Interpretable User Satisfaction Estimation for Conversational Systems with Large Language Models