was not commented on by the authors of the
paper [41]. Taken together, this raises the possibility that the effect
of mercury on plant AQPs might not be attributable to cysteine
residues at all.
Physical properties of the lipid bilayer could explain mercury
activation
If cysteine residues are not responsible for the mercury response of
SoPIP2;1, an alternative could be a secondary effect via changes
to the properties of the lipid bilayer. Mercury has been shown to
decrease the fluidity of the membrane by binding to the amine head
groups of PE (phosphoethanolamine) [42], the major constituent
of the E. coli lipid extracts used for liposomes in the present
study. A decrease in membrane fluidity has been connected to a
lower transport rate of water and other compounds across the lipid
bilayer itself [43] which could explain why in our study control
liposome displays a lower Pf value (15 % decrease) after mercury
treatment. In eukaryotes, membrane fluidity is often regulated
by varying the amount of sterols present, with an increase in
sterols leading to a stiffer membrane [44]. Plasma membranes
from higher plants contain a number of sterol derivatives of which
several have shown to significantly alter membrane fluidity [45].
Many studies have shown that protein-mediated transport is
affected by changes in the physical properties of the membrane. It
is believed that this could occur via a shift in equilibrium between
conformational states [46]. In the case of AQPs, mechanical
stimuli has been proposed to gate AQY1 from P. pastoris,
for which increased lateral pressure as well as bent membrane
induced spontaneous opening of the channel in MD simulations.
In addition, stopped-flow measurements of AQY1-mediated water
transport displayed a significantly higher water transport rate in
liposomes compared with the larger spheroplasts, supporting the
suggestion that membrane curvature affects the conformational
state of AQY1. The MD simulations revealed how external forces
triggering gating could be transmitted from the lipid bilayer to
the protein via coupled movements of TM helices 4, 5 and 6,
resulting in pore widening around residues Leu189 , Ala190 and
Val191 . Together with Tyr31 , these residues form a constriction
site at the cytoplasmic side of the water-conducting pore in an
analogous manner to what is seen for SoPIP2;1. It is therefore
possible that SoPIP2;1 could respond to mechanical stimuli in a
similar way.
497
Mechanical stimuli has also been suggested to gate AQPs in
young corn roots in which case applied pressure pulses resulted
in decreased membrane water transport, most likely due to a
conformational change of the protein [47]. In addition, a study
of human AQP4 revealed that a stiffer membrane due to increased
cholesterol amounts resulted in reduced water permeability [48].
The authors argued that this was due to minor conformational
changes of the TM helices, suggesting that large structural
perturbations are not always necessary in order for mechanical
stress to have an effect on water transport through AQPs.
It should be noted that a structural comparison of AQP0 in 2D
crystals formed using two different sets of lipids (dimyristoyl
phospatatidyl choline and E. coli polar lipids) revealed the
structures to be very much the same [49]. However, as pointed
out by the authors themselves, although the lipids differ in both
head groups and acyl chains, it is possible that the hydrophobic
thickness of the lipid bilayers formed are too similar to induce any
observable changes to the AQP0 structure. It should also be kept
in mind that a crystal environment might not completely mimic
the forces exerted by a lipid bilayer in a cell or a liposome and
that protein conformational changes could be limited by crystal
contacts. Alternatively, it is also possible that different AQPs,
although structurally very similar, vary in their sensitivity towards
structural perturbations due to mechanical stress.
Taken together, several studies have shown that water transport
through AQPs can be affected by mechanical stimuli, one of
which could be the fluidic properties of the membrane. It
is therefore possible that decreased membrane fluidity affects
the conformational state of SoPIP2;1, pushing the equilibrium
towards an open conformation. Although in our experimental setup this could be due to mercury binding to lipid head groups, it
is plausible that in the plant plasma membrane, such mechanical
gating could occur via changes in membrane lipid composition.
Mercury is often seen as a general inhibitor for AQPs and has
traditionally been used to prove AQP-mediated transport. Data
from the present study emphasize that this is an oversimplification
and that there are many exceptions to this rule. In particular for
plant AQPs, it is evident that the effect of mercury is more
complex than what has been thought previously. Although a
cysteine-mediated effect seems immediately obvious, our study
on SoPIP2;1 reveals that this is not necessarily always the case and
that other factors could play important roles. Most importantly,
since mercury inhibition experiments are often performed using
proteoliposomes, the possibility of mercury affecting their
physical properties which, in turn, could have implications for
protein function is a factor that should be taken into consideration.
