Purified Lens Major Intrinsic Protein (MIP) - Tom Walz (Harvard)

Copy and paste this link to your website, so they can see this document directly without any plugins.


with, were, al.,, from, lens, mass, crystals, protein, cell, Biol., AQP1, Figure, structure, membrane, unit, electron, peaks, which, (Figure, density, single, ®bre, freeze-dried, that, containing, membranes, family, 4-fold, proteins, major


Article No. mb981796 J. Mol. Biol. (1998) 279, 855±864Purified Lens Major Intrinsic Protein (MIP) Forms
Highly Ordered Tetragonal Two-dimensional Arrays
by Reconstitution
Lorenz Hasler1, Thomas Walz2, Peter Tittmann3, Heinz Gross3
Joerg Kistler4 and Andreas Engel1*1M. E. MuÈ ller-Institute for
Microscopy at the Biozentrum
University of Basel, Basel
CH-4056, Switzerland
2Krebs Institute for
Biomolecular Research,
Department of Molecular
Biology and Biotechnology
University of Shef®eld, Firth
Court, Western Bank, Shef®eld
S10 2TN, UK
3Institute for Cell Biology
Swiss Federal Institute of
Technology, CH-8093, ZuÈ rich
4School of Biological Sciences
University of Auckland
Auckland, New ZealandAbbreviations used: MIP, major i
decyl-b-D-maltoside; OG, n-octyl-bscanning transmission electron micr
to-protein ratio; AQP1, aquaporin-1
square; cmc, critical micellar concen
0022±2836/98/240855±10 $25.00/0Lens major intrinsic protein (MIP) is the founding member of the MIP
family of membrane channel proteins. Its isolation from ovine lens ®bre
cell membranes and its two-dimensional crystallization are described.
Membranes were solubilized with n-octyl-b-D-glucoside and proteins fractionated by sucrose gradient centrifugation containing decyl-b-D-maltoside. MIP was puri®ed by cation exchange chromatography, and
homogeneity was assessed by mass analysis in the scanning transmission
electron microscope. Puri®ed MIP reconstituted into a lipid bilayer at a
low lipid-to-protein ratio formed highly ordered tetragonal two-dimensional crystals. The square unit cell had a side length of 6.4 nm, and
exhibited in negative stain four stain-excluding elongated domains surrounding a central stain-®lled depression. Projection maps of freeze-dried
crystals exhibited a resolution of 9 AÊ , and revealed a monomer structure
of MIP consisting of distinct densities. Despite signi®cant differences in
the packing of tetramers in the crystals, the projection map of the MIP
monomer was similar to that of aquaporin-1 (AQP1), the ®rst member of
the MIP family which had its structure resolved to 6 AÊ . Our protocols for
the puri®cation and reconstitution of MIP establish the feasibility for
future work to visualize structure elements which determine the diverse
functional properties of the MIP family members.
# 1998 Academic Press Limited
Keywords: major intrinsic protein, membrane protein, 2D crystallization,
electron crystallography, STEM*Corresponding authorIntroduction
The transmembrane channel proteins of the MIP
family are known to form selective pores of water,
small neutral solutes, and possibly ions (Park &
Saier, 1996). More than 100 proteins of this family
have been described in mammals, fungi, plants,
and eubacteria. Its ®rst sequenced member is the
major intrinsic protein (MIP) of vertebrate lens
®bre cell membranes (Broekhuyse et al., 1976;
Gorin et al., 1984). MIP has been demonstrated to
increase the membrane permeability for glycerol,
and to a lesser extent, for water (Mulders et al.,ntrinsic protein; DM,
D-glucoside; STEM,
oscope; LPR, lipid; RMS, root-meantration; 2D, two-1995; Kushmerick et al., 1995, 1997; Chandy et al.,
1997), and to have adhesive properties (Michea
et al., 1994). Mutations of MIP cause lens opaci®cation supporting a crucial role in the development of the transparent lens (Shiels & Bassnett,
Sequence based structure prediction postulates
MIP to be a six-membrane spanning protein (Gorin
et al., 1984). Other members of the MIP family are
thought to have a similar membrane topology.
