TICB-1086; No. of Pages 12

Review

Epigenetics reloaded: the single-cell revolution Poonam Bheda and Robert Schneider Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire (IGBMC), CNRS UMR 7104/Inserm U964/Universite´ de Strasbourg, 67400 Illkirch, France

Mechanistically, how epigenetic states are inherited through cellular divisions remains an important open question in the chromatin field and beyond. Defining the heritability of epigenetic states and the underlying chromatin-based mechanisms within a population of cells is complicated due to cell heterogeneity combined with varying levels of stability of these states; thus, efforts must be focused toward single-cell analyses. The approaches presented here constitute the forefront of epigenetics research at the single-cell level using classic and innovative methods to dissect epigenetics mechanisms from the limited material available in a single cell. This review further outlines exciting future avenues of research to address the significance of epigenetic heterogeneity and the contributions of microfluidics technologies to single-cell isolation and analysis. A single-cell look at epigenetic inheritance mechanisms In a completely gene-deterministic world, all genetically identical cells should have the same phenotype. However, we know that this is not the case; in multicellular organisms, cells undergo differentiation to give rise to various lineages and even when considering single-cell organisms, isogenic cells do not behave the same. While some of the differences between isogenic cells can be attributed to stochastic heterogeneity, others can arise from diversity in their epigenome (see Glossary). Epigenetic mechanisms underlie the transmission of changes in phenotypic traits to progeny independent of alterations in the DNA sequence. Epigenetic phenomena include memory and/or maintenance of distinct transcriptional states that by themselves can arise from ‘predetermined’ programs or environmental signals retained in the absence of the original stimuli. Such processes include gene silencing, position-effect variegation, X-chromosome inactivation, cellular differentiation, and transcriptional memory [1–3]. Key to this regulation are noncoding RNAs (ncRNAs) [4], chromatin modifications [5,6], chromatinmodifying and -binding factors, and chromosome architecture [7]. Thus, chromatin-based mechanisms are the foundation of epigenetic processes and therefore analysis of chromatin states forms a large part of epigenetic studies. Corresponding author: Schneider, R. ([email protected]). Keywords: single cell; epigenetics; live cell; epigenomics; microfluidics. 0962-8924/ ß 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tcb.2014.08.010

Epigenetic mechanisms can be inherited through cellular divisions in single-cell organisms or to cellular and generational progeny in multicellular organisms. Unlike the highly stable genome, epigenetic signatures are metastable, with different epigenetic phenomena having different degrees of stability and variability [8–10]. However, the extent to which the establishment or erasure of specific epigenetic signatures is predetermined, heritable, and/or stochastic is unclear. Some epigenetic phenomena are stably inherited across many cell divisions and/or generations, while others are maintained through only one or a few generations. There is also considerable variability among individual cells in the level of specific epigenetic processes. Position-effect variegation is a classic example of this cell-to-cell variability [1]. This cell heterogeneity, combined with the varying levels of stability, makes it difficult to define heritability in a population of cells. While population experiments provide some mechanistic information for persistent epigenetic processes present in a significant percentage of cells, they miss the intricacies of individual-cell responses. Therefore, it would be of interest to track the maintenance, inheritance, and variability of epigenetic processes in single cells and their progeny (Figure 1). The recent emphasis on single-cell rather than population-based data has pushed the limits of resolution for every type of analysis and has been particularly informative in deciphering epigenetic mechanisms. Variability in transcriptional responses in single cells and the role of chromatin modifications, chromatin-modifying enzymes, binding proteins, and ncRNAs has become a major avenue of research. This, along with the push for high-throughput analyses, has moved the epigenetics frontier to include single-cell ‘omics’ approaches. These studies will be vital to identify carriers of epigenetic information and understand the extent to which an epigenetic process is heritable. In this review, we present techniques geared toward singlecell epigenetics research (Figure 2) and discuss recent insights achieved in our understanding of epigenetic mechanisms that have come from their use. Genomic sequencing-based epigenomic methods The role of chromatin in epigenetic processes has focused research in epigenetics toward multiple techniques involving lysis of populations of cells to analyze gene expression (transcriptomics) and correlations with chromatin states including chromosome architecture, chromatin modifications, and/or nucleosome occupancy. Assays to study these chromatin states on a genome-wide scale all Trends in Cell Biology xx (2014) 1–12

1

TICB-1086; No. of Pages 12

Review Glossary Bisulfite sequencing: a method to differentiate between methylated/hydroxymethylated DNA and unmethylated DNA. Bisulfite converts unmethylated cytosines to uracil while methylated and hydroxymethylated cytosines are unaffected. Conversion is usually followed by PCR amplification and NGS to identify the methylation status [28,29]. Bromodomain: a protein domain found in some modification-reader proteins that recognizes acetylated lysine residues such as those on histones. Although conserved, bromodomains can be selective toward acetylation of specific lysine residues. Cas9: RNA-guided nuclease found in Archaea and some eubacteria for immunity against foreign DNA. These proteins have been exploited for their ability to be directed to a specific genomic locus by synthetic sgRNAs, which hybridize with their complementary DNA for targeted recruitment [55]. See also TALE. Chromatin: histone–DNA complex that packages DNA in the nucleus of a eukaryotic cell. Chromatin immunoprecipitation (ChIP): a technique to identify genomic DNA sequences that are associated with chromatin-associated factors or histone modifications. Antibodies against proteins or histone modifications are used to purify the DNA associated with the protein or modification of interest, followed by locus-specific PCR or NGS methods. Chromatin-modifying enzymes: proteins that catalyze modifications of DNA or histones. DNA adenine methyltransferase identification (DamID): DamID functions by expressing a Dam fused to the protein of interest. Dam is not present in higher eukaryotes but can be expressed to methylate adenines that come in close contact with the enzyme. These methylated adenines can then be mapped, allowing the identification of binding sites in an antibody-independent approach. DNA cytosine methylation: cytosines in DNA can be methylated/hydroxymethylated, which is generally associated with transcriptional repression. Fluorescence in situ hybridization (FISH): can detect DNA regions and nascent transcription of specific RNAs [39]. Fo¨rster resonance energy transfer (FRET): occurs between two chromophores when the donor chromophore excites the second acceptor chromophore when in close proximity [56]. Hi-C: a genome-wide, high-throughput method to detect chromosomal interactions, based on the 3C method, in which DNA interactions are detected by crosslinking of interacting regions followed by restriction digestion of the DNA and ligation of the crosslinked pieces, which are then processed in various ways. Hi-C uses biotinylation before ligation to mark ligation sites, which is subsequently used to purify ligation products, which are then sequenced by NGS [21,22,62]. Histones: small basic proteins important for organization of DNA. The core histones H2A, H2B, H3, and H4 form octameric complexes around which DNA is wrapped, while histone H1 is bound to DNA between these octamers. Histones are modified by post-translational modifications, of which the most well known are acetylation, methylation, phosphorylation, and ubiquitination, which can, for example, affect the affinity of the histones for DNA or recruit specific binding proteins. Immunofluorescence (IF): fluorescently labeled antibodies with specificity for the modification or protein of interest are used for visualization in fixed cells by microscopy. m6 A-Tracer: a methylated adenine (m6A)-binding domain fused to GFP that allows the visualization of adenines in DNA that have been methylated by the Dam methyltransferase [52]. Multiphoton fluorescence-lifetime imaging microscopy–Fo¨rster resonance energy transfer (FLIM–FRET): a highly quantitative type of FRET (see above) in which the fluorescence lifetime of the donor fluorophore provides the readout for FRET such that the lifetime fluorescence of the donor decreases when FRET occurs when the donor and acceptor are in close proximity [53]. Next-generation sequencing (NGS): massively parallelized DNA-sequencing methods for high-throughput analyses [63,64]. Nucleosome: the basic unit of chromatin, a histone–DNA complex comprising approximately 146 bp of DNA wrapped around an octamer of two of each of four histone proteins (H2A, H2B, H3, and H4). Quantitative reverse-transcription PCR (qRT-PCR): an RNA detection and relative quantification method. Conventional qRT-PCR uses a reverse transcriptase to make cDNA from RNA and then applies locus-specific primers to amplify the cDNA of interest. RNA sequencing (RNA-seq): a method to detect RNAs globally using NGS methods [13,14]. Transcription activator-like effectors (TALEs): originally involved in bacterial infection of plants, these proteins recognize DNA sequences by specific amino acids in a variable region of a repeat domain that has been exploited for redirecting TALE fusions (e.g., with GFP) to specific DNA sequences of interest [65]. See also Cas9.