Whether increased membrane stiffness has a physiological role
for the gating of SoPIP2;1 remains speculative, and in order to
confirm or dispute this further experiments are needed.
By showing that SoPIP2;1 is activated by mercury and that
this cannot be attributed to any of its cysteine residues we
have significantly contributed to the understanding of how AQPs
respond to mercury treatment. In addition, the SoPIP2;1–Hg
structure have given important new insights into the function of
plant AQPs by revealing a new Cd2 + -binding site which could
have implications for gating.
AUTHOR CONTRIBUTION
Anna Frick, Michael Järvå and Susanna Törnroth-Horsefield planned the work. Anna Frick
and Susanna Törnroth-Horsefield did the structural work. Anna Frick, Michael Järvå,
Mikael Ekvall, Povilas Uzdavinys and Maria Nyblom prepared the protein and carried out
the functional assay. All of the authors were involved in data analysis and interpretation.
Anna Frick, Michael Järvå and Susanna Törnroth-Horsefield wrote the paper.
c The Authors Journal compilation c 2013 Biochemical Society
498
A. Frick and others
ACKNOWLEDGEMENTS
We thank the staff at National Synchrotron Light Source (NSLS) for help with data collection,
Gerhard Fischer and Ida Lundholm for help with interpreting the stopped-flow data and
Richard Neutze for useful discussions.
FUNDING
This work has been supported by the Swedish Research Council [grant number 2010–5208
(to S.T.-H.)].
REFERENCES
1 Walz, T., Fujiyoshi, Y. and Engel, A. (2009) The AQP structure and functional implications.
Handb. Exp. Pharmacol. 190, 31–56
2 Macey, R. I. (1984) Transport of water and urea in red blood cells. Am. J. Physiol. 246,
C195–C203
3 Preston, G. M., Carroll, T. P., Guggino, W. B. and Agre, P. (1992) Appearance of water
channels in Xenopus oocytes expressing red cell CHIP28 protein. Science 256, 385–387
4 Daniels, M. J., Chaumont, F., Mirkov, T. E. and Chrispeels, M. J. (1996) Characterization
of a new vacuolar membrane aquaporin sensitive to mercury at a unique site. Plant Cell 8,
587–599
5 Roberts, S. K., Yano, M., Ueno, Y., Pham, L., Alpini, G., Agre, P. and LaRusso, N. F. (1994)
Cholangiocytes express the aquaporin CHIP and transport water via a channel-mediated
mechanism. Proc. Natl. Acad. Sci. U.S.A. 91, 13009–13013
6 Lazowski, K. W., Li, J., Delporte, C. and Baum, B. J. (1995) Evidence for the presence of a
Hg-inhibitable water-permeability pathway and aquaporin 1 in A5 salivary epithelial cells.
J. Cell. Physiol. 164, 613–619
7 Fushimi, K., Sasaki, S., Yamamoto, T., Hayashi, M., Furukawa, T., Uchida, S., Kuwahara,
M., Ishibashi, K., Kawasaki, M., Kihara, I. et al. (1994) Functional characterization and cell
immunolocalization of AQP-CD water channel in kidney collecting duct. Am. J. Physiol.
267, F573–F582
8 Preston, G. M., Jung, J. S., Guggino, W. B. and Agre, P. (1993) The mercury-sensitive
residue at cysteine 189 in the CHIP28 water channel. J. Biol. Chem. 268, 17–20
9 Savage, D. F. and Stroud, R. M. (2007) Structural basis of aquaporin inhibition by
mercury. J. Mol. Biol. 368, 607–617
10 Hirano, Y., Okimoto, N., Kadohira, I., Suematsu, M., Yasuoka, K. and Yasui, M. (2010)
Molecular mechanisms of how mercury inhibits water permeation through aquaporin-1:
understanding by molecular dynamics simulation. Biophys. J. 98, 1512–1519
11 Yukutake, Y., Tsuji, S., Hirano, Y., Adachi, T., Takahashi, T., Fujihara, K., Agre, P., Yasui, M.
and Suematsu, M. (2008) Mercury chloride decreases the water permeability of
aquaporin-4-reconstituted proteoliposomes. Biol. Cell 100, 355–363
12 Ishibashi, K., Morinaga, T., Kuwahara, M., Sasaki, S. and Imai, M. (2002) Cloning and
identification of a new member of water channel (AQP10) as an aquaglyceroporin.