Characteristic for all members is that the amino
and carboxyl halves of the molecule share considerable amino acid sequence homology: each half
has three prominent hydrophobic peptide segments and carries the highly conserved NPA motif
(asparagine-proline-alanine). The sequence homologies create an internal structural symmetry with
the NPA motif of the amino-terminal half located
in the cytoplasmic loop between the predicted
membrane spanning helices 2 and 3, and in the# 1998 Academic Press Limited
Figure 1. Puri®cation of MIP from ovine lens ®bre cell
membranes. a, SDS gel (12% (w/v) acrylamide) showing
the various steps in the puri®cation of MIP. Lane 1,
Urea/alkali stripped ®bre cell membranes from the lens
core. Lane 2, Supernatant after OG solubilization and
centrifugation. Lane 3, Sucrose gradient fraction
enriched in MIP. Lane 4, Concentrated MIP after cation
exchange chromatography. Lanes 1 to 4 were silver
stained. Lane 5, Concentrated MIP after cation exchange
chromatography. For this lane, the electrophoresis was
run less far than for the others, and staining was with
Coomassie blue to ensure detection of any lower molecular weight contaminants. Molecular mass standards
marked by ®lled circles are (from top to bottom) 94, 66,
43, 31, 21, and 14 kDa (if present). b, Cation exchange
chromatography of pooled sucrose gradient fractions
enriched in MIP. MIP was eluted in a sharp peak at a
NaCl concentration of approximately 110 mM NaCl in
10 mM Hepes (pH 7.2), 0.3% DM. Continuous line,
Elution pro®le (l ˆ 20 nm); broken line, NaCl-gradient.
856 MIP Isolation and Crystallizationextracellular loop between helices 5 and 6 in the
carboxyl half of the molecule (Reizer et al., 1993).
Structure analysis of members of the MIP family
is most advanced for the erythrocyte water channel
aquaporin-1 (AQP1, Chrispeels & Agre, 1994).
Electron crystallographic analyses of AQP1
revealed six membrane spanning alpha-helices in
agreement with the predicted topology (Jap & Li,
1995; Li et al., 1997; Walz et al., 1997; Cheng et al.,
1997). Considerably less is known about the structure of the founding member of the MIP family.
Solubilized single MIP particles have a tetrameric
structure (Aerts et al., 1990; KoÈnig et al., 1997), consistent with earlier observations that lens ®bre cell
membranes contain square arrays which can be
attributed to MIP (Kistler & Bullivant, 1980;
Zampighi et al., 1982, 1989).
Here we report an improved puri®cation procedure for MIP from ovine lens ®bre cell membranes. The preparations are homogeneous, and
the solubilized MIP oligomers have a molecular
mass which is consistent with tetramers. Using
these preparations, it is now possible to reconstitute highly ordered two-dimensional (2D) crystals
of MIP which are amenable to structure analysis to
a resolution of at least 9 AÊ .
Purification and characterization of MIP
The silver stained SDS gel in Figure 1a documents the outcomes of the different puri®cation
steps for MIP. Lane 1 shows the proteins of urea/
alkali stripped ®bre cell membranes from the lens
core which contain as major integral proteins MIP,
MP20, and connexins 46 and 50 in the cleaved
38 kDa form (Lin et al., 1997). Lane 2 displays the
protein pro®le after solubilization of the membranes with n-octyl-b-D-glucoside (OG). The solubilization ef®ciency was approximately 70% as
estimated by the BCA protein assay. Because MIP
tends to aggregate in OG (KoÈnig et al., 1997), a
sucrose density gradient was prepared with 0.3%
decyl-b-D-maltoside (DM) in 10 mM Hepes (pH 7.2)
to allow gentle detergent exchange. Sucrose gradient fractionation removed the 38 kDa connexin
form (Figure 1a, lane 3). A MiniS column was
®nally used to eliminate MP20 and to concentrate
MIP. Figure 1b shows a sharp elution peak at a
NaCl concentration of 110 mM which contained
highly puri®ed MIP (Figure 1a, lane 4). The purity
of the protein was estimated to be better than 95%
from gels stained with Coomassie blue (Figure 1a,
lane 5).
To assess the homogeneity of the puri®ed MIP
oligomers and to determine their mass, DM-solubilized MIP particles were freeze-dried and
observed unstained with a scanning transmission
electron microscope (STEM). Elastic dark-®eld
images were recorded at a dose of approximately
320 electrons/nm2, 200,000 magni®cation, and
80 kV acceleration voltage. Images indicate ahomogeneous particle size (Figure 2a), and mass
analysis of 1154 particles (Figure 2b) shows a
single peak at 147.5(47.0) kDa. This is compatible
with a tetrameric protein embedded in a detergent
micelle of about 40 kDa.
2D-crystallization of MIP
Puri®ed MIP was concentrated to 1 mg/ml
using a Centricon-50 cartridge and reconstituted
with solubilized Escherichia coli lipids at low lipidto-protein ratios (LPR). Negative stain electron
microscopy of the reconstituted samples revealed
membrane vesicles or sheets which often had a
highly ordered coherent area up to 0.5 mm in diameter (Figure 3a). In most cases, diffraction patterns
contained two sets of diffraction spots, which is
characteristic of crystalline, collapsed vesicles. Consistently but with a lesser frequency, diffraction
patterns (Figure 3b) were of single layered sheetlike 2D crystals. The diffraction spots indicated a
unit cell dimension of a ˆ b ˆ 6.4(0.03) nm
(n ˆ 8) and a resolution of 2.0 nm. Only the single
sheets were used for correlation averaging.
