A structural look at
MscS in comparison to MscL.
Jacob E. Vick
Biophysics of Membrane Proteins
Unlike Eucharea, prokarea and archea are not readily able to control their
environment. Thus, single cells organisms must develop way to quickly and effectively
react to changes to their surroundings. Besides such things as nutrient supply, pH, light
and/or temperature, they must be able to react to osmotic stress. Using patch-clamp
techniques, two distinct mechanosensitive channels have been identified from
Escherichia Coli (1,2). These two channels were given the names Mechanosensitive
channel of small conductance (MscS) and Mechanosensitive channel of large
conductance (MscL) because they had conductance of 1nS and 3nS respectively. Both of
these channels opened as a result of stress on the strength of the cell membrane. As the
membrane was stretched out more, the channels would open. MscS opens at one half the
cell tension of MscL, which opens right before the tension level associated with cell
rupture. MscS has also been shown to be sensitive to cell voltage, but that voltage
dependence has not been shown in MscL.
Recently, both MscS from E. coli (3) and MscL from Mycobacterium tuberculosis
(4) have been crystallized. The Tb MscL is a homolog of MscL from E. coli. These are
the first two mechanosensitive channels to have their structure solved and the first
voltage sensitive channel in the case for MscS. By comparing these channels, a finer
understanding of both the nature of mechanosensitivity and voltage dependence should
be gained. Before comparing these two, we must discuss them individually.
Trans - membrane
Middle – beta domain
C - Terminal domain
fig 1. Whole surface representation of MscS
Trans - membrane
fig 2. Whole surface representation of MscL
MscS is a much larger protein then MscL; this is evidenced by the fact that their
membrane domains should be approximately the same size. MscL is almost entirely
located with in the membrane, while MscS has a large portion of its structure in the
cytoplasm. It is important to note that the first 26 amino acids of the N-terminal end of
MscS did not get resolved during the structural determination. This difference is also
noted in the fact that MscS is a homoheptamer while MscL is homopentamer. Each
monomer of MscS consists of 286 amino acids while MscL monomers are 136 amino
acids, making MscS much larger then MscL. There is almost no sequence homology
between these two monomers, making these possibly two completely different examples
of how a pore can open in response to osmotic stress. Sequence alignments, not shown,
have indicated that MscS and MscL are highly conserved across species. In general, we
will show that the major similarities between MscS and MscL are that they respond to
MscL is a fairly simple protein in comparison to MscS. Since it doesn’t open
until the membrane is about to rupture, it most likely behaves as a stop gap between
extreme osmotic shock and cell death. When it opens, it exhibits a huge non selective
conductance of ions from the interior to the exterior environment. This conductance has
been shown to be approximately 3nS. The reason for the non conductance can be
understood by looking at the two trans-membrane helices that are in the MscL pentamer.
Fig 3. Hydrophobic look of MscL. White balls indicate hydrophobic, green is polar, while blue is positive
and red equals negative charges. This structure was crystallized in the closed state
MscL is dominated by hydrophobic groups, where the majority of the pore hydrophobic
or polar, thus very little specificity is conferred when the channel is opened. The small
size and observed conductance magnitude and lack of selectivity indicates that MscL
should have no major filtering mechanism at all. This is a good trade off to lose some
important small molecules in order to keep the cell from rupturing and dying, implying
that MscL is the last ditch effort by the organism to respond to osmotic shock. Some
groups that should be noted are the positively charged Arganines that are located in at the
ends of the transmembrane helices on the exterior of the protein.
A less drastic response to osmotic shock is the job of MscS. This channel opens at
approximately %50 the shear stress as does MscL. MscS is fundamentally different from
MscL, in that it is opening can be tuned by voltage. As the cell is depolarized, or
becomes more positive on the interior, the channel is more likely to open. This
corresponds to an “e” fold increase for every +15 mV change in the cell voltage (1, 5).
