 |
INTRODUCTION |
Structural and functional studies of biological molecules are
often based on surface-sensitive techniques, among them attenuated total reflection infrared spectroscopy, evanescent field fluorescence microscopy, ellipsometry, and scanning probe microscopies. Hence, a
wide range of techniques would benefit from a generally applicable approach to the immobilization of biological macromolecules. As a key
requirement, any general strategy to immobilization will have to fully
preserve the biological activity of molecules subjected to it; any
adverse effects that immobilization might have on molecules in terms of
their structure and function should be minimized. The sites relevant to
their biological function like ligand-binding sites or active sites
must remain accessible after transfer to a solid surface. This demands
that the orientation of the molecule with respect to the supporting
interface can be fully controlled.
Formation of metal-chelate complexes is a strategy commonly used for
the purification of proteins containing an engineered His tag,
i.e. a stretch of consecutive histidines (1).
Nitrilotriacetic acid (NTA)1
binds nickel ions coordinatively (2), leaving two coordination sites
available for interaction with a His tag. Binding of His-tagged proteins to a Ni-NTA presenting surface is specific (3) with a binding
constant in the micromolar range (4), but easily reversible.
Importantly, via judicious introduction of a His tag, the site through
which molecules bind to a chelating support and, possibly, their
orientation with respect to a chelating support, can be controlled.
The 20 S proteasome is a large multisubunit protease, which plays a key
role in the degradation of misfolded proteins and of short-lived
regulatory proteins (5-7). Its 28 subunits are arranged in four
heptameric rings, which collectively form a barrel-shaped complex 15 nm
in length and 11 nm in diameter. There are two types of subunits, 
and 
; the -type subunits form the two outer rings and the
-type subunits the two inner rings. In most prokaryotes, including
the Archaeon Thermoplasma acidophilum, the rings are homoheptameric; thus, the stoichiometry of these 20 S proteasomes is
7777 (Fig.
1A).
Proteasomes are members of the superfamily of Ntn (N-terminal
nucleophile) hydrolases; a common feature of this group of enzymes is a
single-residue active site, i.e. both the nucleophile and the primary proton acceptor are provided by the same N-terminal residue. In the proteasome this is the Thr1 of the
-subunits (8, 9). During assembly, 20 S proteasomes undergo a
post-translational modification that removes a propeptide and exposes
the -amino group of Thr1 (10). In mature 20 S
proteasomes Thr1 is accessible only from the central, ~5
nm wide cavity, which is formed by the two rings of -subunits (Fig.
1A). Substrates destined for degradation must therefore be
translocated into the interior of the proteasome from either end
through entrance channels situated at the center of the two -rings
(11) (Fig. 1A). Turn forming segments around
Tyr126 of the seven -subunits appear to define the
properties of these 1.3-nm wide channels, although it should be noted
that there are several disordered residues (8) that are likely to
(partially) occlude the channels. In any case, access to the interior
of the proteasome and thus to the active sites is restricted to
completely unfolded polypeptide chains (11). Therefore, to be
functional in vivo, 20 S proteasomes must associate with
regulatory complexes, invariably containing AAA-ATPases, which prepare
substrates for degradation by unfolding them (12-14). Moreover,
regulatory complexes such as the 19 S caps of eukaryotic proteasomes
have been implicated in gating the -ring channels (15, 16).
To explore the use of a His tag as a handle that might allow one to
immobilize functional molecules in a predictable orientation, we have
generated recombinant 20 S proteasomes His-tagged in various positions.
The option to immobilize proteasomes in a uniform and predetermined
orientation allowing regulatory complexes unhindered binding holds
prospect for structural studies addressing the interaction of the
proteasome with cofactors and their role in substrate translocation. Immobilizing proteasomes while fully preserving their proteolytic activity will benefit investigations of proteasome-substrate
interaction in real time for detailed kinetic analysis, possibly at the
single molecule level. Following biochemical characterization of
His-tagged proteasomes, their interaction with metal-chelating lipid
interfaces has been assessed.
| |
EXPERIMENTAL PROCEDURES |
Materials--
PFU-TURBO polymerase from Stratagene (Amsterdam,
The Netherlands) in combination with dNTPs (polymerization mixture)
from Amersham Biosciences were used in polymerase chain
reactions. Sequencing was performed using the BIGDYE terminator cycle
sequencing kit from PerkinElmer Life Sciences. 20 S proteasomes
were purified either on Ni-NTA resin (SUPERFLOW from Qiagen GmbH,
Hilden, Germany) or hydroxyapatite (Bio-Gel HTP from Bio-Rad).
Bicinchoninic acid solution was from Sigma. Protease activity was
assayed using succinyl-Leu-Leu-Val-Tyr-7-amido-4-methylcoumarin (Suc-LLVY-AMC) from Bachem (Heidelberg, Germany) or
14C-methylated -casein from Sigma. Mica was glued to
Teflon sheets with Bindulin two-component epoxy glue
(Fürth, Germany). The chelating lipid NTA-DODA
(N
,N-bis(carboxymethyl)-N
-[(dioctadecylamino)-succinyl]-L-lysine)
was synthesized as described (17). SOPC
(1-stearoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine) was
from Avanti Polar Lipids, whereas DSPC
(1,2-distearoyl-sn-glycero-3-phosphatidylcholine) was from Sigma.
