The basic operation in such a network is the systematic gathering (with or without in-network aggregation) and transmitting of sensed data to a base station for further processing. A key challenging question in WSNs is to schedule nodes’ activities to reduce energy consumption. In this paper, we focus on designing energy efficient protocols for low-datarate WSNs, where sensors consume different energy in different radio states (transmitting, receiving, listening, sleeping, and being idle) and also consume energy for state-transition. We use TDMA as the MAC layer protocol and schedule the sensor nodes with consecutive time-slots at different radio states while reducing the number of state transitions. We prove that the energy consumption by our scheduling for homogeneous network is at most a constant times that of the optimum and the energy consumption by our scheduling for heterogeneous network is at most Θ(log (Rmax/Rmin)) times of the optimum. We also propose effective algorithms to construct data gathering tree such that the energy consumption and the network throughput is within a constant factor of the optimum. Extensive simulation studies show that our algorithms do reduce considerable energy consumption.