The first lineage choice made in human embryo development separates trophectoderm from the inner cell mass, which proceeds to form the pluripotent epiblast and primitive endoderm. Trophectoderm on the other hand gives rise to the placenta. Naïve pluripotent stem cells are derived from the pluripotent epiblast of the blastocyst and offer possibilities to explore how lineage integrity is maintained. Here, we discover that Polycomb repressive complex 2 (PRC2) restricts an intrinsic capacity of naïve pluripotent stem cells to give rise to trophectoderm. Through quantitative epigenome profiling, we find that broad histone H3 lysine 27 trimethylation (H3K27me3) hypermethylation is a common feature of naïve pluripotency across species. We define a previously unappreciated, naïve-specific set of bivalent promoters, featuring PRC2-mediated H3K27me3 concomitant with H3K4me3. Naïve bivalency maintains key trophectoderm transcription factors in a transcriptionally poised state that is resolved to an active state upon depletion of H3K27me3 via inhibition of the enzymatic subunits of PRC2, EZH1/2. Conversely, primed human embryonic stem cells cannot be driven towards trophectoderm development via PRC2 inhibition. While naïve and primed hESCs share the majority of bivalent promoters, PRC2 contributes to the repression of largely non-overlapping subsets of these promoters in each state, hence H3K27me3-mediated repression provides a highly adaptive mechanism to restrict lineage potential during early human development.
Read more in our manuscript with Fredrik Lanner and colleagues on bioRxiv.
Single-stranded genomic DNA can fold into G- quadruplex (G4) structures or form DNA:RNA hybrids (R loops). Recent evidence suggests that such non- canonical DNA structures affect gene expression, DNA methylation, replication fork progression and genome stability. When and how G4 structures form and are resolved remains unclear. Here we report the use of Cleavage Under Targets and Tagmenta- tion (CUT&Tag) for mapping native G4 in mammalian cell lines at high resolution and low background. Mild native conditions used for the procedure retain more G4 structures and provide a higher signal-to-noise ratio than ChIP-based methods. We determine the G4 landscape of mouse embryonic stem cells (ESC), observing widespread G4 formation at active promot- ers, active and poised enhancers. We discover that the presence of G4 motifs and G4 structures dis- tinguishes active and primed enhancers in mouse ESCs. Upon differentiation to neural progenitor cells (NPC), enhancer G4s are lost. Further, performing R- loop CUT&Tag, we demonstrate the genome-wide co- occurrence of single-stranded DNA, G4s and R loops at promoters and enhancers. We confirm that G4 structures exist independent of ongoing transcrip- tion, suggesting an intricate relationship between transcription and non-canonical DNA structures.
Bioorthogonal chemistry allows rapid and highly selective reactivity in biological environments. The copper-catalyzed azide–alkyne cycloaddition (CuAAC) is a classic bioorthogonal reaction routinely used to modify azides or alkynes that have been introduced into biomolecules. Amber suppression is an efficient method for incorporating such chemical handles into proteins on the ribosome, in which noncanonical amino acids (ncAAs) are site specifically introduced into the polypeptide in response to an amber (UAG) stop codon. A variety of ncAA structures containing azides or alkynes have been proven useful for performing CuAAC chemistry on proteins. To improve CuAAC efficiency, biologically incorporated alkyne groups can be reacted with azide substrates that contain copper-chelating groups. However, the direct incorporation of copper-chelating azides into proteins has not been explored. To remedy this, we prepared the ncAA paz-lysine (PazK), which contains a picolyl azide motif. We show that PazK is efficiently incorporated into proteins by amber suppression in mammalian cells. Furthermore, PazK-labeled proteins show improved reactivity with alkyne reagents in CuAAC.
From supplementary information of recent JACS paper
COS-7PylRS-AF cells were transfected with the indicated constructs and incubated for 24 h with 50 µM TCO*K. Cells were subsequently fixed, permeabilized and labeled with met-tet-BDP-FL and analyzed by confocal microscopy. SARS-CoV-2 M (Membrane) protein was used as a control because of its known localization to the Golgi (Klumperman et al., 1994). ORF7b is a type I (luminal N-terminus) transmembrane protein implicated as virulence factor, generated by leaky ribosome scanning in an alternative frame within the main ORF7a (Pfefferle et al., 2009). The homologous microprotein in SARS1 is suggested to localize to the Golgi (Schaecher et al., 2008). Our observation is a wider distribution in Golgi, ER and plasma membrane. ORF9b is an alternative ORF within ORF9a and the SARS1 homolog is reported to localize to mitochondria (Shi et al., 2014). We also observe a distinct speckled pattern in the cytosol compatible with partial mitochondrial localization. ORF9c expression and function is currently unknown but we observe a distinct localization with cellular membranes. ORF3b of SARS-CoV-2 is truncated to a 22 amino acid fragment due to a premature stop codon as compared to the SARS1 homolog, but has been reported to potently suppress host cell antiviral interferon response (Konno et al., 2020). The truncation preserves a single predicted transmembrane domain and our imaging suggests an association with cellular membranes.