Mutagenesis, Expression, and Purification of 20 S
Proteasomes--
pRSET6a plasmids coding for both the -subunit and
-subunit (either wild-type or His-tagged at the C terminus) of the
20 S proteasome from T. acidophilum (referred to simply as
"proteasome" in the following) were provided by Dr. Erika
Seemüller. His tags were introduced into the -subunit starting
from the plasmid coding for the wild-type proteasome using whole
plasmid amplification via the polymerase chain reaction (Pfu-turbo
polymerase) starting with primers including the desired changes.
Sequencing was performed on a PerkinElmer Life Sciences applied
biosystems 373 DNA sequencer. The - and -subunits of the
proteasome were coexpressed in Escherichia coli BL21(DE3).
The assembled complex was purified either on Ni-NTA resin or on
hydroxyapatite if the affinity for Ni-NTA resin was low. In the latter
case, material bound to hydroxyapatite was eluted with a gradient from
10 to 600 mM potassium phosphate buffer, pH 7; proteasomes
eluted at around 500-550 mM potassium phosphate. Purity of
fractions showing proteasomal activity against Suc-LLVY-AMC was tested
by SDS-polyacrylamide gel electrophoresis. Protein concentration was
determined using the bicinchoninic acid assay.
Activity of Proteasomes in Solution--
To follow the
degradation of fluorogenic substrates in solution, Suc-LLVY-AMC was
dissolved at 10 mM in dimethyl sulfoxide (Me2SO) and added at a final concentration of 125 µM (1.25% Me2SO) in 800 µl of buffer A (50 mM Tris, pH 7.5) containing 100 ng of proteasomes
(corresponding to 0.18 nM) for continuous measurement at
60 °C. The released AMC was detected (
ex/em 380 nm/440 nm) in an LS 50 B fluorescence spectrometer (PerkinElmer Life Sciences).
To follow the degradation of proteins in solution, 11 µg of
14C-methylated -casein (21,800 cpm/µg, final
concentration 8.76 µM) was digested with 300-600 ng of
proteasomes (8.6-17.2 nM) in 50 µl of buffer A for 40 min at 60 °C. The reaction was stopped by adding 1 ml of 12%
trichloroacetic acid followed by incubation on ice for 30 min.
Centrifugation for 20 min in the presence of 10 µl of bovine serum
albumin (1%) at 13,000 × g separated acid-soluble degradation products that were quantified in a TRI CARB 1500 liquid scintillation counter (Packard Instrument Co.).
Preparation of Vesicles and Supported Nickel-chelating Lipid
Bilayers--
For preparation of lipid vesicles, NTA-DODA and SOPC
were mixed (1:9 mol/mol) in chloroform. NTA-DODA was loaded with nickel ions (Ni-NTA-DODA) in chloroform/methanol (3:1) by adding equimolar amounts of NiCl2·6H2O dissolved in methanol.
After evaporation of the solvent under a stream of nitrogen, vesicles
were obtained by swelling in buffer B (10 mM Hepes, 150 mM NaCl, pH 7.5) to a final lipid concentration of 2 mM and repeated extrusion (30 times) through 100-nm filters
(LIPOSO FAST-BASIC, Avestin, Ottawa, Canada).
Supported lipid bilayers were prepared on mica sheets glued to Teflon
supports cut to either fit quartz cuvettes (3 mm × 9 mm) or the
fluid cell of an atomic force microscope (6 mm diameter). Freshly
cleaved mica was first incubated with buffer B containing 15 mM Mg2+ for 60 min, then with vesicle solution
(Ni-NTA-DODA/SOPC, 1:9) at 50 °C to yield bilayers. Preincubation of
mica with Mg2+ resulted in increased bilayer stability;
presumably Mg2+ is required to compensate for the negative
surface charge of both mica and chelating lipid film. After 3 h,
the surfaces were thoroughly rinsed with buffer B.
Activity of Proteasomes Immobilized at Nickel-chelating Lipid
Interfaces--
The activity of proteasomes was tested after
immobilization on two kinds of bilayer specimens, that is on vesicle
bilayers and on mica supported bilayers. For the first set of
measurements, vesicles were prepared as described in the previous
section. 11 nM proteasomes were incubated with these
vesicles at a proteasome/chelator lipid ratio of 1:910 for 3 h at
4 °C. Degradation of 125 µM Suc-LLVY-AMC by the
proteasomes immobilized on the vesicles was performed for 15 min at
60 °C. The reaction was stopped by addition of a 9-fold excess of
100 mM chloroacetic acid, 100 mM acetic acid,
pH 4.3, before quantifying fluorescence. The activity of unbound
proteasomes was measured in the presence of 250 mM
imidazole, which leads to desorption of the proteasomes from the
Ni-chelator lipid.
