Appendix

A. Choice of the restriction enzyme
B. Quality Control
C. 3C Primers
D. References

A. Choice of the restriction enzyme

The minimal resolution at which the interactions can be analyzed depends on the size of the fragments generated through digestion and therefore on the restriction enzymes used. Both 6 bp cutters (with an average size of generated fragments of 4 kb; e.g. HindIII, Ncol) and 4 bp cutters (with an average size of fragments of 0.4 kb; e.g. DpnII/Mbo) that generate 5′ overhangs have been successfully used in Hi-C experiments. In 3C experiments where one aims at determining the frequencies of interactions between two given elements separated by a few kb, it is recommended to verify that the genomic distribution of the restriction sites between the targets is sufficient and homogeneous enough to permit the analysis. Some restriction enzymes could not be compatible with the 3C/Hi-C procedure (specific restriction buffer, low activity in presence of detergents, etc). However, the in nuclei version of the protocol allows the user to swap from one buffer to the other by centrifugation therefore allowing more flexible designs. Digestion by HindIII (HindIII site: 5′-AAGCTT-3′) generates 5′ overhangs that can be fill-in by the Klenow fragment of DNA Polymerase I in presence of Biotinylated dCTP in order to produce blunt ends. The ligation of these blunt ends will create a new NheI site, which can be used to analyse the efficiency of the procedure (see below quality control – Figure 5). When using a different restriction enzyme, it is recommended to check the Biotinylated dNTP to be used during the filling reaction as well as the eventual new restriction sites created upon ligation (labeling of the extreme positions of the overhang is not recommended since it could interfere with the ligation).

B. Quality Control

The success of the experiment is dependent on a high quality of cross-linked chromatin, optimal digestion and sufficient repair and ligation. These parameters could be checked throughout the protocol on samples made for this purpose and / or on small aliquots of the sample taken at given steps of the procedure.

Integrity of chromatin
Non-degraded genomic DNA is required since the use of degraded or broken DNA could lead to poor repair efficiency and ligation as well as to the generation of artifacts. It is thus recommended to check that the integrity of the DNA is conserved through the protocol. This can be checked by reversion of a sample or aliquot that has not been submitted to digestion with the restriction enzyme. This non-digested (ND) control should appear as an intact band of high molecular weight on agarose electrophoresis gel (cf. example in Figure 3). Samples showing important degradation with smears of lower molecular weight fragments should be avoided and the presence of cellular or external nulceases should be checked prior to follow. In addition, the Klenow fragment used to repair and label the ends generated by restriction could also repair DNA breaks that are being generated mechanically through the procedure. It is therefore important to minimize such possibilities (e.g. working on ice, pipetting rather than vortexing…).

Efficiency of digestion
Insufficient digestion of the DNA will lead to a poor efficiency of ligation, a low complexity of the libraries and further artifacts in PCR detection in the case of 3C. It is therefore essential to determine the extent of ligation when setting up the protocol. This can be achieved by conserving a control sample / aliquot that will be digested but not submitted to ligation (non-ligated sample – NL). This sample should appear as a smear on the agarose gel, with most of the fragments concentrated close to the average fragment size generated by the enzyme (see Figure 3).

The extent of digestion can also be quantitatively assessed by qPCR comparing ND and NL sample using few primers sets that flank a restriction site (see Figure 4). A digestion efficiency of >70% is recommended.

End-repair / labeling and ligation efficiency
A successful 3C sample migrates as a sharp band of high molecular weight on an agarose gel as compared to a non-ligated sample (cf. Figure 3). The efficiency of ligation is generally higher for cohesive ends than for blunt-ended extremities. A successfully repaired and ligated Hi-C sample will therefore generally show lower ligation efficiency than the corresponding 3C sample, highlighted by a smear of fragments with lower molecular weight (cf. Figure 3).

The efficiency of the ligation can be checked by amplifying a ligation product by standard 3C PCR on the 3C control and Ho-C sample (see Figure 4 and section below for the design of 3C primers and 3C PCR).

The Hi-C sample can contain ligation products emerging from successfully repaired and labeled fragments (Hi-C products) as well as fragments that have not been repaired and conserved their 5′ overhangs (3C products). These last products, although they correspond to products of ligation, will not be purified during the biotin pull-down. It is therefore important to determine the ratio of real Hi-C products over 3C products in the final sample, which depend on the efficiency of repair. As mentioned in section V.1, successful repair and ligation of a HindIII restriction site generate a new site for the Nhel enzyme (non repaired ends will ligate generating a new HindIII site, see Figure 5). After designing 3C primers that allow the amplification of a product that does not contain any endogenous Nhel site, one could submit the amplicon to digestion with HindIII and NheI. Products amplified from the 3C sample can only be digested by HindIII (see Figure 5). In contrast, products amplified from the Hi-C sample should contain a high proportion of products that are digested by NheI. The result of this PCR digest control can be analyzed on agarose gel (see Figure 5).

C. 3C Primers

The results of 3C derived experiments are chimeric DNA molecules resulting from the ligation of DNA fragments that were in close position within the nucleus. These products can be detected by PCR using primers facing to distinct restriction sites (See Figure 4). Although the ligation between fragments can occur independently between the two extremities of each fragment, 3C primers are generally designed in a similar genomic orientation (unidirectional primers) in such a way it prevents the amplification of non-digested or non-informative fragments. Those aspects and additional considerations for the design of 3C primers have been extensively reviewed by Naumova et al. (2012).

D. References