The first step of planet formation is the collisional growth of submicron-sized dust grains to kilometer-sized planetesimals. It is known that there are several "barriers'' which dust should overcome to form planetesimals. One is the "fragmentation barrier", that is, dust aggregates are likely to fragment upon collision at high speeds. This is particularly likely when magnetorotational instability (MRI) is active and strong turbulence is developed in the disks. Interestingly, however, it is also known that dust is not only affected by but also affects the MRI, because dust grains can efficiently absorb ionized gas particles that are needed for the MRI to operate. In this talk, Satoshi will present an empirical basis for modeling the coevolution of dust and MRI-driven turbulence (Okuzumi & Hirose, submitted) and show how the coevolution will alter the outcome of dust growth.