Toward High-Throughput and Multiplexed Imaging of Genome Organization

Introduction

Eukaryotic genomes function as a three-dimensional (3D) chromatin polymer, driven by a complex collection of chromosome interactions. Enhancers, in particular, are often located kilobases (Kb) or even megabases (Mb) away from their target promoters and require chromatin looping to be brought together for transcription initiation. ^{1 –4} The emerging view is that interactions between regulatory elements and promoters further organize into extensive intrachromosomal units termed topologically associated domains (TADs), which engage in long-range interactions with loci in other TADs to package chromosomes into discrete territories. ^{5 –7} Importantly, the spatial organization of genomic loci has been linked to numerous nuclear functions, including transcription, replication, DNA repair, and genome stability. ^{8 –11} Gene positioning itself may also present an explanation for the high frequency of specific translocations that occur in the progression of several different cancers. ¹² Thus, studies of the genome in the context of nuclear architecture are increasing our understanding of genome function in normal development and disease. Still, we remain tremendously limited in our ability to understand the relationship between chromosome arrangement and function. Foremost among our needs are technologies that will permit us to visualize genome-wide chromosome arrangement, nucleus by nucleus, and in a high-throughput manner, to identify the genes that drive the spatial organization of the genome.

The Technical Challenge

One of the most challenging aspects of studying higher order chromosome organization is the labor-intensive nature of the two major approaches: imaging, typically using fluorescence in situ hybridization (FISH), ¹³ and chromosome conformation capture (3C)-based methods. ¹⁴ Both approaches have advantages and limitations. Although a major strength of FISH is its ability to probe the location and physical proximity at single-cell resolution, a shortcoming is its very low throughput nature, which typically permits only a handful of loci to be visualized in a single experiment, thus precluding systematic genome-wide analysis. This limitation is complemented by 3C methods, which rely on chemical cross-linking of interactions throughout the genome, allowing systematic, genome-wide mapping of chromatin interactions in a single experiment. ^15,16 However, both methods are unsuitable for use in unbiased, large-scale screens because of the low throughput nature of FISH and the technical complexity of 3C approaches. These weaknesses has thus stymied any attempt to systematically identify the factors that govern the spatial organization of the genome, forcing researchers to rely almost exclusively on candidate or indirect approaches. Discussed hereunder, we have addressed this gap by developing high-throughput FISH (Hi-FISH), a technology that permits chromosomal positions to be directly visualized in 384-well plates with high accuracy.

Hi-Fish: a Method to Screen for Factors That Directly Regulate Chromosome Interactions

High-content screening (HCS ^17,18 ) and high-throughput imaging ^19,20 approaches bridge the gap between conventional FISH and 3C-based technology. These strategies consist of automated imaging of thousands of samples, automated image analysis, and statistical quantification of multiple localization-based parameters. Hi-FISH brings the benefits of HCS to studies of genome organization by creating a fully automated FISH-based imaging pipeline to quantitatively determine the position and interaction frequency of multiple loci in the nucleus. ²¹ This efficiency is achieved through a protocol that permits FISH to be conducted in multiple 384-well plates simultaneously while reducing the number of manipulations. A similar high-throughput pipeline, HIPMap, ²² has also recently been described. Using these approaches, a large number of cells in thousands of samples can be systematically analyzed, combining the advantages of single-cell data from imaging with the capacity of 3C methods to interrogate genome organization at a large scale.

We have shown the suitability of Hi-FISH for screening by conducting the first FISH-based whole-genome RNAi screen. ²¹ Specifically, we aimed to identify factors important for somatic homolog pairing in Drosophila, using their well-established capacity to form these interchromosomal interactions as a model for chromosome interactions in general. ²³ Importantly, FISH probes were targeted to the repeated sequences of the pericentromeric heterochromatin, which make ideal FISH targets because of their great abundance, simplicity, and chromosome specificity. ²⁴ Collectively, the genome-wide screen consisted of ∼50,000 FISH assays and isolated 105 cellular factors, many of which not previously implicated in genome organization. Our results further revealed candidates with opposing roles, both pairing promoters and antipairing factors, shifting our viewpoint of chromosome interactions toward a more dynamic process. In particular, we and others have isolated the condensin II complex as a major organizing factor that antagonizes chromosome pairing and heterochromatin clustering during interphase. ^21,25 The identification of condensin II as an antipairing factor is in line with the intrachromosomal functions of compaction and chromatin looping being a mechanism by which long-range interactions, such as pairing, are inhibited. ^21,26,27 Consistent with this model, 3C-based studies have found that depletion of condensin, or other architectural proteins, such as cohesin, often shows that long- and short-range chromosomal contact frequencies are inversely correlated. ^{28 –33} Thus, many types of long-range interactions, including pairing, may be precluded by the formation, size, and/or density of small chromatin loops. ^{21,26 –28}