Biochim. Biophys. Acta, Gene Struct. Expression 1576, 335–340
13 Hatakeyama, S., Yoshida, Y., Tani, T., Koyama, Y., Nihei, K., Ohshiro, K., Kamiie, J. I.,
Yaoita, E., Suda, T., Hatakeyama, K. and Yamamoto, T. (2001) Cloning of a new aquaporin
(AQP10) abundantly expressed in duodenum and jejunum. Biochem. Biophys. Res.
Commun. 287, 814–819
14 Yakata, K., Tani, K. and Fujiyoshi, Y. (2011) Water permeability and characterization of
aquaporin-11. J. Struct. Biol. 174, 315–320
15 Hasegawa, H., Ma, T., Skach, W., Matthay, M. A. and Verkman, A. S. (1994) Molecular
cloning of a mercurial-insensitive water channel expressed in selected water-transporting
tissues. J. Biol. Chem. 269, 5497–5500
16 Ishibashi, K., Kuwahara, M., Gu, Y., Kageyama, Y., Tohsaka, A., Suzuki, F., Marumo, F. and
Sasaki, S. (1997) Cloning and functional expression of a new water channel abundantly
expressed in the testis permeable to water, glycerol, and urea. J. Biol. Chem. 272,
20782–20786
17 Calamita, G., Bishai, W. R., Preston, G. M., Guggino, W. B. and Agre, P. (1995) Molecular
cloning and characterization of AqpZ, a water channel from Escherichia coli . J. Biol.
Chem. 270, 29063–29066
18 Daniels, M. J., Mirkov, T. E. and Chrispeels, M. J. (1994) The plasma-membrane of
Arabidopsis thaliana contains a mercury-insensitive aquaporin that is a homolog of the
tonoplast water channel protein TIP. Plant Physiol. 106, 1325–1333
19 Maurel, C. (2007) Plant aquaporins: novel functions and regulation properties. FEBS Lett.
581, 2227–2236
20 Fischer, M. and Kaldenhoff, R. (2008) On the pH regulation of plant aquaporins. J. Biol.
Chem. 283, 33889–33892
21 Johansson, I., Karlsson, M., Shukla, V. K., Chrispeels, M. J., Larsson, C. and Kjellbom, P.
(1998) Water transport activity of the plasma membrane aquaporin PM28A is regulated by
phosphorylation. Plant Cell 10, 451–459
c The Authors Journal compilation c 2013 Biochemical Society
22 Gerbeau, P., Amodeo, G., Henzler, T., Santoni, V., Ripoche, P. and Maurel, C. (2002) The
water permeability of Arabidopsis plasma membrane is regulated by divalent cations and
pH. Plant J. 30, 71–81
23 Alleva, K., Niemietz, C. M., Maurel, C., Parisi, M., Tyerman, S. D. and Amodeo, G. (2006)
Plasma membrane of Beta vulgaris storage root shows high water channel activity
regulated by cytoplasmic pH and a dual range of calcium concentrations. J. Exp. Bot. 57,
609–621
24 Verdoucq, L., Grondin, A. and Maurel, C. (2008) Structure–function analysis of plant
aquaporin AtPIP2;1 gating by divalent cations and protons. Biochem. J. 415, 409–416
25 Tournaire-Roux, C., Sutka, M., Javot, H., Gout, E., Gerbeau, P., Luu, D. T., Bligny, R. and
Maurel, C. (2003) Cytosolic pH regulates root water transport during anoxic stress
through gating of aquaporins. Nature 425, 393–397
26 Tornroth-Horsefield, S., Wang, Y., Hedfalk, K., Johanson, U., Karlsson, M., Tajkhorshid,
E., Neutze, R. and Kjellbom, P. (2006) Structural mechanism of plant aquaporin gating.