2

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

rely on PCR or high-throughput DNA-sequencing methods collectively called ‘next-generation sequencing’ (NGS). With the advent of single-cell sequencing, it is now possible to resolve the transcriptome, genome organization, and some modifications of a single cell on a genomewide scale. Due to the limited sample, single-cell sequencing experiments require the isolation of individual cells (Box 1) and amplification steps in very small volumes. Like all omics methods involving NGS, the techniques described below require significant bioinformatic analysis, which is currently the major bottleneck for these studies. While their drawback in epigenetic inheritance studies is the lack of time-resolved analyses of the same cells, insights gained from single-cell epigenomics is vast and invaluable. RNA analysis for open reading frame (ORF) and ncRNA expression: quantitative reverse-transcription PCR (qRTPCR) and RNA sequencing (RNA-seq) Transcriptomics methods determine locus-specific or whole-genome expression profiles. Since epigenetic changes are associated with heritable effects in ORF and ncRNA expression [11–14], it is vital to compare RNA expression. Single-cell transcriptomics studies differences in transcription profiles between individual cells, for example, to compare different cells of the same embryo to identify cell type-specific changes during development or variability among fully differentiated cells to observe differences between cells of the same cell type. These studies have relied on RNA expression analysis in single cells by either qRT-PCR or RNA-seq, often employing microfluidics platforms (Box 1) [11–14]. qRT-PCR has been used to examine the transcriptional profile of individual mouse cells from the oocyte to early blastocyst stage of embryogenesis, consequently identifying key epigenetic regulators for each developmental stage [15]. RNA expression analyses have also implicated ncRNAs as regulators of gene expression. For example, nuclear long ncRNAs can function either as antisense RNAs or as scaffolds for targeting chromatin-modifying enzymes to specific locations [16]. Single-cell transcript data can afford a deeper understanding of the variability of ncRNA expression, with simultaneous information on downstream effects influencing ORF expression. In the future, single-cell transcriptomics might also be important for understanding disease processes and could be used to reveal differences among cancerous cells and normal cells from the same tissue within a single patient to better understand the initiation of cancer. Chromosome-conformation analysis: Hi-C Chromosomal looping, such as enhancer interactions with promoters or targeting of nuclear lamin-associated domains (LADs) to the nuclear periphery, has been implicated in the regulation of transcription and epigenetic memory. Chromosome-conformation capture (3C) techniques decipher these interactions to study spatial genome organization. These techniques include the original 3C and variations such as 4C, 5C, and Hi-C. Here, DNA interactions are detected by a series of steps including crosslinking of the interacting regions followed by restriction digestion of the DNA and intramolecular ligation of the crosslinked DNA

TICB-1086; No. of Pages 12

Review

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

(A)

12

6

6

3

3

1

1

1

4

2

9

3

8

16 10

8 13

4

5 7

2

14

2

5

2

4

1 11

7 15

Transient smulus Level of chroman modificaon 16 8 14 4 15 7 13 2 12 6 10 3 11 5 9 1 12

6

6

3

3

1

1

1 4

2

16

9

3

8

13 4

5 2

4

1 5

2

8

10

7

14

2 7

11 15

Transient smulus Level of chroman modificaon 16 8 14 4 15 7 13 2 12 6 10 3 11 5 9 1

Number of cells

(B)

Level of modificaon 3 mepoints TRENDS in Cell Biology

Figure 1. Schematic representation of the stability, variability, and heredity of epigenetic processes in cell lineages. (A) Chromatin modifications, noncoding RNAs (ncRNAs), and chromosome architecture underlie many epigenetic mechanisms. Some of these modifications can be maintained through cell divisions even after the initiating stimulus has been removed. Different cells might have varying levels of a modification or express different levels of ncRNAs (depicted by shades of the same color) due to stochasticity or epigenomic variability. Different epigenetic processes (depicted by different colors) can also vary in stability. For example, cell-lineage specification can be associated with stably inherited chromatin modifications (purple, top panel), whereas cycling environmental conditions might be associated with less stable modifications (green, bottom panel). Following these modifications in cell lineages is essential to reveal epigenetic inheritance from progenitor cells to the following generations, here depicted as shades of cells that are maintained after division in the pedigree chart. Some modifications may be inherited for one or a few generations and then gradually lost (green), while others may be more robust (purple). (B) Variability in stability will contribute to the variance/noise level of a chromatin modification, ncRNA, or chromosome interaction and can be observed in single-cell data at a specific time point or as population-averaged data, which will result in the loss of some persistence and heredity information. In these two panels [which correspond to fluorescence-activated cell-sorting (FACS) data, for example, for the level of fluorescence that would be observed from populations of cells depicted in (A)], it can be observed that the noise/variance is maintained over time for some chromatin modifications (top panel), but can increase over time for other modifications (bottom panel); this information and the source of variability is more clearly observed in single cells with a known pedigree as in (A), which allows mathematical modeling as well as predictions.