Figure 2. Mass analysis of soluble MIP oligomers in the STEM. a, Dark ®eld image of freeze-dried, solubilized MIP
particles recorded at 2000 electrons/nm2. b, Mass histogram of 1154 particles recorded at an average dose of 320 electrons/nm2. The Gauss peak ®tted at 147.5(47.0) kDa, omitting a tail of higher mass values resulting from some
MIP Isolation and Crystallization 857Projection maps of negatively stained MIP crystals normally exhibited a pronounced 4-fold symmetry. One unit cell (outlined in Figure 3c) housed
a single tetramer with four stain-excluding regions
(bright) and a central stain-®lled depression (dark),
the tetramers being separated by strongly stained
lipid bilayer domains. The protruding domains
had a length and width of 2.7 nm and 1.8 nm,
Mass measurement of freeze-dried/unstained
MIP sheets with the STEM yielded a mass-perarea histogram with distinct peaks (Figure 4). The
lowest peak at 3.5(0.3) kDa/nm2 (n ˆ 175) was
interpreted as representing single-layered tetragonal MIP crystals. The mass per unit cell of
6.4 nm side length was 144 kDa and thus comparable to the mass of 147 kDa determined for
the soluble MIP/detergent complexes. The
second, most prominent peak at 6.6(0.32) kDa/
nm2 (n ˆ 508) was likely to represent two superimposed MIP square assays in collapsed vesicles.
The peaks at 9.8(0.36) kDa/mn2 (n ˆ 49) and at
13.1(0.38) kDa/nm2 (n ˆ 224) were interpreted
as deriving from the combination of a sheet and
vesicle, and of two vesicles, respectively. The
weighted average from all peaks yielded a mass
per area of 3.35 kDa/nm2, disregarding a minor
mass loss due to electron irradiation.
High resolution projections of freeze-dried 2D
crystals of MIP and comparison with AQP1
Freeze-dried/unstained crystalline sheets of MIP
diffracted to 9 AÊ (Figure 5a) despite some local
defects which were commonly observed in imagesof freeze-dried lattices. Such local defects could be
visualized by the cross-correlation function with a
4-fold symmetrized reference calculated by Fourier
peak ®ltration (Figure 5b). Correlation peaks were
blurred along one direction, suggesting either a
local distortion of the crystal, or a charging artifact.
Since these defects impaired the quality of the
averaged image, the corresponding correlation
peaks were manually eliminated prior to calculating the correlation average displayed in Figure 5c.
4-fold symmetrization (root-mean-square (RMS)
deviation 0.29) ®nally produced the average in
Figure 5d. Alternatively, patches centred around
correlation peaks were extracted from the unprocessed image and aligned angularly and translationally before averaging, using the map in
Figure 5d as reference. While this single particle
averaging was slightly less sharp than that without
angular alignment, the RMS deviation from 4-fold
symmetry was reduced to a value of 0.10
(Figure 5f).
Projection maps revealed distinct and reproducible features after averaging 100 to 400 unit
cells. Each subunit was resolved into eight distinct densities arranged around a central
depression. A quasi 2-fold axis running in the
plane of the membrane at 45C with respect to a
lattice vector, appeared to relate the two halves
of a monomer.
AQP1 is the only member of the MIP family
whose 3D structure had been determined at high
resolution. Direct comparison of the structures of
MIP and AQP1 was, however, hindered by the fact
that projection maps of the latter were from glucose or trehalose embedded samples. Hence, 2D
Figure 3. Negative stain electron microscopy and image analysis of two-dimensional MIP crystals. a, Negatively
stained sheet area of MIP crystals. Scale bar: 100 nm. b, The power spectrum calculated from a selected area in a
shows one set of diffraction spots corresponding to a single-layered crystal. The resolution is approximately 2.0 nm
(diffraction order 3,0). Scale bar: (5 nm)ÿ1. C, The averaged and 4-fold symmetrized unit cell calculated from MIP
crystals exhibits four prominent elliptical domains. Protein is displayed in bright shades, the surrounding stain in
dark shades. The outlined unit cell (a ˆ b ˆ 6.4 nm) includes one tetramer. The dark areas centred at the four corners
of the unit cell represent the lipid bilayer.
858 MIP Isolation and Crystallization
Figure 4. Mass analysis of freeze-dried MIP crystals.
The histogram shows peaks at 3.4(0.3) kDa/nm2
(nˆ175), 6.63(0.39) kDa/nm2 (nˆ508), 9.62(0.44) kDa/
nm2 (nˆ49), and 12.80(0.46) kDa/nm2 (nˆ224). These
peaks represent single sheets, collapsed vesicles, sheets
and vesicles, and two vesicles respectively.