Nature 439, 688–694
27 Frick, A., Jarva, M. and Tornroth-Horsefield, S. (2013) Structural basis for pH gating of
plant aquaporins. FEBS Lett. 587, 989–993
28 Bai, L. Q., Fushimi, K., Sasaki, S. and Marumo, F. (1996) Structure of aquaporin-2
vasopressin water channel. J. Biol. Chem. 271, 5171–5176
29 Kuwahara, M., Gu, Y., Ishibashi, K., Marumo, F. and Sasaki, S. (1997) Mercury-sensitive
residues and pore site in AQP3 water channel. Biochemistry 36, 13973–13978
30 Secchi, F., Maciver, B., Zeidel, M. L. and Zwieniecki, M. A. (2009) Functional analysis of
putative genes encoding the PIP2 water channel subfamily in Populus trichocarpa . Tree
Physiol. 29, 1467–1477
31 Nyblom, M., Frick, A., Wang, Y., Ekvall, M., Hallgren, K., Hedfalk, K., Neutze, R.,
Tajkhorshid, E. and Tornroth-Horsefield, S. (2009) Structural and functional analysis of
SoPIP2;1 mutants adds insight into plant aquaporin gating. J. Mol. Biol. 387,
653–668
32 Karlsson, M., Fotiadis, D., Sjovall, S., Johansson, I., Hedfalk, K., Engel, A. and Kjellbom,
P. (2003) Reconstitution of water channel function of an aquaporin overexpressed and
purified from Pichia pastoris . FEBS Lett. 537, 68–72
33 van Heeswijk, M. P. and van Os, C. H. (1986) Osmotic water permeabilities of brush
border and basolateral membrane vesicles from rat renal cortex and small intestine.
J. Membr. Biol. 92, 183–193
34 Horsefield, R., Norden, K., Fellert, M., Backmark, A., Tornroth-Horsefield, S., Terwisscha
van Scheltinga, A. C., Kvassman, J., Kjellbom, P., Johanson, U. and Neutze, R. (2008)
High-resolution x-ray structure of human aquaporin 5. Proc. Natl. Acad. Sci. U.S.A. 105,
13327–13332
35 Jiang, J., Daniels, B. V. and Fu, D. (2006) Crystal structure of AqpZ tetramer reveals two
distinct Arg-189 conformations associated with water permeation through the narrowest
constriction of the water-conducting channel. J. Biol. Chem. 281, 454–460
36 Herrera, M. and Garvin, J. L. (2011) Aquaporins as gas channels. Pfluegers Arch. 462,
623–630
37 de Groot, B. L. and Hub, J. S. (2011) A decade of debate: significance of CO2 permeation
through membrane channels still controversial. ChemPhysChem 12, 1021–1022
38 Bienert, G. P., Cavez, D., Besserer, A., Berny, M. C., Gilis, D., Rooman, M. and Chaumont,
F. (2012) A conserved cysteine residue is involved in disulphide bond formation between
plant plasma membrane aquaporin monomers. Biochem. J. 445, 101–111
39 Yasui, M., Hazama, A., Kwon, T. H., Nielsen, S., Guggino, W. B. and Agre, P. (1999) Rapid
gating and anion permeability of an intracellular aquaporin. Nature 402, 184–187
40 Hazama, A., Kozono, D., Guggino, W. B., Agre, P. and Yasui, M. (2002) Ion permeation of
AQP6 water channel protein. Single channel recordings after Hg2 + activation. J. Biol.
Chem. 277, 29224–29230
41 Krajinski, F., Biela, A., Schubert, D., Gianinazzi-Pearson, V., Kaldenhoff, R. and Franken,
P. (2000) Arbuscular mycorrhiza development regulates the mRNA abundance of Mtaqp 1
encoding a mercury-insensitive aquaporin of Medicago truncatula . Planta 211,
85–90
42 Delnomdedieu, M., Boudou, A., Desmazès, J. and Georgescauld, D. (1989) Interaction of
mercury chloride with the primary amine group of model membranes containing
phosphatidylserine and phosphatidylethanolamine. Biochim. Biophys. Acta 986,
191–199
43 Lande, M. B., Donovan, J. M. and Zeidel, M. L. (1995) The relationship between
membrane fluidity and permeabilities to water, solutes, ammonia, and protons. J. Gen.