3

TICB-1086; No. of Pages 12

Review

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

Q

PCR RNA mRNA ncRNA

qRT-PCR RNA-seq

cDNA

B

B

Hi-C

B

B

seq Methylated

RSMA bisulfite sequencing C

Nucleosome mapping

C

C C

Bisulfite conversion

U

Me

Me Me

Cell lys

ate

C Me

Unmethylated

Gp C

GpC

GpC GpC

Bisulfite conversion GpC

GpC

GpU

GpC

Fluorescent RNA probe Xist long ncRNA

DNA/RNA - FISH

Single cell

Fluorescent DNA probe Inacve X

Fixed cell ISH - PLA B

Cell type promoter

Live c

HP1

ell

K9me3

CiA

GFP

H3

Promoter

GFP YFP

GFP

FABS/HISTAC

X

FAB CFP

Ac H4

Lamin/DAM

m6

A-tracer

A

GFP

mcherry

Increased FRET measured by FLIM Interphase

Mitosis TRENDS in Cell Biology

Figure 2. Three avenues of single-cell epigenetic research. Single-cell lysate epigenomic techniques include quantitative reverse-transcription PCR (qRT-PCR) and RNA sequencing (RNA-seq) techniques to detect RNA expression and chromosome conformation capture by Hi-C to identify inter/intrachromosomal contacts. Among chromatin modifications, only DNA methylation can be analyzed from lysates of single cells using either restriction enzyme-based single-cell methylation assay (RSMA) for PCR-based differentiation of methylation at selected loci or bisulfite conversion and sequencing [including reduced-representation bisulfite sequencing (RRBS-seq)]. Nucleosome occupancy is determined by protection against methyltransferase activity on single DNA molecules by the presence of nucleosomes. Fixed-cell methods include fluorescence in situ hybridization (FISH) techniques for both DNA and RNA, as well as immunofluorescence methods for global distributions of chromatin modifications. In situ hybridization coupled with proximity ligation assay (ISH–PLA) labels histone modifications at specific loci in single cells. Live-cell techniques using fluorescence microscopy include chromatin in vivo assay (CiA), fluorescently labeled specific antigen-binding fragments (Fabs)/Histacs, and m6A-Tracer technology. CiA employs small molecule-induced proximity for locus-specific, inducible chromatin modification. Fabs and Histacs label histone modifications by fluorescent antibodies or Fo¨rster resonance energy transfer (FRET)-based reporters involving histone fusions with modification-specific readers, respectively. In m6A-Tracer technology, a protein–DNA adenine methyltransferase fusion labels DNA with m6A when in proximity to the fusion, which is visualized by a fluorescent m6A-reader protein.

4

TICB-1086; No. of Pages 12

Review pieces [17–22]. The crosslinks are then reversed and DNA interactions are observed by PCR, microarray analysis, or sequencing. The main limitation of 3C, 4C, and 5C techniques is that, to detect the interactions via PCR, one or both interacting pieces of DNA must be known. Alternatively, HiC bypasses this requirement by subjecting the ligated pieces to NGS, making it a genome-wide, high-throughput method for the detection of inter- and intrachromosomal interactions. These techniques have been widely used for epigenetic analyses in populations of cells [23–26]; however, there is currently only a single report of chromosome-conformation capture in single cells. In this groundbreaking report, Hi-C technology was performed in intact nuclei of single mouse splenic cells to prevent dilution of the DNA, consequently increasing the ligation efficiency of interacting DNAs [27]. The study showed transcriptionally active chromatin domains localizing to the surface of their chromosome territory, while inactive domains present at the surface associated with the nuclear periphery, with variability in some chromosomal contacts and robustness in others on a genome-wide scale between individual cells. It will be

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

interesting to further study chromosome interaction data from various types of single cells to determine tissue- and/or organism-specific differences in chromosomal architecture and conservation of these epigenetic mechanisms. DNA cytosine methylation analysis: locus-specific restriction enzyme-based single-cell methylation assay (RSMA) and bisulfite sequencing Methylation of cytosines within DNA are chromatin modifications that usually mark transcriptionally silent regions of the genome and are involved in many epigenetic processes including genomic imprinting and embryonic development. Methylation status is typically assayed by bisulfite sequencing or RSMA. Bisulfite converts unmethylated cytosines to uracil while methylated and hydroxymethylated cytosines remain unchanged, followed by amplification and NGS to identify methylation sites genome wide [28,29]. In single cells, this application can result in heterogeneity due to incomplete conversion and degradation from bisulfite treatment. Nevertheless, a modified version of this method, reduced-representation

Box 1. Complementary methods: single-cell isolation and analysis techniques Most single-cell analyses require the separation or isolation of individual cells. Here we discuss some techniques for cell isolation that can be coupled with various cell-analysis methods (Figure I). We specifically highlight the use of microfluidics for single-cell experiments, as it is employed not only for isolation but also plays a vital role in cell-analysis techniques. Flow cytometry Flow cytometry such as FACS can rapidly handle and sort (live or fixed) cells in a high-throughput manner and can be used for singlecell isolation [66]. Using a fluorescent reporter, individual cells can be separated by gene expression or cell type. Expression of a nuclear fluorescent reporter in tissue followed by disaggregation of the tissue and purification of fluorescent nuclei by FACS has aided in the purification of tissue-specific cells (BiTS) [67]. Although BiTS was created for isolating populations of cells of the same tissue for ChIP and Hi-C purposes, this method can be adapted for single-cell purification by FACS or microfluidics. Micromanipulation Micromanipulation tools can be used to hold and mechanically separate single cells, such as individual blastomeres of mouse embryos [68]. Typically, a micromanipulator uses the aid of microscopes and joysticks for fine movements of glass pipette tips that can be used, for example, to keep a cell in place or separate individual cells by aspiration. Microdissection Laser-capture microdissection can be used to cut out and extract a single cell from a tissue section using a laser [69]. Microfluidics Microfluidics platforms are a useful accompaniment to single-cell technologies as they have significantly increased the capacity to multiplex single-cell experiments for high-throughput analyses. The main advantages of microfluidics are the automation of highthroughput experiments, reduced amounts of samples and reagents, and lower variability, making single-cell experiments feasible [70]. Single-cell microfluidics systems can visualize gene expression in live cells and can accompany cell-lysate assays where the ability to contain reactions within a small volume is vital to minimize loss of sample. Thus microfluidics has made PCR and even high-throughput sequencing analyses possible from single cells [14,15,33]. Microfluidics has even enabled the integration of cell-sorting/compartmentalization, phenotype-observation, and lysate-based analyses [70–72]. For example, fluorescence analyzers can be coupled to microfluidics set-ups to sort cells, which are then lysed and analyzed all within the microfluidics device. These systems are highly

customizable such that they can be used with many different types of cells and for many applications [73]. The two main types of microfluidics set-up that are used for single-cell epigenetics analyses are chamber based and droplet based. In chamber-based microfluidics, a single liquid phase flows through a solid chamber comprising many spatially separated compartments, or microchannels, that allow the compartmentalization of cells [63,74]. Chamber-based microfluidics systems are especially suitable for both long-term live-cell imaging and single-cell lysate analysis. The continuous flow of media/buffer is laminar, allowing rapid and precise exchange of conditions with minimal mixing. Cells can be trapped in microchannels and grown in a monolayer to allow for tracking of cell lineages over several generations [75]. Chamber microfluidics designs have also enabled high-throughput screening with time-lapse microscopy of hundreds of individual live cells in >103 individual microchannels, allowing large-scale, simultaneous screening of mutants [76]. Depending on the function, microfluidics chambers can be simple in design or can contain sophisticated systems of valves for on-chip handling of single-cell capture and reagent flow [77]. These developments have facilitated single-cell multiplexed qRT-PCR and RNA-seq experiments. However, the current maximal number of compartments for single-cell analysis in a single device is approximately 4  104 and thus throughput remains lower than in droplet-based microfluidics systems. In droplet-based microfluidics, two immiscible liquids are used such that one liquid forms a droplet within the other continuous phase, which is used as a carrier through the microfluidics apparatus and maintains compartmentalization of the droplets [71,78]. Each droplet can encapsulate individual cells and different reagents can be mixed by fusing multiple input droplets together. Turbulent flow within droplets ensures fast mixing such that each droplet functions as an independent microreactor. Typically, the volume of a droplet ranges from a few nanoliters down to a few picoliters, thus conferring an advantage over chamber-based systems. Using droplet-based microfluidics to select for antibody-secreting cells, it was estimated that 106 mammalian cells could be analyzed within a single experiment over a span of several hours [78]. However, not all assays are suitable for droplet-based microfluidics, for example, long-term imaging of live cells with pedigree analysis. Since individual cells are suspended within a droplet, they can move, and analyzing these cells as they grow and divide is difficult. Also, because of the lack of continuous flow, reagents in the media can be consumed in a relatively short time. In addition, it is difficult to exchange liquids in the droplet if the assay requires washing steps. 5

