Cognitive Radio has emerged as a promising solution to the spectral congestion problem so as to enable the opportunistic users to share the unoccupied portions of licensed frequency bands. For the purpose of spectrum pooling, OFDMA based multi-carrier techniques such as NC-OFDM offers a good choice due to flexibility in deactivating the subcarriers occupied by primary licensed users to avoid interference, but suffer from high PAPR, not desirable particularly at the user equipment (UE) end. Single Carrier FDMA with localized and interleaved mapping gives lower PAPR at the UE end in the uplink. While Interleaved SC-FDMA has least PAPR along with higher frequency diversity, it requires the free subcarriers to be in a certain pattern, i.e. equi-spaced which is not guaranteed in spectrum pooling due to the fact that the total number of free subcarriers and their locations might change continuously in a dynamic spectrum access network. In this paper, we propose a spectrum allocation algorithm that gives an optimum DFT size and the bandwidth spreading factor to decide the number of users that can be allocated depending on specified required spectral efficiency so that the UE can employ Interleaved SC-FDMA to improve PAPR performance at its transmitter. The simulation results show that the algorithm aims to find an optimum DFT size that balances the required spectral efficiency with significantly reduced PAPR at the UE end as compared to NC-OFDMA and localized SC-FDMA.