For the measurements with proteasomes bound to solid supported
bilayers, first the activity of a known amount of proteasomes in
solution was assayed. Quartz cuvettes holding 300 ng of proteasomes in
2970 µl of buffer C (100 mM Tris, pH 7.5) were
equilibrated at 40 °C for 15 min. Following addition of 30 µl of
Suc-LLVY-AMC (final concentration 100 µM Suc-LLVY-AMC,
1% Me2SO, 0.14 nM proteasome), cuvettes were
kept at 40 °C for 30 min, then chilled on ice for 5 min. Released
AMC was detected (ex/em 380 nm/440 nm) in a fluorescence spectrometer.
An equal amount of His-tagged proteasomes (300 ng) was adsorbed onto a
mica-supported nickel-chelating lipid bilayer (Ni-NTA-DODA/SOPC, 1:9)
overnight to allow them to bind. Proteasomes that did not bind were
transferred using a pipette into a cuvette holding 2970 µl of buffer
C. The rate at which AMC was released by the unbound proteasomes was
quantified as before. Subtracting this rate from the rate at which 300 ng of proteasomes in solution released AMC equals the rate at which
those proteasomes bound to the bilayer would release AMC if their
activity was not affected by immobilization at all (the "ideal
rate"). The rate at which proteasomes immobilized on the supported
bilayer overnight actually released AMC was measured as before after
placing the mica-supported bilayer with bound proteasomes at the bottom
of a cuvette holding 2970 µl of buffer C. Dividing this observed rate
by the "ideal rate" indicates the extent to which the activity of
His-tagged proteasomes bound to chelating lipid bilayers is preserved.
The system mica-bilayer-proteasome was stable during incubation with
Suc-LLVY-AMC: upon removal of the mica-supported lipid bilayer with
bound proteasomes from the cuvette during hydrolysis, the fluorescence
level ceased to rise.
Electron Microscopy--
To investigate the interaction of
proteasomes with lipid films, 1.5 µg of proteasomes in 15 µl of
buffer B were placed in a Teflon well (4 mm diameter, 1 mm deep). The
surface of the solution was coated with 1 µl of 1 mM
lipid in chloroform/hexane (1:1, v/v). The lipid was, as indicated,
SOPC, a mixture of NTA-DODA (no nickel) and SOPC (1:9), a mixture of
Ni-NTA-DODA and SOPC (1:9), or a mixture of Ni-NTA-DODA and DSPC (1:9).
Following incubation for either 1 h or overnight, the
lipid-proteasome assembly was transferred onto a carbon-coated grid by
placing the grid onto the droplet for 3 min. Such samples as well as
proteasomes adsorbed directly onto carbon-coated grids were stained
with 2% uranyl acetate for 30 s and inspected in a transmission
electron microscope (CM12, Philips, Eindhoven, The Netherlands) at 120 kV. A Silicon Graphics work station with SEMPER software (18) was used
for averaging images using a correlation-based approach (19). An area
comprising three unit cells in diameter was chosen as a reference and
cross-correlated with the image.
Atomic Force Microscopy--
To obtain AFM (atomic force
microscope) topographs of the lipid-proteasome assembly, His-tagged
proteasomes were immobilized on supported nickel-chelating lipid
bilayers. Between 1 and 5 µg of His-tagged proteasomes in buffer B
were adsorbed overnight on the mica-supported bilayer. Care was taken
to always keep the solid supported membranes submerged in buffer. After
rinsing with buffer B the samples were mounted in the fluid cell of a
NANOSCOPE AFM (Digital Instruments, Santa Barbara, CA) and imaged in
buffer B either in tapping mode using a drive frequency around 9 kHz and 5-10 nm amplitude of the cantilever, or in contact mode. In either
mode, forces exerted by the tip onto the sample were minimized by
adjusting the set point manually while scanning. Silicon nitride cantilevers (OMCL-TR400PSA-2, Olympus, Japan) with nominal spring constants of 0.09 newton/m were used; their oxide sharpened tips were cleaned by UV irradiation for ~30 min prior to use.
| |
RESULTS |
Proteasomes His-tagged in Various Positions Can be Efficiently
Expressed and Purified--
The main goal of this study was to
immobilize the cylindrical-shaped 20 S proteasome with predetermined
orientation, end-on and side-on. To this end, recombinant T. acidophilum 20 S proteasomes (referred to simply as
"proteasome" in the following) were engineered, carrying His
tags at positions favoring one or the other orientation. Proteasomes
His-tagged at the sides were obtained by fusing six histidines to the C
termini of -subunits, referred to as C-His6 proteasomes (Fig. 1B).
Proteasomes His-tagged at the ends were obtained by fusing six
histidines to the N termini of -subunits (N-His6
proteasomes) (Fig. 1C), or alternatively, by integrating a
stretch of histidines into a solvent-accessible turn of -subunits connecting helix H0 and strand S1 between Lys33 and
Gly34 (Fig. 1D). Proteasomes with six or eight
consecutive histidines within this turn are referred to as turn-His6 or turn-His8 proteasomes,
respectively.
View larger version (39K):
[in this window]
[in a new window]
|
Fig. 1.
