Venice Italian Treebank
VIT is a treebank which includes 275,000 tokens composed by different representation layers: the first is of constituent tree strctures with tagged words; the second is derived from the first, and is composed by dependency structures organized in 8 columns which describe the morphological and semantic features, the lemma, the functional marker which differentiates topic/focus; other layers refer to the orthograohical version with multiwords and in canonical form.
The full treebank is available against payment, while a smaller version can be downloaded for free.
Here is a PDF presentation and the link to the website of the project for Venice Italian Treebank 
