DMRT (DM domain)
DMRT proteins are the DM-domain transcription factors, named after Drosophila Doublesex and C. elegans MAB-3, the sexual regulators in which the domain was found. Humans have seven well-studied genes, DMRT1, DMRT2, DMRT3, DMRTA1, DMRTA2, DMRTB1 and DMRTC2, plus DMRTC1 paralogues. The DM domain is not a classical zinc finger: two zinc ions are chelated by an intertwined arrangement of cysteines and histidines that cannot be split into independent fingers, and the module ends in a recognition helix. The proteins bind a pseudopalindromic element built from two ACA/TGT half-sites, canonically ACAWTGTW, as dimers and higher-order oligomers. DMRT1 is required for testis differentiation and sustains it after birth; it is the sex-determining gene in chicken, and the 9p24 region containing it is deleted in human 46,XY sex reversal.
The factors
The DM domain was identified in Drosophila Doublesex (DSX) and C. elegans MAB-3 and is conserved across animals 9490411. Vertebrates have a small family: the seven human genes usually discussed are DMRT1, DMRT2, DMRT3, DMRTA1 (DMRT4), DMRTA2 (DMRT5), DMRTB1 (DMRT6) and DMRTC2 (DMRT7), with additional DMRTC1 paralogues that also have measured binding. Outside the DM domain the proteins share almost no sequence 17605809, so the family is defined by its DNA-binding module and not by any common architecture. That module is structurally unusual. Rather than the independent zinc fingers of C2H2 or GATA proteins, the DM domain chelates two zinc ions through an intertwined set of cysteine and histidine ligands that cannot be separated into discrete fingers, followed by a C-terminal recognition helix and a basic tail 10898790,26005864. DMRT proteins do not bind as simple monomers: the crystal structures of human DMRT1 on DNA show tetrameric, trimeric and dimeric modes on the same kind of site, and ChIP with exonuclease treatment indicates that several of these modes are used in cells 26005864. Expression is restricted. DMRT1 is essentially gonad-specific (Sertoli cells and germ cells) 9490411,20616082; DMRT2 acts in somites; DMRT3 is in spinal interneurons and gonad; DMRTA2 is in developing cortex; DMRTB1 and DMRTC2 are germ-cell proteins.
The motif
The element is a pseudopalindrome assembled from two trinucleotide half-sites that are reverse complements of one another, ACA and TGT, separated by one or two variable bases 17605809,26005864. Written as the literature usually does, the core is ACAWTGTW; the equivalent forward-strand rendering with the half-sites in the other order is TGTWACA, and this is what the data consensus here, AATGTAACAA, represents. The two are the same element, not a disagreement. Information is concentrated on the two ACA/TGT triplets; the spacer base and the flanks are weakly constrained, which is why individual database models range from a bare six-base TGTADC to a full 13-base AATGTATCAATTT. Structurally, the early NMR work on the DSX DM domain described contacts in the minor groove of an undistorted duplex 10898790, whereas the later crystal structures of DMRT1 showed two adjacent antiparallel recognition helices inserted into a widened major groove making base-specific contacts 26005864; the current picture is major-groove reading with minor-groove stabilization, and this shift is worth knowing when reading older descriptions. In vitro selection across six of the seven vertebrate proteins returned sites resembling the DSX site for all of them, differing only in efficiency and in preference at individual positions 17605809. The family is correspondingly cohesive here (0.86 across 20 motifs) and no paralogue is separable by sequence. Some DMRT models drift toward pure AT-richness: the Vierstra archetype AC0092:DMRT/HOXB is an ATATATT-like model shared with HOXB, and its nearest neighbours lie outside the family, a reminder that AT-rich DM sites and AT-rich homeodomain sites are hard to tell apart.
Where it acts
DMRT1 occupancy in mouse testis is largely promoter-proximal and promoter-distal in roughly equal measure, with thousands of sites, and the strongest targets are genes of the male pathway (SOX9, PTGDS) and genes that DMRT1 must keep off (FOXL2, and the retinoic-acid-responsive meiotic initiator STRA8). DMRT1 works with and against other gonadal regulators: it opposes FOXL2 and ESR1/ESR2 in the ovarian program 21775990, and cooperates with SOX9 and SF1/NR5A1 in the testis. DMRT3 acts in the spinal cord in circuits that coordinate limb movement 22932389. Because the factors are tissue-restricted, the genomic distribution of DM sites is much broader than the set that is ever occupied: in a cell that expresses no DMRT protein, the sequence is inert.