MIP Isolation and Crystallization 859tetragonal AQP1 crystals prepared as described
previously (Walz et al., 1996), were freeze-dried,
recorded, and processed in the same way as the
crystalline sheets of MIP. The projection maps of
half-tetramers of MIP and AQP1 are displayed
adjacent to each other in Figure 6 to facilitate comparison. Several features are similar. First, the lowest density is at the 4-fold axis of the tetramers,
while the second lowest density is within the
monomers. Second, the low density regions at the
interface between monomers appear similar. Third,
a striking similarity is further found in the three
density peaks lining one side of this interface
(labelled A to C). On the other hand, signi®cant
differences are observed at the periphery of the tetramers. The latter is not unexpected as the tetramer
packing in the MIP crystal has p4 symmetry and
differs from the p4221 symmetry for AQP1 (Walz
et al., 1995). The differences seen in the projection
maps could be directly attributed to primary structure differences in the molecules and different
interactions between tetramers.
Here we present an improved protocol for the
puri®cation of lens ®bre cell MIP, and for the
®rst time, the formation of highly ordered 2D
crystals with structural details resolved to 9 AÊ .
A recently published puri®cation procedure for
MIP (KoÈnig et al., 1997) used cation-exchange
chromatography (MonoS) followed by gel ®ltration chromatography (Superdex 200). It wasnoted that while the detergent OG was ef®cient
for lens ®bre membrane solubilization, it tended
to aggregate MIP. We made similar observations,
and concluded that detergent exchange was
indeed necessary. DM had a suitable cmc value
(0.1%) for dialysis driven reconstitution, and did
not destabilize the MIP tetramers. We exchanged
OG with DM immediately following the initial
solubilization. The exchange was carried out ef®ciently but most gently by centrifugation of the
solubilized membrane components through a
sucrose gradient containing DM (Kistler et al.,
1994). Consequently, MIP was exposed to OG for
4one hour throughout the entire process of solubilization and puri®cation. Sucrose gradient fractionation also removed the connexins and most
MP20. While MIP was considerably enriched,
chromatography was still required for further
puri®cation. But in contrast to KoÈnig et al. (1997),
only a single cation-exchange chromatographic
step was necessary to achieve >95% purity
according to SDS-PAGE (Figure 1).
The soluble MIP preparation was homogeneous
as demonstrated in Figure 2, and the oligomers
were stable in DM. The mass value obtained with
the STEM correlates well with those determined by
others using sedimentation or gel ®ltration analysis
(KoÈnig et al., 1997), and is consistent with a
detergent-protein complex containing four MIP
During reconstitution, MIP tetramers were again
brie¯y exposed to OG during the initial dialysis
step but this time in the presence of lipids which
are known to have a stabilizing effect on the solubilized protein. The most highly ordered crystals
were obtained when the second dialysis cycle was
performed at temperatures up to 37C. This
ensured ef®cient removal of DM despite its cmc of
0.1%. The formation of protein aggregates was a
risk at elevated temperatures but was not
observed, probably because the protein was incorporated in the bilayer within the ®rst few hours of
the dialysis (Engel et al., 1992).
Projection maps of the negatively stained tetragonal MIP arrays revealed one tetrameric structure
per unit cell, composed of 2.7 nm long and 1.7 nm
wide elliptical domains. Disregarding the faint
differences in staining, the unit cell exhibited 2-fold
and 4-fold axis perpendicular to the membrane
plane, but no screw axes in the membrane plane.
Therefore, the MIP crystal belongs to the plane
group symmetry p4.
STEM mass measurement yielded a total mass of
either 144 kDa per unit cell (single layered sheets
only) or 137 kDa per unit cell (all mass peaks in
the histogram shown in Figure 5). Taking the latter
value and considering the beam induced mass loss
(i.e. 2.4% at 320 electron/nm2; MuÈ ller et al., 1992),
the overall mass per unit cell is 140 kDa. Contouring the projection map of negatively stained tetragonal crystals at the steepest contrast gradient
suggested that 37% of the unit cell surface is comprised of lipids. Assuming a lipid bilayer mass per
Figure 5. Correlation and single particle averaging of freeze-dried 2D MIP crystals. a, The diffraction pattern shows
strong diffraction spots to about (9 AÊ )ÿ1 in one direction. The transfer function had its ®rst zero value at a resolution
of (8.5 AÊ )ÿ1 and does not indicate any major drift. Scale bar: (2 nm)ÿ1 b, The cross-correlation function of the 2D crystal with a 4-fold symmetrized reference containing two tetramers exhibits sharp peaks as well as peaks blurred in
one direction (area marked by a circle). The latter peaks suggest local defects, possibly resulting from charging. Scale
bar: 20 nm. C, The correlation average calculated from 395 ®elds centred at the sharp, unblurred peaks in b still exhibits some residual asymmetry, explaining the RMS deviation of 0.29 for the 4-fold symmetrized tetramers in d.
e, single particle averaging of 138 ®elds that exhibited a correlation coef®cient >0.14 after angular and translational
alignment with respect to a reference containing nine tetramers as shown in d. The symmetry is improved as documented by the RMS deviation of 0.1 from the 4-fold symmetrized tetramer displayed in f. Scale bar: 1 nm.
