Structural basis for sequence-specific recognition of guide and target strands by the Archaeoglobus fulgidus Argonaute protein

Abstract

Argonaute (Ago) proteins are found in all three domains of life. The best-characterized group is eukaryotic Argonautes (eAgos). Being the structural core of RNA interference machinery, they use guide RNA molecules for RNA targeting. Prokaryotic Argonautes (pAgos) are more diverse, both in terms of structure (there are eAgo-like ‘long’ and truncated ‘short’ pAgos) and mechanism, as many pAgos are specific for DNA, not RNA guide and/or target strands. Some long pAgos act as antiviral defence systems. Their defensive role was recently demonstrated for short pAgo-encoding systems SPARTA and GsSir2/Ago, but the function and action mechanisms of all other short pAgos remain unknown. In this work, we focus on the guide and target strand preferences of AfAgo, a truncated long-B Argonaute protein encoded by an archaeon Archaeoglobus fulgidus. We demonstrate that AfAgo associates with small RNA molecules carrying 5′-terminal AUU nucleotides in vivo, and characterize its affinity to various RNA and DNA guide/target strands in vitro. We also present X-ray structures of AfAgo bound to oligoduplex DNAs that provide atomic details for base-specific AfAgo interactions with both guide and target strands. Our findings broaden the range of currently known Argonaute-nucleic acid recognition mechanisms.

The missing part: the Archaeoglobus fulgidus Argonaute forms a functional heterodimer with an N-L1-L2 domain protein

“…Subsequent analysis revealed that all proteins associate with short RNA molecules (the predominant length being ∼20–30 for, fAfAgo and scfAfAgo, and ∼15–25 nt for AfAgo) that contain a 5′-phosphate (Figure 2A and B ) and display a pronounced preference for the 5′-AU terminal dinucleotide (Figure 2B ). These findings are in line with our previous study, where we demonstrated that AfAgo preferentially uses gRNA to recognize complementary tDNA ( 21 ).…”

“…We find that for both fAfAgo and scfAfAgo the optimal combination is an RNA guide and a complementary DNA target since all other guide-target combinations resulted in suboptimal target binding (Figure 3A ). In line with our previous results, AfAgo does not exhibit such clear selectivity, but its target binding is evidently weaker in comparison to fAfAgo and scfAfAgo ( 21 ). The affinity of all tested binary pAgo:gRNA complexes to DNA targets also depends on the guide:target sequence complementarity, as the low complementarity tDNA strands yielded no detectable ternary complexes (Figure 3B ).…”

“…The affinity of all tested binary pAgo:gRNA complexes to DNA targets also depends on the guide:target sequence complementarity, as the low complementarity tDNA strands yielded no detectable ternary complexes (Figure 3B ). Taken together, our EMSA studies suggest that fAfAgo and scfAfAgo are both more efficient at RNA-guided binding of complementary DNA targets than the previously studied standalone AfAgo ( 21 ).…”

“…Such NA duplex binding greatly differs from the overall ds NA position in the crystal structures obtained previously with an isolated AfAgo protein, e.g. PDB ID: 6t5t and 6tuo ( 16 , 17 , 18 , 19 , 21 ) ( Supplementary Figure S11A ), where only the 5′-end of the guide/target duplex contacts AfAgo, the remaining part pointing away from the protein. As the fully formed fAfAgo NA-binding cleft may protect longer RNAs from cellular ribonucleases, this may also explain why fAfAgo and scfAfAgo co-purify from E. coli bound to longer RNAs (∼20–30 nt) than the stand-alone AfAgo (∼15–25 nt) (Figure 2 ).…”

“…However, the biological role of AfAgo remains elusive. The recently demonstrated AfAgo ability to form homodimeric assemblies that can bring together two copies of the guide-target duplex ( 20 ), and its intrinsic specificity to 3 nucleotides of both RNA guide and DNA target strands ( 21 ) distinguish it from previously characterized monomeric Argonaute proteins, that limit specific recognition of terminal nucleotides either to the guide strand (e.g. RsAgo, PDB ID: 6d8p ( 22 )), or to the target strand ( 23 ), raising further questions regarding AfAgo functions in vivo .…”