Uncovering the “ZIP code” for bZIP dimers reveals novel motifs, regulatory rules and one billion years of cis-element evolution

Abstract

Many eukaryotic transcription factors (TF) form homodimer or heterodimer complexes to regulate gene expression. For example, dimerization properties of the basic leucine zipper (bZIP) family play a critical role in regulating the unique biological functions in all eukaryotes. However, the molecular mechanism underlying the binding sequence and functional specificity of homo- versus heterodimers remains elusive. To fill this gap, we developed a double DNA Affinity Purification sequencing (dDAP-seq) technique that maps heterodimer DNA binding sites in an endogenous genome context. Our genome-wide binding profiles of twenty pairs of C/S1 bZIP heterodimers and S1 homodimers in Arabidopsis revealed that heterodimerization significantly expands the DNA binding preferences of bZIP TFs. Analysis of the heterodimer target genes in stress response and development suggest heterodimerization gives rise to regulatory responses that are distinct from the homodimers. In addition to the classic ACGT elements recognized by plant bZIPs, we found that the C/S1 heterodimers bound to motifs that might share an origin with the GCN4 cis-elements in yeast that diverged from plants more than one billion years ago. Importantly, heterodimer binding specificities can be distinguished by their relative preference for ACGT motifs versus GCN4-related motifs. More broadly, our study demonstrates the potential of dDAP-seq in deciphering the DNA binding specificities of interacting TFs that are key for combinatorial gene regulation.

A Single-Nucleus Atlas of Seed-to-Seed Development in Arabidopsis

“…For instance, basic leucine zipper (bZIP) TFs recognize target DNA sequences by forming various homo– and hetero-dimers 45 . A previous study showed that Group C bZIP TFs (bZIP9, bZIP10, bZIP25, bZIP63) do not bind DNA as homodimers but form heterodimers with Group S1 members (bZIP1, bZIP2, bZIP11, bZIP44, bZIP53) to bind unique DNA sequences 46 . We found that genes in each group showed varying cell type/tissue specificity (Figure 6K).…”

Principles of Computation by Competitive Protein Dimerization Networks

“…4 A median of 42 human bZIP proteins appear to be co-expressed in any given cell type (Figure S7), and previous work has already demonstrated that different cell types express components of these protein families at different abundances. 2,24 Thus, the bZIP dimerization network exhibits the potential for complex and cell-type-specific computation. Further, similar statistics characterize the mouse and Arabidopsis thaliana bZIP families, the mouse and human nuclear receptor families, and the Arabidopsis thaliana MADS-box family (Figure S7B, Figure S7C).…”

“…The dataset was filtered to include only responses with a dynamic range greater than 10-fold; all other dimers either did not form at all or did not change significantly in response to the input monomer(s). The remaining responses were categorized into “unique” functions by discretizing the space of possible outputs into ten-fold bins (i.e., bin edges at 10 - 3 , 10 - 2 , 10 - 1 , 1, 10 1 , 10 2 , and 10 3 , for both input and output). For each input bin (e.g., input between 10 - 3 and 10 - 2 ), the response points were averaged in log space, and this value was categorized into one of the output bins (e.g., 0.5 categorized into the 10 - 1 to 1 bin).…”

Assessing Combinatorial Diversity of Aureochrome Basic Leucine Zippers through Genome-Wide Screening

“…The bipartite α-helical structure of the dimer bZIPs forms a "Y"-shaped complex with the DNA [Bader et al, 2006]. In fact, dimerization among compatible bZIP family members has resulted in a genome-wide dimermediated regulatory network of bZIPs [Li et al, 2022]. It is still expanding with the gene duplication of the different bZIP sequences.…”