Von Neumann outlined two rules.

1) Self-reproduction

2) Simulation of a Turing machine

I have seen other summaries of this description as replacing step 2 with "mutation" or using simulation of a Turing machine as being synonymous with mutation

I am not a fan of this analysis, I think two rules is too many. Really there should be only one rule: no rules.

Plus, I think the second rule necessarily implies the first, since the theory of the universal Turing machine sufficiently allows for self-reproduction which later outlined the sketches for cellular automata and subsequent models such as Conway's Game of Life.