One-loop correction to non-leptonic weak interaction vertices In chapter 18, fig 18.3 Peskin & Schroeder work out the QCD corrections to the strength of the six nonleptonic weak interaction vertices 
The third and fourth diagrams give a contribution
$$
 i\mathcal{M}= \int \frac{d^4k}{(2\pi)^4}(ig)^2\frac{-i}{k^2}\left(\bar{d}_L \gamma^{\nu}t^a\frac{i\not k}{k^2}\gamma^{\mu}u_L\right) \left(\bar{u}_L \color{red}{\gamma_{\nu}t^a \frac{-i \not{k}}{k^2}\gamma_{\mu}}s_L\right) 
\tag{18.35}$$
while the fifth and sixth give
$$
 i\mathcal{M}= \int \frac{d^4k}{(2\pi)^4}(ig)^2\frac{-i}{k^2}\left(\bar{d}_L \gamma^{\nu}t^a\frac{i\not k}{k^2}\gamma^{\mu}u_L\right) \left(\bar{u}_L \color{red}{ \gamma_{\mu}t^a \frac{i \not{k}}{k^2}\gamma_{\nu}}s_L\right) \;.
\tag{18.45}$$
I put in red the difference between the two. The second contribution makes sense to me, but I don't understand the first. Why does a minus sign appear here, and why $\gamma_{\nu}$ and $\gamma_{\mu}$ are inverted? Do these two contributions have a different physical meaning?
Cheers!
 A: First thing that you have do to when writing down a Feynman diagram is to figure out your momenta. The computation here is performed in assumption that the momentum $k$ running in the loop is much larger than any other momentum - check the discussion around Eq (18.10). You only need to pay attention to the loop momentum $k$. I've sketched the momentum direction below:

The expression in the first curly brackets is the same for both diagrams:
$$\left(\bar{d}\gamma^\nu t^a \frac{i\not k}{k^2}\gamma^\mu u\right)$$
You build it by going "backwards" along the upper fermionic $ud$ line. Here you encounter the Dirac propagator with momentum $k$ since, so you add the $\frac{i\not k}{k^2}$ factor.
The second curly bracket is built it by going "backwards" along the lower fermionic $su$ line. The contribution is different for two diagrams:

*

*on the left  you encounter $\gamma_\nu$, the propagator with reversed momentum $-k$ and then $\gamma_\mu$


*on the right you encounter $\gamma_\mu$, the propagator with positive momentum $k$ and then $\gamma_\nu$
That explains both the sign difference and the ordering of $\gamma$ matrices.
