SUPPLEMENTARY MATERIALS
Fig. S1.TAD organization with respect to Hox clusters
A,B. ChIP-seq tracks for CTCF across a region surrounding the HoxA (A) and HoxC (B) clusters. TAD annotation is from Dix on et. al. (Dixon et al. 2012). ChIP-seq is from GSE60232(Narendra et al. 2015).
Fig. S2. Analysis of biological replicates of Fig. 1 show that the 5C experiment is highly reproducible
A. 5C chromatin interaction heatmap for biological replicate of WT MN cells depictsinteraction frequency between restriction fragments across a 4Mb region surrounding the HoxA cluster (data is binned in 15 Kb windows, step size 5Kb; and the median is shown). Darker colors represent increasing interaction frequency. Grey lines represent missing data. Dotted magenta line bounds the extent of the HoxA cluster within the heatmap. Topological boundaries within the region of the genome that form loops to interact with the a5|6 position are labeled in green.
B. Insulation profiles (log2) of biological replicate of WT and HoxaΔ5|6:7|9 ESCs and MNs. Dotted blue line indicates the TAD boundary present in WT MNs at the a5|6 position. The dotted green line indicates the TAD boundary present in HoxaΔ5|6:7|9 MNs near the a10|11 position, that also matches the boundary found in WT and HoxaΔ5|6:7|9 ESCs. The genomic region that interacts with the HoxA cluster is highlighted, and the associated insulation profile is magnified below. CTCF binding profiles within TAD boundaries are displayed, along with motif orientation. Convergent motifs point towards each other, and the corner peaks identified in the 5C heatmap are represented by green loops.
C,D,E. ChIP-seq tracks for indicated proteins and histone modifications across the HoxA cluster in the indicated cell types and genetic backgrounds (biological replicate of Fig. 1). ChIP-seq data was obtained from GSE60232 (Narendra et al. 2015). Insulation scores (log2), as calculated in Fig. 1b, are depicted above. Blue and green arrows point to TAD boundary positions within the HoxA cluster, which lie at the minima of the insulation score. The relative orientation of the CTCF motifs within the HoxA cluster aredisplayed.
Fig. S3. 5C interaction heatmap in WT and HoxaΔ5|6:7|9 ESCs and MNs
5C chromatin interaction heatmapsdepict the interaction frequency between restriction fragments across a 4Mb region surrounding the HoxA cluster(data is binned in 20 Kb windows, step size 5Kb; and the median is shown). Darker colors represent increasing interaction frequency. Grey lines represent missing data. Blue arrows point to corner-peaks that engage in long-range interactions with the HoxA CTCF-delimited TAD boundary.
Fig S4. Ratio of 5C interaction heatmaps between WT and HoxaΔ5|6:7|9ESCs and MNs
A. Heatmap of log2(MN observed/expected) – log2(ESC observed/expected) in WT cells (left) and HoxaΔ5|6:7|9cells (right). Blue pixels represent interactions that are more frequent in the HoxaΔ5|6:7|9setting. The expected is the average signal for a pair loci given its genomic distance, calculated by LOESS smoothing with alpha = 0.05.
B. Heatmap of log2(HoxaΔ5|6:7|9 observed/expected)- log2(WTobserved/expected) in ESCs (left) and MNs (right).
Fig. S5. HoxA and HoxC deletion genotypes
A,B.Sequencing chromatograms depicting genotypes of HoxaΔ5|6, HoxaΔ7|9, and HoxcΔ5|6mutant alleles. These lines were previously derived (Narendra et al. 2015).
MATERIALS AND METHODS (continued)
Cell culture and MN differentiation
ESCs were grown in standard ESC medium containing Lif, 1 μM MEK1/2 inhibitor (PD0325901) and 3 μM GSK3 inhibitor (CHIR99021). ESC differentiation was previously described(Wichterle et al. 2002; Mazzoni et al. 2013).
