Multilevel mediation in gsem using interaction terms.

I am running a MLM with mediators using gsem. I've been able to use the SEM Builder to set up my models, but I have an interaction term for "X." The interaction is AgeXStress. However I am unsure how to model Age and Stress and AgeXStress in this path analysis. Here is my suggested model, which happens to provide the same results whether or not I include the singular variables, Age and Stress.

gsem (gender) (WPstress)(genderXWPstress -> WPPos, ) (genderXWPstress -> AUC, ) ///
(WPPos -> WPPneg, ) (WPPneg -> AUC, ) (M1[NASID] -> WPPos, ) ///
(M2[NASID] -> AUC, ), covstruct(_lexogenous, diagonal) latent(M1 M2 ) ///
nocapslatent

These data are for 161 older adults who provided cortisol data over a 4 day period. They also provided diary entries on those four days that included measures of stress and positive and negative affect. I am trying to ascertain stress response differences between genders on days when people reported stressors (GenderXWPstress). I have developed a model where cortisol outcomes on stressor are mediated by positive and negative affect (higher positive affect --> lower negative affect --> lower cortisol). That's a little background on this model, but my main concern is where to include the "gender" and "stress" variables in this model so I can estimate the interaction term. Thanks so much for anyone who may have an answer.