Biochimica et Biophysica Acta 1808 (2011) 1581–1586
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b b a m e m
Water permeation dynamics of AqpZ: A tale of two states
Lin Xin a,b,f, Haibin Su c, Claus Helix Nielsen d,e, Chuyang Tang a,f, Jaume Torres b, Yuguang Mu b,⁎
a
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
School of Biological Sciences, Nanyang Technological University, Singapore
c
School of Material Sciences, Nanyang Technological University, Singapore
d
DTU Physics, Technical University of Denmark, DK2800 Lyngby, Denmark
e
Aquaporin A/S, Ole Maaløes Vej 3 DK2200 Copenhagen, Denmark
f
Singapore Membrane Technology Centre, Nanyang Technological University, Singapore
b
a r t i c l e
i n f o
Article history:
Received 31 August 2010
Received in revised form 27 January 2011
Accepted 1 February 2011
Available online 18 February 2011
Keywords:
AqpZ dual permeation states
PMF
Structure–permeability relation
MD simulations
a b s t r a c t
Molecular dynamics simulations of aquaporin Z homotetramer which is a membrane protein facilitating rapid
water movement through the plasma membrane of Escherichia coli were performed. Initial configurations
were taken from the open and closed states of crystal structures separately. The resulting water osmotic
permeability (pf) and diffusive permeability (pd) displayed distinct features. Consistent with previous studies,
the side chain conformation of arginine189 was found to mediate the water permeability. A potential of mean
force (PMF) as a function of the distance between NH1 of R189 and carbonyl oxygen of A117 was constructed
based on the umbrella sampling technique. There are multiple local minima and transition states on the PMF.
The assignment of the open or closed state was supported by the permeability pf, calculated within
trajectories in umbrella sampling simulations. Our study disclosed a detailed mechanism of the gated water
transport.
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
Aquaporins are a ubiquitous family of intrinsic membrane proteins
which are discovered in a wide range of organisms from archae,
bacteria, plants, insects to mammals [1]. Water molecules pass
through aquaporin channels at almost the diffusion rate yet the
pore retains high selectivity [2–8]. With different solute transportation ability, aquaporins perform a variety of functions [9–11] in
different organisms. In prokaryotic and other microorganisms,
aquaporins provide protection to cell against osmotic shock and
rapid freezing [12]. In eukaryotes, aquaporins have been found to
perform a diverse range of physiological functions, such as maintaining lens transparency in the eye [13], concentrating urine in the
kidneys [14], maintaining water homeostasis within the brain [15],
facilitating rapid response to osmotic shock in yeast [16], and
regulating cell osmolarity within plants [17].
Pioneered by the electron crystallography study on the human
water channel AQP1 [18], there are 41 aquaporin crystal structures
in total [19]. They are from 10 different aquaporin subfamilies:
AQP0 [8,18], GlpF [20,21], AQPZ [22,23], SoPIP2;1 [24,25], AQPM [26],
PfAQP [27], AQP4, AQP5 [28] and Aqy1 [29]. All the reported structures share a conserved structural fold which consists of five loops and
six transmembrane α-helices surrounding a single narrow pathway
⁎ Corresponding author.
E-mail address: [email protected] (Y. Mu).
0005-2736/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.bbamem.2011.02.001
through the membrane. Loops B and E which contain the aquaporin
signature NPA-motif (asparagine, proline, and alanine), form half
transmembrane helices and dip into the channel from opposite sides
of the membrane, creating a seventh pseudo-helix with the NPAmotifs placed at the center of the channel. A constriction region
contains a conserved aromatic/arginine (ar/R) motif which exists at
the extracellular side of the channel and forms the selective filter [30].
These structures provide solid bases for molecular dynamics (MD)
simulation study of the underlying mechanism for water permeation,
solute selectivity, ion/proton exclusion and channel gating. A few
theoretical works have been carried out addressing gating mechanism
of different aquaporins [31,32]. A recent MD study showed that
starting from the closed state, by either truncating the D-loop or
introducing R190A and D191A double mutations, the conformational
equilibrium of SoPIP2;1, a plant water channel, shifted toward the
open state [31]. Furthermore, the relation between the calculated
single channel osmotic permeability pf and the distance between the
D-loop and the N-terminus supported the notion that the open and
closed states mainly differed on the D-loop conformation.
