What should I do to lower down the fa_sol scores?

Greetings,

I ran a design protocol and I find my results score has high fa_sol scores like 250 around. What should I do to lower down the fa_sol scores? 

I used ResidueSelector to separate my alpha-helixs into three parts: buried, semi-buried and solvent exposure. And use resfiles to assign the amino acids, buried------hydrophobic, semi-buried------less hydrophobic, solvent exposure-------hydrophilic. 

And I am wondering if the fa_sol score of a model is too high, will it be insoluble in water after experimental expression?

Any suggestions will help!