Deep Mutagenesis of a Transporter for Uptake of a Non-Native Substrate Identifies Conformationally Dynamic Regions

The serotonin transporter, SERT, catalyzes serotonin reuptake at the synapse to terminate neurotransmission via an alternating access mechanism, and SERT inhibitors are the most widely prescribed antidepressants. Here, deep mutagenesis is used to determine the effects of nearly all amino acid substitutions on human SERT surface expression and transport of the fluorescent substrate analogue APP+, identifying many mutations that enhance APP+ import. Comprehensive simulations of the entire ion-coupled import process reveal that while binding of the native substrate, serotonin, reduces free energy barriers between conformational states to promote SERT dynamics, the conformational free energy landscape in the presence of APP+ instead resembles Na+ bound-SERT, with a higher free energy barrier for isomerization to an inward-facing state. The deep mutational scan for SERT-catalyzed import of APP+ finds mutations that promote the necessary conformational changes that would otherwise be facilitated by the native substrate. Indeed, hundreds of gain-of-function mutations for APP+ import are found along the permeation pathway, most notably mutations that favor opening of a solvent-exposed intracellular vestibule. The mutagenesis data support the simulated mechanism in which the neurotransmitter and a symported sodium share a common cytosolic exit pathway to achieve coupling. Furthermore, the mutational landscape for SERT surface trafficking, which likely filters out misfolded sequences, reveals that residues along the permeation pathway are mutationally tolerant, providing plausible evolutionary pathways for changes in transporter properties while maintaining folded structure. Human SERT was encoded by a synthetic codon-optimized gene. The SERT sequence was diversified by overlap-extension PCR to generate two single-site saturation mutagenesis libraries spanning the full length of the gene. The libraries were expressed in human Expi293F cells (a HEK293 derivative) and sorted by FACS for ability to express on the cell surface and the ability to transport a fluorescent neurotransmitter analog, APP+. Enrichment ratios were calculated for all single amino acid substitutions by comparing the frequencies of sequence variants in the sorted cells (from RNA transcripts) with the naive libraries (plasmid DNA). Evolution experiments were duplicated.