Similar to the SoPIP2;1, a gating mechanism has been proposed for
AqpZ, a membrane protein facilitating rapid water movements
through the plasma membrane of Escherichia coli. The open and
closed states of AqpZ were related to the side chain conformation of a
residue R189 which is within a highly conserved sequence box of
NPAR (Asn–Pro–Ala–Arg) resolved by a comprehensive evolutionary
analysis of aquaporins [33]. In one crystal structure of AqpZ, two
protomers were obtained with different side chain conformations of
1582
L. Xin et al. / Biochimica et Biophysica Acta 1808 (2011) 1581–1586
R189 [23]. Not surprisingly MD simulations of AqpZ observed that
R189 flipped between the two distinct yet stable conformations
[34,35]: one in which the guanidinium group of R189 is in up
configuration according to the ∠ Cγ − Cδ − Nε − Cς dihedral angle, and
the channel is open; the other in which the guanidinium group of
R189 is in down configuration and renders a closed channel. Recent
crystal structure of AqpZ also supported these observations [22].
Putative voltage-regulated water channels have been suggested for
human AQP1 and AQP4 [32]. However, so far there are no
experimental evidences for the physiological significance of the
AqpZ gating. Furthermore, unlike the SoPIP2;1 the conformational
states of the R189 side chain have never been related to the water
permeation dynamics inside the channel. The open and closed states
were directly linked to the observation of different water profiles
inside the channel, and have been argued to reflect a dynamic
conformational balance [22]. Since water transportation is a dynamic
process, it is important to establish a direct relationship between the
water permeation dynamics and R189 side chain conformation. Such
relationship would be a strong support to the open-closed dual states
suggested by the structural information and channel water profile.
In this study, water permeation dynamics through AqpZ channel
was studied by extensive MD simulations and two-state behavior of
water transportation was clearly shown. Distinctive diffusive and
osmotic permeability of water in these two states characterized the
open and closed states of the channel. By applying umbrella sampling
scheme, free energy profile of the open and closed states was obtained
as the function of R189 side chain conformation. Simultaneously, the
water permeation dynamics under different side chain conformations
of R189 were analyzed which provided strong support to the
mechanism of gating through side chain conformation change.
2. Results and discussions
2.1. Dual permeability states of AqpZ
Crystal structures unambiguously resolved two different conformations of AqpZ [22,23]. Starting from tetramers formed by either
conformer A or B, two simulations showed different water diffusive
permeation rate k0 (Table 1). The permeation rate, k0, of individual
protomers displays significant differences: protomer 1, 3 and 4 of
tetramer B showed nearly zero water permeation while other
protomers showed significant permeation with values of k0 ranging
from 0.67 to 1.19 water/ns (see Fig. 1). It is noteworthy that, starting
from the closed conformation protomer 2 of tetramer B was observed
to spontaneously transform to the open channel. In the first 12 ns, no
water permeation events were observed for protomer 2. After 12 ns
continuous permeation happened. Based on the permeation rate,
8 single channel from tetramers A and B displayed clearly open/closed
behaviors in simulations. This was a dynamic demonstration of the
dual-state of the crystal structures.
Through equilibrium MD simulations the single channel osmotic
permeabilitypf, an important quantity to characterize water permeability, can be computed [36]. The pf data obtained in the present study
are listed in Table 2. The data range is comparable to other simulation
studies [31,34,37–39]. In a water permeability matrix analysis [39],
Fig. 1. Cumulative permeation events. Linear regression results are shown in lines. Panel A,
cumulative permeation events through 4 protomers from tetramer A simulation.
Panel B, cumulative permeation events through 4 protomers from tetramer B simulation.
pf was predicted to be 15 * 10−14 cm3/s for the critical region which
locates around the ar/R selective filter in the open state. In another
study, when both open and closed states were sampled [34], the pf of
AqpZ monomers ranged from 2.7 to 6.7 in the unit of 10−14 cm3/s. In
the present simulations, the closed channels always had pf lower than
2.5 while open channels had pf varying from 5.8 to 11.4.
In an idealized single-file water transportation, the continuous
time random walk model [36,40] predicted that pf = pd = N + 1,
where N is the average number of water molecules occupying the
channel. pd is the single channel diffusive permeability linked to the
permeation rate k0 through the equation: pd = vwk0 [36], where
vw = Vw/NA with Vw being the molar volume of water and NA being the
Avogadro's number. The pf/pd and N are also summarized in Table 2.