TICB-1086; No. of Pages 12

Review

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

Flow cytometry e.g. FACS

-

Microdissecon e.g. laser capture

Micromanipulaon e.g. mechanical aspirator

Microfluidics e.g. chamber or droplet

+

ChIP-qPCR/seq: histone modificaons

Single-cell separaon/isolaon

pooled cells Subpopulaon analysis

Cell-lysate analysis

Live-cell analysis

qRT-PCR/RNA-seq: transcriptomics Hi-C: 3D chromosome conformaon RSMA/bisulfite seq: DNA methylaon locus-specific nucleosome mapping

DamID: chroman-associated factors Genome-wide nucleosome mapping

Fluorescent reporter for gene expression CiA: inducon of chroman modificaons Fabs/Histac: global distribuon of chroman modificaons m6 A-Tracer: tracking of protein-DNA interacons FLIM-FRET: chroman compacon TRENDS in Cell Biology

Figure I. Single-cell isolation techniques.

bisulfite sequencing (RRBS-seq) has been used with some success in single cells to determine cytosine methylation statuses at cytosine residues with an adjacent guanine (CpGs), which form the majority of DNA-methylation sites in vivo [30]. This method also uses bisulfite sequencing and thus has the same caveats; however, in contrast to wholegenome bisulfite sequencing, DNA is digested before conversion with a unique restriction enzyme that cuts CpGs regardless of methylation status. This ensures that CpGs are located at the end of digested fragments, reducing the amount of sequencing required. RRBS-seq has been used to compare methylation at CpGs in single haploid sperm cells and individual male and female pronuclei isolated from the same zygote at various stages, revealing faster demethylation in male pronuclei and demethylation of genic regions occurring first [30]. Another recent protocol also modifies traditional bisulfite treatment protocols for improvements in single-cell genome-wide methylation analysis by performing bisulfite conversion before sequencing adapter ligation, to minimize loss of sequenced fragments from adapter degradation during bisulfite treatment [31]. To avoid the limitations of bisulfite conversion, RSMA can assay DNA methylation with higher accuracy in single cells, but only at a limited number of loci. RSMA employs DNA methylation-sensitive restriction enzymes followed by PCR to detect the number of cleavage products and deduce whether DNA is methylated at a particular locus [32]. This assay has been used to probe single cells of mouse embryos for maintenance of DNA methylation, revealing that imprint chimerism due to methylation errors from aberrant epigenetic reprogramming at selected loci leads to unpredictable phenotypes and developmental arrest 6

[33]. Since DNA methylation can vary significantly between different organisms and can change during development, studies in single cells will provide new insight into DNA methylation dynamics. Nucleosome occupancy on single DNA molecules: locus-specific nucleosome mapping The positions of nucleosomes, the fundamental units of chromatin comprising histone-protein octamers wrapped by DNA, play an important role in gene expression, for example, by regulating DNA accessibility. Specifically positioned nucleosomes or nucleosome-free regions may be epigenetically inherited to maintain specific gene-expression patterns and therefore nucleosome mapping continues to be a widely used technique. There have been a few reports of nucleosome mapping on single stretches of DNA at specific loci [34–36] but these have not yet been extended genome wide from the same cell. These methods mostly rely on cytosine methylation by a GpC-specific methyltransferase in unprotected GpCs (i.e., in nucleosome-free parts of DNA), followed by bisulfite conversion of the remaining unmethylated (protected by nucleosomes) cytosines. The bisulfite-converted fragments can then be cloned and sequenced for nucleosome positioning at specific loci. Analysis of a single locus of the inducible Pho5 gene in individual yeast cells has revealed significant cell-to-cell variability for nucleosome positioning at this promoter as well as a correlation between nucleosome positioning and gene expression, providing insight into heterogeneous gene expression [34,36]. Future studies should focus on extending nucleosome mapping to genome-wide studies in single cells.

TICB-1086; No. of Pages 12

Review Chromatin analysis for localization of chromatinassociated factors and histone modifications: chromatin immunoprecipitation–quantitative PCR/sequencing (ChIP–qPCR/seq) and DNA adenine methyltransferase identification (DamID) ChIP–qPCR/seq is used to determine the presence of proteins or histone modifications at specific genomic sequences. In these methods, cells are lysed and antibodies against proteins or histone modifications of interest are used to purify the associated DNA, followed by PCR or NGS to identify genomic regions of localization. ChIP– qPCR and ChIP-seq are among the most widely used chromatin assays in population-based studies. However, their application in single cells is currently lacking because the efficiency of antibody-based immunoprecipitation is limited and the samples are not abundant enough to amplify. Currently, low cell-number ChIP-seq and ChIP– qPCR require populations of at least 104 or 103 cells, respectively, although further developments in microfluidics may be able to reduce this number [37]. Thus, cellspecific ChIP has so far been confined to either homogeneous subpopulations of sorted cells or isolated, single cells that give rise to clonal populations. DamID is another method to identify protein–DNA interactions. For this, the Dam enzyme is fused to the protein of interest to methylate adenines of DNA that come in proximity to this fusion [38], which can then be identified by NGS. DamID offers a single snapshot of the history of interactions between the protein of interest and DNA, while ChIP provides a snapshot at a single time point. Although DamID can have a high background from random interactions of Dam with DNA and lower resolution than ChIP, this method is useful when an antibody against the protein of interest is unavailable. Although this has not yet been reported, DamID has a higher potential to be applied to single cells since amplification of methylated DNA is possible by PCR, ensuring that enough signal is available for analysis; however, it is not a suitable approach for histone modifications, as modifications cannot be fused with the Dam protein. Locus-specific in situ hybridization methods In addition to cell-lysate analyses, microscopy-based visualization methods are useful for studying epigenetic mechanisms in single cells. These tools, which have been used for decades, include in situ hybridization (ISH) and immunofluorescence (IF) to allow the visualization of nucleic acids or proteins, respectively. To some extent, these methods have been replaced by the sequencing-based genome-wide techniques discussed above, because they are limited to visualizing a few specific molecules or interactions at a time. However, these methods are advantageous because single cells can be quickly isolated and fixed for microscopy or can be easily detected within a histological sample. Since data collection and analysis are often faster and simpler than the applications discussed above, many cells can be examined in a short time, providing statistically relevant numbers for analysis. We discuss ISH and IF below for use in epigenetic studies including a recently modified version of ISH that detects locus-specific histone modifications in single cells.

