Firstly, this paper presents an overview of DoD code  against progressive collapse and points out the shortcomings of the current design approaches. After that, seven experimental tests of common types of bolted steel beam-column joints under a middle-column removal scenario are presented. This study provides the behaviour and failure modes of different types of connections, including their resistances and rotational capacities in catenary action. The test results indicate that the web cleat connection has the best performance in the development of catenary action. The flush end plate, fin plate and top and seat with web angle (TSWA) connections could also deform in a ductile manner and develop catenary action prior to failure. Numerical simulations have also been conducted. Both static and explicit dynamic solvers were employed to overcome problems of non-convergence, contact, large deformation and fracture simulations. It is demonstrated that the finite element analyses give reasonable accuracy compared to the test results. In addition, an extensive parametric study was undertaken using these validated models to obtain the rotation capacities of various types of connections under catenary action. Finally, some practical design implications have been drawn up from the experimental tests and the parametric study. A new tying resistance expression is proposed to consider the effect of large rotation. If large rotation capacity is not considered in the design stage, the joints with poor rotation capacities would fail to achieve the design tying resistances. In addition, four new connection acceptance criteria of rotation capacities have been proposed to incorporate catenary action under a middle column removal scenario. The work shows that current acceptance criteria of rotation capacities for steel joints such as web cleat, fin plate, flush end plate and TSWA connections, are probably too conservative as they only consider pure flexural resistance.