For open channels, the pf/pd ratios are comparable to N + 1. For
closed channels, due to the extremely low pd value, the pf/pd ratios
turn out to be much larger values which suggests that open channels
provide nearly idealized single-file water transportation while closed
channels do not.
2.2. Time resolution of pf
The osmotic permeability pf is estimated from the linear regression
of the dimensionless collective coordinate b n(t)2 N [41], when t is
much longer than the velocity correlation time of n. The velocity
correlation time of n is usually in the order of 10 ps. To further
characterize osmotic permeability for open and closed channels,
instantaneous pf is calculated in small time windows (10 ns) moving
along the simulation time (Fig. 2).
For closed channels, pf display values smaller than 3.5 throughout
the simulation (Fig. 2 upper panel B protomers 1, 3 and 4). For open
channels, pf values show dramatic fluctuations. In protomer 4 of
tetramer A, the minimum pf = 2.3 is observed at 56 ns. This pf value is
in the range of a closed channel. However, as shown in Fig. 1, between
Table 2
Single-channel osmotic and diffusive permeability (pf, pd in units of 10−14 cm3 s−1) for all
protomers (P1–P4) in tetramer A (80 ns), tetramer B (100 ns) and R189S mutant tetramer
B (60 ns) simulations. N denotes the average number of molecules in the channel.
Tetramer B
Table 1
Water permeation events and permeation rate for all monomers from tetramer A and
tetramer B simulations.
Tetramer A, 80 ns
Tetramer B, 100 ns
pf
pd
N
pf/pd
99(1.08/ns)
98(1.12/ns)
56(0.69/ns)
65(0.60/ns)
1(0.04/ns)
108(1.19/ns)
1(0.014/ns)
5(0.01/ns)
pf
pd
Number of permeation events and (permeation rate)
Monomer
Monomer
Monomer
Monomer
1
2
3
4
Tetramer A
P1
P2
P3
P4
P1
P2
P3
P4
1.7
0.06
12.7
28.3
11.4
0.9
8.8
12.7
2.5
0.01
11.4
250
2.4
0.08
11.4
300
9.94
0.8
10.1
12.4
10.1
0.9
7.9
11.9
5.8
0.5
7.2
11.4
6.1
0.5
7.4
13.5
Tetramer B R189S mutant
13.1
0.9
11.02
1.5
9.7
1.2
6.3
1.1
L. Xin et al. / Biochimica et Biophysica Acta 1808 (2011) 1581–1586
1583
Fig. 3. Evolution of RMSD. Red squares and black dots are for the closed and open
channel, respectively. RMSDs are calculated relative to the conformer A structure.
conformational change. Our current simulations provided a consistent
picture: both the open and closed channels have similar overall
structural fluctuation profiles.
Fig. 2. Instantaneous pf as function of time for individual monomers in tetramer A
(upper left) and B (upper right). Number of waters occupying the channel as function of
time in individual protomers from tetramer A (lower left) and B (lower right).
56 and 66 ns, there were still water permeation events observed.
Therefore, at this point, the channel was not in the closed state.
As mentioned, pf is estimated as the slope of linear regression of
b n(t)2 N against time using chi-square fitting, the variance in the
estimation of pf can be evaluated [42]. In all calculations, the variance
is extremely small with the maximum being less than 10−2% of the
estimated pf. Hence the fluctuation of pf is related to the intrinsic
property of the channel itself. Many factors may affect the
microscopic fluctuation and dissipation of water in the channel,
and would directly affect the estimation of the diffusion coefficient
Dn. It is hard to use pf as the only criterion to gauge the open and
closed states of a channel, especially when the calculated pf has a
small value. For long enough simulations, like in this study, when a
good fitting of pd is available, the diffusive permeability pd is a better
criterion to discriminate the open and closed states. In order to
understand better water diffusion in irregularly shaped microscopic
channel and to improve the pf calculation, further study is needed.
