PLAG / PLAGL
The PLAG proteins - PLAG1, PLAGL1 (also called ZAC1 or LOT1) and PLAGL2 - are C2H2 zinc-finger factors with seven fingers that read a bipartite G-rich site: a GRGGC core and a GGG cluster separated by roughly six to eight base pairs. Finger 3 contacts the G-cluster and fingers 6 and 7 contact the core, so the two halves are read by different parts of one protein. PLAG1 is the classic translocation target in salivary-gland pleomorphic adenoma, where promoter swapping with CTNNB1 or LIFR switches it on; PLAGL2 is also transforming, while PLAGL1/ZAC1 is an imprinted, maternally silenced growth suppressor whose loss of imprinting causes transient neonatal diabetes. All three activate IGF2.
The factors
Three human genes make up this family. PLAG1 (8q12) was identified as the gene activated by promoter swapping in pleomorphic adenomas of the salivary gland carrying t(3;8)(p21;q12), which fuses the CTNNB1 promoter to the PLAG1 coding region 9020842; a second recurrent translocation, t(5;8)(p13;q12), puts PLAG1 under the LIFR promoter. PLAGL1 (6q24), independently named ZAC1 for zinc finger with apoptosis and cell-cycle arrest 9184226 and LOT1 for lost on transformation, and PLAGL2 (20q11) complete the set. All three carry seven Cys2-His2 zinc fingers in a conserved N-terminal DNA-binding domain plus a C-terminal serine/threonine/proline-rich transactivation domain, and all bind DNA as monomers. The DNA-binding domains are highly similar in sequence 11888928, which is the central problem for interpreting a hit. PLAG1 and PLAGL2 are broadly expressed in embryonic tissues and largely silenced after birth; PLAGL1 is expressed from the paternal allele only 10655556, in pituitary, brain, pancreatic islets and many embryonic tissues. PLAG1 activity is further tuned by SUMOylation of its transactivation domain.
The motif
The literature consensus is bipartite: a GRGGC core and a G-cluster. The binding-site selection that defined it placed seven random nucleotides between the core and a G-cluster written RGGK 10646861; the same group's later target-gene screen quoted it as GRGGC(N)6-8GGG, allowing six to eight 14712223. How far the spacer can vary has not been measured directly, but to the extent it does, the element is better described as two half-sites than as a fixed-width matrix. Mutational and binding work assigned the two halves to different fingers: finger 3 engages the G-cluster, fingers 6 and 7 engage the GRGGC core, so a single molecule bridges the gap 10646861.
The data in this atlas are consistent with that but compress it. The JASPAR PLAG1 model (GGGGCCCAAGGGGG, 14 bp) and the equivalent CIS-BP and Vierstra models (CCCCCTTGGGCCCC) 25215497,32728250 span the whole bipartite element with the short end of the spacer; the PLAGL2 and PLAGL models collapse to an 8-bp GGGCCCCC, effectively just the G-rich halves. The atlas medoid, GGGCCCCC, is therefore shorter than the literature consensus rather than in conflict with it: fixed-width position weight matrices cannot represent a variable spacer, so the databases store the tightest-spaced version. Expect real sites to be longer and more variably spaced than any of these matrices.
The three paralogues are not equivalent in vitro. PLAG1 and PLAGL2 bind the same bipartite site and both transform NIH3T3 cells, whereas PLAGL1/ZAC1 binds a distinguishable, more homogeneously G-rich sequence and arrests rather than promotes growth 11888928,9184226. The CIS-BP PLAGL1 model here is a 4-bp CCCC inferred from mouse Plagl1 at 88% DNA-binding-domain identity 25215497, so it carries almost no information and should not be used for scanning.
