Structural and dynamic insights into the HNH nuclease of divergent Cas9 species

Abstract

CRISPR-Cas9 is a widely used biochemical tool with applications in molecular biology and precision medicine. The RNA-guided Cas9 protein uses its HNH endonuclease domain to cleave the DNA strand complementary to its endogenous guide RNA. In this study, novel constructs of HNH from two divergent organisms, G. stearothermophilus (GeoHNH) and S. pyogenes (SpHNH) were engineered from their respective full-length Cas9 proteins. Despite low sequence similarity, the X-ray crystal structures of these constructs reveal that the core of HNH surrounding the active site is conserved. Structure prediction of the full-length GeoCas9 protein using Phyre2 and AlphaFold2 also showed that the crystallographic construct of GeoHNH represents the structure of the domain within the full-length GeoCas9 protein. However, significant differences are observed in the solution dynamics of structurally conserved regions of GeoHNH and SpHNH, the latter of which was shown to use such molecular motions to propagate the DNA cleavage signal. Indeed, molecular simulations show that the intradomain signaling pathways, which drive SpHNH function, are non-specific and poorly formed in GeoHNH. Taken together, these outcomes suggest mechanistic differences between mesophilic and thermophilic Cas9 species.

Keywords: CRISPR-Cas9; HNH; MD simulations; NMR; Protein dynamics.

Figure 1.

(A) Arrangement of the GeoCas9 and SpCas9 domains and primary sequence. Cartoons are scaled to approximate the molecular weight of each domain. (B) Architecture of the GeoCas9 model and the SpCas9 crystal structure (PDB ID 4UN3). A full-length GeoCas9 homology model was generated with Phyre2.(Kelley, Mezulis, Yates, Wass, & Sternberg, 2015) RNA and DNA from PDB 6JDV were docked and aligned into the model. The similarity of the HNH domains of these proteins is shown in the X-ray crystal structures of SpHNH (light green, PDB 6O56) and GeoHNH (dark green, PDB ID 7MPZ) solved at 1.9 Å and 2.0 Å, respectively.

Figure 2.

AlphaFold2 model of GeoCas9 and comparison with Phyre2. (A) Structural alignment of our newly crystallized GeoHNH domain (orange) with the model of GeoCas9 obtained using AlphaFold2 (left, HNH in green) and Phyre2 (right). (B) Architecture of GeoCas9 modeled using AlphaFold2 (left) and Phyre2 (right). RNA and DNA from PDB 6JDV were docked and aligned into the models. The two models mainly differ for the orientation of the HNH domain with respect to the recognition (REC) lobe.

Figure 3.

(A) R₁R₂ values determined from NMR spin relaxation experiments on GeoHNH-ΔN (top) and SpHNH (bottom). R₁R₂ parameters were measured at 600 (black) and 850 (blue) MHz. (B). Order parameters (S²) for GeoHNH-ΔN (top) and SpHNH (bottom) determined from Model-free analysis of T₁, T₂, and ¹H-[¹⁵N] NOE measurements. Average S² for the entire domain is also indicated. Cartoons of the GeoHNH (dark blue) and SpHNH (light blue) secondary structures are shown above the plots. (C) Order parameters are mapped onto the GeoHNH (top) and SpHNH (bottom) structures, where the heat map legend reports the magnitude of S². (D) Representative CPMG relaxation dispersion profiles for residues in GeoHNH and SpHNH collected at 25 °C at 600 (solid line) and 850 MHz (dashed line). Flat profiles for GeoHNH indicate the absence of detectable ms motions at these sites.

Figure 4.

Temperature dependence of GeoHNH-ΔN NMR spectra and dynamics. (A) ¹H-¹⁵N HSQC of GeoHNH-ΔN collected at 25 (green), 35 (yellow), and 40 °C (red), showing temperature-dependent resonance shifts. (B) R₁R₂ relaxation parameters collected at 25 (blue) and 40 °C (maroon), showing very similar overall profiles. Average values of R₁R₂ and ±1.5σ of the 10% trimmed mean (statistical significance cutoffs) are indicated. (C) ¹H-[¹⁵N] heteronuclear NOEs collected at 25 (blue) and 40 °C (maroon) also appear to be very similar. Small differences in temperature-dependent profiles of (B) and (C) are confined to loops within the protein core and a solvent exposed alpha helix.

Figure 5.

(A) Sequence alignment of HNH domains from divergent Cas9 species; NmeHNH (N. meningitidis), CjHNH (C. jejuni), GeoHNH (G. stearothermophilus), StHNH (S. thermophilus), SaHNH (S. aureus), SpHNH (S. pyogenes), FtHNH (F. novicida), and CdHNH (C. diphtheriae). The level of sequence conservation is indicated by the bars below each row and by color (blue = low conservation; red = high). (B) Plot of NMR order parameters (S²) for GeoHNH highlighting residues that are ≥65% conserved across all Cas9 species with gray bars. (C) HNH sequence conservation within the active site (circled) is mapped onto the GeoHNH structure and colored according to the legend.

Figure 6.

(A) Representative structures of GeoHNH at 40°C. The flexible N-terminal helix is represented in pink and the core of the domain is shown in green. (B) Residue-averaged root mean square fluctuation for GeoHNH at 25°C (green) and 40°C (blue). (C) The allosteric pathways identified for GeoHNH at 25°C (left) and 40°C (right). Cyan and dark blue spheres represent residues proximal to Rec3 (source of the signal) and RuvC (sinks of the signal), respectively. Pink spheres represent the residues that are involved in the allosteric signal transmission.