ESCs were trypsinized and seeded at 2.5×104cells/ml in ANDFK medium (Advanced DMEM/F12:Neurobasal (1:1) Medium, 15% Knockout Serum Replacement, Pen/Strep, 2 mM L-Glutamine, and 0.1 mM 2-mercaptoethanol) to initiate formation of embryoid bodies (Day 0). Medium was exchanged on Day 2 of differentiation. Patterning of embryoid bodies was induced by supplementing the media on Day 2 with 1 μM all-trans-Retinoic acid (RA, Sigma) and 0.5 μM smoothened agonist (SAG, Calbiochem). For immunocytochemistry and RT-qPCR analysis, a γ-secretase inhibitor, DAPT, was added to all cell lines two days after RA/SAG treatment at a 5μM final concentration, leading to an increase in the proportion of spinal MNs obtained during RA/SAG differentiation (Maury et al. 2015; Tan et al. 2016). Doxycycline was added two days after RA/SAG treatment at 3μg/ml to the inducible Hoxc6 cell line (iHoxc6) to induce the overexpression of Hoxc6. Embryoid bodies were harvested 4 days after treatment with RA/SAG and processed for 5C, immunocytochemistry or RT-qPCR analysis. For quantification analysis, embryoid bodies were dissociated into single MNs 4 days after treatment with RA/SAG. The MNs were fixed and stained a day later. The HoxcΔ5|6 mutant line yields lower total cell numbers than control after differentiation, however, the percentage of motor neurons is comparable to control.
5C primer design
Alternating forward and reverse 5C primers were designed to tile across a 4Mb region surrounding the HoxA cluster on chromosome 6 (mm9), with primers annealing to the sequence adjacent to HindIII restriction sites. These were designed using the my5Csuite primer design tools (Dostie and Dekker 2007). Primers annealing to repetitive regions were discarded. Universal T7 and T3 sequences were tethered to forward and reverse primers, respectively. In total, 479 forward and 479 reverse primers were designed. The 5’ end of each of the reverse primers was modified with a phosphate group by T4 PNK. Primers were pooled such that each 5C primer was represented in an equimolar distribution at a concentration of 5 fmol/ul.
Preparation of 3C library
5C was performed as described in Dostie et al (Dostie and Dekker 2007). Two biological replicates were separately cultured/differentiated, trypsinized, and counted. The replicates were processed separately. 5 x 107 cells were resuspended in fresh growth media and crosslinked in 1% formaldehyde for 10’ at room temperature. The reaction was quenched with glycine. Cells were lysed in 1ml lysis buffer (10mM Tris-HCl pH 8.0, 10mM NaCl, 0.2% Igepal (NP-40)) for 15’ in the presence of protease inhibitors, and disrupted using a dounce homogenizer. Nuclei were spun, washed and resuspended in HindIII restriction buffer. Cells were sequentially incubated in 0.1% SDS for 1hr at 37⁰C, and then 1% Triton X-100 for 1hr at 37⁰C. Extracts were digested o/n at 37⁰C in a thermal shaker with 1000U of HindIII (total 450μL reaction). Restriction enzyme was inactivated with 10% SDS at 37⁰C for 1hr. The reaction was added to a 15mL conical tube containing 7.61mL of ligation buffer (10X: 500mM Tris-HCl pH7.5, 100mM MgCl2, 100mM DTT) and incubated for 30’ at 37⁰C. The reactions were then placed on ice and 80μL of 100mM ATP was added to each tube. The ligation reaction was carried out for 4hrs at 16⁰C. 25μl 20mg/ml proteinase K was then added and incubated o/n at 65⁰C. The following morning, a fresh 25μl 20mg/ml of proteinase K was added and incubated for a further 2hrs, and the DNA was precipitated and purified. The purified DNA was treated with 1μl of 10mg/ml RNAse A for 15’ at 37⁰C.
5C Library Preparation
5C libraries were prepared separately for each biological replicate. 5C primers were annealed at 48⁰C for 16hrs atop 2 separate biological replicates of 3C libraries from each sample. 1fmol of each primer was used in the annealing reaction with 1μg of 3C template and 1μg of salmon sperm DNA. 16 separate annealing reactions were performed per sample, along with control reactions with individual components removed. Forward and reverse primers that annealed to adjacent regions of the 3C template were ligated with 10U of Taq ligase for 60’ at 48⁰C. Successfully ligated forward-reverse primer pairs were then amplified in 6 separate PCR reactions per annealing reaction, using primers specific to the T7 and T3 overhangs. PCR reactions from the equivalent initial sample were then pooled, purified and run on a gel to ensure the control reactions did not show an amplification product. Libraries were then generated from the purified PCR product to allow for deep sequencing.