2.4. Potential of mean force and the structure–permeability relationship
There were few open–closed state transitions in our equilibrium
simulations. Thus it was rather difficult to construct the global free
energy profile including the open and closed states. In the previous
work [39] and current simulations, the hydrogen bond between
amide hydrogen, NH1, of R189 and carbonyl oxygen, O, of A117 was
observed to be stably maintained in all open-state trajectories but is
missing from all closed state trajectories. Therefore, we chose the
distance between these two atoms, doh, as the reaction coordinate. To
quantify the energy barrier between the open and closed states, the
umbrella sampling technique and the weighted histogram analysis
method [43] were employed to get the potential of mean force (PMF)
as a function of doh which is shown in Fig. 4. On the PMF, there are
three local minima, I, II and III whose representative structures are
shown in Fig. 5. Interestingly the three states, I, II and III are also
relating to three configuration states, up, middle and down of the side
chain dihedral angle, ∠ Cβ − Cγ − Cδ − Nε of R189, respectively [34].
There is a free energy barrier between I and II states, 4.8 kcal/mol,
around doh = 0.4 nm, and a smaller barrier between states II and III,
2.3 kcal/mol around doh = 0.6 nm.
2.3. Structure fluctuations of open/closed states
Two conformers of AqpZ from the crystal structure are very similar
to each other with RMSD of 0.44 Å [23]. However, our simulations
starting from these two conformers exhibit clearly open and/or closed
states. Such different water permeation behaviors do not resort to
overall structural change. After the initial relaxation, Cα RMSDs of
both the open and closed channels fluctuate between 0.8 and 1.5 Å
(Fig. 3). In fact they are indistinguishable. We have also measured
other collective quantities such as the helices' tilting/crossing angles,
as well as the RMSD of the pore-lining residues only. No direct link
was observed between these quantities and the open or closed states
(data not shown).
Both X-ray structure [23] and molecular dynamics simulations [34]
suggested that the side chain conformation of a conserved residue
R189 in the selective filter region gauges the open and closed state.
Different from the case of SoPIP2:1, the transition between open and
closed states of AqpZ seems localized and requires much less
Fig. 4. The potential of mean force as function of the distance between R189 NH1 and
A117O (doh).
1584
L. Xin et al. / Biochimica et Biophysica Acta 1808 (2011) 1581–1586
Fig. 5. Simulation snapshots of local structure around the R189 and A117 region. 1, 2, 3
panels correspond to 3 local minimums, TR1 and TR2 correspond to 2 transition states
in Fig. 4. Residues R189, A117 and H174 are shown in sticks. Dotted lines denote the
distances between R189 and H174, and R189 and A117.
In the umbrella sampling simulations, several side chain dihedral
angles change concomitant with the change of doh (Table 3). The most
dramatic change was found in χ4 (∠ Cγ − Cδ − Nε − Cς), which
represents the rotation around the Cδ − Nε bond and defines the
placement of the guanidinium group. The difference of χ4 between
state I and state II is about 160° which indicates the flip of the
guanidinium group. The snapshots of states I and II shown in Fig. 5
indicate that this rotation renders the guanidinium group come into
the vicinity of the atom Nδ of H174. In the open state, both NHs from
the guanidinium group of R189 are in close contact with the atom O of
A117. With the flip of guanidinium group, the distance between atoms
Nδ of H174 and NH1 of R189 drops from 0.74 to 0.42 nm that leads to
the side-chain contacts between these 2 residues and also causes the
drop in pore minimum radius from 0.1 to 0.05 nm (calculated using
HOLE [44]). Thus the channel is blocked. Going from state III to state II,
all the χ angles undergo mild change for about 20° to 60°. The major
difference is that in state III the guanidinium group of R189 turns
away from the atom O of A117 and forms hydrogen bond with Nδ of
H174 which stabilizes the structure. It is clear that the transition
between the open and closed states is multi-variant dependent. The
projection of the intrinsic high dimensional free energy surface to a
low dimensional PMF (Fig. 4) will definitely lose much detailed
information of the system. As a result the barrier between the open
and closed states is considered better in a qualitative rather than a
quantitative way. Constructing a high dimensional free energy
surfaces is under way and related discussion is beyond the scope of
the current study.
In all previous MD studies of AqpZ [34,35,37,39], the open and
closed states are inferred from water profile in the channel lumen
instead of water permeation characteristics of the channel. It is
desirable to explore the channel property by relating the water
permeability to the side chain conformation of the residue R189.
Fig. 6. pf as a function of the distance between R189 NH1 and A117O (doh).