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

Fixed-cell imaging of nucleic acids and proteins/ chromatin modifications: fluorescent ISH (FISH) and IF DNA and RNA FISH take advantage of complementary oligos for the fluorescent labeling of specific DNA or RNA sequences in single fixed cells for visualization of their spatial organization or expression [39]. Classically, IF has been used to visualize the global distribution of chromatin modifications in single fixed cells. Various adaptations and combinations of ISH and/or IF are used to detect expression and (co)localization [39]. For example, ISH methods have been adapted to differentiate between methylated and unmethylated DNA, providing insight into DNA methylation patterns at major and minor satellite repeats and aberrant hypermethylation of cytosines in cancerous cells [40,41]. Furthermore, combined RNA–DNA-FISH can reveal differences in allelic expression, while RNA-FISH combined with IF can be used to correlate transcription or ncRNA localization with chromatin modifications. FISH/IF studies have played a critical role in understanding X inactivation; specifically, Xist RNA expression and X chromosome coating are followed by transcriptional silencing of X-linked genes in cis. RNA-FISH combined with dual IF revealed that the long ncRNA Xist, which coats the inactive X chromosome, colocalizes with the repressive histone modification H3K27me3, while RNA polymerase is largely excluded from these regions in single cells [39]. Fixed-cell imaging of histone modifications at specific loci: the ISH–proximity ligation assay (PLA) method As discussed above, single-cell analysis of histone modifications with ChIP is currently not possible. To bypass this limitation, a method to observe site-specific histone modifications in single cells of a fixed sample was developed by combining ISH with a PLA [42]. Biotinylated complementary probes are used for ISH at a specific DNA sequence and then antibodies against biotin and the histone modification of interest are combined with secondary antibodies that have DNA oligos attached to them. These oligos are ligated together when the antibodies bind in close proximity. Rolling-circle amplification then produces several hundred copies of the ligated oligos, which can be detected at single-cell resolution using fluorescent complementary oligos. The ISH–PLA method has allowed visualization of histone H3K4 dimethylation at specific promoters of active genes in single smooth muscle cells in mammalian tissue samples. Despite their utility, ISH and IF techniques do not allow time-lapse studies of the same cell because the cells must be fixed. However, these methods have been used to compare different single cells over time and through development; therefore, time-resolved variations of ISH–PLA may provide new insight into chromatin modifications in epigenetic processes. Real-time visualization methods Many reports have used the term ‘epigenetic’ for chromatin modifications that are correlated with a transcriptional state that could be maintained through cell division; however, few studies have been able to address the inheritance of locus-specific chromatin modifications and transcriptional states because current techniques are unable to detect chromatin modifications over time. Fulfilling this 7

TICB-1086; No. of Pages 12

Review criterion has been perhaps the most challenging, because most single-cell techniques require cell fixation or lysis. To truly establish an epigenetic phenomenon, there must be experimental evidence that a process occurs over time and through divisions, and this requires live-cell methods. In the past few years, methods to visualize chromatin modifications at the live, single-cell level have been published; however, these methods have been limited to visualizing global levels of modifications. Understanding epigenetic phenomena: microscopy using fluorescent reporters to detect gene expression Currently, live-cell visualization of chromatin modifications at a particular locus is impossible. To circumvent this problem, epigenetic processes can be monitored by time-lapse microscopy using fluorescence-based in vivo reporters such as GFP to visualize gene expression and inheritance of expression patterns in single live cells. A GFP-tagged sporulation factor in Bacillus subtilis revealed that epigenetic memory of sporulation factor expression is inherited by the next generations to promote faster spore differentiation in certain lineages [43]. To search for novel factors/modifications involved in the inheritance and regulation of gene expression, the underlying mechanism can be determined by either a bottom-up approach, which could use large-scale screening for mutants that disrupt or enhance the epigenetic process, or a top-down approach that first correlates the epigenetic phenomenon with potential transcription and epigenetic mechanisms and then tests selected candidates. This type of screening was used in a recent report in an important area of epigenetic research striving toward the controlled dedifferentiation, or ‘reprogramming’, of differentiated cells into pluripotent cells [44]. Using a Nanog-GFP reporter, which is expressed in pluripotent but not differentiated cells, a small interfering RNA (siRNA)-based knockdown screen with fluorescence-activated cell-sorting (FACS) analysis was performed (Box 1) for factors that increase the expression of Nanog-GFP in individual cells, which positively correlates with epigenetic reprogramming. The screen identified Mbd3, a component of the NuRD complex that deacetylates histones and remodels chromatin to regulate gene repression, providing further evidence of the importance of epigenetic chromatin marks in development. Live-cell imaging: chromatin in vivo assay (CiA) for induction of chromatin modifications to initiate gene silencing The recently developed CiA is one of a few techniques that addresses the dynamics of histone modifications and their effects on transcriptional states in vivo [45]. Based on chemically induced proximity, this breakthrough technique allows the initiation of genomic locus-specific and temporally controlled recruitment of chromatin-associated factors to induce specific chromatin states, allowing the examination of histone modification patterns and downstream gene expression. With this approach, Heterochromatin Protein 1 (HP1), a known interactor with histone methyltransferases, was recruited to a specific promoter, resulting in silencing of a fluorescent reporter in single 8

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

cells. Furthermore, this silencing was inherited through multiple generations. In addition, the recruitment of HP1 and transcriptional silencing correlated with H3K9 trimethylation at this locus detected by ChIP–qPCR of populations of cells. The method also revealed that the gradual acquisition of DNA methylation helped to stably maintain repression and silencing. As CiA has only recently been developed, its application in inducing other chromatin modifications at specific sites is appealing. This technique would provide further characterization of the sufficiency of modifications in epigenetic mechanisms for single-cell studies in vivo. Live-cell imaging of chromatin modifications: fluorescently labeled specific antigen-binding fragments (Fabs) and Histacs Although ISH–PLA and CiA have made progress toward locus-specific visualization of histone modifications, they are not yet suitable for direct visualization of histone modifications at specific loci with live single-cell imaging. ISH–PLA requires cells to be fixed, while CiA can provide only indirect evidence of histone modifications since the readout is a transcriptional reporter and not the modification in the same cell at the targeted locus. Techniques to visualize histone modifications at the live, single-cell level remain limited, with the most common techniques being used only to monitor the global distribution of histone modifications with fluorescent antibodies or proteins that bind to the modification (‘readers’). Fabs can be introduced into live cells either mechanically (beads; e.g., for cells in culture) or by injection (e.g., in mouse embryos) to label endogenous histone modifications [46]. However, Fabs are diluted during cell division, making long-term imaging of these cells infeasible. To overcome this limitation, modification-binding single-chain antibody fragments (chromobody, mintbody) or modification-specific readers can be expressed constitutively in vivo as fusions with fluorescent proteins (e.g., GFP) [47–49] such that division events do not affect the levels of the fluorescent proteins because all cells are constantly expressing new protein. However, this approach can result in a high signal-to-noise ratio due to fluorescence from expressed but unbound fluorescent proteins. To circumvent the high background, real-time measurements of histone modifications can be made with Fo¨rster resonance-energy transfer (FRET)-based probes called Histacs, which are used to visualize the acetylation status of histone H4. Histacs are expressed as histone H4 fusions with modification-specific bromodomains, both of which are linked to FRET partners. The acetylation status of H4 determines the efficiency of FRET, such that when H4 is unmodified the FRET pair is in close proximity and produces a FRET signal. However, when the bromodomain binds to acetylated H4, FRET efficiency is reduced due to a conformational change in the fusion protein that increases the distance between the FRET partners. Specificity for certain acetylation sites is governed by the bromodomain used. For example, the first report of a Histac used the BRDT protein, which binds to acetylated histone H4 at K5 and K8 [50], whereas a subsequent report used the BRD2 bromodomain for specificity toward acetylated H4K12