Physiol. 106, 67–84
44 Bastiaanse, E. M., Hold, K. M. and Van der Laarse, A. (1997) The effect of membrane
cholesterol content on ion transport processes in plasma membranes. Cardiovasc. Res.
33, 272–283
45 Schuler, I., Milon, A., Nakatani, Y., Ourisson, G., Albrecht, A. M., Benveniste, P. and
Hartman, M. A. (1991) Differential effects of plant sterols on water permeability and on
acyl chain ordering of soybean phosphatidylcholine bilayers. Proc. Natl. Acad. Sci. U.S.A.
88, 6926–6930
Mercury activation of SoPIP2;1
46 Levitan, I., Christian, A. E., Tulenko, T. N. and Rothblat, G. H. (2000) Membrane
cholesterol content modulates activation of volume-regulated anion current in bovine
endothelial cells. J. Gen. Physiol. 115, 405–416
47 Wan, X., Steudle, E. and Hartung, W. (2004) Gating of water channels (aquaporins) in
cortical cells of young corn roots by mechanical stimuli (pressure pulses): effects of ABA
and of HgCl2 . J. Exp. Bot. 55, 411–422
499
48 Tong, J., Briggs, M. M. and McIntosh, T. J. (2012) Water permeability of aquaporin-4
channel depends on bilayer composition, thickness, and elasticity. Biophys. J. 103,
1899–1908
49 Hite, R. K., Li, Z. and Walz, T. (2010) Principles of membrane protein interactions
with annular lipids deduced from aquaporin-0 2D crystals. EMBO J. 29,
1652–1658
Received 13 March 2013/7 June 2013; accepted 3 July 2013
Published as BJ Immediate Publication 3 July 2013, doi:10.1042/BJ20130377
c The Authors Journal compilation c 2013 Biochemical Society
Biochem. J. (2013) 454, 491–499 (Printed in Great Britain)
doi:10.1042/BJ20130377
SUPPLEMENTARY ONLINE DATA
Mercury increases water permeability of a plant aquaporin through
a non-cysteine-related mechanism
Anna FRICK1 , Michael JÄRVÅ1 , Mikael EKVALL, Povilas UZDAVINYS, Maria NYBLOM and Susanna TÖRNROTH-HORSEFIELD2
Department of Chemistry and Molecular Biology, University of Gothenburg, PO Box 462, SE-405 30 Gothenburg, Sweden
METHODS
Crystallization
Generation of mutants
A protein solution of 15 mg/ml was incubated with 5 mM HgCl2
overnight on ice followed by dilution with buffer (20 mM Tris,
pH 7.5, 100 mM NaCl and 1 % β-OG) to wash away excess
HgCl2 . The protein solution was again concentrated to 15 mg/ml
and crystallization set-ups were made using the hanging drop
vapour diffusion method. The protein solution was mixed in a 1:1
ratio with reservoir solution containing 26 % PEG 400, 25 mM
Tris, pH 7.0, and 100 mM NaCl, to which 100 mM CdCl2 had been
added as an additive in a 1:4 ratio. The crystallization set-ups were
left to equilibrate at 4 ◦ C. Boat-shaped crystals appeared after a
few weeks and were fished and flash-frozen in liquid nitrogen
without any additional cryo-protectant.
The mutants were generated from the wild-type construct through
vector PCR following the protocol from QuikChange® SiteDirected Mutagenesis Kit (Stratagene). All constructs contained
an N-terminal 6×His-tag followed by a thrombin cleavage site
and were transformed into an AQY1-deficient strain of P. pastoris
GS115 aqy1Δ [1] using electroporation.
Heterologous membrane protein overproduction
Wild-type and mutant SoPIP2;1 were overproduced in the
methylotrophic yeast P. pastoris. The strain was grown in a
3 litre fermentor typically resulting in 280 g cells/l culture after
24 h of methanol induction. An X-press (Biox AB) was used to
break the cells and after debris and unbroken cells were removed
the membranes were collected using centrifugation (150 000 g,
90 min, 4 ◦ C). The membranes were washed two times, first in
urea (4 M urea, 50 mM Tris/HCl, pH 9.5, 2 mM EDTA and 2 mM
EGTA), then in 20 mM NaOH to get rid of peripheral membrane
proteins. As a last step the membranes were resuspended
in resuspension buffer (20 mM Hepes/NaOH, pH 7.0, 50 mM
NaCl, 10 % glycerol and 2 mM 2-mercaptoethanol) to a final
concentration of approximately 10 mg membrane proteins/ml
buffer and stored at − 80 ◦ C until further use.
