T4 ligase doesn't like to ligate ssDNA to dsDNA so sticky-to-blunt ligations are extraordinarily inefficient, so basically that won't work on its own. Traditionally what would be done in this case is to blunt the sticky ends with a polymerase (usually T4 polymerase) and do a blunt end ligation. In this case you will have to do some screening to find clones with the desired orientation since it can ligate in either direction. Blunt end ligations are also prone to having high background so you might need to screen a lot of colonies. However, if you don't already have a blunting kit laying around and would need to order it anyway, I would recommend trying something else, like NEBuilder/Gibson Assembly or add some different compatible restriction sites with PCR. You'll have to design primers for those but in the end it will probably be a lot less work.

Hi Alexandra, thank you very much for your answer. Sorry if I was not clear. Both my vector and insert are dsDNA. The idea od making the sticky ends to blunt looks very good. I might try that. The issue is I am trying to do a homologous recombination to delete a gene and the plasmid I use has very specific restriction site and the protocol has been optimised, hence I wanted to just stick in the insert in the most efficient way possible so that my downstream application would not require much optimization.

There should be no problem in performing the ligation of a ScaI-XhoI fragment into a vector cut with SmaI and SalI. Both ScaI and SmaI yield compatible blunt ends, and so do SalI an XhoI, which yield the same 5' sticky end with the sequence TCGA. This will force the orientation of the insert relative to the vector. Don't worry about bluntifying your vector and fragment or adding new restriction sites, unless the forced orientation of the fragment into the vector doesn't suit you for some reason

Pierre Béguin is correct that you should have no problem with this ligation and don't need to do anything special. Just be aware that once ligated, none of the 4 restriction enzymes will cut the hybrid sites any longer, so you need to use other flanking restriction sites for analysis (or PCR).