TICB-1086; No. of Pages 12

Review [51]. This technique can be applied to visualize the dynamics of H4 acetylation in vivo over time since the reader affinity for binding is dependent on the acetylation status. While this methodology offers in vivo expression and reversible, low-background advantages, it also requires expression of nonendogenous histones. Since epigenetic inheritance of histone-acetylation patterns is controversial, techniques that offer live, single-cell visualization through cellular division in single cells will be useful to address these questions, but will require further enhancements to study the presence of these modifications at specific genomic loci. Live-cell tracking of protein–DNA interactions: m6ATracer technology The recently developed m6A-Tracer technology builds on the DamID method [52], allowing visualization of protein–DNA interactions in vivo by fluorescence microscopy of a DNA methylation-specific reader. The m6A-Tracer is a methyladenine-binding domain–GFP fusion protein expressed in cells that also express the protein of interest fused to Dam, to label interacting DNA with methylated adenine (m6A). This approach has been used to follow genomic regions that interact with the nuclear lamina (NL), also known as LADs. A lamin–Dam fusion was expressed in cells with the m6A-Tracer to visualize the history of LAD association with the NL via time-lapse microscopy in single, live cells. Further ChIP studies of cell populations confirmed that these LADs correspond to heterochromatin marked with histone H3K9 dimethylation. The m6A-Tracer is currently the only technique suitable for following the history of protein–DNA interactions over time in an individual cell. The m6A-Tracer could be used in various combinations of Dam fusions to quantify interactions between DNA and chromatin regulators or their association with specific nuclear structures such as the nucleolus, chromocenters, or nuclear bodies. Chromatin compaction in single cells: histone multiphoton fluorescence-lifetime imaging microscopy (FLIM)–FRET Densely packed nucleosomes can abrogate binding by transcription factors and components of the RNA polymerase complex. Visualizing chromatin compaction in live cells has mostly relied on fluorescent reporters. A FLIM–FRET technique has been used to directly visualize chromatin compaction in single living cells [53]. Histone H2B was fused with GFP (donor) or mCherry (acceptor) such that FRET occurred when H2B molecules from adjacent nucleosomes were in close proximity due to chromatin compaction. However, visualization by fluorescence microscopy could not distinguish between the free fluorophore signal and the FRET signals; thus, proximity was measured by FLIM, the fluorescence lifetime of the donor fluorophore, which decreases when FRET occurs between the donor and its acceptor when they are in close proximity [53]. FLIM revealed that interphase chromatin is not found in two distinct levels of compaction (i.e., heterochromatin and euchromatin) but rather exists as a spectrum of compaction. This method can so far provide information only on

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

the global level of chromatin compaction in a cell and thus is likely to be most useful for testing the effects of different treatments on global levels of gene expression. Single-cell epigenetics: the next generation To truly label a phenomenon as epigenetic requires the ability to follow it through cellular division. Currently, most single-cell experimental techniques provide snapshots of transcription profiles or chromatin states at a specific time because they depend on cell lysis. To overcome these limitations, individual cells can be analyzed at several time points. While this has produced interesting results, some of the conclusions are, ultimately, population-averaged data because each time point is averaged with other time points from different cells. Most persistence and hereditary information is lost and dynamic events such as transcriptional bursting or stochastic switching are overlooked. To fully understand epigenetic mechanisms, we propose developing integrated approaches to live-cell visualization of locus-specific chromatin and RNA dynamics in single cells. While some techniques have been adapted for live single cells, very few are able to observe these effects at specific loci. Although it has not yet been reported, we can envision the further development of current methods toward locus-specific epigenetic studies. For example, Cas9 proteins and transcription activator-like effectors (TALEs) have enabled genomic locus targeting [54,55]; therefore, it may be possible to utilize a FRET pair [56] harboring a nuclease-deficient Cas9/TALE for specific locus targeting and a Fab/chromobody/mintbody [46–48] or reader [49] for modification binding to achieve FRET when a particular modification is found at the specified locus. This method would allow dynamic, locus-specific, time-resolved visualization of a chromatin modification in vivo (Figure 3). However, a major limitation may be the low sensitivity of FRET, which would probably not provide enough signal for single-locus, single-cell visualization. For a Cas9-based system, this could be overcome with tiled arrays of singleguide RNAs (sgRNAs), which direct Cas9 to its genomic locus, across the locus of interest [57]. Another possibility would be to use fluorescent protein complementation (e.g., split GFP) with one half fused to Cas9/TALE and the other containing a modification-binding fragment or modification reader. However, fluorescent protein halves often have high affinity for each other, making them interact irreversibly, and are therefore unsuitable for visualizing dynamic interactions [58]. As each technique has its own caveats, pursuing novel methods will be important for studying chromatin-associated factor and RNA-based epigenetic inheritance mechanisms in single live cells. These methods will be particularly important for epigenetic studies as they will allow the tracking of locus-specific chromatin-based mechanisms through multiple generations. For example, a combination of live-cell analytical techniques as proposed above and lineage markers could track the inheritance of transcriptional and/or chromatin states in a pedigree of cell lineages to approach a detailed description of the stability, variability, and inheritance of epigenetic processes as in Figure 1. 9

TICB-1086; No. of Pages 12

Review

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

FRET e.g. CFP

e.g. YFP Chroman modificaon-specific anbody fragment e.g. Fabs/chromobodies/mintbodies

Cas9

m sgRNA

Genomic locus target

Fluorescent protein complementaon e.g. split GFP

me Chroman modificaon “reader”

TRENDS in Cell Biology

Figure 3. Proposed method to visualize locus-specific chromatin modifications in single live cells. Cas9 fusion with a fluorescent protein could be directed to a genomic sequence by a synthetic single-guide RNA (sgRNA) and a fluorescent signal could be observed when there is fluorescent protein complementation (e.g., split GFP) or Fo¨rster resonance energy transfer (FRET) with, for example, modification-specific antibodies (e.g., Fabs, chromobodies, mintbodies).

