Interdiction for the direct current optimal power flow problem is commonly modeled in two stages, where an attacker disables components in the first stage and an operator makes remedial adjustments in the second stage. The resulting bilevel mathematical program is often solved via transformation into a single-level mixed-integer program using duality. A drawback of this approach is that the single-level program often has big-M values that must be chosen large enough to ensure optimality. We present a single-level reformulation along with theoretical bounds on big-M values that allows for interdiction of branches, buses, and generators. Additionally, we provide a general approach to bounding big-M values for combinatorial problems on graphs, leveraging techniques from spectral graph theory. In doing so, we derive several tight lower bounds on the minimal eigenvalues of grounded Laplacians. We show computational results on standard test problems using the theoretical bounds and on real power systems using heuristic bounds. Finally, we demonstrate that our analysis applies to other interdiction problems, such as the decoupled linear power flow problem and a second-order cone formulation.