Reasoning Backwards Through a Markov Chain

A server is in one of two states each day, Up or Down, and the daily transitions form a Markov chain with

$P = \begin{pmatrix} P(U \to U) & P(U \to D) \\ P(D \to U) & P(D \to D) \end{pmatrix} = \begin{pmatrix} 0.75 & 0.25 \\ 0.25 & 0.75 \end{pmatrix}$

On Monday the server is known with certainty to be Up.

You are later told only that on Wednesday the server was Up. Given that, what is the probability the server was Down on Tuesday?