Concluding remarks The epigenetics field is influenced by the current trends for larger scale and higher throughput, but also by the need to readdress these questions at the level of individual cells to understand epigenetic and chromatin mechanisms. Hence, the field finds itself breaking ground in single-cell epigenetics. Ultimately, the benefit of single-cell epigenetic analyses will be the ability to simultaneously address the inheritance of specific chromatin states or epigenetic signatures at specific loci as one cell divides into two and to clarify the stability through generations. This information will allow a more detailed understanding of epigenetic mechanisms, for example, by identifying factors responsible for establishing and maintaining epigenetic processes. Thus, single-cell analyses are at the forefront of epigenetics research. Single-cell epigenetics and chromatin research has not only addressed current topics in the field, but has also raised new questions. Single-cell data have shown that there is variation in gene expression among genetically identical single-cell organisms and even phenotypically similar cells from the same cell lineage in multicellular organisms [59,60]. These differences have mechanistically been linked to ncRNA, chromatin architecture, and chromosome dynamics [4–7]. It will be interesting to see how these variabilities affect epigenetic inheritance and whether they are an important epigenetic mechanism. Single-cell epigenetics during the development of multicellular organisms is intriguing, as epigenetic reprogramming is required to give rise to a totipotent zygote [61]. Yet, transgenerationally inherited epigenetic traits/ modifications must bypass this reprogramming. Studies of 10

transgenerational inheritance are tracing epigenetic mechanisms from somatic cells in parents to individual germ cells as they mature and become fertilized, and further in individual cells within an embryo as it goes through developmental transitions from totipotent to pluripotent to differentiated cells, until germ cells are formed in the new organism. Single-cell epigenetic studies are likely to also have clinical significance. It is increasingly clear that many diseases result from a series of transitions. Disease states are often accompanied by changes to the transcriptional profile. Changes in gene expression in individual cells that give rise to clonal populations of cells might be epigenetically inherited and might play a major role in disease progression (e.g., privileged subpopulations in certain tumors that drive tumor propagation). Anticipating the epigenetic transitions of individual cells from a normal to a diseased state would have clear diagnostic value, especially if those epigenetic changes precede any disease phenotype. Thus, identifying disease- and development-associated epigenetic modifications in single-cell epigenetic analyses in vivo would aid in understanding disease processes. Acknowledgments The authors thank Sandra Bour and Elodie Legrand for the graphic designs and the R.S. laboratory for helpful discussions. P.B. is a Marie Curie Fellow [International Incoming Fellowship (IIF)]. Work in the R.S. laboratory is supported by the Fondation pour la Recherche Me´dicale, the Agence Nationale de Recherche (CoreAc), La Ligue National Contre La Cancer (Equipe Labellise), and a European Research Council (ERC) starting grant. The authors apologize to those whose work they were unable to cite due to space limitations.

TICB-1086; No. of Pages 12

Review References 1 Elgin, S.C. and Reuter, G. (2013) Position-effect variegation, heterochromatin formation, and gene silencing in Drosophila. Cold Spring Harb. Perspect. Biol. 5, a017780 2 Saxena, A. and Carninci, P. (2011) Long non-coding RNA modifies chromatin: epigenetic silencing by long non-coding RNAs. Bioessays 33, 830–839 3 Foster, S.L. et al. (2007) Gene-specific control of inflammation by TLRinduced chromatin modifications. Nature 447, 972–978 4 Collins, L.J. et al. (2011) The epigenetics of non-coding RNA. In Handbook of Epigenetics (Tollefsbol, T.O., ed.), pp. 49–61, Elsevier 5 Rothbart, S.B. and Strahl, B.D. (2014) Interpreting the language of histone and DNA modifications. Biochim. Biophys. Acta. 1839, 627– 643 6 Petruk, S. et al. (2012) TrxG and PcG proteins but not methylated histones remain associated with DNA through replication. Cell 150, 922–933 7 Misteli, T. (2013) The cell biology of genomes: bringing the double helix to life. Cell 152, 1209–1212 8 Kouzarides, T. (2007) Chromatin modifications and their function. Cell 128, 693–705 9 Daxinger, L. and Whitelaw, E. (2010) Transgenerational epigenetic inheritance: more questions than answers. Genome Res. 20, 1623– 1628 10 Heard, E. and Martienssen, R.A. (2014) Transgenerational epigenetic inheritance: myths and mechanisms. Cell 157, 95–109 11 Sanchez-Freire, V. et al. (2012) Microfluidic single-cell real-time PCR for comparative analysis of gene expression patterns. Nat. Protoc. 7, 829–838 12 White, A.K. et al. (2011) High-throughput microfluidic single-cell RTqPCR. Proc. Natl. Acad. Sci. U.S.A. 108, 13999–14004 13 Tang, F. et al. (2010) RNA-seq analysis to capture the transcriptome landscape of a single cell. Nat. Protoc. 5, 516–535 14 Wu, A.R. et al. (2014) Quantitative assessment of single-cell RNAsequencing methods. Nat. Methods 11, 41–46 15 Burton, A. et al. (2013) Single-cell profiling of epigenetic modifiers identifies PRDM14 as an inducer of cell fate in the mammalian embryo. Cell Rep. 5, 687–701 16 Fatica, A. and Bozzoni, I. (2014) Long non-coding RNAs: new players in cell differentiation and development. Nat. Rev. Genet. 15, 7–21 17 Dekker, J. et al. (2002) Capturing chromosome conformation. Science 295, 1306–1311 18 Zhao, Z. et al. (2006) Circular chromosome conformation capture (4C) uncovers extensive networks of epigenetically regulated intra- and interchromosomal interactions. Nat. Genet. 38, 1341–1347 19 Simonis, M. et al. (2006) Nuclear organization of active and inactive chromatin domains uncovered by chromosome conformation captureon-chip (4C). Nat. Genet. 38, 1348–1354 20 Dostie, J. et al. (2006) Chromosome Conformation capture carbon copy (5C): a massively parallel solution for mapping interactions between genomic elements. Genome Res. 16, 1299–1309 21 van Berkum, N.L. et al. (2010) Hi-C: a method to study the threedimensional architecture of genomes. J. Vis. Exp. http://dx.doi.org/ 10.3791/1869 22 Belton, J.M. et al. (2012) Hi-C: a comprehensive technique to capture the conformation of genomes. Methods 58, 268–276 23 Nora, E.P. et al. (2012) Spatial partitioning of the regulatory landscape of the X-inactivation centre. Nature 485, 381–385 24 Tsai, C.L. et al. (2008) Higher order chromatin structure at the Xinactivation center via looping DNA. Dev. Biol. 319, 416–425 25 Sexton, T. et al. (2012) Three-dimensional folding and functional organization principles of the Drosophila genome. Cell 148, 458–472 26 Jin, F. et al. (2013) A high-resolution map of the three-dimensional chromatin interactome in human cells. Nature 503, 290–294 27 Nagano, T. et al. (2013) Single-cell Hi-C reveals cell-to-cell variability in chromosome structure. Nature 502, 59–64 28 Hernandez, H.G. et al. (2013) Optimizing methodologies for PCR-based DNA methylation analysis. Biotechniques 55, 181–197 29 Laird, P.W. (2010) Principles and challenges of genomewide DNA methylation analysis. Nat. Rev. Genet. 11, 191–203 30 Guo, H. et al. (2013) Single-cell methylome landscapes of mouse embryonic stem cells and early embryos analyzed using reduced representation bisulfite sequencing. Genome Res. 23, 2126–2135

