It is unclear why some proteins need GroE for correct folding and others do not. Computational studies have shown that obligatory substrates, for example, have lower folding propensities and are more aggregation-prone (Azia et al., 2012). GroE substrates are also more conserved than other proteins and tend to utilize more optimal codons (Noivirt-Brik et al., 2007), although this latter feature is less pronounced in the case of obligatory substrates.
Currently, we are trying to convert GroE substrates into non-substrates and vice versa in order to identify features that render proteins GroE-dependent.