Where it acts
PLAG-family sites are G-rich and land preferentially in GC-rich, CpG-island promoters. The best-validated target is IGF2: PLAG1 and PLAGL2 both bind the IGF2 P3 promoter region and up-regulate it, and IGF2 induction accompanies PLAG1- and PLAGL2-driven transformation 10646861,11888928. PLAGL1/ZAC1 sits at the top of an imprinted gene network that includes IGF2, H19, CDKN1C, DLK1 and MEST, and coordinates their expression during embryonic growth 17084362. Because the sites are G-rich, they overlap the binding preferences of KLF/SP factors, EGR, MAZ, PATZ1 and other GC-box readers, and PLAG matches are common in the CpG-island promoters of housekeeping genes where none of these factors is necessarily the functional occupant.
Biology and disease
PLAG1 is the defining oncogene of salivary-gland pleomorphic adenoma 9020842 and also of lipoblastoma, where the HAS2 or COL1A2 promoter is fused to the entire PLAG1 coding sequence, again a promoter swap rather than a chimeric protein 10987300. PLAGL2 is amplified and overexpressed in glioma and colorectal cancer and promotes proliferation partly through WNT-dependent 20478531 and partly through IGF2/ASCL2-dependent routes 37310750. PLAGL1/ZAC1 has the opposite behaviour: it induces apoptosis and G1 arrest 9184226, is imprinted with maternal silencing 10655556, and relaxation of that imprint at the 6q24 HYMAI/PLAGL1 locus causes transient neonatal diabetes mellitus 10655556,11448939. Zac1-null mice are growth-restricted 17084362, consistent with a network role in embryonic growth. Outside disease, a regulatory variant at the bovine PLAG1-CHCHD7 interval is one of the largest single-locus effects on stature and carcass weight in cattle 22607022,29215042, a reminder that PLAG1 dosage controls growth in normal physiology.
Reading a hit
A PLAG match is a G-rich, GC-box-like element, so the honest conclusion from sequence alone is "a G-rich promoter element," not "a PLAG site." KLF, SP, EGR, MAZ, PATZ1 and ZNF263 all compete for the same matches, and G-rich matrices produce many chance hits in CpG islands. Within the family, PLAG1 and PLAGL2 cannot be told apart by sequence; PLAGL1 is somewhat distinguishable in vitro, but the PLAGL1 matrices available here are too short (4-8 bp) to exploit that. Because the true element is bipartite, with a spacer of about seven bases that later work quotes as six to eight, a fixed-width matrix will miss differently spaced real sites and may score contiguous G-runs that are not functional sites. Check for the two halves explicitly if the call matters. Expression is the strongest discriminator: PLAG1 and PLAGL2 are largely embryonic, so in adult tissue a PLAG match is more likely to be a generic GC-box unless a fusion or amplification is known.
How this was worked out
PLAG1 is identified as the 8q12 gene activated in pleomorphic adenomas with t(3;8)(p21;q12), where promoter swapping with CTNNB1 drives expression of an intact PLAG1 coding sequence.
Still open then: PLAG1 was a predicted zinc-finger protein with no known DNA-binding specificity.
Zac1 (PLAGL1) is cloned as a zinc-finger protein that induces apoptosis and G1 arrest and suppresses tumor formation, the opposite of PLAG1's oncogenic behavior.
- 2000mediumBinding-site selection; zinc-finger mutagenesis; gel shift; transactivation; tumor expression10646861
The PLAG1 site is defined as a bipartite element, a GRGGC core and a G-cluster separated by seven random nucleotides, contacted by fingers 6-7 and finger 3 respectively; IGF2 promoter 3 is a bound and activated target.
Still open then: Whether PLAGL1 and PLAGL2 read the same element was untested.
- 2000mediumRestriction landmark genome scanning; allele-specific RT-PCR; methylation analysis; reporter assays10655556 11448939
ZAC/PLAGL1 is shown to be imprinted and paternally expressed at 6q24, making it the leading candidate for transient neonatal diabetes; loss of methylation at the HYMAI/ZAC imprinting control region is then found in most patients with a normal karyotype.