This also corresponds to a movement of approximately 1.7 “+” charges of the protein
moving towards the extracellular side across the membrane. Thus, as more charged
molecules are flowing in then is the norm, the MscS is more likely to open. Also to be
noted is that MscS tends to prefer anions, which makes sense. A cell normally has a
overall negative charge on the interior, if the cell is becoming depolarized or more
positive, the Cell will want to preferable remove the anions, so that it can return to its
negatively charged interior. MscS can be divided into three distinct regions. Some
insight into MscS can possibly by looking at the three distinct regions.
Trans-membrane domain residues 29-127
A B ASP 62
fig 4. Transmembrane region of MscS. A is a look from the extracellular side of MscS in the
transmembrane domain, through the pore. B is a look at a single chain of MscS. The Amino Acids that
comprise the channel are the right edge of B, where ARG88 and GLN 112 are hidden. An arm of a MscS
monomer is facing outwards in B. The dashed line indicated the membrane. MscS was crystallized in an
This region seems to be the main business end of the MscS. Just like in MscL,
this region is occupied by a lot of hydrophobic amino acids. The pore is almost totally
hydrophobic with the exception of ARG 88 and GLN 112. Both groups present –NH3
groups to the pore surface, and it could be these positively charged groups that confer the
preference for anions moving through the channel. MscS is different from MscL in that
that MscS has three transmembrane(TM) helices. If both MscS and MscL react directly to
membrane stretching, then the difference in the amount of TMs could infer a difference
in opening mechanisms. The Most notable thing about MscS is outer arms of the TM
helices. The membrane/solvent interface is lined with ARG59, LYS 60, SER 58 and ASP
62, while the protein actually presents ARG 47, 54, 74 and ASP 67 directly to the lipids
of the membrane. Also located here are the polar groups ANS 50, 53 and TYR 75 (not
labeled). It is possible that the membrane/solvent (M/S) amino acids confer
mechanosensitivity while the lipid presented (LP) amino acids allow for the voltage
gating. There has to be a reason to expose this many charged groups directly to the
lipids, because there is a huge energetic penalty. The area between the outer arm and the
amino acids that comprise the pore are almost exclusively hydrophobic (not shown).
fig 5 Whole TM domain of MscS. The dashed line is the membrane
A look at the whole TM domain reveals that these LPs are not buried in monomer
Middle-Beta domain residues 132 to 177
The next region of the MscS is the Middle-Beta domain, named for its domination
by 5 beta strands packed tightly together in each monomer. MscL has no region that
even compares to this.
fig 6. Middle-Beta domain of MscS. A is a sideways view of the TM domain in comparison to the Middle-
Beta domain. B is a view looking down the center.
The Beta-domain is the first intracellular region of MscS. This region has a giant
interior region that is most likely too large to be either a gate or a selectivity filter. This
region may be necessary to give stability to the heptamer, which is important to have in
order to possible offset the LP amino acids of the TM domain.
Carboxy terminal domain residues 188 to 265
Fig 7 Carboxy Terminal domain of MscS. A side shot. B. center shot looking from intracellular out
Just the like the Beta-Middle domain, there is not much opportunity for selectivity
or pore formation in the carboxy terminal (CT) domain of MscS. That is the majority of
the CT domain is a alpha beta barrel type structure with a huge internal open space. But,
the carboxy terminal is ended with a very small Beta barrel, much like the intracellular
portion of MscL. A comparison of the smallest distance in the middle of the TM domain
and the CT domain shows that the CT domain actually has the smallest pore width.
Fig 8 Pore Widths. A shows leucine 109 of the TM domain, the smallest width is 10.56 B illustrates that
Met 273 actually has a width of 8.51.
This small width located with in the CT domain may play an important role in filtering
out larger molecules, allowing only small anions and cations to enter the large area
formed by the rest of the CT domain and middle-Beta domains. Together, the CT and
middle-beta domains may act as a place for regulatory proteins, but such possible
proteins have not been identified yet.
