Profiling Thermus thermophilus Argonaute Guide DNA Sequence Preferences by Functional Screening

Abstract

Prokaryotic Argonautes (pAgo) are an increasingly well-studied class of guided endonucleases, and the underlying mechanisms by which pAgo generate nucleic acid guides in vivo remains an important topic of investigation. Recent insights into these mechanisms for the Argonaute protein from Thermus thermophilus has drawn attention to global sequence and structural feature preferences involved in oligonucleotide guide selection. In this work, we approach the study of guide sequence preferences in T. thermophilus Argonaute from a functional perspective. Screening a library of 1,968 guides against randomized single- and double-stranded DNA substrates, endonuclease activity associated with each guide was quantified using high-throughput capillary electrophoresis, and localized sequence preferences were identified which can be used to improve guide design for molecular applications. The most notable preferences include: a strong cleavage enhancement from a first position dT independent of target sequence; a significant decrease in activity with dA at position 12; and an impact of GC dinucleotides at positions 10 and 11. While this method has been useful in characterizing unique preferences of T. thermophilus Argonaute and criteria for creating efficient guides, it could be expanded further to rapidly characterize more recent mesophilic variants reported in the literature and drive their utility toward molecular tools in biology and genome editing applications.

Keywords: Thermus thermophilus; capillary electrophoresis; nucleic acid guide; prokaryotic argonaute; targeted endonuclease.

FIGURE 1

In a population of fully-complementary 17 nucleotide-long guides targeting randomized-sequence single- and double-stranded substrates of various GC contents, the first position of the guide (g1) was strongly correlated with overall guided-endonuclease activity of TtAgo. In particular, guides with g1T performed better than those with g1V. Each g1 grouping is divided to show the overall GC content of the substrate targeted, however, the significance levels shown are for each g1 grouping against the others. Boxplots have been annotated for statistical significance between relevant groups using the following labels: ns (not significant) for p > 0.05; ^∗ for p ≤ 0.05; ^∗∗∗ for p ≤ 0.001; and ^**** for p ≤ 0.0001.

FIGURE 2

Guided-endonuclease activity of TtAgo using fully-complementary guides, where g1 is complementary to t1 (g1:t1), and partially complementary guides, where g1 is not complementary to t1 (g1Xt1) acting on (A) single-stranded and (B) double-stranded DNA substrates. Significance levels shown only for g1:t1 to g1Xt1 comparisons. Boxplots have been annotated for statistical significance between relevant groups using the following labels: ns (not significant) for p > 0.05; ^∗ for p ≤ 0.05, ^∗∗ for p ≤ 0.01; ^∗∗∗ for p ≤ 0.001; and ^**** for p ≤ 0.0001.

FIGURE 3

(A) Comparison of fully-complementary g1B/D/H/V guides to partially-complementary (g1 not complementary to t1) g1A/C/G/T guides. (B) A heatmap comparing the average activity of all g1-t1 combinations with single- and double-stranded substrates. (C) Comparing partially-complementary g1C and g1T guides to the same population of fully-complementary g1R guides. Boxplots have been annotated for statistical significance between relevant groups using the following labels: ns (not significant) for p > 0.05; ^∗∗ for p ≤ 0.01; and ^**** for p ≤ 0.0001.

FIGURE 4

(A) Aligning guides which, when used with TtAgo, exhibit high activity (>80% product after 30 min at 75°C) highlights some sequence preferences for specific locations along the guide (g1 excluded from the alignment). (B) The preference for g12B is exemplified by the low activity of guides possessing g12A on single- and double-stranded substrates. (C, D) Comparing g1R and g1Y populations, the identity of g12 appears to be decoupled from the lower activity observed with the less favored g1R identity. (E) The effect of g12 on activity remains for g1Y and single-stranded substrates only, which are more readily accessible by TtAgo. Boxplots have been annotated for statistical significance between relevant groups using the following labels: ns (not significant) for p > 0.05; ^∗ for p ≤ 0.05; ^∗∗∗ for p ≤ 0.001; and ^**** for p ≤ 0.0001.

FIGURE 5

TtAgo shows a preference for low GC content dinucleotides at the g10/g11 positions which align with the enzyme active site. (A) Populations expressed as box plots individually for single- and double-stranded substrates. (B) Single- and double-stranded substrates combined and expressed as a heatmap.

FIGURE 6

(A–D) Application of guide design principles to achieve highly active guides for in vitro use.