Fig 1.
PBM analysis of the DNA-binding specificity of MYC/MAX complexes.
(A) The custom PBM design contained immobilized DNA probes in triplicates with the indicated sequences (N and n indicate the varying positions having any of the 4 nucleotides). The total number of different sequences is shown in brackets. (B) Coomassie-stained SDS-PAGE showing the purified recombinant MYC/MAX complexes. (C) Images of fluorescent spot signals on PBM slides bound by MYC:MAX and MAX:MAX complexes and probed with anti-MAX antibody. (D) Scatter plot illustrating the correlation between MYC:MAX and MAX:MAX binding scores (normalized fluorescent intensities) for each individual probe spotted on the PBM (R2 = 0.764). Dashed lines indicate the threshold values for binding scores above 80% of random probes. (E) Ranked order list of top-scoring motifs in PBM probes bound by MYC:MAX and MAX:MAX (with binding scores above 95% of random probes). Double arrows indicate identical motifs. Itallics are the low complexity G/C-rich motifs. Asterisks indicate non-E-box motifs, including the AACGTT motif (underlined). Motifs highlighted in bold were bound by both MYC:MAX and MAX:MAX above the 80% random threshold.
Fig 2.
Position weight matrices of ranked core 6-mers bound by MAX:MAX and MYC:MAX.
Core 6-mer probes of the PBM were ranked according to their binding scores (log2 of normalized fluorescent intensities) for MAX:MAX (A) and MYC:MAX (B). Only probes scoring above 80% of random sequences are shown. The highest scoring probes containing the CME (CACGTG) and the NE (AACGTT) are indicated with arrows. Position weight matrices of 6-mer motifs present in the top 1–50, 51–100, and 101–350 bound probes were obtained with MEME and are represented as logos. The 8-mer sequences of bound probes (i.e., aNNNNNNg) were used as input for motif discovery.
Fig 3.
Specific binding of MYC/MAX complexes to the non-E-box (NE) motif AACGTT.
(A) Sequences of the probes used in EMSA. Nucleotides in red indicate positions in the NE probe that deviate from an E-box. (B) EMSA with the CME, NE and Ctrl probes and with the indicated amounts (pmol) of MAX:MAX and MYC:MAX. The position of the unbound/free probe is shown (F). (C) EMSA competition assays were performed with the labeled CME probe and the indicated MYC/MAX complex and the indicated amounts (pmol) of either CME, NE or Ctrl competitor oligos or no competitor (-). Two representative gels are shown (left). Results from three independent replicates were quantitated and the means (± S.D.) were plotted (right) and significantly differed (p<0.05) between the competitor oligos.
Fig 4.
Sequence requirements for MYC/MAX binding to the NE motif AACGTT.
(A) Sequences of the oligo competitors. (B, C) EMSA competition assays with MAX:MAX and the labeled CME probe and the indicated amounts (pmol) of competitor oligos. Three independent experiments were quantitated as in Fig 3C. Significant differences were observed between NE and NE-V1/-V2/V3/V4 (p<0.005 by two-tailed t-test).
Fig 5.
The NE motif AACGTT and several E-box variants have a similar affinity for MAX:MAX.
(A) Sequences of the oligo competitors. (B) EMSA competition assays with MAX:MAX and the labeled CME probe and the indicated amounts (pmol) of competitor oligos. Three independent experiments were quantitated as in Fig 3C. No statistically significant difference was observed between NE and the different competitors (p>0.05 by two-tailed t-test).
Fig 6.
Summary of relative affinities of MYC:MAX and MAX:MAX complexes for different E-box and non-E-box sequences.
Affinities were normalized relative to the affinity of the NE sequence, which was arbitrarily set to 1 (*). Relative affinities from the PBM scores are indicated between brackets. Sequences not analyzed by EMSA are indicated by “n.a.” and non-detectable binding is indicated by “-“.
Fig 7.
Normalized Frequencies of E-boxes and of the NE motif AACGTT at MYC-occupied genomic loci in human P493-6 B-cells.
MYC-bound sequences were obtained from ChIP-seq data of P493-6 cells overexpressing MYC [22]. The frequencies of all ten possible E-boxes (CANNTG), the NE motif AACGTT, and two control (Ctrl) sequences under MYC ChIP-seq peaks (A) and summits (B) were normalized to the occurrence of each motif in the human genome. The summits are the ±100 bp region from the apex of ChIP-seq peaks. The normalized frequencies are shown for all MYC peaks/summits (left) and for the most significant MYC peaks/summits (top 33% based on p-value; right).
Fig 8.
Characterization of MYC-occupied CME (CACGTG) and NE (AACGTT) genomic loci and influence of MYC expression levels in P493-6 cells.
(A) Venn diagrams show the number of MYC ChIP-seq peaks and summits that contain the CME, the NE or both motifs in high MYC-expressing P493-6 cells. (B, C) Frequency distribution plots show the fraction (%) of all MYC ChIP-seq peaks and summits and those specifically containing the CME or NE motifs that are located within promoters (± 2 kb from a TSS), intergenic or intragenic regions. (D) Frequency distribution of the CME and NE motifs in the human genome. (E) Effect of MYC overexpression on the number and average (mean) height of MYC ChIP-seq summits is shown as fold change (high MYC vs. low MYC) for the summits containing the CME or NE motifs. (F) Effect of MYC overexpression on the number of MYC ChIP-seq summits at promoters, intergenic and intragenic regions is shown as fold change (high MYC vs. low MYC) for CME- or NE-containing summits. (G) Examples of NE and CME ChIP-seq peaks from human genome browser in P493-6 cells expressing low and high MYC levels.