Atomic Force Microscopy of Polyhydroxyalkanoate Inclusions from Ralstonia eutropha Polyhydroxyalkano

1
Atomic Force Microscopy of Polyhydroxyalkanoate
Inclusions from Ralstonia eutrophaPolyhydroxyalk
anoates (PHAs) are intracellular storage polymers
that are made by some bacteria when environmental
levels of carbon are high, but levels of another
essential element (such as nitrogen, sulfur, etc)
are low. This allows the bacteria to store
carbon as a reserve. When the carbon is needed,
the intacellular polymer is degraded. PHAs are
packaged in inclusions, about which we know very
little. We know that they have a major
structural protein ,PhaP, as well as the
polymerizing enzyme, PhaC, at their surface.
PhaP is present in high levels and may comprise
as much as 5 of the cell protein under
PHA-accumulating conditions. The arrangement of
PhaP is not known, but some researchers have
suggested that it it a lattice, based on electron
micrographic studies. At least two groups have
found a 4-nm thick boundary layer. Once of the
groups presented arguments as to why this is most
likely a lipid monolayer. The purpose of our
atomic force microscopy studies is to enhance out
understanding of PHA inclusions, particularly
with respect to the putative boundary layer and
any protein structures that may be found on the
surface of the inclusion
2
  To image PHA inclusions we use a very simple
procedure in which 1-5 microliters of cells are
added to 5 ml of Tris-NaCl.  The cells are then
sonicated for 3 or 4 minutes with a probe
sonicator and immediately added to 10 ml of
Tris-NaCl.  By using a very small amount of cells
and diluting them immediately upon sonication,
the amount of cellular debris adhering to the
inclusions is minimized.  A small portion
(100-500 microliters) of the diluted cells is
filtered onto polycarbonate filters (200-nm pore
size) and the filter is attached to a metal puck
and the inclusions are immediately imaged.  From
the time of sonication to the acquisition of the
first image is no more than 90 minutes. 
Therefore, the inclusions are quite fresh. We
generally image the inclusions in both height
imaging and phase imaging tapping modes.  The
phase image tends to give a more realistic
looking image, but can be prone to artifacts. 
Therefore the height image is used as a control
to check for structures that are seen in the
phase image.  Height image is also very useful in
cross-section analysis for measuring the
thickness, depth, and width of structures.  Phase
image is a relative measure of the elasticity of
the surface.  Harder objects are distinguished
from softer objects.  We have found that the
quality of the image is directly linked to the
quality of the cantilever tip.
3
Below is an example of the inclusions situated on
the filter, showing that they are not
contaminated with debris, etc.
4
Using these techniques, we generally are able to
image two types of inclusions.  One is more ovoid
and has a "rougher" appearance.  The other is
rounder, looks more like a mushroom, is not as
tall, and has a smoother surface.  The large
majority of inclusions are of the rough type. 
However, we can increase the smooth inclusion
percentage by using Tris-EDTA instead of
Tris-NaCl, sonicating longer, or using cells that
have been stored in the refrigerator for a
while.  These two types of cells can be seen
below.
5
The image on the left is a height image and the
image on the right is a phase image.  The terrace
effect at the perimeter of the inclusions in the
phase image is an artifact.  It higher resolution
pictures are taken of the smooth and rough
inclusions, distinct differences are seen.
6
The smooth inclusions have very little surface
structure, other than a series of parallel arrays
that can be imaged.  The spacing of these
parallel arrays is about 7-nm between each
parallel line.  They are seen on the surface of
the inclusions as if they were painted there by
the strokes of a paintbrush.  Once series of
parallel lines will go for a while, and then be
intersected by another series of parallel lines
that a traveling in another direction.  A
close-up of this can be seen below. 
7
The AFM can also image this in surface mode,
which looks something like this
8
This image is taken at a larger format and also
shows the parallel lines on the surface of a
smooth inclusion.
9
We can use the cross-sectional analysis of the
AFM to measure the distance between the linear
strands. The cursors are set a little wide on
this one (9.638 nm).