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

31 Smallwood, S.A. et al. (2014) Single-cell genome-wide bisulfite sequencing for assessing epigenetic heterogeneity. Nat. Methods 11, 817–820 32 Kantlehner, M. et al. (2011) A high-throughput DNA methylation analysis of a single cell. Nucleic Acids Res. 39, e44 33 Lorthongpanich, C. et al. (2013) Single-cell DNA-methylation analysis reveals epigenetic chimerism in preimplantation embryos. Science 341, 1110–1112 34 Jessen, W.J. et al. (2006) Active PHO5 chromatin encompasses variable numbers of nucleosomes at individual promoters. Nat. Struct. Mol. Biol. 13, 256–263 35 Kelly, T.K. et al. (2012) Genome-wide mapping of nucleosome positioning and DNA methylation within individual DNA molecules. Genome Res. 22, 2497–2506 36 Small, E.C. et al. (2014) Single-cell nucleosome mapping reveals the molecular basis of gene expression heterogeneity. Proc. Natl. Acad. Sci. U.S.A. 111, E2462–E2471 37 Gilfillan, G.D. et al. (2012) Limitations and possibilities of low cell number ChIP-seq. BMC Genomics 13, 645 38 van Steensel, B. and Henikoff, S. (2000) Identification of in vivo DNA targets of chromatin proteins using tethered Dam methyltransferase. Nat. Biotechnol. 18, 424–428 39 Chaumeil, J. et al. (2008) Combined immunofluorescence, RNA fluorescent in situ hybridization, and DNA fluorescent in situ hybridization to study chromatin changes, transcriptional activity, nuclear organization, and X-chromosome inactivation. Methods Mol. Biol. 463, 297–308 40 Li, Y. et al. (2013) Sequence-specific microscopic visualization of DNA methylation status at satellite repeats in individual cell nuclei and chromosomes. Nucleic Acids Res. 41, e186 41 Nuovo, G.J. et al. (1999) In situ detection of the hypermethylationinduced inactivation of the p16 gene as an early event in oncogenesis. Proc. Natl. Acad. Sci. U.S.A. 96, 12754–12759 42 Gomez, D. et al. (2013) Detection of histone modifications at specific gene loci in single cells in histological sections. Nat. Methods 10, 171–177 43 Veening, J.W. et al. (2008) Bet-hedging and epigenetic inheritance in bacterial cell development. Proc. Natl. Acad. Sci. U.S.A. 105, 4393– 4398 44 Rais, Y. et al. (2013) Deterministic direct reprogramming of somatic cells to pluripotency. Nature 502, 65–70 45 Hathaway, N.A. et al. (2012) Dynamics and memory of heterochromatin in living cells. Cell 149, 1447–1460 46 Hayashi-Takanaka, Y. et al. (2011) Tracking epigenetic histone modifications in single cells using Fab-based live endogenous modification labeling. Nucleic Acids Res. 39, 6475–6488 47 Rothbauer, U. et al. (2006) Targeting and tracing antigens in live cells with fluorescent nanobodies. Nat. Methods 3, 887–889 48 Sato, Y. et al. (2013) Genetically encoded system to track histone modification in vivo. Sci. Rep. 3, 2436 49 Yamagata, K. (2010) DNA methylation profiling using live-cell imaging. Methods 52, 259–266 50 Sasaki, K. et al. (2009) Real-time imaging of histone H4 hyperacetylation in living cells. Proc. Natl. Acad. Sci. U.S.A. 106, 16257–16262 51 Sasaki, K. et al. (2012) Development of live-cell imaging probes for monitoring histone modifications. Bioorg. Med. Chem. 20, 1887–1892 52 Kind, J. et al. (2013) Single-cell dynamics of genome–nuclear lamina interactions. Cell 153, 178–192 53 Lleres, D. et al. (2009) Quantitative analysis of chromatin compaction in living cells using FLIM–FRET. J. Cell Biol. 187, 481–496 54 Doyle, E.L. et al. (2013) TAL effectors: highly adaptable phytobacterial virulence factors and readily engineered DNA-targeting proteins. Trends Cell Biol. 23, 390–398 55 Jinek, M. et al. (2012) A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 337, 816–821 56 Truong, K. and Ikura, M. (2001) The use of FRET imaging microscopy to detect protein–protein interactions and protein conformational changes in vivo. Curr. Opin. Struct. Biol. 11, 573–578 57 Chen, B. et al. (2013) Dynamic imaging of genomic loci in living human cells by an optimized CRISPR/Cas system. Cell 155, 1479–1491 58 Remy, I. and Michnick, S.W. (2007) Application of protein-fragment complementation assays in cell biology. Biotechniques 42, 137 139, 141 passim 11

TICB-1086; No. of Pages 12

Review 59 Raser, J.M. and O’Shea, E.K. (2005) Noise in gene expression: origins, consequences, and control. Science 309, 2010–2013 60 Torres-Padilla, M.E. and Chambers, I. (2014) Transcription factor heterogeneity in pluripotent stem cells: a stochastic advantage. Development 141, 2173–2181 61 Torres-Padilla, M.E. (2013) Generating different epigenotypes. Reprod. Biomed. Online 27, 624–628 62 Dekker, J. et al. (2013) Exploring the three-dimensional organization of genomes: interpreting chromatin interaction data. Nat. Rev. Genet. 14, 390–403 63 Mensaert, K. et al. (2014) Next-generation technologies and data analytical approaches for epigenomics. Environ. Mol. Mutagen. 55, 155–170 64 Sims, D. et al. (2014) Sequencing depth and coverage: key considerations in genomic analyses. Nat. Rev. Genet. 15, 121–132 65 Miyanari, Y. et al. (2013) Live visualization of chromatin dynamics with fluorescent TALEs. Nat. Struct. Mol. Biol. 20, 1321–1324 66 Piyasena, M.E. and Graves, S.W. (2014) The intersection of flow cytometry with microfluidics and microfabrication. Lab Chip 14, 1044–1059 67 Bonn, S. et al. (2012) Cell type-specific chromatin immunoprecipitation from multicellular complex samples using BiTS–ChIP. Nat. Protoc. 7, 978–994 68 Behringer, R. et al. (2014) Disaggregating cleavage-stage embryos and the inner cell mass of blastocysts into individual cells. In Manipulating

12

Trends in Cell Biology xxx xxxx, Vol. xxx, No. x

69 70 71 72 73

74 75 76

77 78

the Mouse Embryo: A Laboratory Manual. (4th edn), pp. 547–548, CSH Press Decarlo, K. et al. (2011) Laser capture microdissection: methods and applications. Methods Mol. Biol. 755, 1–15 Yin, H. and Marshall, D. (2012) Microfluidics for single cell analysis. Curr. Opin. Biotechnol. 23, 110–119 Tian, W.C. and Finehout, E., eds (2008) Microfluidics for Biological Applications, Springer Autebert, J. et al. (2012) Microfluidic: an innovative tool for efficient cell sorting. Methods 57, 297–307 Crane, M.M. et al. (2010) Microfluidics-enabled phenotyping, imaging, and screening of multicellular organisms. Lab Chip 10, 1509–1517 Mehling, M. and Tay, S. (2014) Microfluidic cell culture. Curr. Opin. Biotechnol. 25, 95–102 Charvin, G. et al. (2010) Long-term imaging in microfluidic devices. Methods Mol. Biol. 591, 229–242 Denervaud, N. et al. (2013) A chemostat array enables the spatiotemporal analysis of the yeast proteome. Proc. Natl. Acad. Sci. U.S.A. 110, 15842–15847 Unger, M.A. et al. (2000) Monolithic microfabricated valves and pumps by multilayer soft lithography. Science 288, 113–116 Mazutis, L. et al. (2013) Single-cell analysis and sorting using dropletbased microfluidics. Nat. Protoc. 8, 870–891