The mDNMT1 and hDNMT1 complexed with DNA are shown below, these structures have been solved at a resolution of 3 angstroms:
hDNMT1 (3.6 angstroms resolution)
hDNMT1 (3.6 angstroms resolution)
mDNMT1
As you can see they are quite similar in structure but with mouse DNMT1 is complexed with 19bp DNA and AdoHcy (S-adenosyl homocysteine). The DNA contains two unmethylated CpG dinucleotides seperated by 8bp. The following domains we are able to see:
The CXXC domain is connected to the BAH1 domain by a long linker sequence at the opposite ends of the methyltransferase domain. The paper also shows BAH1 and BAH2 are seperated by a long alpha helical linker with both BAH positioned on the surface remote from the bound DNA. Another interesting sequence is the (GK)n linker, this connects the BAH2 domain to the catalytic domain but is disordered in the complex, shown below.
The catalytic domain contacts both BAH domains and DNA forming the core of the complex, it also contains a molecule of AdoHcy in the active site.
The mDNMT1 is also complexed with metal cations, 4 Zn(2+) in the structure where two in the Cys(4) coordination in the CXXC domain, and the other two are involved in single coordination Zn ions in Cys(3)His coordination in BAH1 and in the target recognition domain (TRD) of the methyltransferase domain. The 19bp DNA adopts a B-form duplex in the complex.
The mDNMT1 lacking the CXXC domain and the CXXC-BAH1 linker in the free state at 2.5 angstroms is shown below, it is very similar to the DNMT1 with DNA:
add image
mDNMT1 and hDNMT1 share 85% sequence identity and there it has been suggested there is a common mechanism of action. Differences occur in the methyltransferase domain which is repositioned relative to the CXXC domain and DNA by a 1bp translation the DNA axis.
The paper suggests these differences could arise from the flexibility of the CXXC-BAH1 liner and may also be due to the different packing environment between the two complexes.