10
The rough-surfaced inclusions look quite a bit
different. Sometimes we can image the parallel
arrays, but not always. Instead we see
structures at the surface that are somewhat
reminiscent of porins in that they have central
pores. These can be seen scattered over the
surface of the inclusion. Scale is in
nanometers.
11
Here is a closeup of the pores. The have a
central pore about 15 nm in diameter, and a
collar that is also about 15 nm wide. The entire
complex is about 35-40 nm wide.
12
This is a cross-sectional analysis of the pore,
showing the width of the pore and that the collar
sticks above the surface of the inclusion. Note
that the cross-section line crosses several of
the pore structures
13
Occasionally, we are able to image the parallel
array under the pore-like structures (arrow 1).
It appears that these pore-like structures are
connected via a network of linear structures
(arrow 2). These linear structures are enhanced
as the radius of curvature increase, because of
tip convolution (arrow 3).
3
2
1
14
We have hypothesized that the surface proteins
form an interlacing network that covers the
inclusion. At the intersection of these network
proteins are protein structures that are likely
to be responsible for ingress of substrates and
egress of depolymerized products. The central
pore of the pore-like structures would be a
central site for these events. We also propose a
lipid monolayer boundary layer, but are unsure at
this time whether it covers the protein network,
or whether the network is underneath the boundary
layer. We have found proof of this network using
a very light tapping force combined with a slow
rastering speed of the cantilever tip.
15
This is an example of the network that we have
imaged. Because of the very light tapping force
used the resolution is not as good as in some
prior images, but it clearly shows the network,
with the central pores (arrows) being seen as
some of the intersections of the network. The
parallel polymer strands can also barely be seen
beneath the network.
16
We have also attempted to image the putative
boundary layer that covers the membrane. We have
done this by using agents that would disrupt the
boundary layer, with the idea of imaging the
disruption as it occurs. In our first
experiments, we have used sodium lauryl sulfate
at concentrations ranging between 0.05 and 0.1
(w/v). In the early stages of deterioration,
holes are made in the boundary layer (arrows).
Not the pore at the bottom center of the image.
17
As the deterioration progresses, the holes get
larger. Debris starts to appear on the surface
of the inclusion. Note that pore can still be
seen
Debris?
Holes
Pores
18
As the holes get larger, the surface of the
inclusion begins to look like Swiss cheese.
19
Toward the end of the deterioration the only
thing that is left of the boundary layer is
stringy remnants.
20
These must be quite sticky because they can be
seen connecting inclusions that are next to each
other. Also, 35-nm wide spherical objects can be
seen littering the surface of the filter around
the inclusion, but not in other areas of filter,
indicated that they have been washed off the
inclusion. In the inclusions in the lower left,
the parallel strands can still be seen.
21
Farther away view showing spherical objects and
connections between inclusions
22
Cross-sectional analysis of the deteriorated
boundary layer indicates that it is 4 nm in
thickness.
23
We have also been able to measure the boundary
layer using untreated rough inclusions that have
been dried. These form cracks, which can be used
to measure the boundary thickness.
24
These and other AFM images have allowed us to
construct a model for PHA inclusions that is more
advanced than previous models. In our model
there is a boundary layer that is 4-nm thick,
that is a lipid monolayer. In addition, there is
a web of linear protein structures that
traverse the surface of the boundary layer. At
the intersection of some of these linear
structures can be found protein complexes that
have a central pore. This pore is likely the
site of ingress for substrates and egress for
depolymerized products through the boundary
layer. It is also possibly a site for attachment
of synthesizing and depolymerizing enzymes. The
spatial relationship of the boundary layer and
the protein web has not been determined from the
AFM images. However, it appears that the protein
network may be located on the surface of the
boundary layer. This would explain why
inclusions that have lost their boundary layer
(smooth inclusions) do not have the protein
network. The protein/boundary layer complex
appears to be easily removed from the surface of
the inclusion, which may explain why it has not
been seen using electron micrographic techniques.
25
Here are some images of miscellaneous
structures. Cells with inclusions inside.
26
An image showing what inclusions look like when
they are heavily contaminated with proteins
27
Remnants of cells.