DNA G-Quadruplex Recognition In Vitro and in Live Cells by a Structure-Specific Nanobody

Abstract

G-quadruplexes (G4s) are four-stranded DNA secondary structures that occur in the human genome and play key roles in transcription, replication, and genome stability. G4-specific molecular probes are of vital importance to elucidate the structure and function of G4s. The scFv antibody BG4 has been a widely used G4 probe but has various limitations, including relatively poor in vitro expression and the inability to be expressed intracellularly to interrogate G4s in live cells. To address these considerations, we describe herein the development of SG4, a camelid heavy-chain-only derived nanobody that was selected against the human Myc DNA G4 structure. SG4 exhibits low nanomolar affinity for a wide range of folded G4 structures in vitro. We employed AlphaFold combined with molecular dynamics simulations to construct a molecular model for the G4-nanobody interaction. The structural model accurately explains the role of key amino acids and K_d measurements of SG4 mutants, including arginine-to-alanine point mutations that dramatically diminish G4 binding affinity. Importantly, predicted amino acid-G4 interactions were subsequently confirmed experimentally by biophysical measurements. We demonstrate that the nanobody can be expressed intracellularly and used to image endogenous G4 structures in live cells. We also use the SG4 protein to positionally map G4s in situ and also on fixed chromatin. SG4 is a valuable, new tool for G4 detection and mapping in cells.

Figure 1

(A) Schematic representation of a G-tetrad (four guanines in green), stabilized by a monovalent cation (M⁺), and of a G-quadruplex. (B) Workflow of phage display screening of a nanobody library and validation through nonadsorbent phage ELISA, leading to the selection of G4-binding SG4. Sequences of the oligonucleotides used in the screening are reported. A counterselection step against single-stranded and double-stranded DNA negative controls was followed by three rounds of incubation with MycG4. (C) The amino acid sequence of the SG4 nanobody. Complementarity-determining regions (CDRs) highlighted in red (CDR1), green (CDR2), and blue (CDR3). The nanobody is tagged with 6xHIS (orange) and with a 3xFLAG tag (purple). The amino acid position is reported above the sequence. (D) SG4 binding curves to MycG4 and negative controls ssDNA mutant Myc and 8-aza-7-deazaguanine corresponding oligonucleotide, determined by ELISA. (E) ELISA SG4 binding curves to different G4 topologies: Kit1G4 and VegfG4 (parallel), TbaG4 (antiparallel), and hTeloG4 (hybrid). (F) ELISA mSG4 R105A binding curves to MycG4 and negative control ssDNA mutant Myc. The sequence of CDR3 carrying the mutation from arginine to alanine is reported. Dissociation constants (K_d) are indicated in nanomolar; in some cases, they could not be determined (ND). Error bars represent the standard error of the mean (s.e.m.) calculated from two replicates.

Figure 2

(A) SG4, mSG4 R105A, and mSG4 R107A secondary structures determined by CD spectroscopy (200–280 nm) showing a characteristic β-sheet with a negative peak at 218 nm and a positive peak at 200 nm. Units are measured in molar ellipticity. (B) Overlap of Alphafold 2.0 top 5 structure predictions, CDRs in red and scaffold in blue. (C) First four principal components and projections between 1–2 and 3–4, and time-lagged cross-correlation of the combined 5 μs simulation of SG4 with a time lag of 5 ns. (D) MMGBPA analysis of the binding mode of SG4 with MycG4 throughout MD simulations of 500 ns (in 50 frames) with the individual contribution of residues. (E) 3D frame with residues colored according to energy contribution, G4 in light green.

Figure 3

(A) Venn diagrams of the overlap of SG4 regions identified through CUT&Tag (SG4 CUT&Tag peaks) with sequences previously identified as capable of folding into a G4 structure in vitro (so-called observed G4 sequences and referred to as OQs) (4) in HEK293T. (B) Fold enrichment over random (black bars) and proportion (gray) of SG4 CUT&Tag consensus regions across different genomic features in HEK293T. (C) Genome browser screenshots for biological replicates of CUT&Tag for SG4-GFP-FLAG-expressing HEK293T cells (Rep1 and Rep2), and control HEK293T (cells not expressing SG4-GFP-FLAG) using an anti-FLAG primary antibody. Strong peaks are shown for the MYC and KRAS promoter G4 regions in cells expressing SG4-GFP-FLAG upon doxycycline treatment. (D) Confocal live cell imaging of SG4 foci in the nuclei of HEK293T cells expressing SG4-GFP. The scale bar is 20 μm.