There is a sequence difference right before the diglycine between the different SUMOs in human (QTGG) and yeast (QIGG).
Do you have any hands-on experience with cross-specificities between the human and yeast SUMO-system, or perhaps is only efficiency affected?
1) Does yeast Ulp1 (and 2) protease cleave human SUMO-1, 2 and 3 (some papers reports activity)?
2) Does human SenP proteases (several) cleave yeast smt3?
all eukaryotic cells seem to have enough SUMO protease activity to cleave the fusion proteins. The incubation time inside the cell is definitely long enough even if the activity is 1% of the specific acitivity.
SenP2 is more specific for Sumo 2 & 3 and SenP1 for Sumo1 in human cells, the difference is not dramatic if you cleave e.g. Sumo1 and Sumo3 with the catalytic domain of SenP2 (i.e. our construct). In our hands the SenP2 cleaves the Sumo1-fusion a bit less efficient compared to a Sumo3-fusion. This is not really an issue since you see the difference in the activity assay when incubating in ratios >1:1000, in real preparative experiments you usually have more of the enzyme or incubate a little longer and will see no major difference.
The only benefit of human Sumo1 might be that it is 98% identical to the rabbit Sumo1 and could be used as fusion partner for immunisation in rabbits. I don't know if it would cause a strong immune response in the rabbit against Sumo1, I didn't want to sacrifice a rabbit for finding that out, so the high conservation might not even be a real advantage.
For more details on specificities, have a look here: http://www.sciencedirect.com/science/article/pii/S0969212604002436