Schematic representation of the T. acidophilum proteasome showing the three different
positions of the His tags. A, low resolution
model (~1.2 nm) derived from the atomic coordinates of the T. acidophilum proteasome (8). One subunit per heptameric ring is
highlighted; -subunits form the two outer rings, -subunits form
the two inner rings (top). The proteasome cut open shows the
two entrance channels (one per -ring) and the central cavity. The
N-terminal, proteolytically active threonines are highlighted in
red (bottom). B, ribbon drawing of the
-subunit with its C terminus (top). Proteasomes
His-tagged at the C termini of their -subunits display His tags
around their sides and are referred to as C-His6
proteasomes. White circles indicate the location of the His
tags (bottom). C, ribbon drawing of the
-subunit with its N terminus (top). Proteasomes
His-tagged at the N termini of their -subunits display His tags at
their ends around the two cylinder openings and are referred to as
N-His6 proteasomes (bottom). D,
ribbon drawing of the -subunit with Lys33. Helix H0 is
located between Lys33 and the N terminus (top).
Proteasomes including six or eight consecutive histidines between
Lys33 and Gly34 of their -subunits display
His tags at the ends and are referred to as turn-His6
or as turn-His8 proteasomes, respectively
(bottom). Drawings were produced with MOLSCRIPT (49) and
RASTER3D (50).
|
|
Expression levels were similar for all four constructs.
C-His6, N-His6, and turn-His8 proteasomes had high affinity for nickel-chelating nitrilotriacetic acid facilitating their efficient purification in a single step on Ni-NTA resin. turn-His6 proteasomes, however, could not be purified on
Ni-NTA resin, which is indicative of a low affinity for the Ni-NTA
group, but could be purified on hydroxyapatite (see "Experimental
Procedures").
The Activity of the Proteasome Can Be Both Decreased and
Increased--
The activity of all constructs with high
affinity for Ni-NTA was first assayed using Suc-LLVY-AMC, a standard
fluorogenic substrate of the proteasome. Under the experimental
conditions chosen (60 °C), hydrolysis of Suc-LLVY-AMC proceeded at
rates of 9.5 µmol/h per mg of proteasome for C-His6,
N-His6, and turn-His8 proteasomes, close
to the rates measured with wild-type proteasomes and in good agreement
with previous data (20). Degradation of proteins proceeded at almost
identical rates for C-His6 and wild-type proteasomes, as
judged from hydrolysis of 14C-methylated -casein (Table
I). However, N-His6 and
turn-His8 proteasomes differed in their activities from
wild-type proteasomes, indicating that the His tags might affect
translocation of long polypeptide chains across the -ring channel.
Interestingly, N-His6 proteasomes degraded radiolabeled
-casein at lower rates, whereas turn-His8
proteasomes degraded this substrate faster than wild-type proteasomes
(Table I).
Proteasomes Can Be Oriented at Metal-chelating Lipid Interfaces via
the Location of His Tags--
To test whether site-specific binding
between His-tagged proteasomes and nickel-chelating interfaces would
permit one to control orientation, proteasomes were allowed to interact
with nickel-chelating lipid layers. Proteasomes were adsorbed to lipid
films containing 10% Ni-NTA-DODA and 90% SOPC or DSPC for either
1 h or overnight. The lipid films were then transferred onto
carbon grids and examined by transmission electron microscopy.
In standard transmission electron microscopy preparations where
20 S proteasomes from T. acidophilum are applied to carbon film, they show up in two characteristic orientations (21): ring-shaped
end-on views with a diameter of 11 nm and rectangular side-on views (15 nm × 11 nm) with a characteristic pattern of four striations,
representing the four proteasomal rings; a typical example is seen in
Fig. 2. Likewise, when adsorbed
unspecifically to SOPC lipid films, proteasomes are seen randomly
oriented, regardless of the presence of His tags.
View larger version (190K):
[in this window]
[in a new window]
|
Fig. 2.
Electron micrograph of cylindrical wild-type
20 S proteasomes adsorbed on carbon film. End-on views appear
ring-shaped (11 nm), side-on views appear rectangular (15 nm × 11 nm) with four striations, representing the four proteasomal rings
(scale bar: 70 nm).
|
|
In contrast, the orientation of isolated, His-tagged proteasomes bound
to metal-chelating lipid films was highly uniform and in line with
expectations: C-His6 proteasomes His-tagged at their sides displayed exclusively side-on views (Fig.
3A), whereas
N-His6 and turn-His8 proteasomes
His-tagged at their ends displayed exclusively end-on views (Fig.
3B). turn-His6 proteasomes adsorbed randomly
oriented, which is not surprising given their low affinity for Ni-NTA
resin.
View larger version (215K):
[in this window]
[in a new window]
|
Fig. 3.
According to electron micrographs the
orientation of His-tagged proteasomes bound to metal-chelating lipid
films depends on the location of the His tags and is highly uniform.
A, C-His6 proteasomes His-tagged at their
sides show exclusively side-on views when adsorbed to nickel-chelating
lipid films (scale bar, 100 nm in A-C).
