Advances in ChIP-Seq analysis with small samples

A look at the advances made in ChIP-seq analysis with small samples, including ChIP-seq using limited samples, optimizing amplification for small ChIP samples and reducing samples needed for ChIP-seq.

Given the heterogeneity that exists in tumors and other cell types, researchers continue to seek experimental techniques that enable efficient analysis of small, homogeneous samples.  

In 2007 epigenetic analysis was revolutionized by ChIP-seq. It delivered immense benefits, including higher resolution, broader genome coverage, fewer artifacts, and dramatically reduced the cost per interaction profiled, but it required a significant amount of input sample (~10 million cells) (Barski et al, 2007).

Such amounts of material are challenging to acquire when studying specific cell-types, sub-populations, or when attempting to use clinical samples. Fortunately, a number of improvements have been made to the ChIP methodology, pre-amplification approaches, and sequencing library construction reagents that make working with small sample amounts much more feasible.

Today, running comprehensive ChIP-seq experiments with 1 million cells is fairly routine, but groups continue to push the limits of technology to squeeze out even better performance. Here we discuss several recent advancements that continue to reduce the sample input for ChIP sequencing studies.

N-ChIP shines with limited samples

ChIP-seq using limited cells is a challenge because the immunoprecipitated DNA recovered often has reduced complexity, which may adversely affect sensitivity and reproducibility.

This creates a need for immunoprecipitation approaches that are optimized for low input settings.

Attempts to develop low cell ChIP-seq protocols using the more common, formaldehyde cross-linked chromatin (X-ChIP) have yielded encouraging results; based on standard Illumina library prep procedures, ChIP-seq was successfully performed from 1 million human stem cells (Hitchler et al, 2011).

However, while X-ChIP is especially well suited for investigating weakly binding transcription factors or chromatin associated proteins, the cross-linking process makes it much less effective when limited sample material is involved. Some researchers have turned to Native ChIP (N-ChIP) approaches as a way around those shortcomings.

N-ChIP eliminates the need for cross-linking, which can deliver better resolution and reduced non-specific interactions, as compared to X-ChIP, (O’Neil et al, 2003). This can be a major advantage when attempting to work with small cell numbers.

Recently, a highly optimized N-ChIP method was applied to genome-wide analysis by ChIP-seq (Gilfillan et al, 2012) and demonstrated very reliable results using 100,000 cells per immunoprecipitation (IP). The low input oriented N-ChIP implemented a number of tweaks to the standard immunoprecipitation protocol.

It begins with the dilution of chromatin in concentrated IP buffer, eliminating the dialysis step, which reduces handling and sample loss. Next, the addition of a specialized sonication step using TPX plastic ware minimizes sample loss and contamination, and ensures consistent, unbiased fragmentation.

Then, an Agilent 2100 Bioanalyzer is used to monitor the subsequent MNase digestion, in order to visualize the tiny amounts. After digestion, the Illumina library construction was adjusted to capture nucleosome-sized fragments based on size selection. Finally, they swapped out column-based cleanup with SPRI-beads for better DNA retention at each step.

Although the optimized N-ChIP requires more cells than some other methods, it minimizes the need for amplification and is particularly useful when investigating histone modifications throughout the genome.

According to investigators, further enhancements are likely possible by increasing the efficiency and robustness of the immunoprecipitation, DNA purification, and sequencing library construction protocols.

Optimizing amplification for small ChIP samples

Once chromatin immunprecipitation is complete, the next stage in the process that is ripe for optimization is amplification of the ChIP’ed material. Standard library preparation protocols used to make ChIP DNA ready for next-generation sequencing involve several purifications and inefficient enzymatic steps, which can lead to sample loss.

In addition, the sample recovery from ChIP is often very low, particularly if the input sample heading into the ChIP reaction is minimal. Therefore it is critical to obtain enough material from the ChIP reaction for the subsequent sequencing library preparation.

In such instances, sample amplification may be necessary prior to library construction in order to ensure enough material is available. The following techniques accomplish this while taking steps to minimize the biases, inherent in nucleic acid amplification (Ponzielli et al, 2008), in the subsequent ChIP-seq data analysis.

Employing alternative, amplification based strategies that deploy either linear amplification or primer extension followed by PCR before moving forward with the library generation; optimized techniques have been able to greatly reduce the input cell requirements for ChIP-seq to a mere 5,000-10,000 cells. (Adli et al, 2010 and Shankaranarayanan et al, 2011):

Nano-ChIP-seq (Adli et al, 2010) is an amplification method, which was specifically developed to work with rare cell types. Nano-ChIP-seq takes a three-step process to create a ChIP-seq library from as few as 10,000 cells. First, a random hairpin primer is used, followed by an exonuclease digestion step that generates amplified sample from small initial quantities of ChIP DNA.

This process also minimizes non-specific products. Next, high fidelity amplification of ChIP DNA, maintaining proper representation of GC-rich sequences, is achieved using optimized additives, cycling conditions and PCR enzymes.

Finally, amplified DNA is digested at BciVI sites placed near the ends of the ChIP fragments. This digestion creates double-stranded DNA products with 3′ A overhangs appropriate for direct ligation to Illumina adapters prior to sequencing.

The LinDA protocol (Shankaranarayanan et al, 2011) uses linear DNA amplification techniques to boost the amount of DNA available after the ChIP procedure. LinDA can start with as little as 30 picograms of chromatin immunoprecipitated DNA, the equivalent of 5,000-10,000 cells, to enable genome-wide ChIP-seq analysis of transcription factors and chromatin-associated proteins. LinDA has some other important features including:

• Unlike earlier low cell number ChIP protocols, LinDA is completely compatible with next generation sequencing.
• A reduced number of ligation steps and PCR cycles to keep the amount of potential artifacts introduced to a minimum.
• Compared to other T7 protocols developed for small input DNA amounts as the entire reaction is carried out using only one buffer within a single tube.

Going lower with ChIP sequencing

The recent improvements to ChIP protocols, pre-amplification methods and next-generation sequencing library preparation methods have greatly reduced the minimum amount of sample needed for successful ChIP sequencing, from in the neighborhood of 10 million cells, to as low as 5 or 10 thousand cells.

Although this represents a huge improvement, the goals of the research community are to eventually be able to study a single cell, or even single DNA molecule. Based on the big leaps forward to date, it seems promising that single cell analysis could be a technological reality within the next several years.