Large amount of data was produced in our umbrella sampling
simulations with specific configuration of R189 side chain constrained from external harmonic potentials. These data allowed us to
make an extensive analysis of permeability–conformation relationship. Osmotic permeability as a function of doh was calculated from
the doh restrained trajectories from umbrella sampling simulations
and are displayed in Fig. 6. In simulations with doh larger than 0.4 Å,
the pf of the channel is always smaller than 3 and no diffusive
permeation of water was observed which clearly indicated a closed
state. All the doh restrained simulation lasted for 12 ns. To further
confirm the observation, simulation for the case of doh = 0.5 nm was
prolonged to 40 ns. Still no diffusive permeation of water was
observed. As illustrated in Fig. 5, in both states I and II the
guanidinium group of R189 is in contact with atom O of A117 and
seems to leave space for water to occupy the surrounding region. In
previous work, the up and middle configurations (I and II state in
Fig. 6) were considered to be in open state since the channel lumen
was consistently occupied with water molecules [34]. However our
current study unambiguously showed that the middle configuration
(with doh = 0.5 nm) which belongs to state II is in fact in a closed
state. When the doh is smaller than 0.4 nm, the pf of the channel falls
into the range of open state. There is a clear relation between the
side chain conformations of R189 and the channel permeability,
which also validates that doh is a good reaction coordinate for
discriminating the open and closed states.
The important role of R189 in modulating the open and closed
states has been further addressed by the simulation of R189S mutant.
By mutating the R189 to a serine residue, the long side chain of R189 is
removed, with the environment around the 189 position remaining
hydrophilic. Simulation (40 ns) of the R189S mutant starting from
conformer B showed high osmotic permeability as well as diffusive
permeability (Table 2). The pf of all 4 protomers are in the range of
open state and the average is slightly higher than that of the open
channel formed by AqpZ wild-type.
Table 3
Structural features characterized by doh, dihedral angles χ1 (∠ N − Cα − Cβ − Cγ), χ2 (∠ Cα − Cβ − Cγ − Cδ), χ3 (∠ Cβ − Cγ − Cδ − Nε), χ4 (∠ Cγ − Cδ − Nε − Cς), distance between
H174ND and R189NH1, pore radius for the 3 local minima and 2 transition states displayed by the PMF.
State
1
tr1
2
tr2
3
doh (nm)
χ1 (°)
χ2 (°)
χ3 (°)
χ4 (°)
174ND-189NH1 (nm)
Pore radius (Å)
0.28
176 ± 8
−96 ± 12
−176 ± 10
−102 ± 19
0.74 ± 0.01
0.8
0.4
−178 ± 11
−105 ± 16
140 ± 38
−152 ± 73
0.67 ± 0.12
0.8
0.48
176 ± 7
−108 ± 26
169 ± 8
103 ± 27
0.42 ± 0.04
0.5
0.61
177 ± 9
−159 ± 30
174 ± 11
162 ± 31
0.38 ± 0.01
0.5
0.73
152 ± 42
−165 ± 23
−168 ± 43
158 ± 28
0.29 ± 0.00
0.2
L. Xin et al. / Biochimica et Biophysica Acta 1808 (2011) 1581–1586
3. Conclusion
In current study, the open and closed states of AqpZ have been
shown to directly relate to the side chain conformation of residue
R189. For the first time, the open and closed states have been directly
related to water permeation dynamics through the channel. Our study
confirmed previous suggestions about the gating mechanism for
water transportation in AqpZ. The high level of genetic conservation
of R189 [33] also attests to its functional role. The free energy profile
obtained in this study characterizes the two states energetically and
implies fast transition between them with thermal fluctuation. This
indicates the complexity in functional characterization of the two
states by experimental tools and highlights the call for more detailed
computational studies targeting at the possible role of the two states
in proton exclusion or other function of AqpZ.
4. Method
Modeling and MD simulations: Simulations were performed for
AqpZ tetramers embedded in palmitoyloleyl phosphatidylcholine
(POPC) lipid bilayers. The AqpZ tetramer from the crystal structure
(PDB ID: 1RC2) were incorporated into POPC bilayers and fully
solvated. The system consists of 4 AqpZ monomers, 330 POPC and
24937 water molecules. MD simulations were carried out using
GROMACS [45] with CHARMM 27 force field [46,47]. The systems
were initially minimized and equilibrated for 1 ns with protein heavy
atom coordinates restrained, and subjected to 80 and 100 ns NPT
(T = 300 K, P = 1 atm) simulations respectively for tetramers A and B.