B, in contrast, proteasomes His-tagged at their ends
uniformly display end-on views when adsorbed to nickel-chelating lipid
films, as shown here for turn-His8 proteasomes.
C, N-His6 proteasomes His-tagged at their
ends form extended, close packed arrays upon overnight incubation with
nickel-chelating lipid films. Large two-dimensional crystalline arrays
only form with constructs yielding a uniform orientation. D,
the power spectrum calculated from the two-dimensional crystal in
C. The two crystallographic vectors are indicated.
Reflections go out to 1/2.5 nm 1 (indicated by a
white circle).
|
|
All components of the His tag-nickel-NTA system are required to achieve
uniform orientation. In fact, proteasomes were randomly oriented if (i)
wild-type proteasomes lacking His tags were adsorbed on
Ni-NTA-DODA/SOPC (1:9), (ii) His-tagged proteasomes were adsorbed (in
the presence of nickel ions) on SOPC films lacking NTA-DODA, or (iii)
His-tagged proteasomes were adsorbed on NTA-DODA/SOPC (1:9) in the
absence of nickel ions (in the presence of EDTA). The strict dependence
of orientation on all components of the His tag-nickel-NTA system
confirms that binding of His-tagged proteasomes to metal-chelating
lipid interfaces is specific.
The Activity of Proteasomes Bound to Metal-chelating Lipid
Interfaces Is Well Preserved--
To demonstrate that immobilization
of His-tagged proteins at metal-chelating lipid interfaces is
biocompatible, the activity of immobilized proteasomes was compared
with the activity of proteasomes in solution. One set of measurements
was performed at 60 °C with the C-His6 and the
N-His6 mutants bound directly on vesicles formed by
nickel-chelating lipid bilayers (Ni-NTA-DODA/SOPC, 1:9). By this
immobilization, activity of the proteasomes was reduced by about 8%
for both mutants in comparison to their activity in solution.
Additionally, proteasome activity was measured for
C-His6 proteasomes bound to a mica-supported bilayer.
First the activity of a known amount of these proteasomes in solution
was assayed. At the suboptimal temperature of 40 °C, which ensured
stability of the complete mica-lipid bilayer-proteasome system,
hydrolysis of Suc-LLVY-AMC by 300 ng of proteasomes proceeded at a rate
of 0.89 nmol/h, which corresponds to 2.97 µmol/h/mg of proteasome.
An equal amount of proteasomes was adsorbed onto a supported
nickel-chelating lipid bilayer (Ni-NTA-DODA/SOPC, 1:9) overnight to
allow them to bind. Beforehand, the presence of a continuous lipid
bilayer was routinely confirmed by AFM. Proteasomes that did not bind
to the bilayer released AMC at a rate of 0.2 nmol/h. Subtracting this
rate from the rate at which 300 ng of proteasomes in solution released
AMC equals the rate at which those proteasomes bound to the bilayer
would release AMC if their activity was not affected by immobilization
at all: 0.69 nmol/h (the ideal rate). The rate at which the proteasomes
immobilized on the supported bilayer overnight actually released AMC
was measured to be 0.35 nmol/h, which amounts to 51% of the ideal rate.
Oriented, His-tagged Proteasomes Crystallize in Two Dimensions
on Fluid Metal-chelating Lipid Films--
Proteasomes bound to fluid
nickel-chelating lipid films end-on (N-His6 and turn-His8) showed a high propensity to self-organize in two
dimensions. Small close packed arrays of proteasomes were already
observed after incubation for 1 h with Ni-NTA-DODA/SOPC (1:9)
films. Much larger arrays were observed after overnight incubation
(e.g. Fig. 3C). Under the same conditions,
C-His6 proteasomes also formed densely packed
protein layers, but only a few small ordered patches were
observed. His-tagged proteasomes were also found to orient on
nickel-chelating lipid films when SOPC was replaced by DSPC
(Ni-NTA-DODA/DSPC, 1:9). However, no two-dimensional
organization was observed on such lipid films. Possibly, DSPC restricts
the lateral mobility of the lipid-bound proteasomes.
Power spectra calculated from negatively stained two-dimensional arrays
of N-His6 proteasomes indicated that the organization of
proteasomes within the two-dimensional arrays went beyond simple close
packing and showed diffraction spots to 1/2.5 nm1 (Fig.
3D). Correlation averaging revealed that the rotation of proteasomes about their 7-fold symmetry axis is limited to
discrete angles (Fig. 4), giving rise to
crystals of pg symmetry.
View larger version (182K):
[in this window]
[in a new window]
|
Fig. 4.