Despite highly homologous DNA-binding domains, the three PLAG proteins split in DNA binding: PLAG1 and PLAGL2 both need the GRGGC core and the G-cluster, with finger 3 reading the cluster, whereas PLAGL1 recognizes only an extended core; PLAG1 and PLAGL2 transform NIH3T3 cells and induce IGF2, in contrast to growth-suppressive PLAGL1.
Still open then: Which residues give PLAGL1 its different preference was not identified.
Zac1 knockout causes intrauterine growth restriction, and Zac1 is placed in a network of co-regulated imprinted growth genes, directly regulating Igf2/H19 through a shared enhancer.
- 2010mediumIntegrated copy-number and expression genomics; neural stem cell differentiation assays; transcriptome analysis20478531
PLAGL2 is identified as an oncogene targeted by amplification or gain in malignant glioma, where it suppresses neural stem cell differentiation partly by modulating Wnt/beta-catenin signaling.
- 2023mediumGain- and loss-of-function in CRC lines and intestinal cells; reporter assays; rescue with ASCL237310750
In colorectal cancer, PLAGL2 drives transformation with only modest effects on canonical Wnt signaling, acting instead through direct targets IGF2 and ASCL2, which revises the Wnt-centered model for this tissue.
Still open then: Whether the glioma and colorectal mechanisms differ by tissue or reflect a disagreement is unresolved.
Still unresolved
- The spacer is stated inconsistently: the original PLAG1 selection placed seven random nucleotides between the GRGGC core and the G-cluster, while the consensus usually quoted allows six to eight, and no genome-wide measurement has established how much spacer variation PLAG1 tolerates in vivo.
- How PLAGL1 acquires a DNA preference different from PLAG1 and PLAGL2, despite a highly homologous DNA-binding domain, has not been explained structurally, and the PLAGL1 matrices in these databases (CCCC and TGGGGCCA) are too short to show the difference.
- Whether PLAGL2's oncogenic activity runs through WNT signaling is disputed: in glioma and neural stem cells it was attributed in part to Wnt/beta-catenin modulation, whereas in colorectal cancer its effect on canonical WNT signaling was minimal and IGF2 and ASCL2 carried the effect.
The real element is bipartite, a GRGGC core and a G-cluster seven bases apart in the original selection and quoted as six to eight in later work, and a fixed-width matrix represents only one spacing, so the matrices here are truncated or tightest-spaced versions that will both miss differently spaced sites and over-call plain G-runs. Matches overlap KLF/SP, EGR, MAZ and other GC-box factors, and G-rich models hit CpG-island promoters frequently by chance. PLAG1 and PLAGL2 share a specificity and cannot be separated by sequence; PLAGL1/ZAC1 differs in vitro, but its models here (CCCC, TGGGGCCA) are too short to use, and the CIS-BP PLAGL1 entry is inferred from mouse at 88% DBD identity. PLAG1 and PLAGL2 are largely embryonic, so adult-tissue matches deserve extra scepticism.
