On the specificity of starch debranching enzymes