Periodic boundary conditions were applied and the particle-mesh
Ewald method [48] with cutoff = 9 Å was utilized for electrostatic
potential calculation. The Lennard–Jones interactions were switched
off beyond the range of 12 Å. An integration step of 2 fs was used, and
the simulated structures of the system were recorded every 1 ps.
Umbrella sampling simulations: Sampling along the reaction
coordinate (distance between R189 NH1 and A117O) was performed
by applying umbrella potentials to the reaction coordinate at each
sampling window. Center of sampling windows started from 2.0 Å and
ended at 8.0 Å with a 0.5-Å interval. For each sampling window, 12-ns
simulations were performed and the first 2-ns trajectory was removed
from the analysis. In total 13 trajectories were generated, and for each
of them, the population histograms as a function of the reaction coordinate were obtained. All these population functions were subjected to
1D weighted histograms analysis to generate the free energy profile.
For checking the convergence of the free energy calculation, last 2 ns of
all trajectories was truncated and the same analysis was performed.
Almost identical free energy profile was obtained under such condition.
Single-channel water permeability: The number of water traversing the channel is recorded as a function of time. By linear regression
of the cumulative water traverse events as a function of time, the
water permeation rate k0 for a given trajectory was obtained. And the
single-channel diffusive permeability pd is obtained as: pd = vwk0
[36,41] where vw is the volume of a single water molecule. By
cumulating the individual water displacement inside the channel,
the collective motion of water inside the channel is described as:
dn = ∑ dz = L where S(t) denotes the water in the channel and L was
i∈Sðt Þ
the channel length. When t is much longer than the velocity
correlation time of n, the mean square displacement of n, 〈n2(t)〉
obeys the Einstein relationship: 〈n2(t)〉 = 2Dnt where Dn is defined as
the diffusion coefficient of n. From Dn the single channel osmotic
permeability is calculated as: pf = vwDn [36,41].
Acknowledgment
We thank Dr. Bert de Groot (Gottingen) for reading and comments
on our manuscript and the Environment and Water Industry
1585
Programme Office (EWI) under the National Research Foundation of
Singapore (grant MEWR MEWR651/06/169) for the financial support
of the work. Generous allocation of CPU time from HPC NTU is
gratefully acknowledged.
References
[1] G.M. Preston, T.P. Carroll, W.B. Guggino, P. Agre, Appearance of water channels in
Xenopus oocytes expressing red cell CHIP28 protein, Science 256 (1992)
385–387.
[2] B. Wu, E. Beitz, Aquaporins with selectivity for unconventional permeants, Cell.
Mol. Life Sci. 64 (2007) 2413–2421.
[3] L.S. King, D. Kozono, P. Agre, From structure to disease: the evolving tale of
aquaporin biology, Nat. Rev. Mol. Cell Biol. 5 (2004) 687–698.
[4] Z. Liu, J. Shen, J.M. Carbrey, R. Mukhopadhyay, P. Agre, B.P. Rosen, Arsenite
transport by mammalian aquaglyceroporins AQP7 and AQP9, Proc. Natl. Acad. Sci.
USA 99 (2002) 6053–6058.
[5] Y.J. Ko, J. Huh, W.H. Jo, Ion exclusion mechanism in aquaporin at an atomistic level,
Proteins 70 (2008) 1442–1450.
[6] D. Fu, M. Lu, The structural basis of water permeation and proton exclusion in
aquaporins, Mol. Membr. Biol. 24 (2007) 366–374.
[7] N. Chakrabarti, E. Tajkhorshid, B. Roux, R. Pomes, Molecular basis of proton
blockage in aquaporins, Structure 12 (2004) 65–74.
[8] H. Sui, B.G. Han, J.K. Lee, P. Walian, B.K. Jap, Structural basis of water-specific
transport through the AQP1 water channel, Nature 414 (2001) 872–878.
[9] M. Hara-Chikuma, A.S. Verkman, Aquaporin-1 facilitates epithelial cell migration
in kidney proximal tubule, J. Am. Soc. Nephrol. 17 (2006) 39–45.
[10] O. Bloch, M.C. Papadopoulos, G.T. Manley, A.S. Verkman, Aquaporin-4 gene
deletion in mice increases focal edema associated with staphylococcal brain
abscess, J. Neurochem. 95 (2005) 254–262.
[11] P. Agre, D. Kozono, Aquaporin water channels: molecular mechanisms for human
diseases, FEBS Lett. 555 (2003) 72–78.
