Redundancy analysis

Hi,
I'm trying to run a RDA but I have some doubts....
Should I check for the collinearity within both explanatory and predicted variables and between explanatory and predicted variables ? Should I centre both explanatory and predicted variables?
Many thanks