860 MIP Isolation and Crystallizationarea of 2.8 kDa/nm2, the protein mass comes to
approximately 98 kDa, compatible with one tetramer per unit cell.
The application of freeze-drying for electron
crystallography of a membrane protein is novel.
Our images of MIP and AQP1 are the ®rst documented examples where a resolution better than
10 AÊ was obtained. Embedding of crystals in glucose or ice decreases the contrast with which the
protein can be imaged because of the background
noise of the embedding matrix. This led us to
explore the applicability of a single particle averaging for processing ill-ordered crystals or even
patches of densely packed membrane protein complexes. While the feasibility of this application is
demonstrated here, we have evaluated neither the
signal-to-noise ratio of differently prepared crystals
nor the resolution achievable. Nevertheless, the
projection map of freeze-dried AQP1 is in excellentagreement with the map at 6 AÊ resolution previously obtained in glucose (Walz et al., 1995)
thereby validating the freeze-dry method.
It is now interesting to ®rst compare the projection map of negatively stained MIP tetramers with
that of freeze-dried ones. In the latter the molecular
boundaries within the tetramer run parallel to the
lattice vectors (Figure 5 c to f), analogous to the
boundaries of AQP1 (Walz et al., 1997). The negative stain map, however, suggests a molecular
interface at 45C with respect to the lattice vectors
(Figure 3c). Hence, the hydrophilic elliptical
domain of 2.7 nm length and 1.8 nm width either
spans the gap between monomers or consists of
two unresolved domains, each belonging to a
different monomer.
Secondly, a detailed comparison of the MIP and
AQP1 density maps is possible at 9 AÊ resolution.
Projections of both channel proteins show distinct
Figure 6. Comparison of projection maps from freeze-dried MIP (left) and AQP1 seen from the extracellular side
(right). Pronounced similarities are recognized in the low density regions as well as the three density maxima (A,B,C)
that line the interface between monomers. Differences are noticeable to the periphery of the tetramer as well as in the
density L (for explanation see the text). The full width of the image corresponds to 6.4 nm.
MIP Isolation and Crystallization 861and similar density peaks A to C. They can be
interpreted from the 3D map of AQP1 (Walz et al.,
1997) as the superposition of the tilted helices 5
and 6 (peak A), helices 4 and 5 (peak B), and
helices 3 and 4 (peak C; see Heymann et al. (1998)
for the helix assignment). Furthermore, the similar
disposition of density maxima and minima at the
molecular interfaces within the tetramer suggests a
helix packing arrangement in MIP similar to that
in AQP1. Differences, however, are seen at the periphery of the tetramer. This region is due to the
projection of helices 2 and 3 and part of theimportant loops B and E containing the NPA
motifs (Heymann et al., 1998) and is denoted as
density peak L in Figure 6. These differences at the
periphery are likely to be the cause of the different
tetramer packing in the MIP (p4) and AQP1 crystals (p4221). The more subtle differences in the
interior of the monomers are likely to be related
with the differences in functional properties of MIP
and AQP1.
Further work aimed at obtaining higher resolution structural data may open the way to identify
different structure features which correlate with
862 MIP Isolation and Crystallizationdifferences in functional properties between members of the MIP family. The MIP crystals presented
here appear to be of suitable size and order to start
such an undertaking.
Materials and Methods
Purification of MIP
Membranes were isolated from sheep lens cores and
prepared as described by Kistler et al. (1994). Brie¯y, lens
tissue (100 lenses) was homogenized in 300 ml 5 mM
Tris (pH 8.0), 5 mM EDTA, 5 mM EGTA. Crude membranes were pelleted at 12,000 g for 20 minutes and
washed twice in the same buffer. Proteins adhering to
the membranes were removed by washing with 100 ml
4 M urea, 5 mM Tris (pH 9.5), 5 mM EDTA, 5 mM
EGTA, and centrifugation at 120,000 g for four minutes.
Membranes were further extracted twice with 100 ml
20 mM NaOH (pH 12.0) and washed once in 100 ml
5 mM Tris (pH 8.0), 100 mM NaCl. The membranes (normal protein concentration 3-4 mg/ml) were frozen in
5 mM Tris (pH 8.0), 100 mM NaCl and stored at ÿ70C
until further use.