Correlation average of the two-dimensional
crystal with N-His6 proteasomes
(see Fig. 3C) indicating pg
symmetry. The unit cell outlined by brackets
measures 19.5 × 11.2 nm. Within rows running
diagonally from the left bottom to the right
top, proteasomes interlock with their neighbors.
|
|
Images of the two-dimensional crystals of N-His6
proteasomes on supported nickel-chelating lipid bilayers were also
taken by AFM (Fig. 5). Crystals could be
repeatedly scanned without apparent degradation. AFM topographs
revealed that N-His6 proteasomes alternated in height
within the crystals; every second row of proteasomes stood out by about
2 nm, adding a doubled period to the close packing (Fig. 5). This
pattern of alternating rows was continuous throughout scan areas as
large as 2 × 2 µm. The diameter of N-His6
proteasomes in the two-dimensional crystals was 11 nm with their ends
appearing dome-shaped in AFM topographs. Possibly, the central channel
in -rings was not visible because of the presence of seven His tags
per -ring. The appearance of different proteasomes was slightly
heterogeneous, the arrangement of the seven His tags may vary for
different proteasomes. However, the appearance of proteasomes did not
change in time during repeated scanning.
View larger version (132K):
[in this window]
[in a new window]
|
Fig. 5.
AFM topograph showing a crystal of
N-His6 proteasomes. Proteasomes
within crystals are arranged at 2 levels; rows of proteasomes
running diagonally from the left top to the
right bottom alternate in distance from the support by about
2 nm. The cartoon at the bottom illustrates the
staggered arrangement of proteasomes along the drawn line.
Neighboring proteasomes nest ring-to-waist into each other by up and
down shifts along the proteasomal cylinder axis. The picture was taken
in contact mode (software zoom, 150 × 150 nm; scan area, 400 × 400 nm).
|
|
| |
DISCUSSION |
The Orientation of His-tagged Proteins with Respect to
Metal-chelating Lipid Interfaces Can Be Controlled via a Judicious
Choice of His Tag Position--
Recombinant DNA techniques allow one
to introduce specific sites of attachment into proteins. Cysteines (22)
and, more commonly, amino acid sequences including His tags (1) are
used for site-specific immobilization. To demonstrate, however, that
site-specific immobilization can lead to uniform orientation it is
necessary to vary the location of an affinity site within otherwise
identical molecules and to show that thereby their orientation can be changed.
Previously, cysteine residues have been used to manipulate the
orientation of proteins on solid supports. To this end, the location of
an engineered cysteine has been varied in cytochrome b5. Two mutants having either Thr8
or Thr65 replaced by cysteine (T8C or T65C) have been shown
to differ in their mean orientation with respect to their support upon
site-specific immobilization (23, 24). Linear dichroism measurements
revealed that the heme group of cytochrome b5
(T8C) adopted an average angle of 78° with respect to the support,
whereas the heme group of cytochrome b5 (T65C)
adopted an average angle of 45° with respect to the support (24).
However, the distribution of orientations has been shown to be rather
broad when molecules were immobilized via a single cysteine: cytochrome
c immobilized on a self-assembled monolayer with terminal
thiol groups displayed a broad distribution of heme orientations (25),
as judged from a combination of fluorescence intensity and polarization
measurements (26, 27). Thus, whereas site-specific immobilization will
influence orientation, it will not necessarily lead to a uniform orientation.
Taking advantage of the chelating lipid concept (17, 28), the effect of
the location of the His tag on the orientation of proteins at a
chelating lipid film has been investigated using two variant RNA
polymerase I molecules, His-tagged on either subunit AC40
or subunit ABC23. Within two-dimensional crystals,
orientation was found to correlate with the location of the His tag
(29); both variants contacted the lipid film with their respective
tagged subunits, thus interacting with the lipid film in reversed
orientation to yield different crystal forms. However, isolated
polymerase molecules outside the crystal patches appeared in several
different orientations. This indicates that the uniform orientation
observed within the crystal patches may at least in part result from
lateral protein-protein interactions.
To investigate the potential use of the His tag as a handle to exert
control over the orientation of a protein, 20 S proteasomes His-tagged
at different sites were examined by transmission electron microscopy
and AFM after binding to nickel-chelating lipid films. When incubation
periods were short enough, proteasomes were mostly well separated from
one another; therefore, we can assume that their orientation is
determined exclusively by their interaction with the lipid film.
Proteasomes His-tagged at their sides (C-His6) uniformly
displayed side-on views when bound to nickel-chelating lipid films
(Fig. 3A). Proteasomes His-tagged at their ends
(N-His6 and turn-His8) were bound with
their ends to the nickel-chelating lipid film, resulting exclusively in
end-on views (Fig. 3B).
Interestingly, six histidines are sufficient to control orientation
when fused to the C terminus of -subunits or the N terminus of
-subunits, whereas a stretch of eight histidines are required to
control orientation when the His tag is inserted between
Lys33 and Gly34 of -subunits. Proteasomes
having only six consecutive histidines between Lys33 and
Gly34 of their -subunits were found to be randomly
oriented on nickel-chelating lipid films in line with their low
affinity for Ni-NTA. Possibly, a stretch of six histidines "pinned"
at both ends is not sufficiently accessible or conformationally too
restricted to tightly coordinate Ni-NTA. It is interesting to note that
isolated histidines on the surface of the -subunit of the proteasome
(His109) have no effect on the orientation, as is evident
from the random orientation of wild-type proteasomes when incubated
with nickel-chelating lipid films.