Purification of protein for functional assays and crystallization
Membranes containing SoPIP2;1 were solubilized in resuspension buffer by drop-wise addition of β-OG (Anatrace) to a
final concentration of 5 %. After stirring at room temperature
(20 ◦ C) for 30 min, unsolubilized material was spun down at
180 000 g for 30 min at 4 ◦ C. The supernatant was filtered through
a 0.45 μm filter and loaded on to a Resource S (GE Healthcare)
cation exchange column pre-equilibrated with resuspension buffer
plus 1 % β-OG. After washing away unbound material, SoPIP2;1
was eluted using a linear gradient of 50 mM–0.5 M NaCl over
five-column volumes. After SDS/PAGE (4–12 % gel) analysis,
fractions containing SoPIP2;1 was concentrated in a VivaSpin
concentrator with 10 000 molecular mass cut-off (Sartorius
Stedim Biotech). The concentrated protein solution was filtered
through a 0.2 μm filter and injected on to a Superdex 200
10/300HK gel filtration column (GE Healthcare) equilibrated with
20 mM Tris/HCl, pH 7.5, 100 mM NaCl and 1 % β-OG. Fractions
were analysed, pooled and concentrated as described above. The
resulting protein concentration was measured using NanoDropTM
(typically 15 mg/ml).
Data collection and processing
Complete data at 100 K from a frozen crystal was collected on
beamline X25 of the NSLS, Brookhaven National Laboratory,
U.S.A. Image data were processed and scaled using the HKL2000
package [2] to a resolution of 2.15 Å. Crystals belonged to space
group P21 21 21 with two tetramers in the asymmetric unit. The
cell dimensions were a = 123.5 Å, b = 141.8 Å, c = 186.5 Å,
α = β = γ = 90◦ .
Structural determination
Molecular replacement was carried out using the program
MOLREP from the CCP4 program suite [3] with a tetramer of
the closed structure of SoPIP2;1 at pH 8 (PDB code 1Z98 as the
search model). A clear molecular replacement solution was found
with a scoring function and R-factor of 56.7 % and 51.0 %
respectively was found. The crystal packing was checked using
the program Coot [4] and showed no overlaps between molecules.
The model was subjected to iterative rounds of refinement
in REFMAC5 [5] and CNS [6] with manual rebuilding into
CNS-generated composite omit maps between each round. NCS
restraints between the eight monomers in the asymmetric unit
were used. The model consists of eight protein chains (A, B, C,
D, H, J, L and N) with the following residues: A 23–274, B 23–
274, C 23–272, D 23–273, H 23–274, J 24–274, L 23–273 and N
23–272. In addition, there are 17 cadmium ions, 32 mercury ions,
two chloride ions, 14 β-OG molecules and 621 water molecules.
The final Rwork and Rfree for this model is 19.5 % and 22.6 %
respectively. The model quality was checked using Procheck [7].
For statistics on data collection, refinement and the model, see
Table 1 in the main text. The co-ordinates have been deposited in
the PDB under accession code 4JC6.
1
These authors contributed equally to this work.
To whom correspondence should be addressed (email [email protected]).
The structural co-ordinates of the SoPIP2;1–mercury complex will appear in the PDB under accession code 4JC6.
2
c The Authors Journal compilation c 2013 Biochemical Society
A. Frick and others
Figure S1
Sequence alignment of plant and mammalian AQPs discussed in the main text
Cysteine residues and NPA motifs are highlighted in orange and blue respectively. The alignment was carried out using the program Clustal W (http://www.genome.jp/tools/clustalw/).