[12] A. Tanghe, P. Van Dijck, J.M. Thevelein, Why do microorganisms have aquaporins?
Trends Microbiol. 14 (2006) 78–85.
[13] A.S. Verkman, Role of aquaporin water channels in eye function, Exp. Eye Res. 76
(2003) 137–143.
[14] Y.C. Chen, M.A. Cadnapaphornchai, R.W. Schrier, Clinical update on renal
aquaporins, Biol. Cell 97 (2005) 357–371.
[15] M. Amiry-Moghaddam, O.P. Ottersen, The molecular basis of water transport in
the brain, Nat. Rev. Neurosci. 4 (2003) 991–1001.
[16] I. Johansson, M. Karlsson, U. Johanson, C. Larsson, P. Kjellbom, The role of
aquaporins in cellular and whole plant water balance, Biochim. Biophys. Acta
1465 (2000) 324–342.
[17] C. Maurel, H. Javot, V. Lauvergeat, P. Gerbeau, C. Tournaire, V. Santoni, J. Heyes,
Molecular physiology of aquaporins in plants, Int. Rev. Cytol. 215 (2002)
105–148.
[18] K. Murata, K. Mitsuoka, T. Hirai, T. Walz, P. Agre, J.B. Heymann, A. Engel, Y.
Fujiyoshi, Structural determinants of water permeation through aquaporin-1,
Nature 407 (2000) 599–605.
[19] S. Törnroth-Horsefield, K. Hedfalk, G. Fischer, K. Lindkvist-Petersson, R. Neutze,
Structural insights into eukaryotic aquaporin regulation, FEBS Lett. 584 (2010)
2580–2588.
[20] D. Fu, A. Libson, L.J. Miercke, C. Weitzman, P. Nollert, J. Krucinski, R.M. Stroud,
Structure of a glycerol-conducting channel and the basis for its selectivity, Science
290 (2000) 481–486.
[21] E. Tajkhorshid, P. Nollert, M.Ø. Jensen, L.J.W. Miercke, J. O'Connell, R.M. Stroud, K.
Schulten, Control of the selectivity of the aquaporin water channel family by
global orientational tuning, Science 296 (2002) 525–530.
[22] J. Jiang, B.V. Daniels, D. Fu, 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 (2006) 454–460.
[23] D.F. Savage, P.F. Egea, Y. Robles-Colmenares, J.D. O'Connell III, R.M. Stroud,
Architecture and selectivity in aquaporins: 2.5 a X-ray structure of aquaporin Z,
PLoS Biol. 1 (2003) E72.
[24] S. Törnroth-Horsefield, Y. Wang, K. Hedfalk, U. Johanson, M. Karlsson, E.
Tajkhorshid, R. Neutze, P. Kjellbom, Structural mechanism of plant aquaporin
gating, Nature 439 (2006) 688–694.
[25] T. Gonen, P. Silz, J. Kistler, Y. Cheng, T. Walz, Aquaporin-0 membrane junctions
reveal the structure of a closed water pore, Nature 429 (2004) 193–197.
[26] J.K. Lee, D. Kozono, J. Remis, Y. Kitagawa, P. Agre, R.M. Stroud, Structural basis for
conductance by the archaeal aquaporin AqpM at 1.68 Å, Proc. Natl. Acad. Sci. USA
102 (2005) 18932–18937.
[27] Z.E.R. Newby, J. O'Connell Iii, Y. Robles-Colmenares, S. Khademi, L.J. Miercke, R.M.
Stroud, Crystal structure of the aquaglyceroporin PfAQP from the malarial parasite
Plasmodium falciparum, Nat. Struct. Mol. Biol. 15 (2008) 619–625.
[28] R. Horsefield, K. Nordén, M. Fellert, A. Backmark, S. Törnroth-Horsefield, A.C.
Terwisscha Van Scheltinga, J. Kvassman, P. Kjellbom, U. Johanson, R. Neutze, Highresolution x-ray structure of human aquaporin 5, Proc. Natl. Acad. Sci. USA 105
(2008) 13327–13332.
[29] G. Fischer, U. Kosinska-Eriksson, C. Aponte-Santamaria, M. Palmgren, C. Geijer, K.
Hedfalk, S. Hohmann, B.L. de Groot, R. Neutze, K. Lindkvist-Petersson, Crystal
structure of a yeast aquaporin at 1.15 angstrom reveals a novel gating mechanism,
PLoS Biol. 7 (2009) e1000130.
