A research team from Cold Spring Harbor Laboratory last week published new details about human Argonaute proteins, further revealing why only one of these four proteins is able to slice target messenger RNA during RNA interference.
Small RNA binding to an Argonaute is a defining step in the RNAi process given that the proteins are critical components of the RNA-induced silencing complex. In humans, however, only Argonaute2 slices despite its structural similarity to the other three Argonaute proteins.
Previously it had been thought that the source of Ago2’s slicing capabilities lay in the four amino acids located in the active sites of the protein family, Leemor Joshua-Tor, a CSHL researcher and senior author of the new study, told Gene Silencing News.
But while Ago1 and Ago4 contain slightly different sequences at their active sites, both Ago2 and the non-slicing Ago3 have the same one — aspartate-glutamate-aspartate-histidine — called the DEDH tetrad.
As such, “it was clear from the beginning that [the active site components were] not the full story” and slicing was not just a result of “having the active site residues intact,” Joshua-Tor said.
To solve this problem, Joshua-Tor and colleagues first determined the structure of human Ago1, both on its own and complexed with the microRNA let-7.
They found that the protein displays the same domain architecture found in all Argonaute proteins: the four primary domains of N, PAZ, Mid, and PIWI, and the two linker regions L1 and L2, according to the study, which appeared in Cell Reports.
The investigators identified “subtle differences” between Ago1 and Ago2, the structure of which had previously been determined, “but those couldn’t really explain to us why one would slice and one wouldn’t,” Joshua-Tor said.
Importantly, Ago1 contains an arginine rather than a histidine in its DEDH tetrad, leading the team to hypothesize that this difference might be behind the protein’s lack of slicing. However, when the team substituted a histidine for an arginine in Ago1, the protein did not function as a slicer, indicating that there were other factors behind Ago2’s unique activity.
The next step was to conduct domain-swapping experiments in which the scientists substituted each functional domain of human Ago2 with the equivalent one in Ago1. The N, PAZ, and Mid domains of Ago1 could individually substitute for their Ago2 counterparts and support Ago2 slicing, as could the Ago1 linkers L1 and L2, they wrote in Cell Reports.
In contrast, exchange of the PIWI domain of either Ago1 into Ago2 eliminated slicing, “providing strong evidence that other factors, in addition to the incomplete DEDH tetrad, are responsible for the slicer defect” observed in Ago1, they noted.
They then undertook a reciprocal experiment, swapping an active PIWI domain from Ago2 into Ago1, and found that this was sufficient to trigger slicing in the ordinarily non-slicing protein. Still, this now-active Ago1 only showed around 10 percent of the slicing activity of Ago2.
Joshua-Tor and her colleagues undertook double domain-swapping experiments, exchanging the N, PAZ, and Mid domains in combination with the human Ago2 PIWI domain. By substituting the N and PIWI domains of Ago2 into Ago1, they were able to convert the latter into an active slicer that performed on par with the former.
In a bid to uncover the “minimal defect” in human Ago1 slicing, the CSHL team then mutated every non-conserved amino acid in the Ago1 PIWI domain into the corresponding amino acid found in human Ago2. In doing so, they identified a single mutation in a loop adjacent to the active site — dubbed PIWI loop 3, or PL3, to avoid confusion with linkers L1 and L2 — that rescued Ago1 slicer activity.
Despite the findings, the exact reasons behind Ago1’s slicing deficiency remain unclear, although it may be due to a catalytic defect, Joshua-Tor said.
“Where [PL3] is sitting, it’s probably interacting with the guide and target strands,” she said. “It might be some kind of turnover thing, but we can’t tell yet.” Work into this question is ongoing in her lab.
Overall, the data presented in Cell Reports indicate that Argonaute slicing activity “has many more nuances than previously appreciated,” the study’s authors wrote. In addition, they show that “events distant from the active site play equally important roles.”
To Joshua-Tor, the findings are expected to also help researchers gain further insights into the Argonaute family.
In humans, she noted, the roles of Ago1, Ago3, and Ago4 are “not clear cut.” Even Ago2, which has been extensively studied, has its secrets.
For example, “Argonaute2 can take a slicing pathway or a non-slicing pathway depending on the hybridization state of the guide with the target,” she noted.