Taken together, our data demonstrate the feasibility of Hi-FISH and argue for a parallel technology for targeting single-copy euchromatic regions, which would be applicable to the study of a wide variety of aspects of nuclear organization. Here, the challenge lies in obtaining FISH signals from euchromatin that are as strong as those obtained from heterochromatin, a goal that would require hundreds or thousands of probe species, either individually synthesized or generated from cloned chromosomal fragments. In brief, the difficulty of generating probes to single-copy regions, compounded with the quantities of probe required for high-throughput assays, has been prohibitive in terms of labor and cost. These challenges, however, are being overcome by Oligopaint FISH probes, ³⁴ which are especially attractive for use in high-throughput approaches. Oligopaint probes are made from PCR-amplifiable oligomers derived from custom-designed microarrays, a strategy that gives precise control over target sequences and allows for single and multicolor imaging of regions ranging from a few Kb to several Mb with the same basic protocol. ³⁴ Our recent work has greatly reduced the time and cost of generating Oligopaints, ³⁵ and has shown that these probe libraries are amenable to a modified version of our Hi-FISH protocol. ³⁴ This brings probe costs for large-scale studies within reason, providing the feasibility of Oligopaints-based whole-genome screening. Potential applications of Hi-FISH and Oligopaints are discussed hereunder.

Future Directions for Oligopaints-Based Hi-Fish

Oligopaint libraries can either be directly labeled with fluorophores or designed to include adapter sequences that can be used as secondary hybridization sites for multiplexing the number of genomic targets detected from a single library. ^35,36 Such an approach, in combination with Hi-FISH, will enable multiplexed screening of large sets of genomic targets in a single experiment. Indeed, Hi-FISH could become a valuable tool to validate 3C-based interaction data. ³⁷ For example, Hi-FISH could be used in combination with Oligopaints to measure 3D distance distributions between thousands of pair-wise loci at varying genomic distances. Interactions of similar signal strength by 3C-based could then be systematically analyzed at the single-cell level to determine how signal strength and interaction frequencies are related.

Multicolor and whole-chromosome Oligopaint FISH also provides an opportunity to systematically detect rare chromosomal breakage and translocation frequencies in a cell population, which can be scaled in a high-throughput manner using Hi-FISH. An example of such an application has recently been described, revealing an association between translocation frequencies and the presence of specific histone modifications. ^38,39 Hi-FISH can also be combined with immunofluorescence (IF) to measure the proximity of genomic loci to nuclear structures such as the nucleolus or lamina. This would test whether particular nuclear structures promote the clustering of loci and whether gene activity is linked to the position of loci relative to those structures. In addition, combined Hi-FISH and IF will permit the frequency of specific and/or rare interactions to be measured in subpopulations of cells defined by specific protein markers. Finally, coupling Hi-FISH with multiplexed RNA FISH protocols ^40,41 would provide an allele-specific, high-throughput view of genome organization and transcriptional activity. For instance, direct analysis of both the gene expression profile and the long-range chromatin interaction frequency can provide insights into the relationship between stochastic gene expression and cell-to-cell variability of chromosome conformation. ⁴²

In conclusion, Hi-FISH is a highly modular technology that combines the power of single-cell imaging with the ability to interrogate genomes at a large scale. The combination of Hi-FISH with the development of complex probe libraries using Oligopaints will further enhance the impact of the method, which can be widely applied to the analysis of chromosome positioning and nuclear organization.