recently read a couple of articles claiming that DOLS can also be applied as a co integrating equation even when there's a mixture of I(1 ) dependent variable and I(0) and I(1) independent variables.

Reading some part of Stock and Watson (1993),I excerpted an assertion on page 784 " In the case with single cointegrating vector,one simply regress one of the variable (I.e I(1) dependent variable) onto the contemporaneous level of the remaining variables, lead and lags of their first difference and constant using either OLS or GLS, the resulting Dynamic OLS is asymptotically equivalent to the Johansen/Ahn-Reisel estimator". Does this assertion validate the claims in these papers...?