SSE(Yhat)=SSE(X1,X2)???

#1
I've been attempting at this question (along with other friends) for some time now. However, we either don't understand the question (have been to office hours and emailed), or simply can't see it. We understand it intuitively but can't get the calculations to work:

The question asks to show SSE(Yhat)=SSE(X1,X2) where

Y regressing on Yhat (SLR): r^2 = 1-SSE(Yhat)/SST
Y regressing on X1 & X2 (MLR): R^2 = 1-SSE(X1,X2)/SST

I'd prefer to do it through matrix algebra (supposedly it's easier) but can't seem to understand it:

SSE(X1,X2) = (Y-Xb)'(Y-Xb)?
SSE(Yhat) = (Y-Yhat)'(Y-Yhat)?

Since Yhat=Xb, that's why the SSE's are equal? I have no idea...

At this point, any help will do. Please, if you can!! :(