Table 1 – Plasmids are available on request and via Addgene:
COS-7 cell lines were maintained in Dulbecco’s Modified Eagle Medium (DMEM), containing high glucose, GlutaMAXTM and pyruvate (Gibco). Medium was supplemented with 10% fetal bovine serum (FBS, Sigma-Aldrich). All cell lines were cultured in an ambient-controlled incubator at 37°C, 5% O2 and 5% CO2. Cells were forward transfected using Lipofectamine LTXTM with PLUSTM reagent (Invitrogen) according to manufacturer’s protocol. Axial trans-cyclooct-2-ene-l-lysine (TCO*K) was added at the time of transfection as indicated and cells were harvested after 24 h. TCO*K (SiChem, SC-8008) stock solution was prepared at 100 mM in 0.2 M NaOH/H2O, 15% DMSO. 6-Methyl-Tetrazine-BODIPY®-FL (me-tet-BDP-FL, Jena Bioscience) stocks were prepared in DMF and further diluted in either RIPA buffer (lysate labeling), TBS-T (fixed cells labeling) or the appropriate growth medium (live cell labeling).
Immunofluorescence and live cell imaging
For immunofluorescence, cells were grown and transfected in 96-well µ-Plates (ibidi). After withdrawing the ncAA for 4 h, cells were fixed in 4% formaldehyde for 10 min at room temperature and permeabilized for 15 min with 0.1% (v/v) triton/PBS. Prior to incubation with the appropriate antibodies, cells were click-labeled with 500 nM 6-Methyl-tetrazine-BODIPY-FL (Jena Bioscience), washed 3 times with PBS and then blocked for 1 hour in 2% BSA in TBS supplemented with 0.1% Tween-20 (TBS-T). Cells were incubated with the HA-probe antibody (F-7, Santa Cruz Biotechnology) overnight at 4°C. After washing with TBS-T, cells were stained with Alexa555-conjugated secondary antibodies (Life Technologies) for 60 min at room temperature and counterstained with 1 mg/ml DAPI (Sigma-Aldrich). After washing, cells were imaged on a Zeiss LSM780 confocal laser scanning microscope using a 40x/1.3 oil objective.
For live cell imaging, cells were grown and transfected in 96-well imaging plates (BD Falcon). After withdrawing ncAA for 1 hour, cells were incubated 30 minutes at 37°C in presence of 500 nM me-tet-BDP-FL. Where stated, cells were co-stained with either 4 µM Hoechst, 10 mM ER-trackerTM Red or 250 nM MitoTrackerTM Orange CMTMRos (Invitrogen). After washing 2 times with PBS, cells were immediately imaged in Live Cell imaging Solution (Molecular Probes), on a Nikon eclipse Ti2 inverted widefield microscope equipped with a heated imaging chamber. Images were acquired using a 20×/0.75 air objective or a 40x/1.15 water objective. For the long-term imaging experiment shown in Supplementary Figure 3, cells were maintained in Leibovitz’s L-15 Medium (Gibco) and imaged using a 10x/0.45 air objective.
Proteins are the molecular machines of life, performing a myriad of functions inside every cell of our body. Proteins are assembled from small building blocks, the amino acids, by large protein factories called ribosomes. Understanding how proteins work is a quest of basic biology and medical research.
To study proteins inside of human cells, researchers have been using light microscopes for more than one hundred years. Thirty years ago, the cloning of the green fluorescent protein GFP, together with genetic engineering tools, revolutionized the field by enabling researchers to fuse a fluorescent ‘beacon’ to any protein of interest so that it can be directly observed in living cells using fluorescence microscopy. Fast forward, today’s microscopes achieve live imaging, at nanometer resolution, in multicolor, allowing researchers to resolve even the smallest subcellular structures and essentially watch protein at work.
Fluorescent proteins and other tools that are available to researchers have however one limitation: the size of the fluorescent tag is often equivalent to the size of a typical folded protein, thus adding a considerable molecular ‘cargo’ to the protein under study and potentially impacting its function. This can become a particular obstacle for the study of microproteins, a newly appreciated class of proteins that are much smaller than average. Such tiny proteins have often been overlooked in the past but seminal discoveries of microproteins with important biological functions have sparked growing interest by the research community.
We have made it a focus of our laboratory to tackle the challenges of discovering and characterizing microproteins. Here, we developed a method which allows fluorescent tagging of proteins with the smallest imaginable perturbation – a single amino acid – added genetically on either end of a (micro)protein of interest.
For the method, termed STELLA, a synthetic building block (a non-canonical “designer” amino acid, rather than one of the 21 canonical ones) is incorporated together with a larger tag using a technique termed genetic code expansion. The tag however is swiftly removed by the cell, leaving a single terminal designer amino acid on the protein of interest. As an advantage over existing labeling techniques relying on genetic code expansion, STELLA can thus be used to conveniently and universally label the termini of any proteins. While very similar to its natural counterpart, the designer amino acid introduces a peculiar chemical group into the protein that subsequently allows conjugation with a small organic fluorescent dye, now lighting up the protein of interest inside of the living cell.