References
- Kas K, Voz ML, Röijer E, et al. Promoter swapping between the genes for a novel zinc finger protein and beta-catenin in pleiomorphic adenomas with t(3;8)(p21;q12) translocations. Nat Genet 1997;15:170-4. PMID 9020842
- Spengler D, Villalba M, Hoffmann A, et al. Regulation of apoptosis and cell cycle arrest by Zac1, a novel zinc finger protein expressed in the pituitary gland and the brain. EMBO J 1997;16:2814-25. PMID 9184226
- Voz ML, Agten NS, Van de Ven WJ, et al. PLAG1, the main translocation target in pleomorphic adenoma of the salivary glands, is a positive regulator of IGF-II. Cancer Res 2000;60:106-13. PMID 10646861
- Kamiya M, Judson H, Okazaki Y, et al. The cell cycle control gene ZAC/PLAGL1 is imprinted--a strong candidate gene for transient neonatal diabetes. Hum Mol Genet 2000;9:453-60. PMID 10655556
- Arima T, Drewell RA, Arney KL, et al. A conserved imprinting control region at the HYMAI/ZAC domain is implicated in transient neonatal diabetes mellitus. Hum Mol Genet 2001;10:1475-83. PMID 11448939
- Hensen K, Van Valckenborgh IC, Kas K, et al. The tumorigenic diversity of the three PLAG family members is associated with different DNA binding capacities. Cancer Res 2002;62:1510-7. PMID 11888928
- Varrault A, Gueydan C, Delalbre A, et al. Zac1 regulates an imprinted gene network critically involved in the control of embryonic growth. Dev Cell 2006;11:711-22. PMID 17084362
- Nishimura S, Watanabe T, Mizoshita K, et al. Genome-wide association study identified three major QTL for carcass weight including the PLAG1-CHCHD7 QTN for stature in Japanese Black cattle. BMC Genet 2012;13:40. PMID 22607022
- Utsunomiya YT, Milanesi M, Utsunomiya ATH, et al. A PLAG1 mutation contributed to stature recovery in modern cattle. Sci Rep 2017;7:17140. PMID 29215042
- Fischer AD, Veronese Paniagua DA, Swaminathan S, et al. The oncogenic function of PLAGL2 is mediated via ASCL2 and IGF2 and a Wnt-independent mechanism in colorectal cancer. Am J Physiol Gastrointest Liver Physiol 2023;325:G196-G211. PMID 37310750
- Weirauch MT, Yang A, Albu M, et al. Determination and inference of eukaryotic transcription factor sequence specificity. Cell 2014;158:1431-1443. PMID 25215497
- Vierstra J, Lazar J, Sandstrom R, et al. Global reference mapping of human transcription factor footprints. Nature 2020;583:729-736. PMID 32728250
- Hibbard MK, Kozakewich HP, Dal Cin P, et al. PLAG1 fusion oncogenes in lipoblastoma. Cancer Res 2000;60:4869-72. PMID 10987300
- Zheng H, Ying H, Wiedemeyer R, et al. PLAGL2 regulates Wnt signaling to impede differentiation in neural stem cells and gliomas. Cancer Cell 2010;17:497-509. PMID 20478531
- Voz ML, Mathys J, Hensen K, et al. Microarray screening for target genes of the proto-oncogene PLAG1. Oncogene 2004;23:179-91. PMID 14712223
8 motifs in this family — JASPAR 3 · CIS-BP 3 · Vierstra 2
| motif | database | consensus | bits |
|---|---|---|---|
| PLAG1 (MA0163.1) | JASPAR | GGGGCCcAAGGGGG | 19.4 |
| M06465_2.00 | CIS-BP | CCCCCTTgGGCCCC | 19.4 |
| AC0503:PLAG:C2H2_ZF | Vierstra | CCCCCTTgGGCCCC | 19.4 |
| M01171_2.00 | CIS-BP | GGGCCCCC | 12.1 |
| PLAGL2 (MA1548.2) | JASPAR | GGGCCCCC | 12 |
| AC0501:PLAGL:C2H2_ZF | Vierstra | GGGCCCCC | 12 |
| Plagl1 (MA1615.2) | JASPAR | TGGGGCCA | 10.9 |
| M00130_2.00Mus musculus 0.88 | CIS-BP | CCCC | 6.3 |
HOCOMOCO tags read as assay + quality grade: P ChIP-seq, S HT-SELEX, M methyl-HT-SELEX, G GHT-SELEX, I SMiLE-seq, B PBM; grade A means two or more assay types agreed. A P-only matrix is a record of where the protein was crosslinked, which need not be a sequence it binds itself.
The factors, elsewhere
These entries are written from the primary literature and every citation is checked against PubMed, but this one can still be out of date, too confident, or simply wrong — especially where a motif’s published site and its database matrices disagree. Corrections and additions are welcome, and they are the fastest way to make this page better for everyone.