Genetic code expansion in stable cell lines enables encoded chromatin modification

Abstract

Genetically encoded unnatural amino acids provide powerful strategies for modulating the molecular functions of proteins in mammalian cells. However, this approach has not been coupled to genome-wide measurements, because efficient incorporation of unnatural amino acids is limited to transient expression settings that lead to very heterogeneous expression. We demonstrate that stable integration of the Methanosarcina mazei pyrrolysyl-tRNA synthetase (PylRS)/tRNA(Pyl)CUA pair (and its derivatives) into the mammalian genome enables efficient, homogeneous incorporation of unnatural amino acids into target proteins in diverse mammalian cells, and we reveal the distinct transcriptional responses of embryonic stem cells and mouse embryonic fibroblasts to amber codon suppression. Genetically encoding N-ɛ-acetyl-lysine in place of six lysine residues in histone H3 enables deposition of pre-acetylated histones into cellular chromatin, via a pathway that is orthogonal to enzymatic modification. After synthetically encoding lysine-acetylation at natural modification sites, we determined the consequences of acetylation at specific amino acids in histones for gene expression.

Figure 1

Transient transfection leads to heterogeneous levels of an unnatural amino acid containing protein, with many cells untransfected, this problem is solved by PiggyBac integration of the unnatural amino acid incorporation machinery. a PiggyBac targeting vectors used in transient transfection and stable cell line generation, Insulators (Ins), inverted terminal repeats (pink triangles). Amber codon between mCherry and EGFP is indicated as ‘TAG’ b FACS analysis of transiently transfected HEK293 and mouse ES cells (mESC), 48 h after transfection and addition of 0.5 mM CpK. Cells were gated for live and single cell population. c Representative FACS analysis of a HEK293 cell line with the unnatural amino acid incorporation machinery integrated, cultured 48 hours with or without addition of 0.5 mM CpK.

Figure 2

PiggyBac-mediated generation and differentiation of mouse embryonic stem cell lines for unnatural amino acid mutagenesis. a PiggyBac targeting vectors used to generate stable mouse embryonic stem cell (ESC) lines from an E14 line, containing bidirectional EF1 and U6 expression cassette, internal ribosome entry site (IRES) and resistance markers (NeoR, PuroR), Insulators (Ins) and inverted terminal repeats (pink triangles). b Representative images of a clone grown in the presence or absence of 0.2 mM CpK for 48 hours. Scale bar: 100 µm. c Embryoid body differentiation protocol to produce beating cardiomyocyte aggregates in the presence (bottom) or absence (top) of 0.2 mM CpK. Scale bar: 100 µm. d FACS analysis of cells differentiated via embryoid body protocol in the absence of CpK and incubated with 0.5 mM CpK for 24 h.

Figure 3

RNA-Seq analysis of E14 ESC cell lines incorporating the unnatural amino acids CpK or AcK. a RNA-Seq analysis of E14 cell line bearing PylS/4xPylT and GFP^150TAG/4xPylT. Whole transcriptome fragments per kilobase of exon per million fragments mapped (FPKM) values in the presence versus absence of 0.2 mM CpK for 48h are plotted. Significantly (P <0.005) up- and downregulated genes from two biological replicates are colored in red and blue, respectively. Further analysis and biological replicates are shown in Supplementary Fig. 6. b RNA-Seq analysis of 3T3 mouse embryonic fibroblast (MEF) cell line bearing PylS/4xPylT and GFP^150TAG/4xPylT. Whole transcriptome FPKM values in the presence versus absence of 0.2 mM CpK for 48 h are plotted. Up- and downregulated gene sets defined in ESC (a) are colored in red and blue, respectively‥ c RNA-Seq analysis of E14 cell line bearing AcKS-^TAGDendra2/4xPylT. Whole transcriptome FPKM values in the presence versus absence of 10 mM AcK for 24h are plotted. Significantly up- and downregulated genes from (a) are colored in red and blue, respectively. d RNA-Seq analysis of wild type E14 cell line, in the presence and absence of 10 mM AcK for 24h.

Figure 4

Site-specific incorporation of acetyl-lysine into the histone H3 in ESC. a PiggyBac targeting vectors used to generate stable mouse ESC lines, Insulators (Ins) , inverted terminal repeats (pink triangles). b Western blot panel of histone cell lines bearing AcKS-^TAGDendra2/4xPylT and H3.2, H3.2(XXTAG)-HA, H3.3 or H3.3 (XXTAG) - designated WT or KXX^TAG in the figure, respectively - with a C-terminal triple HA-tag and 4xPylT. Synthetic histone expression is detected in acid-extracted chromatin by western blot against HA tag, an endogenous histone H3 loading control is shown. Expression of histone H3.2 and H3.3 genes containing TAG codons is dependent on addition of AcK (10 mM for 24 hours). c Amber suppression efficiency in all cell lines from panel b, as measured by FACS analysis of Dendra2 fluorescence, in the presence of 10 mM AcK d, e Expression of HA-tagged Histone H3 without (d) and with amber codon (e) in single cells as measured by fluorescent staining with anti-HA Alexa647 antibody and FACS analysis. Data for the entire panel of H3 variants is shown in Supplementary Fig. 10.

Figure 5

Genetically encoded, site-specific histone acetylation in chromatin regulates gene expression. a RNA-Seq analysis of E14 ESC cell lines bearing AcKS-Dendra2/4xPylT and H3.3(XXTAG)/4xPylT. Heatmap of genes that were found to be significantly ( P <0.005) upregulated in at least one H3.3(XXTAG) cell line, with respect to both the Parental AcKS-Dendra2/4xPylT and the matched control AcKS-Dendra2/H3.3/8xPylT cell lines. b Reverse transcription-quantitative PCR (RT-qPCR) analysis of noncoding XIST RNA expression in H3.3(56TAG) cell lines, with and without addition of 10 mM AcK (24 h) to the medium, and a matched H3.2(56TAG) control cell line, showing that XIST upregulation is dependent on K56 acetylation of histone H3.3. Error bars indicate 95% confidence interval of three technical replicates. c locations of qPCR primers used in (d). d. ChIP assay validating variant-specific incorporation of H3.3K56^TAG-HA₃, in the presence of 10 mM AcK for 24 h, at the XIST locus at primer sites tiling the XIST locus. ChIP assay was performed using α-HA magnetic beads. Error bars indicate 95% confidence interval of three technical replicates.