Abstract To improve the specific production rate of Rhizopus oryzae lipase (ROL) in Pichia pastoris, a protein that triggers the unfolded protein response in P. pastoris, the effect of sorbitol/methanol mixed substrates was tested in batch and fed-batch cultures. Remarkably, a different substrate consumption behaviour was observed depending on the host's phenotype (Mut + or Mut s) in batch cultures: when the methanol assimilation capacity is genetically reduced (Mut s phenotype), both substrates were consumed simultaneously, allowing not only a higher specific growth rate but also higher lipase levels (8.7-fold) compared to those obtained by cells growing on methanol as a sole carbon source in batch culture. This effect was not observed in Mut + phenotype, where the two substrates were consumed sequentially and the levels of heterologous product were only slightly higher (1.7-fold). A mixed substrate strategy was also applied to a Mut s fed-batch culture at a low methanol concentration set-point (0.5 g l −1). This resulted in a 2.2-fold increase in the heterologous protein level achieved, compared with the methanol-only feeding strategy. In addition, sorbitol co-feeding permitted the achievement of higher specific growth rates, and avoided the drastic decrease of the specific production rate observed after the start of the induction phase when methanol was used as sole carbon source This resulted in a significant increase in the overall bioprocess volumetric productivity (2.2-fold) and specific productivity (1.7-fold). Moreover, whereas increased ROL gene dosage in Mut s strains have been previously reported to be deleterious for P. pastoris cells growing on methanol, sorbitol co-feeding allowed for sustained cell growth and lipase production.