Both MscS and MscL display mechanosensitivity, but they have unique TM
domains. The most common feature besides the use of helices is that both show the use or
ARGs on the border between the membrane and the cytoplasim. It is reasonable to
imagine that these ARGs may be the key to the mechanosensitivity. It is possible that an
ARG, which has a large positively charged sidechain, could associate with head groups
of membrane lipids. MscS also adds to this by the use of a LYS and ASP adjacent to the
ARG. As those lipids are spread thinner do to osmotic shock stretching the membrane,
the arganines could be pulled apart from the center of the channel. The groups could act
as a handle to open the channels.
Fig 9. Cartoon representation os MscL and the TM domain of MscS
The difference is what those handles are attached too. In the case of MscL, it is a
2 helix TM domain. This has been theorized to open much like a camera lens where as
the ARG move out word, the helices rotate against each other causing the pore to open
(4). MscS has a much different possible mechanism, do to the three TM helices. Most
likely, the helix comprising the pore of MscS doesn’t move, but as The other two helices
open do to mechanic stress effecting the M/S amino acids, the extracellular side is opened
or closed. This can not be totally understood, since the first 26 amino acids were not
stable enough to resolve, but if these groups are highly mobile and attached directly the
outer two helices, then they could act as a way to close the pore.
MscS has a reactivity to changes in voltage. This is most likely a result of the LP
amino acids of the TM domain. These charge groups could react to depolarization with
in the cell and shift the helices to a confirmation that is more likely to open the pore (3).
This may involve more then just the ARGS mention by Rees et al. Mutagenesis studies to
see the effect of voltage gating in the region would probably resolve this question. MscL
has no sensitivity to voltage, which is reasonable to understand if the LP AAs are the
cause in MscS, because MscL has no LP AAs of a charged nature.
Implications for function
As patch clamp tests have shown, MscS and MscL behave quite differently. MscL
acts as a last ditch stop gap for cell rupture, while MscS acts a tunable response to bad
conditions. Since MscL takes a lot of stress to open, the presence of only two ARGs to
act as handles is easy to live with. Therefore it would take a lot of pulling the ARGs in
order to rotate the TM domain helices against each other. MscS is a much more
complicated protein; it is not only mechanosensitive, but voltage sensitive. Since it takes
less stress to open, it has a much large group M/S amino acids to react to lesser stress,
while it can add to chance of open by reacting to a change in voltage, which of course
means that there has been an influx of positively charged ions into the cell, and hence
will soon be followed water per osmosis. This all adds up to more likely opening event
Selectivity is not readily apparent in either MscL or MscS. This is understandable
for MscL, since it is the last ditch effort, selection gets thrown out in order to keep the
cell from rupturing. This is also why the pore of MscL is predicted to be hydrophobic in
the open position (4). MscS on the other hand does show some selectivity. MscS tend to
only allow the smallest of charged groups to exit the cell. Large groups like NTPs do not
flow out. The exclusion of the larger charged groups is most likely explained by MET
273, which is a smaller opening then the pore it self, allowing only small molecules to
flow into the area formed by the middle-Beta domain and the carboxy terminal domain.
MscS also seems to prefer to let anions through over cations. It has been reason by Rees,
that this is caused by ARG88, but GLN112 actually presents a smaller diameter to the
pore and only presents its –NH3 groups to the pore. Therefore, it may be the GLN112
and ARG88 working to slow down the cations that allows MscS to prefer anions. This
preference for the anions goes hand in hand with the voltage sensitivity. As a cell is
depolarized, anions move into the cell, so opening in reaction to the voltage and osmotic
stress, u would want to get more anions out so, u could be helping to return the cell to its
proper electronic state.
1. B. Martinac, M. Buechner, A. Delcour, J. Adler and C. Kung, Proc. Natl. Acad.
Sci. U.S.A 84, 2297 (1987).
2. S. Sukharev, B. Martinac, V. Arshavsky, and C. Kung, Biophysical Jour. 65, 177
3. R. Bass, P. Strop, M Barclay and D. Rees, Science. 298, 1582 (2002)
4. G. Chang, R. Spencer, A Lee, M Barclay and D Rees, Science. 282, 2220 (1998)
5. C. cui, D. O. Smith, J. Adler, J. Membr. Biol. 144, 31 (1995)