Uniform orientation was only achieved when all components of the His
tag-nickel-NTA system were present. The lack of any one component
resulted in the coexistence of side-on and end-on views. Thus, the
uniform orientation is because of specific interactions between His
tags and the metal-chelating lipid layer. Both the end-on and the
side-on orientation proteasomes are probably bound to chelating lipid
films via several His tags. For the end-on orientation, as many as
seven His tags might participate in the interaction. This multipoint
attachment may play a critical role in turning site-specific binding
into oriented immobilization.
Uniform Orientation of His-tagged Proteasomes on Nickel-chelating
Lipid Films Promotes Their Two-dimensional
Crystallization--
Incubation of His-tagged proteasomes with fluid
metal-chelating lipid films for longer periods of time often resulted
in two-dimensional crystalline arrays (see Fig. 3C). When
the 20 S proteasomes occurred both in side-on and end-on orientations,
as was the case with wild-type proteasomes on SOPC lipid films, only
very small crystal patches were observed. Possibly, the nonuniform
orientation prevents the growth of larger two-dimensional crystals.
With C-His6 proteasomes, and more often with
N-His6 and turn-His8 proteasomes, larger crystalline patches were observed. The orientation of the proteasomes in these two-dimensional crystals was the same as found for the corresponding isolated proteasomes. Thus, the uniform orientation achieved with these constructs appears to promote two-dimensional crystallization. The potential of His tags to be used in combination with metal-chelating lipid layers to generate two-dimensional crystalline arrays has been recognized some time ago (17, 29-33). It
is seen as a widely applicable and rational strategy to produce two-dimensional crystals of soluble and membrane proteins suitable for
electron crystallography.
When crystals formed by N-His6 proteasomes were analyzed
in more detail, it was found that the close packed proteasomes optimize their packing by a rotation about their 7-fold symmetry axis so as to
form rows of interlocking proteasomes. The directionality of
interlocking alternates between neighboring rows (Fig. 4). The
resulting two-dimensional crystals of pg symmetry are built from
a unit cell containing two proteasomes related by glide reflection. The
parameters of the unit cell are a = 19.5 nm,
b = 11.2 nm, 
= 90°.
After very long periods of incubation, multilayered structures were
occasionally observed. Their formation may be promoted by the fact that
bound proteasomes not only display His tags toward the lipid layer, but
also in the direction pointing away from it. Other His-tagged molecules
may bind to the primary arrays decorated with His tags, possibly
mediated by small amounts of nickel released from the Ni-NTA resin
during purification.
Binding of His-tagged Proteins to Metal-chelating Lipid Interfaces
Is Biocompatible--
Although it has been shown for several enzymes
that they retain their activity after immobilization (see Refs. 34 and
35), this cannot be taken for granted. Not only the
immobilization might affect the activity, also the mere presence of His
tags can, in principle, have an influence. Therefore, it was important to show that the activity of His-tagged proteasomes
(C-His6, N-His6, and turn-His8) against Suc-LLVY-AMC in solution was identical
to that of wild-type proteasomes. Obviously, the presence of His tags
on the proteasomal exterior has no strong effects on peptide hydrolysis.
A critical step in the degradation of proteins is their uptake through
the channel formed by the -rings that controls access to the
interior. In the crystal structure of the T. acidophilum proteasome this channel appears to be open with a diameter of 1.3 nm
(8), in the yeast proteasome it is closed by the interdigitating N-terminal residues of several -type subunits (36). However, because
N-terminal residues 1 to 12 were disordered in the case of the
Thermoplasma proteasome, it is likely that here too the channel is partially occluded. Interestingly, extending the N termini
of the -subunits by six additional His residues causes a reduction
in the rate of protein degradation, suggesting that they impose a
further steric hindrance. Conversely, the degradation of radiolabeled
-casein by turn-His8 proteasomes was accelerated compared with wild-type proteasomes (Table I). Possibly, a His tag
inserted between helix H0 and -strand S1 of the -subunits disrupts the contact of helix H0 with the cleft formed by the -sandwich, relieving the blockage of the -ring channels by the residues N-terminal to H0, and thus facilitates the uptake of protein substrates.
It has been shown previously that proteasomes immobilized on supported
nickel-chelating lipid bilayers via His tags located at the C termini
of their -subunits retain proteolytic activity (37). Surface plasmon
resonance measurements allowed the authors of Ref. 37 to monitor the
loss of mass from the interface because of dissociation of proteolytic
fragments generated from oxidized insulin -chain by proteasomes. In
the experiments described in this article we have measured the activity
of immobilized proteasomes by direct observation of the products of
hydrolysis. We observed only a small reduction of hydrolytic activity
against Suc-LLVY-AMC after immobilization.
Hydrolysis by C-His6 proteasomes bound side-on to
vesicles at low surface coverage was reduced by 8%. When immobilized
on mica-supported bilayers instead, the activity of
C-His6 proteasomes was reduced by 50%. However, several
factors may cause this more pronounced reduction of activity compared
with proteasomes bound to vesicles. Whereas in the latter case, surface
coverage was only 3%, surface coverage was close to complete for the
supported bilayers. Clearly, access to the entrance channels of
proteasomes will be partly blocked at such high coverage. Additionally,
the substrate concentration at the interface will be different for a
flat, extended bilayer and for small vesicles for geometric reasons.
