You could do phenol:chloroform:isoamyl alcohol extraction followed by ethanol precipitation or add proteinase K or run it on a denaturing acrylamide gel and cut out your ligated DNA band.  Only the gel purification would separate ligated from unligated DNA.

Thanks for suggestions. Is it possible to use phenol: chloroform method to handle ~100 ng DNA? What is the typical recovery rate? The native PAGE gel shows that ligation is very efficient and over 95% percent  .  

Is it necessary to remove the ligase? What is the down side of leaving it in the sample?

The downside is to form complex for sequencing. It is kind of sensitive.