CAMTA

The factors

CAMTA stands for calmodulin-binding transcription activator. Humans have two genes, CAMTA1 and CAMTA2, and they are listed in HOCOMOCO under the older symbols CMTA1 and CMTA2 — worth knowing when searching, because the two spellings are not interchangeable in database queries. The domain layout is the same in plants and animals: an N-terminal CG-1 domain, which carries the DNA-binding activity, followed by a TIG (transcription factor immunoglobulin) domain, two ankyrin repeats that mediate protein-protein interaction, and a C-terminal set of IQ motifs that bind calmodulin ^{11925432,17689537}. The name comes from that last feature: calcium-loaded calmodulin binding to the IQ region is the regulatory input, which makes CAMTAs one of the few direct links from calcium signaling to a sequence-specific DNA-binding domain ^11925432. The CG-1 domain has no similarity to any other DNA-binding class and was first characterized in plants, where the family is larger — Arabidopsis has six members (AtSR1-6) induced by cold, salt, UV-B, wounding and hormone signals ^{8075408,12218065}. In humans both genes are expressed most strongly in heart and brain ^{16678093,17689537}. Notably, CAMTA2 was characterized biochemically not as a direct DNA-binding activator but as a coactivator recruited by the homeodomain factor NKX2-5, so part of the family's activity at chromatin is tethered rather than motif-driven ^16678093.

The motif

The element established for this domain family is the CGCG box, a 6-bp site written (A/C/G)CGCG(G/T/C), defined for the Arabidopsis CG-1 protein AtSR1 by direct binding-site analysis ^12218065. This is the consensus given above. A mammalian consensus was established separately: four rounds of gel-shift selection from random oligonucleotides with a recombinant mouse CAMTA2 fragment gave (T/C)GCANTGCG, a palindrome around one variable central base, bound by a CAMTA dimer and dependent on an adjacent A/T-rich stretch; mutating the core abolished both binding and reporter activation ^25049392. The HOCOMOCO CAMTA2 matrix, CGCAATGCG — a GC-rich flank on either side of a central AT, architecturally a short inverted repeat rather than a tandem CGCG — is an exact instance of that consensus, while the CAMTA1 matrix is shorter and weaker, CGCAtt, with only 9.4 bits of information. The two agree with each other only moderately (family cohesion 0.79), and for the CAMTA1 model the most similar motif anywhere across the four databases lies outside this family rather than inside it.

The honest reading is that the plant and mammalian sites are related but distinct, as the mammalian work itself concluded ^25049392: both are short CG-rich elements read by the same conserved domain, but the mammalian consensus carries a central A/T and GCG arms where the plant box is a tandem CGCG. Both human models carry HOCOMOCO evidence code PSGI at grade A, meaning they combine ChIP-seq with three in vitro assays: HT-SELEX, GHT-SELEX and SMiLE-seq; the in vitro component is what keeps them credible, but the CAMTA1 motif remains low in information content, and there is no published vertebrate CG-1-DNA structure to explain which base contacts matter.

Where it acts

With only two models, and both of low information content, there is little reliable information about genomic distribution for the human factors. What is documented is functional context rather than motif context: CAMTA2 acts at cardiac gene promoters through NKX2-5, in opposition to class II histone deacetylases (notably HDAC5), and its nuclear activity is regulated by HDAC5 export ^16678093. In plants, CGCG boxes are found in promoters of ethylene-signaling, abscisic-acid-signaling and light-perception genes, often in multiple copies ^12218065 — the multiplicity is part of how a weak site becomes functional. A 6-mer consensus like VCGCGB will match hundreds of thousands of positions in a 3-Gb genome, so occurrence carries essentially no information on its own; matches are concentrated in CG-rich promoter sequence for compositional reasons rather than biological ones.

Biology and disease

The mouse phenotypes are neurological and cardiac. *Camta1*-null mice develop ataxia with Purkinje cell degeneration ^25049392, and CAMTA2 overexpression drives cardiac hypertrophy while its loss blunts the hypertrophic response ^16678093. In humans, intragenic CAMTA1 rearrangements cause non-progressive congenital ataxia, with or without intellectual disability ^22693284, and a described CAMTA1 mutation produces a characterized neuropsychological and neuroimaging phenotype ^24145135. CAMTA1 lies within the commonly deleted 1p36 region in neuroblastoma and behaves as a tumor suppressor candidate there, with reduced expression associated with worse outcome ^21385898. The most striking cancer connection is a fusion: WWTR1 (TAZ)-*CAMTA1* is the defining driver rearrangement in epithelioid hemangioendothelioma ^21885404, where the CAMTA1 portion supplies constitutive nuclear localization and transcriptional output to the TAZ coactivator ^35740643.

Reading a hit

Treat a CAMTA match as weak evidence. The two available models are short and low in information, they agree with each other only moderately, and for CAMTA1 the nearest neighbour across the databases is outside the family — so a hit may equally reflect a different CG-rich motif. CAMTA1 and CAMTA2 cannot be distinguished from one another. The consensus quoted at the top of this entry is the plant CGCG box; the mammalian site defined by selection is (T/C)GCANTGCG ^25049392, so scanning human sequence with the plant box is an assumption rather than a result.

Two structural cautions about the evidence. First, CAMTA2's best-documented mode of action in mammals is as an NKX2-5 coactivator ^16678093, so ChIP-derived signal for CAMTA2 may reflect a homeodomain site rather than a CG-1 site — the classic indirect-binding case, where the tagged protein "has" the motif of whatever recruits it. Second, both matrices are partly ChIP-derived (PSGI), and even at grade A that mixture cannot separate direct from tethered binding. Use a CAMTA hit to raise a hypothesis in a calcium-responsive, heart- or brain-expressed promoter, and verify with a direct binding assay before relying on it.