Lurking in our genomes are thousands of uncharacterized protein-coding genes. Even more genes can be included in this uncharacterized category if we were to exclude general one- or two-word descriptors of “function.” Furthermore, a significant portion of these genes are conserved across a range of eukaryotes, highlighting their important and ancient potential functions. How do we begin addressing these uncharacterized proteins and uncovering novel cellular mechanisms?
The advent of AlphaFold has led to the production of many descendant programs enabling high-confidence predictions for a number of topics, but especially protein complexes. By selecting a “bait” protein of interest, AlphaFold could be used to fish out the “prey” protein interactors, allowing for downstream hypothesis generation based on interactor identity. Our lab applied this principle to a functional category we believe is rife with uncharacterized factors: protein folding chaperones. Chaperones (or proteins that help other proteins achieve their native folds) are typically transient and early interactors in a substrate protein’s life cycle, which has most likely led to these interactors being difficult to capture with general approaches. AlphaFold reduces this bias by capturing the full breadth of a protein’s interactome.
To uncover novel chaperones, we turned to a protein we had worked on in the past. Our previous work on the bespoke chaperone Zpr1 and its co-chaperone Aim29 revealed a folding sequence tailored to the biogenesis of a single protein known as EF1A, a universal translation factor crucial for the delivery of aminoacylated tRNAs to the ribosome (Sabbarini, Reif, McQuown et al., 2023; McQuown, Nelliat, Reif et al., 2023). Briefly, the post-translational maturation of eukaryotic EF1A (eEF1A) is carried out by a series of steps contingent on the nucleotide state of eEF1A, in which Zpr1 and Aim29 harness eEF1A’s energy cycle to drive folding and chaperone release/recycling. Due to a number of previous insights, we hypothesized that folding factors for eEF1A may be required for its biogenesis prior to its reception by Zpr1, namely at the co-translational level (while eEF1A is still being produced by the ribosome). We used an AlphaFold-derived program termed AlphaPulldown to scour the yeast proteome for novel physical interactors using eEF1A’s protein sequence as bait. Interestingly, the third hit from the top was the conserved and uncharacterized protein Ypl225w. Even more surprisingly, the predicted complex revealed a non-native conformation of eEF1A’s N-most terminal domain, the GTP-binding (G) domain. This in-silico screen led us to hypothesize that Ypl225w was an eEF1A chaperone, owing to the fact that Ypl225w’s predicted complex with eEF1A clashed with the structure of a native, mature eEF1A molecule (suggesting this interaction occurred prior to complete maturation of eEF1A).
However, our history with Ypl225w began earlier during our studies on the Zpr1 co-chaperone Aim29. Deletion of AIM29 had been revealed to be the top uncharacterized stressor of the folding stress response in yeast (Brandman et al., 2012). Tailing directly behind was the deletion of YPL225W. Synthesizing these data led to a hypothesis that Ypl225w was a novel chaperone for eEF1A. But at what stage did Ypl225w act? Before, after, or concurrent with Zpr1/Aim29? To address this, we drew upon a myriad of assays from in vitro translation setups to co-translational folding simulations (in collaboration with Kibum Park and the Shakhnovich Lab from the Department of Chemistry and Chemical Biology). Our complementary approaches arrived at the same conclusion: Ypl225w seemed to bind eEF1A’s G domain prior to translation of its subsequent two domains. Mediating this ribosomal tether was another factor (similarly uncovered via AlphaPulldown), the ribosome-associated nascent polypeptide-associated complex or NAC. Subsequent assays revealed a satisfying conclusion to Ypl225w’s role. First, Ypl225w uses NAC as a platform to dock onto eEF1A-translating ribosomes where it then inserts its N-terminal alpha helix into eEF1A’s GTP-binding-incompetent G domain, preventing unproductive intra- and inter-domain interactions until further translation/folding produces the GTP-binding pocket. Binding of GTP then drives a folding switch that produces a near-native G domain, while also leading to release and recycling of Ypl225w.
In conclusion, the top two (previously) uncharacterized stressors of the folding stress response were in fact deletions of genes whose products encoded two eEF1A chaperones. Ypl225w acts prior to Zpr1 and Aim29 by enabling loading of the pioneer GTP into eEF1A’s G domain, which stabilizes the native G domain fold.
Our work reveals the first example of a substrate-derived energy cycle being used to mediate a co-translational folding process. It also highlights the power of AlphaFold to uncover new chaperones and folding factors. How many more of those uncharacterized genes in our genomes are actually dedicated chaperones?
by Ibrahim (Abe) Sabbarini