After thawing an appropriate aliquot, lens membranes
were washed twice in 10 mM Hepes (pH 7.2) and solubilized with 4% OG (Biomol, Germany) in 10 mM Hepes
(pH 7.2) for 30 minutes at 4C. Following centrifugation
(110,000 g, 4C, 40 minutes) the supernatant was directly
loaded onto a sucrose density gradient (5% to 20%
(w/w) in 10 mM Hepes (pH 7.2) containing 0.3% DM
(Biomol, Germany), and centrifuged overnight at
160,000 g (4C). Fractions containing MIP were identi®ed by SDS gel electrophoresis by Laemmli (1970). Subsequently, pooled fractions were loaded on a MiniS
column (based on MiniBeads2 Pharmacia) connected to
a Smart2 System (Pharmacia). The column was equilibrated with 10 mM Hepes (pH 7.2) containing 0.3%
DM, and the protein was eluted with 110 mM NaCl.
Highly puri®ed MIP was concentrated in Centricon-50
cartridges to 1 mg/ml as determined by the BCA assay
(Pierce, USA) and kept in DM containing 10 mM Hepes
(pH 7.2) at 4C for immediate use.
2D crystallization
A 50 ml aliquot of puri®ed MIP (1 mg/ml) was mixed
with puri®ed E. coli lipids (70% phospahtidylethanolamine, 15% phosphatidylglycerol, 15% cardiolipin (w/w);
Ambudkar & Maloney, 1986) in 10 mM Hepes (pH 7.2)
containing 2% OG at a lipid-to-protein ratio (LPR; w/w)
of 1.0. After pre-incubation for one hour at room temperature the sample was dialysed against one litre
10 mM Mes (pH 6.0), 50 mM MgCl2, 100 mM NaCl,
0.005% (w/v) NaN3, for 24 hours at 23
C, 24 hours at
37C and another 24 hours at 32C. The dialysis buffer
was exchanged once after 24 hours.
The puri®cation of AQP1 and reconstituted into crystalline sheets followed the procedures described by Walz
et al. (1994).
Mass determination in the STEM
Solubilized MIP particles or reconstituted 2D crystals
were adsorbed to thin carbon ®lms supported by fenestrated plastic ®lms, washed in double distilled water,
and freeze-dried for mass analysis of solubilized protein
oligomers, or just air dried for mass analysis of crystal-line sheets. Annular dark-®eld images were recorded in
a STEM (VG-HB5) at 80 kV and doses of 320 electrons/
nm2 for mass measurements. The particle mass and the
mass-per-area of crystalline sheets, were evaluated as
described by MuÈ ller et al. (1992).
Imagining 2-D crystals in the conventional
transmission electron microscope
For negative staining, reconstituted MIP crystals were
adsorbed to a glow-discharged carbon-coated parlodion
®lm supported by a copper grid and stained with 0.75%
(w/v) uranyl formate. Micrographs were recorded at
50,000magni®cation in a Hitachi H-7000 transmission
electron microscope operated at 100 kV using a dose of
2000 to 3000 electrons/nm2.
For freeze-dried samples, grids with MIP or AQP1 2D
crystals were prepared as described by Walz et al. (1996).
Brie¯y, MIP and AQP1 crystals were adsorbed to glowdischarged carbon coated grids (400 mesh), washed with
bidistilled water, blotted on a ®lter paper, and immediately frozen in liquid nitrogen. Grids were transferred to
the pre-cooled stage of the MIDILAB which was permanently attached to a Philips CM12-electron microscope
(Gross et al., 1990). Freeze drying was carried out in the
MIDILAB for two hours at ÿ80C and at a pressure of
p410ÿ7 mbar. Images were recorded at 100 kV acceleration voltage, at a magni®cation of 96.000 (corresponding to a pixel size of 2.5 AÊ ) under cryo (ÿ175C) and low
dose conditions (4500 electrons/nm2) with a 1 k Gatan
694 slow scan CCD camera.
Digital image processing
Micrographs of negatively stained MIP 2D crystals
were inspected by optical diffraction (Aebi et al., 1973),
and well ordered areas were digitized with a Zeiss
SCAI scanner. Averaged projections of negatively
stained sample were determined using the MRC program suite for image processing (Crowther et al., 1996).
Correlation averaging (Saxton & Baumeister, 1982) and
single particle averaging of images from freeze-dried
crystals was performed by the SEMPER image processing system (Saxton, 1996). To this end, a suitable
reference was obtained by Fourier peak ®ltration and
4-fold symmetrization of the motif. For re®nement
cycles references containing two or nine tetramers were
used. While the smaller reference was suf®cient for
translational alignment, the larger reference was
required to achieve a suf®ciently precise angular alignment (during single particle averaging). Averages contained 100 to 400 motifs, depending on the threshold
value for the cross-correlation peak of the aligned
motif with the reference.