Biology and disease
Dmrt1-null mice have severely dysgenic testes 11040213, and postnatal loss of DMRT1 in Sertoli cells reprograms them toward granulosa cells with activation of FOXL2 and ESR2, so DMRT1 is needed continuously, not just at the sex-determining step 21775990. In chicken, the Z-linked DMRT1 is the male sex-determining gene, and knockdown feminizes male gonads 19710650. In humans, deletions of distal 9p that remove DMRT1 cause 46,XY gonadal dysgenesis and sex reversal 9490411,11040213, and point mutations in residues that contact DNA have the same effect 26005864; common variants near DMRT1 are associated with testicular germ cell tumor risk 20543847. Outside sex determination, a premature stop codon in horse DMRT3 underlies the ability to perform alternate gaits, and the same mutation modeled in mice disturbs spinal circuit function 22932389.
Reading a hit
The element is short and AT-rich in its flanks, so chance matches are common and the AT-rich tails are readily confused with homeodomain and other minor-groove AT binders. A match cannot be assigned to a paralogue: the seven vertebrate DM proteins select nearly the same sequence in vitro 17605809, and most models in these databases come from HT-SELEX or methyl-HT-SELEX on isolated domains 23332764, which measure intrinsic specificity and cannot distinguish family members. The strongest interpretive constraint is expression: DM sites are only meaningful in tissues where a DMRT protein is present, which for DMRT1 means gonad. Note also that the JASPAR mouse Dmrt1 matrix MA1603.2 (TACAAAGT) is an outlier here, with nearest neighbours outside the family, so a hit scored against that specific model deserves a second look.
How this was worked out
The C. elegans male regulator mab-3 is related to Drosophila doublesex; their shared DNA-binding motif is named the DM domain, and a human DM gene (DMT1, now DMRT1) is testis-specific and maps to distal 9p, a region implicated in XY sex reversal.
Still open then: What DNA sequence the domain reads, and whether vertebrate DM genes act in sex development.
NMR structure of the DSX DM domain: an intertwined CCHC/HCCC zinc module with a disordered tail that acts as a nascent recognition helix, described as binding the minor groove without sharp DNA bending.
Still open then: How a minor-groove binder achieves the unusually high sequence specificity measured for DSX and MAB-3.
Dmrt1-null mice fail postnatal testis differentiation, with phenotypes like those of human 9p deletions, while ovaries develop normally: the first functional evidence for a vertebrate DM gene in male development.
Random-oligonucleotide selection on six of the seven vertebrate DMRT proteins returns DSX-like sites for all of them, differing in efficiency and per-position preference, and some DMRTs bind DNA as heterodimers.
Still open then: How paralogues with near-identical sites reach different targets.
ChIP in juvenile mouse testis identifies almost 1,400 DMRT1-bound promoter-proximal regions, enriched for a motif closely resembling the in vitro site; DMRT1 activates some targets and represses others, differently in germ cells and Sertoli cells.
Deleting DMRT1 in Sertoli cells, even in adults, activates Foxl2 and reprograms them into granulosa-like cells, so testis identity has to be maintained continuously.
Crystal structures of DMRT1 on DNA show two adjacent antiparallel recognition helices in a widened major groove making base-specific contacts, with tetramer, trimer and dimer binding modes, and ChIP-exo suggests several modes are used in vivo. This replaced the minor-groove picture drawn from the DSX NMR structure.
Still open then: Which stoichiometry operates at which genomic sites.
Still unresolved
- Which binding stoichiometry operates at functional sites is unresolved: DMRT1 binds DNA as a tetramer, trimer or dimer in crystals and ChIP-exo suggests more than one mode in testis, but which mode a given genomic site uses, and whether the mode changes the transcriptional outcome, is not known.
- How DMRT paralogues with near-identical in vitro preferences reach different targets is unclear: the proposed explanations, subtle per-position preferences and DMRT heterodimers, compete with the simpler view that non-overlapping expression does all the work.
- The JASPAR mouse Dmrt1 matrix MA1603.2 (TACAAAGT) breaks the second half-site of the pseudopalindrome (ACAAAGT against ACAWTGT) and its nearest neighbours lie outside the family; whether it records a real in vivo variant of the site or a misassigned motif is not settled by the data.
The informative part of a DM site is two ACA/TGT triplets separated by one or two bases, so an effective width of about seven bases plus AT-rich flanks: chance matches are frequent. All seven vertebrate paralogues select nearly identical sequences in vitro, so a hit identifies the family, never the member, and expression is the only practical discriminator. AT-rich renderings of the site overlap homeodomain space, which is visible in the Vierstra AC0092:DMRT/HOXB archetype. Most models here derive from in vitro selection on isolated DM domains, which supports the sequence preference but says nothing about genomic occupancy. The JASPAR mouse Dmrt1 model MA1603.2 does not match the family consensus.
