Fractionating Ca++ accumulation into different pools

NMODL and the Channel Builder.
Post Reply
Site Admin
Posts: 5795
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine

Fractionating Ca++ accumulation into different pools

Post by ted »

On 11/19/2004 a NEURON user wrote:
> we are trying to create separate ions for ca++ ions that go through different
> channels, eg. the L-type vs. the N-type channels because we want to couple
> these channels with respective ca-dependent K+ channels (sAHP and fAHP).
> The way we have done it thus far was to create separate mod files where we
> define new ions, i.e. the "can" and the "cal" ion, which we in tern use as input
> to respective calcium pools that measure the "cani" and the "cali" concentrations.
> These concentrations are then used as input for the sAHP and fAHP channels
> respectively.

Presumably ca that enters through different ca channels accumulates
in the local vicinity of different ca-gated k channels. Defining can and cal
ions, as you have done, is a perfectly valid part of a strategy for what
you want to do. You also need to modify your ca accumulation mechanism,
because ca from whatever source is buffered, diffuses, and is pumped out
of the cell.

For the sake of illustration, let's say that your can channels are described
by a mechanism whose NEURON block has a USEION statement like this
Let's say that there are also some other ca channels, but for the moment
we'll ignore cal.

Your can-dependent gk should have a
USEION can READ cani
statement, where cani is the concentration of ca in the vicinity of the
can-dependent k channels.

Your ca accumulation mechanism should contain at least two USEION
statements--one for "ordinary calcium"
and another for ca that entered through can channels

Probably the best way to represent ion accumulation, pump, and diffusion is
with kinetic schemes. The BREAKPOINT block will be

Code: Select all

  SOLVE state METHOD sparse
and there will be a KINETIC block with the name "state" that will look very
much like what you find on page 48 of ... xedref.pdf

It is probably reasonable to assume that the ca that affects can-dependent
k channels is contained in a volume that is a small fraction of the volume of
the outermost shell. It is also probably reasonable to assume that this
volume depends on the number of can channels or can-dependent k
channels, since these two channel species tend to be closely associated
with each other.

So your ca accumulation mechanism needs some extra things.

1. A new STATE cani which is the concentration of ca in this volume.
There will also be two other states to represent buffer in this volume,
which you might call Buf and canBuf (free and ca-complexed buffer).
As a first assumption, this buffer could have identical properties to the
buffer in the rest of the cell.

2. There will also be an ASSIGNED variable canvol that is initialized in
the INITIAL block by a statement something like this
canvol = PI*diam*gbarcan*kcanvol
gbarcan is equal to gbar_can (density of can channels)
kcanvol is related to the volume per cluster of can channels
(of course if you are using a GHK formulation for i_can, we'd be talking
about permeability rather than conductance).

3. A COMPARTMENT statement for cani and its buffer
COMPARTMENT canvol {can Buf canBuf}

4. A "reaction" that represents the effect of ican on cani
~ cani << -ican*PI*diam/(2*FARADAY)

5. Another reaction for buffering of can
~ cani + Buf <-> canBuf (k1buf*canvol, k2buf*canvol)

6. Another "reaction" that represents exchange between the can
compartment and the rest of the cytoplasm. It's probably safe to
assume that this exchange is really between the can compartment
and the outermost shell.
~ ca[0] <-> cani (kx, kx)
where kx is a rate constant that governs the exchange between these
two ca pools. You'll have to experiment with this to see what effect it has.

7. If canvol is very small compared to the volume of the outermost shell,
you might not have to include any pump reaction for the canvol
compartment--especially if the flux between canvol and the shell
is fast compared to the pump flux.

This is only a sketchy outline, and I may have left something out. The way
to develop it is to start with a one compartment model that contains only
two mechanisms--an ordinary calcium channel, and this ca accumulation
mechanism. Test and revise until it works properly when ican is 0. Then
add your can channel and test and revise again.

You'd handle cal in much the same way.

It will probably be helpful to read about initializing ion concentrations--see
chapter 8 of the NEURON book.
Post Reply