The authors are indebted to Dr Shirley MuÈ ller for critical proofreading of the manuscript and helpful discussion. The work of Ms Bettina Wolpensinger who
operated the STEM to record dark-®eld micrographs for
mass determination, and the expert photographic assistance of Ms Marlies Zoller are gratefully acknowledged.
T.W. thanks EMBO for a postdoctoral fellowship. This
work was supported by the Swiss National Foundation
for Scienti®c Research grant No. 31-42435.94 to A.E., and
MIP Isolation and Crystallization 863grant no. 31-36324.92 to H.G.), the State of Basel-Stadt,
the Maurice E. MuÈ ller Foundation of Switzerland, the
Health Research council of New Zealand, the Marsden
Fund, and the University of Auckland.
Aebi, U., Smith, P. R., Dubochet, J., Henry, C. &
Kellenberger, E. (1973). A study of the structure of
the T-layer of Bacillus brevis. J. Supramol. Struct. 1,
Aerts, T., Xia, J.-Z., Slegers, H., de Block, J. &
Clauwaert, J. (1990). Hydrodynamic characterization of the major intrinsic protein from the bovine
lens ®bre membranes. J. Biol. Chem. 265, 8675±8680.
Ambudkar, S. & Maloney, P. C. (1986). Anion exchange
in bacteria: reconstitution of phosphate: hexose 6phosphate antiport from Streptococcus lactis. Methods
Enzymol. 125, 558±563.
Broekhuyse, R. M., Kuhlmann, E. D. & Stols, A. L. H.
(1976). Lens membranes. II. Isolation and characterization of the main intrinsic polypeptide (MIP) of
bovine lens ®ber membranes. Exp. Eye Res. 23, 365±
Chandy, G., Zampighi, G., Kreman, M. & Hall, J. (1997).
Comparison of the water transporting properties of
MIP and AQP1. J. Membrane Biol. 159, 29±39.
Cheng, A., van Hoek, A. N., Yeager, M., Verkman, A. S.
& Mitra, A. K. (1997). Three-dimensional organization of a human water channel. Nature, 387, 627±
Chrispeels, M. J. & Agre, P. (1994). Aquaporins: water
channel proteins of plant and animal cells. Trends
Biochem. Sci 19, 421±425.
Crowther, R. A., Henderson, R. & Smith, J. M. (1996).
MRC image processing programs. J. Struct. Biol.
116, 9±16.
Engel, A., Hoenger, A., Hefti, A., Henn, C., Ford, R. C.,
Kistler, J. & Zulauf, M. (1992). Assembly of 2-D
membrane protein crystals: dynamics, crystal order,
and ®delity of structure analysis by electron
microscopy. J. Struct. Biol. 109, 219±234.
Gorin, M. B., Yancey, S. B., Cline, J., Revel, J.-P. &
Horowitz, J. (1984). The major intrinsic protein
(MIP) of the bovine lens ®bre membrane: characterization and structure based on cDNA cloning. Cell,
39, 49±59.
Gross, H., Krusche, K. & Tittmann, P. (1990). Recent
progress in high resolution shadowing for biological
transmission electron microscopy. In XIIth International Congress of Electron Microscopy, San Francisco Press Inc., Seattle.
Heymann, J., Agre, P. & Engel, A. (1998). Progress on
the structure and function of aquaporin-1. J. Struct.
Biol. In the press.
Jap, B. K. & Li, H. (1995). Structure of the osmo-regulated H2O-channel, AQP-CHIP, in projection at
3.5 AÊ resolution. J. Mol. Biol. 251, 413±420.
Kistler, J. & Bullivant, S. (1980). Lens gap junctions and
orthogonal arrays are unrelated. FEBS Letters, 111,
Kistler, J., Goldie, K., Donaldson, P. & Engel, A. (1994).
Reconstitution of native-type noncrystalline lens
®ber gap junctions from isolated hemichannels.
J. Cell Biol. 126, 1047±1058.
KoÈnig, N., Zampighi, G. A. & Butler, P. J. G. (1997).
Characterisation of the major intrinsic proteins
(MIP) from bovine ®bre membranes by electronmicroscopy and hydrodynamics. J. Mol. Biol. 265,
Kushmerick, C., Rice, S. J., Baldo, G. J., Haspel, H. C. &
Mathias, P. T. (1995). Ion, water and neutral solute
transport in Xenopus oocytes expressing frog lens
MIP. Exp. Eye Res. 61, 351±362.
Kushmerick, C., Varadaraj, K. & Mathias, R. T. (1997).