Continues. . .
c The Authors Journal compilation c 2013 Biochemical Society
Mercury activation of SoPIP2;1
Figure S1
Continued.
c The Authors Journal compilation c 2013 Biochemical Society
A. Frick and others
Preparation of proteoliposomes
Liposomes were created from E. coli polar lipid extract (Avanti
Polar Lipids) by sonication in reconstitution buffer (50 mM NaCl
and 50 mM Tris/HCl, pH 8.0). A β-OG stock solution (50 mM
NaCl, 50 mM Tris/HCl, pH 8.0, and 10 % β-OG) was added to a
concentration of 1.1 % and the different SoPIP2;1 constructs were
added to a theoretical LPR (lipid-to-protein ratio) of 50. After
30 min incubation on an orbital shaker (130 rev./min) at room
temperature the solutions were diluted in a two-step manner. First,
reconstitution buffer was added to lower the β-OG concentration
to 0.6 % for 1 min. They were further diluted to a total dilution
factor of at least 50. The samples were centrifuged for at least
60 min at 138 000 g. The supernatant was discarded and the
pellet slowly resuspended in reconstitution buffer to 2 mg lipid/ml
using reconstitution buffer and then filtered through a 0.2 μm
filter.
Activity assays
The osmotic permeability of the vesicles was measured
using stopped-flow spectroscopy (μSFM-20, BioLogic Science
Instruments). Liposome solution (74 μl) and sucrose solution
(74 μl) were mixed and the light scattering at 90◦ was measured
at 436 nm for 1.6 s. HgCl2 (500 μM) was added 15 min before
the experiment. Vesicle size was measured using dynamic light
scattering (Malvern DLS).
Data analysis
At least three traces were averaged and fitted to either single
(control liposomes) or double (proteoliposomes) exponential
functions. The Pf (cm/s) was calculated according to
Pf = k/[(S/V 0 ) · V w · Cout ], where k (s − 1 ) is the fitted rate constant,
Received 13 March 2013/7 June 2013; accepted 3 July 2013
Published as BJ Immediate Publication 3 July 2013, doi:10.1042/BJ20130377
c The Authors Journal compilation c 2013 Biochemical Society
(S/V 0 ) is the vesicle external surface to internal initial volume
ratio, V w is the partial molar volume of water (18 cm3 /mol) and
Cout is the external osmolality (350 mOsm/kg) [8]. All curve fits
was done in the interval 0.016–1.0 s and had an R2 >0.99. A t
test was performed to evaluate the difference in transport rate
between wild-type SoPIP2;1 and its mutants. Since the variances
are assumed to be unequal a Welch’s t test was used. Differences
was assumed to be significant when P < 0.05. When applicable,
error bars are shown as S.E.M.
REFERENCES
1 Fischer, G., Kosinska-Eriksson, U., Aponte-Santamaria, C., Palmgren, M., Geijer, C.,
Hedfalk, K., Hohmann, S., de Groot, B. L., Neutze, R. and Lindkvist-Petersson, K. (2009)
Crystal structure of a yeast aquaporin at 1.15 angstrom reveals a novel gating mechanism.
PLoS Biol. 7, e1000130
2 Otwinowski, Z. and Minor, W. (1997) Processing of X-ray diffraction data collected in
oscillation mode. In Methods in Enzymology (Carter, Jr, C. W. and Sweet, R. M., eds),
pp. 307–326, Academic Press, New York
3 Vagin, A. and Teplyakov, A. (2010) Molecular replacement with MOLREP. Acta Crystallogr.,
Sect. D: Biol. Crystallogr. 66, 22–25
4 Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 60, 2126–2132
5 Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P., Evans, P. R., Keegan,
R. M., Krissinel, E. B., Leslie, A. G., McCoy, A. et al. (2011) Overview of the CCP4 suite
and current developments. Acta Crystallogr., Sect. D: Biol. Crystallogr. 67, 235–242
6 Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve,
R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S. et al. (1998) Crystallography &
NMR system: a new software suite for macromolecular structure determination. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 54, 905–921
7 Laskowski, R. A., Macarthur, M. W, Moss, D. S. and Thornton, J. M. (1993) PROCHECK: a
program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26,
283–291
8 van Heeswijk, M. P. and van Os, C. H. (1986) Osmotic water permeabilities of brush border
and basolateral membrane vesicles from rat renal cortex and small intestine. J. Membr.
Biol. 92, 183–193