ERM model for Endogenous treatment effect with nointeract option

[ram shrestha]

I am trying estimate treatment effect models using ERM command in Stata 15. However, it has been shown that
etregress y x, treat(t1 = x z) is equivalent to eregress y x, entreat(t1 = x z, nointeract)

I confuse the role nointeract term. What will be the difference in result if I do not include nointeract term ?

I am using
estat teffects command to estimate average treatment effects.