So-called transposons are abundant DNA-elements found in every eukaryotic organism as a consequence of their ability to jump and multiply within the host genome. Their activity represents a threat to the integrity of the host genome and thus the host cell engages a number of protective mechanisms to silence the expression of transposons. It is known that some of these mechanisms fail in cancer cells and also ageing cells, leading to a mobilization of transposons with largely unknown consequences. Histones, the proteins that package the genome in the eukaryotic nucleus, are key to the most fundamental line of defense to transposons. By forming a highly compacted array, so-called heterochromatin, they render the associated DNA sequence inert to being read and expressed. Heterochromatin is defined by characteristic modifications to histone proteins and DNA, such as histone H3 K9 trimethylation and DNA CpG methylation.
Here, we studied endogenous retroviral elements (ERVs), a particularly active and abundant family of transposable elements in the mouse genome, which are in fact remnants of once-active viruses. Curiously, while we found all the hallmarks of heterochromatin to be employed in the silencing mechanism, ERV chromatin was highly enriched in a histone variant, termed histone H3.3, which has previously been invariably associated with active regions of the genome. Following up on this observation, wecould elucidate an unexpected mechanism involving a continuous loss of ‘old’ histones and replenishment with newly synthesized histones H3.3 molecules. By genetic manipulation, we were able to deduce a mechanism explaining this dynamic process: the ATP-dependent chromatin remodeler Smarcad1 evicts histones within heterochromatin, thus creating gaps in the chromatin fibre that could render parts of the ERV gene accessible. Following suit, the histone chaperone DAXX seals these gaps by facilitating reassembly of nucleosomes with histone variant H3.3. The concerted process of eviction of one and deposition of another histone is so smooth and efficient that it leaves no apparent trace of accessible DNA.
The result is puzzling because active remodeling and nucleosome eviction is expected to counteract a compacted chromatin structure, inert to transcriptional activation. But we believe that dynamic heterochromatin is an adaption of a ubiquitous silencing mechanism to the specific requirements of a pluripotent chromatin state. The highly transient opening of heterochromatin may allow sequence-specific co-repressors to find their target DNA sequence within the transposable element, in turn recruiting more repressive factors to propagate and amplify the silent state.
In our new bioRxiv preprint, we reveal the surprisingly dynamic nature of interstitial heterochromatin in mouse embryonic stem cells. Interstitial heterochromatin is formed over relatively small (typically 10kb or less) domains harbouring endogenous retroviral elements.
Fundamentally, or results suggest that H3.3 incorporation, and thus exchange of core histones is compatible with the hallmarks of heterochromatin, H3K9me3 and HP1-binding.
Our method provides a proportional measurement of true quantities as exemplified by an artificial gradient of H3K27me3 set up through mixing two samples with 100% and 0% at defined ratios. A quantitative western blot validates the mixing ratios:
The MINUTE-ChIP quantification reproduces the same quantities from the sequencing read counts:
Resulting ChIP-Seq tracks accurately reflect the true quantities:
Why is it important to use quantitative ChIP? Quantitative ChIP is true to changes in peaks and background, as exemplified by an analogy of a volcanic peak rising up above sea level (background):
Here is the boat again. For the observer, the peak now appears pretty small.
But we as an outside observer can see that it is not the volcano height that changed, it is the sea level that rose up to the peak.
Traditional ChIP normalisation assumes a constant background both on the technical and the biological level. Just like the observer in the boat who does not know the change in sea level, the method is blind to global alterations in histone modification levels.
In a paper in press at Cell Reports, we revisit promoter bivalency in naive mouse embryonic stem cells (mESC) using multiplexed quantitative MINUTE-ChIP. Transitioning between Serum-primed and naïve state, both H3K27me3 and H3K4me3 are subject to global alterations. These changes are uncoupled and, in fact, controlled by different external stimuli in the growth media.
The ultra-low technical background of our method allows us to quantify the true distribution of histone H3K27me3 above the technical background. Strikingly, naive mouse ESC have twice as much H3K27me3 methylated nucleosomes as ESC in Serum, and the gained H3K27me3 modifications distributes broadly across the genome while traditional Polycomb targets, such as bivalent domains retain approximately equal levels.
H3K27me3 is an abundant and extremely broad modification in the ground state of pluripotency. Essentially no new H3K27me3 peaks are formed upon transition to a primed state; instead the ambient H3K27me3 ‘background’ is more and more suppressed, ‘revealing’ existing H3K27m3 peaks at all cognate PcG. We further show that the ground state is characterized by particularly low levels of H3K4me3 (more than two-fold lower than Serum-primed), leading to a new model:
Bivalent domains start out in the ground state mainly covered by H3K27me3, and only upon priming, accumulate H3K4me3 levels comparable to those at active genes. We hypothesize that low H3K4me3 together with high H3K27me3 levels at bivalent promoters act to safeguard the ground state of pluripotency.