Right.
First, you did copy the first equation correctly. In Equation (4.4) Carroll defines
$$a = -\nabla \Phi$$
which is ok, but a bit different than I have seen it before, but makes sense, in light of the way forces and potentials are defined in electromagnetism. Perhaps this is the proper way to define gravitational potential, so that the analogy between electrical charges and gravitational "charges" is clear.
Second, your equation $$F = Mg\nabla\Phi$$ is still wrong. The correct Equation, from the lecture notes Eq. (4.2) is $$F = - M_g \nabla\Phi$$, which is the same as the first equation. For future reference, $$M_g \neq M g$$.
to all who are following, note that $$M_g$$ has units of mass, while $$M g$$ has units of force. This is important, because Reiku's next equation
$$\sum_i \gamma M_i c^2 = I_t = M_g$$
is definitely wrong. No matter WHAT $$I_t$$ is, the left hand side has units of energy, and the right hand side has units of mass.
I think you can continue using it without the minus sign. If not, it's not a remarkable mistake.
And the left hand side must contain units of energy, if assumed $$I_{t}$$ (which is the total inertia) is related by: $$I_{t}=\Delta M_{i}$$ for the total inertial mass. And if the total inertial mass is a measure of how much relativistic mass is contained within the system, $$E= \gamma M_{i}c^{2}$$ says that there is a relativistic mass in inertial matter. It was clear to involve energy on one side if it relates to the total inertia of a system. Of course, the last equation is a small modification of the normal $$E=\gamma Mc^{2}$$, and i have involved it to mean inertial systems as well. If you don't like this, then its pure semantics.