5C data normalization
5C interaction maps were balanced using the ICE strategy (Imakaev et al. 2012). Briefly, each pairwise interaction between a 5C forward fragment and reverse fragment was divided by the sum of all interactions detected by that forward and the sum of all interactions detected by that reverse. This process was iterated 100 times, which produced a balanced forward-reverse interaction map where the sum of all interactions in each row and each column was the same.
Insulation analysis.
Insulation was calculated by summing all interactions occurring across pairs of adjacent forward-reverse primers for loci located up to 18 interrogated interactions upstream of the pair and 18 interrogated fragments downstream of the pair. This corresponds to an ~160Kb window around each forward – reverse pair. The sum was then divided by the average sum calculated across the entire locus. The insulation score is then defined as the log ratio of this value and is assigned to the coordinates of the forward-reverse fragment pair. Local minima in the insulation profile correspond to loci that have low levels of chromatin interactions across them and point to the presence of domain boundaries.
CRISPR zygotic injection
HoxA mutant mice were generated by zygotic injection (Yang et al. 2014). A T7 promoter was introduced by PCR upstream of the gRNAs targeting the a5|6 motif and a7|9 motif within the HoxA cluster (HoxaΔ5|6:7|9). The gRNA template was then transcribed using the MEGAshortscript T7 kit. Similarly, polyadenlyated Cas9 mRNA was generated using the mMESSAGEmMACHINE T7 ULTRA kit (Life Technologies). A mix containing 100 ng/μl Cas9 mRNA and 25 ng/μl of each of the a5|6 and a7|9 gRNAs was injected into the cytoplasm of ~100 C57BL/6 zygotes. Surviving embryos were transferred into 2 pseudopregnant females, and a total of 9 pups were born. These pups were genotyped by PCR using primers annealing to flanking genomic DNA, revealing the described genomic alterations in a number of pups.Similarly, HoxcΔ5|6 mutant mice were generated by using a gRNA targeting the c5|6 motif within the HoxC cluster.Mouse studies were approved by NYUMC IACUC.
SUPPLEMENTAL REFERENCES
Dixon JR, Selvaraj S, Yue F, Kim A, Li Y, Shen Y, Hu M, Liu JS, Ren B. 2012. Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature485: 376-380.
Dostie J, Dekker J. 2007. Mapping networks of physical interactions between genomic elements using 5C technology. Nature protocols2: 988-1002.
Imakaev M, Fudenberg G, McCord RP, Naumova N, Goloborodko A, Lajoie BR, Dekker J, Mirny LA. 2012. Iterative correction of Hi-C data reveals hallmarks of chromosome organization. Nature methods9: 999-1003.
Maury Y, Come J, Piskorowski RA, Salah-Mohellibi N, Chevaleyre V, Peschanski M, Martinat C, Nedelec S. 2015. Combinatorial analysis of developmental cues efficiently converts human pluripotent stem cells into multiple neuronal subtypes. Nat Biotechnol33: 89-96.
Mazzoni EO, Mahony S, Peljto M, Patel T, Thornton SR, McCuine S, Reeder C, Boyer LA, Young RA, Gifford DK et al. 2013. Saltatory remodeling of Hox chromatin in response to rostrocaudal patterning signals. Nature neuroscience16: 1191-1198.
Narendra V, Rocha PP, An D, Raviram R, Skok JA, Mazzoni EO, Reinberg D. 2015. Transcription. CTCF establishes discrete functional chromatin domains at the Hox clusters during differentiation. Science (New York, NY)347: 1017-1021.
Tan GC, Mazzoni EO, Wichterle H. 2016. Iterative Role of Notch Signaling in Spinal Motor Neuron Diversification. Cell Rep16: 907-916.
Wichterle H, Lieberam I, Porter JA, Jessell TM. 2002. Directed differentiation of embryonic stem cells into motor neurons. Cell110: 385-397.
Yang H, Wang H, Jaenisch R. 2014. Generating genetically modified mice using CRISPR/Cas-mediated genome engineering. Nature protocols9: 1956-1968.