Erik Procko, Ph.D. - US grants

ABSTRACT The norepinephrine transporter (NET), dopamine transporter (DAT), and serotonin transporter (SERT) catalyze the reuptake of their respective monoamine neurotransmitters from the extracellular space surrounding synaptic membranes, terminating neurotransmission. These transporters are critical regulators of neuronal signaling. Non-synonymous sequence polymorphisms of NET, DAT and SERT are implicated in numerous psychiatric disorders, including major depression, anxiety disorders and obsessive compulsive disorder (OCD), and the transporters are molecular targets for widely prescribed mental health drugs. Deep mutational scanning is a new technology that combines directed evolution of large, diverse libraries of sequence variants with deep sequencing to track the phenotypes of potentially thousands of mutants in a single experiment. This makes it possible to experimentally determine a comprehensive fitness landscape of a protein, but this technique has yet to be applied to human transmembrane proteins evolving in human cell culture. The specific aims of this project are Aim 1: to experimentally determine the fitness landscapes of monoamine neurotransmitter transporters. Single site-saturation mutagenesis libraries encoding all amino acid substitutions of NET, DAT and SERT will be transfected into human cells, which will subsequently be sorted for variants that are expressed on the cell surface and actively transport a fluorescent neurotransmitter analogue. Enrichment scores for each mutation during in vitro evolution will be calculated, and from these fitness landscapes insight will be gained into critical residues for transporter folding, surface trafficking, oligomerization, substrate binding, transport and regulation. In addition to learning how the transporter sequences relate to their basic molecular functions, the data will pre-emptively describe molecular phenotypes for all possible single amino acid substitutions that may exist in the human population. This is a radical change from existing approaches where non-synonymous SNPs are identified first, perhaps in clinical populations with psychiatric disease, and then molecular activity is investigated. Aim 2: Develop a deep mutational scanning- based assay for mapping drug interaction sites to protein sequences of neurotransmitter transporters. A diverse library of SERT mutants will be screened by deep mutational scanning for variants with decreased response to the drugs (S)-citalopram and paroxetine. Mutations that `rescue' transport will be mapped to crystal structures of human SERT bound to (S)-citalopram and paroxetine, which will validate whether this method can successfully identify known orthosteric and allosteric binding sites. In addition to providing a new technology platform for mapping drug interaction sites with residue-level resolution, the data will again pre- emptively describe how all possible non-synonmous sequence variants of the transporter impact drug potency. We envisage in the future deep mutational scanning data will inform interpretation of a patient's genotype, even for rare alleles not previously observed in the clinic, and influence practitioner drug choice.

The HIV-1 surface glycoprotein Env engages host cell receptors, including either the CCR5 or CXCR4 chemokine receptors, to drive the necessary protein rearrangements that mediate virus entry into the cell. Therapies blocking HIV-1 Env interactions with chemokine receptors are clinically effective, underscoring their importance. This proposal uses the new technology of deep mutational scanning to comprehensively determine sequence-function relationships in CCR5, CXCR4 and Env. Deep mutational scanning combines unbiased, diverse libraries of mutations with in vitro evolution and deep sequencing, making it possible to determine the relative phenotypes of many thousands of mutations in a single experiment. From this unprecedented mutational data, a protein's sequence-fitness landscape can be experimentally mapped, from which functional sites and important residues for stabilizing discrete conformations can be inferred. The sequence-fitness landscape also reveals mutations that can be combined to engineer variants with new or enhanced properties. Deep mutational scanning has primarily been limited to proteins that are expressed in phage, bacteria or yeast, but in this proposal, libraries encompassing all single amino acid substitutions of CCR5, CXCR4 and Env expressed in human cells will be evolved. The specific aims of this proposal are Aim 1: To determine the oligomeric organization of CCR5 and CXCR4 by deep mutational scanning. When libraries of CCR5 and CXCR4 are sorted for high affinity to antibodies recognizing resting conformations, conserved residues in the sequence-fitness landscapes map to transmembrane surfaces of the receptors. We hypothesize that these conserved surfaces are dimerization sites, which will be validated using biochemical methods. Residue conservation scores from the mutational scans will guide computational modeling of the dimeric states. Aim 2: To comprehensively map the sequence-fitness landscapes of CCR5 and CXCR4 during signaling responses to agonists. A cell sorter will be adapted for continuous mixing and sorting of Ca2+- indicator stained libraries with chemokines. Critical residues for chemokine interactions, G protein coupling, and adopting an active conformation will be conserved in the sequence-fitness landscapes. Aim 3: To characterize the interaction between chemokine receptors and HIV-1 gp120-CD4 by deep mutational scanning. CCR5 and CXCR4 sequence-fitness landscapes for tight affinity to gp120-CD4 will reveal similarities and differences in how these chemokine receptors are engaged by R5 and X4 HIV-1 strains, and how maraviroc- resistant Env clones have altered CCR5 interaction footprints. Aim 4: To comprehensively determine the sequence-fitness landscape of HIV-1 Env interacting with soluble CD4 and broadly neutralizing antibodies VRC01 and PG16. These protein ligands recognize distinct Env quaternary structures, despite CD4 and VRC01 sharing a common binding site. Deep mutational scanning, covering over 17,000 Env mutations, will guide engineering of trimeric Env to pre-stabilize conformations, with implications for immunogen design.