Whereas both entrance/exit channels are obviously accessible in the
C-His6 proteasomes, this is not the case with
N-His6 proteasomes in the end-on orientation. One could
envisage that the channel directed to the lipid layer is less
accessible than the opposite one and a quantification of the activity
of proteasomes in this orientation could provide some insight into the
functional interplay of both channels. In fact, such measurements show
that the activity of the N-His6 proteasomes in end-on
orientation and at low coverage was reduced by only 8%, the same small
reduction as found with C-His6 proteasomes bound
side-on. This observation is consistent with a scenario in which, at
least in the case of small substrates, both substrates and products can
pass through the same pore. However, it is also possible that the
entrance facing the lipid film is not obstructed sufficiently. Complete sealing of this entrance has to be expected for two-dimensional crystals on lipid films, but the corresponding measurements of activity
are very problematic because the percentage of two-dimensional crystals
in the preparation cannot be quantified accurately enough. Utilizing
the His-tag/Ni-chelator system for analysis of the functional interplay
of both channels thus needs more complex
investigations.2
Supported Chelating Lipid Bilayers Are a Versatile Support for
Scanning Probe Microscopy of Proteins--
AFM allows one to visualize
biological structures in action under near physiological conditions
and, in favorable cases, with subnanometer resolution (38). In many
cases, closely packed arrays or two-dimensional crystals have been
found to facilitate attaining high resolution (39-41). One strategy
that has been explored to obtain such arrays is the use of streptavidin
films grown on biotinylated lipid films (42) as a matrix for growing
oriented arrays of biotinylated proteins (43, 44). However, the
self-organization of the underlying two-dimensional crystalline
streptavidin matrix imposes severe steric limitations on the
two-dimensional arrangement of proteins bound to it and, moreover,
these proteins need to be biotinylated in a site-specific manner.
Similar restrictions apply to self-assembled monolayers carrying NTA
groups (45); nevertheless, they have been used successfully to monitor
the transcription of circular, single-stranded DNA templates by AFM
(46). The His tag / metal-chelating lipid layer approach described
here has two important advantages. First, proteins are frequently
expressed with a His tag in the first instance to facilitate their
purification. Second, lipid monolayers are more flexible as a matrix
than self-assembled monolayers; they allow one to influence the lateral
organization via the phase behavior of the chosen lipid. If desired, a
condensed lipid film will keep immobilized molecules separated, whereas
a fluid lipid film will allow lateral protein-protein interactions and
may thus favor dense arrays or even two-dimensional crystallization.
The N-His6 proteasome two-dimensional crystals were
found to be well suited for imaging with the AFM. The appearance of the proteasomes in the crystals did not change with time during repeated scanning. Based on AFM recordings, yeast 20 S proteasomes have been
reported to switch between 2 distinct conformations (47). In contrast,
the AFM images of proteasomes in two-dimensional arrays did not reveal
the existence of different conformational states. The AFM images gave
insight into the vertical packing of crystalline proteasomes, which
cannot be deduced from (two-dimensional) electron micrographs.
Proteasomes were found to be arranged in 2 levels in AFM topographs.
Proteasomes in one row were elevated with respect to proteasomes in
neighboring rows by ~2 nm, half the height of a single proteasomal
ring (Fig. 5). Apparently, proteasomes within the two-dimensional
crystals optimize their packing not only by interlocking laterally, but
also by a vertical shift along their cylinder axis. As a result, one
ring of the proteasome fits into its neighbors waist. To pack in that
manner, proteasomes have to be able to move away from the interface by 2 nm. Assuming that all proteasomes within a crystal are actually bound
to the lipid film requires that their linkage be rather flexible.
Possibly, this flexibility can be attributed to the last, disordered
N-terminal residues 1-12, to which the His tag is appended in
N-His6 proteasomes. Such a flexible linker might prove
useful in many systems and could similarly be realized in the form of a
short stretch of disordered residues leading up to a His tag.
In addition to facilitating high resolution studies, immobilizing
proteins in a predetermined orientation will also enable dynamic
studies of their interaction with substrates or cofactors in real time.
Binding and dissociation of individual GroES cofactors from GroEL
chaperonin molecules on a mica support has been followed with the AFM
(48). GroEL molecules oriented spontaneously with one GroES binding
face directed toward the AFM tip. However, the orientation of
unspecifically adsorbed molecules will often be random or unfavorable.
In contrast, the His tag-nickel-NTA system allows one to bind molecules
in an orientation conducive to further investigation. For example, the
interaction of regulatory particles with proteasomes bound end-on will
be readily accessible to an AFM tip as height fluctuations.
§,
TOP
ABSTRACT
INTRODUCTION
To whom correspondence should be addressed. Tel.:
49-89-8578-2651; Fax: 49-89-8578-2641; E-mail:
guckenbe@biochem.mpg.de.
