I think I got it: he added 5x to the denominator of the first term and to the numerator of the second term, which both cancel out so is the same as multiplying by 1. He then multiplied the first term by 3/3, factored out 3/5, and was left with an easily solvable limit?