Effects of lens MIP on glycerol permeability and
metabolism. Invest. Ophthal. Vis. Sci. 38, 3266.
Laemmli, U. K. (1970). Cleavage of structural proteins
during the assembly of the bacteriophage T4.
Nature, 227, 680±688.
Li, H., Lee, S. & Jap, B. K. (1997). Molecular design of
aquaporin-1 water channel as revealed by electron
crystallography. Nature Struct. Biol. 4, 263±265.
Lin, J., Fitzgerald, S., Dong, T., Knight, C., Donaldson,
P. & Kistler, J. (1997). Processing of the gap junction
protein connexin50 in the ocular lens is accomplished by calpain. Eur. J. Cell Biol. 73, 141±149.
Michea, L. F., de la Fuente, M. & Lagos, N. (1994). Lens
major intrinsic proteins (MIP) promotes adhesion
when reconstituted into larger unilamellar liposomes. Biochemistry, 33, 7663±7669.
Mulders, S. M., Preston, G. M., Deen, P. M. T., Guggino,
W. B., van Os, C. H. & Agre, P. (1995). Water channel properties of major intrinsic proteins of lens.
J. Biol. Chem. 270, 9010±9016.
MuÈ ller, S. A., Goldie, K. N., Buerki, R., Haering, R. &
Engel, A. (1992). Factors in¯uencing the precision of
the quantitative scanning transmission electron
microscopy. Ultramicroscopy, 46, 317±334.
Park, J. H. & Saier, M. H., Jr (1996). Phylogenetic characterization of the MIP family of transmembrane
channel proteins. J. Membrane Biol. 153, 171±180.
Reizer, J., Reizer, A. & Saier, M. H., Jr (1993). The MIP
family of integral membrane channel proteins:
sequence comparisons, evolutionary relationships,
reconstructed pathway of evolution, and proposed
functional differentiation of the two repeated halves
of the proteins. Crit. Rev. Biochem. Mol. Biol. 28,
Saxton, W. O. (1996). Semper: distortion, compensation,
selective averaging, 3-D reconstitution, and transfer
function correlation in a highly programmable system. J. Struct. Biol. 116, 230±236.
Saxton, O. & Baumeister, W. (1982). The correlation
averaging of a regularly arranged bacterial cell
envelope protein. J. Microsc. 127, 127±138.
Shiels, A. & Bassnett, S. (1996). Mutations in the founder
of the MIP gene family underlie cataract development in the mouse. Nature Genet. 12, 212±215.
Walz, T., Smith, B. L., Zeidel, M. L., Engel, A. & Agre,
P. (1994). Biologically active two-dimensional crystals of aquaporin CHIP. J. Biol. Chem. 269, 1583±
Walz, T., Typke, D., Smith, B. L., Agre, P. & Engel, A.
(1995). Projection map of aquaporin-1 determined
by electron crystallography. Nature Struct. Biol. 2,
Walz, T., Tittmann, P., Fuchs, K., MuÈ ller, D., Smith, B.,
Agre, P., Gross, H. & Engel, A. (1996). Surface topographies at subnanometer-resolution reveal asymmetry and sidedness of aquaporin-1. J. Mol. Biol.
264, 907±918.
Walz, T., Hirai, T., Murata, K., Heymann, J. B.,
Mitsuoka, K., Fujiyoshi, Y., Smith, B. L., Agre, P. &
Engel, A. (1997). The three-dimensional structure of
aquaporin-1. Nature, 387, 624±627.
864 MIP Isolation and CrystallizationZampighi, G., Simon, S. A., Robertson, J. D., McIntosh,
T. J. & Costello, M. J. (1982). On the structural
organization of isolated bovine lens ®ber junctions.
J. Cell Biol. 93, 175±189.Zampighi, G. A., Hall, J. E., Ehring, G. R. & Simon, S. A.
(1989). The structural organization and protein composition of lens ®ber junctions. J. Cell Biol. 8, 2255±
2275.Edited by W. Baumeister(Received 8 October 1997; received in revised form 10 March 1998; accepted 15 March 1998)

PDF Document reader online

This website is focused on providing document in readable format, online without need to install any type of software on your computer. If you are using thin client, or are not allowed to install document reader of particular type, this application may come in hand for you. Simply upload your document, and Docureader.top will transform it into readable format in a few seconds. Why choose Docureader.top?

  1. Unlimited sharing - you can upload document of any size. If we are able to convert it into readable format, you have it here - saved for later or immediate reading
  2. Cross-platform - no compromised when reading your document. We support most of modern browers without the need of installing any of external plugins. If your device can oper a browser - then you can read any document on it
  3. Simple uploading - no need to register. Just enter your email, title of document and select the file, we do the rest. Once the document is ready for you, you will receive automatic email from us.

Previous 10

Next 10