References
- Zhu L, Wilken J, Phillips NB, et al. Sexual dimorphism in diverse metazoans is regulated by a novel class of intertwined zinc fingers. Genes Dev 2000;14:1750-64. PMID 10898790
- Murphy MW, Zarkower D, Bardwell VJ. Vertebrate DM domain proteins bind similar DNA sequences and can heterodimerize on DNA. BMC Mol Biol 2007;8:58. PMID 17605809
- Murphy MW, Lee JK, Rojo S, et al. An ancient protein-DNA interaction underlying metazoan sex determination. Nat Struct Mol Biol 2015;22:442-51. PMID 26005864
- Raymond CS, Shamu CE, Shen MM, et al. Evidence for evolutionary conservation of sex-determining genes. Nature 1998;391:691-5. PMID 9490411
- Raymond CS, Murphy MW, O'Sullivan MG, et al. Dmrt1, a gene related to worm and fly sexual regulators, is required for mammalian testis differentiation. Genes Dev 2000;14:2587-95. PMID 11040213
- Matson CK, Murphy MW, Sarver AL, et al. DMRT1 prevents female reprogramming in the postnatal mammalian testis. Nature 2011;476:101-4. PMID 21775990
- Smith CA, Roeszler KN, Ohnesorg T, et al. The avian Z-linked gene DMRT1 is required for male sex determination in the chicken. Nature 2009;461:267-71. PMID 19710650
- Andersson LS, Larhammar M, Memic F, et al. Mutations in DMRT3 affect locomotion in horses and spinal circuit function in mice. Nature 2012;488:642-6. PMID 22932389
- Turnbull C, Rapley EA, Seal S, et al. Variants near DMRT1, TERT and ATF7IP are associated with testicular germ cell cancer. Nat Genet 2010;42:604-7. PMID 20543847
- Kopp A. Dmrt genes in the development and evolution of sexual dimorphism. Trends Genet 2012;28:175-84. PMID 22425532
- Jolma A, Yan J, Whitington T, et al. DNA-binding specificities of human transcription factors. Cell 2013;152:327-39. PMID 23332764
- Weirauch MT, Yang A, Albu M, et al. Determination and inference of eukaryotic transcription factor sequence specificity. Cell 2014;158:1431-1443. PMID 25215497
- Murphy MW, Sarver AL, Rice D, et al. Genome-wide analysis of DNA binding and transcriptional regulation by the mammalian Doublesex homolog DMRT1 in the juvenile testis. Proc Natl Acad Sci U S A 2010;107:13360-5. PMID 20616082
20 motifs in this family — JASPAR 5 · HOCOMOCO 6 · CIS-BP 6 · Vierstra 3
| motif | database | consensus | bits |
|---|---|---|---|
| DMRT3.H14CORE.0.S.CSC | HOCOMOCO | AATGTATCAATTT | 17.4 |
| DMRT1.H14CORE.0.PS.APSA | HOCOMOCO | tGctACAtTGTatC | 14.2 |
| DMRTA.H14CORE.0.SM.BSMB | HOCOMOCO | AATGTAACAAt | 14 |
| AC0283:DMRTC:DM | Vierstra | TGTATCGTT | 13.9 |
| DMRTD.H14CORE.0.SM.BSMB | HOCOMOCO | ATTGATACAT | 13.5 |
| M04685_2.00 | CIS-BP | TGTATCgTT | 12.9 |
| DMRTC2 (MA1479.2) | JASPAR | AATTGaTACAT | 12.1 |
| DMTA2.H14CORE.0.SM.BSMB | HOCOMOCO | AaTGTAaC | 12 |
| Dmrt1 (MA1603.2) | JASPAR | TACAAAGTa | 11.6 |
| M01927_2.00 | CIS-BP | TGTATCA | 11.2 |
| M01931_2.00 | CIS-BP | AATgTAT | 11 |
| M04694_2.00 | CIS-BP | aaTGTAaCAa | 11 |
| DMRT3 (MA0610.2) | JASPAR | TGTATCA | 10.5 |
| DMRTA1 (MA1707.2) | JASPAR | tTGtTACAtt | 10.5 |
| DMRT1.H14CORE.1.M.CMC | HOCOMOCO | tCGcTACATT | 10.5 |
| AC0284:DMRT/DMRTA:DM | Vierstra | TGTAtCAA | 9.5 |
| AC0092:DMRT/HOXB:DM | Vierstra | ATaTATT | 9.3 |
| DMRTA2 (MA1478.2) | JASPAR | GtTACA | 8 |
| M01928_2.00 | CIS-BP | TGTadC | 4.5 |
| M01930_2.00 | CIS-BP | gT | 1.9 |
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.