Change coefficient of an assembled matrix

Hi all,

I am trying to implement a well model in a program for biphasic flow in porous media.
I assemble the matrix for the potential (pressure) problem as below:

 def Galerkin(self, delfineVar, parameter):
        """Assembles the elliptic problem in a Galerkin variational form with Dolfin."""

        # Get data from case initialization
        .
        .
        .

        # Create function space
        V = FunctionSpace(mesh, "CG", order)

        # Define distributed source terms (not wells)
        f = Source()

        # Define Neumman boundary conditions (vel.n=0 on borders)
        g = Expression("0.0")

        # Define variational problem
        v = TestFunction(V)
        u = TrialFunction(V)

        a = inner(grad(v), K*mob*grad(u))*dx
        L = v*f*dx - g*v*ds

        # Assemble matrices and vectors
        A, rhs = assemble_system(a, L)

After that, I would like to change the coefficient of the 'A' matrix relative to a specific vertex representing the well. I already managed to find the index of this vertex, so everything I would like to do is something like:
A[index][index] += value

For example, if my vertex index is 3, in finite differences it would look like:
-- --
|1 1 0 0 0|
|1 -2 -1 0 0|
|0 1 (-2+value) 1 0|
|0 0 1 -2 1|
|0 0 0 1 1|
-- --

How could I do that?? I am using uBlas as the linear algebra backend.

Thank you in advance,
Bruno