1586
L. Xin et al. / Biochimica et Biophysica Acta 1808 (2011) 1581–1586
[30] B.L. De Groot, H. Grubmüller, Water permeation across biological membranes:
mechanism and dynamics of aquaporin-1 and GlpF, Science 294 (2001)
2353–2357.
[31] H. Khandelia, M.O. Jensen, O.G. Mouritsen, To gate or not to gate: using molecular
dynamics simulations to morph gated plant aquaporins into constitutively open
conformations, J. Phys. Chem. B 113 (2009) 5239–5244.
[32] J.S. Hub, C. Aponte-Santamaría, H. Grubmüller, B.L. de Groot, Voltage-regulated
water flux through aquaporin channels in silico, Biophys. J. 99 (2010) L97–L99.
[33] R. Zardoya, Phylogeny and evolution of the major intrinsic protein family, Biol.
Cell 97 (2005) 397–414.
[34] M.O. Jensen, O.G. Mouritsen, Single-channel water permeabilities of Escherichia
coli aquaporins AqpZ and GlpF, Biophys. J. 90 (2006) 2270–2284.
[35] Y. Wang, K. Schulten, E. Tajkhorshid, What makes an aquaporin a glycerol
channel? A comparative study of AqpZ and GlpF, Structure 13 (2005) 1107–1118.
[36] F. Zhu, E. Tajkhorshid, K. Schulten, Theory and simulation of water permeation in
aquaporin-1, Biophys. J. 86 (2004) 50–57.
[37] M. Hashido, M. Ikeguchi, A. Kidera, Comparative simulations of aquaporin family:
AQP1, AQPZ, AQP0 and GlpF, FEBS Lett. 579 (2005) 5549–5552.
[38] M.O. Jensen, R.O. Dror, H. Xu, D.W. Borhani, I.T. Arkin, M.P. Eastwood, D.E. Shaw,
Dynamic control of slow water transport by aquaporin 0: implications for
hydration and junction stability in the eye lens, Proc. Natl. Acad. Sci. USA 105
(2008) 14430–14435.
[39] M. Hashido, A. Kidera, M. Ikeguchi, Water transport in aquaporins: osmotic
permeability matrix analysis of molecular dynamics simulations, Biophys. J. 93
(2007) 373–385.
[40] A. Berezhkovskii, G. Hummer, Single-file transport of water molecules through a
carbon nanotube, Phys. Rev. Lett. 89 (2002) 064503.
[41] F. Zhu, E. Tajkhorshid, K. Schulten, Collective diffusion model for water
permeation through microscopic channels, Phys. Rev. Lett. 93 (2004)
224501.
[42] W.H. Press, S.A. Teukolsky, W.T. Vetterling, B.P. Flannery, Numerical Recipes in C:
The Art of Scientific Computing, The Press Syndicate of the University of
Cambridge, 1992, p. 663.
[43] S. Kumar, D. Bouzida, R.H. Swendsen, P.A. Kollman, J.M. Rosenberg, The weighted
histogram analysis method for free-energy calculations on biomolecules: I. The
method, J. Comput. Chem. 8 (1992) 1011–1021.
[44] O.S. Smart, J.M. Goodfellow, B.A. Wallace, The pore dimensions of gramicidin A,
Biophys. J. 65 (1993) 2455–2460.
[45] D. Van Der Spoel, E. Lindahl, B. Hess, G. Groenhof, A.E. Mark, H.J. Berendsen,
GROMACS: fast, flexible, and free, J. Comput. Chem. 26 (2005) 1701–1718.
[46] D.M. Alexander Jr., B. Nilesh, F. Nicolas, Development and current status
of the CHARMM force field for nucleic acids, Biopolymers 56 (2000)
257–265.
[47] P. Bjelkmar, P. Larsson, M.A. Cuendet, B. Hess, E. Lindahl, Implementation of the
CHARMM force field in GROMACS: analysis of protein stability effects from
correction maps, virtual interaction sites, and water models, J. Chem. Theory
Comput. 6 (2010) 459–466.
[48] T. Darden, L. Perera, L. Li, L. Pedersen, New tricks for modelers from the
crystallography toolkit: the particle mesh Ewald algorithm and its use in nucleic
acid simulations, Structure 7 (1999) R55–R60.