File Name : figure s1.jpg

Caption : figure s1. the structure-based sequence alignment for hcgas (pbd id: 6ct9) and mcgas (pdb id: 4o6a). the secondary structure analysis from the aligned sequence was completed by espript 3.0 tool.3 the hcgas from hghd and hgmd possess the residues 161-521 and mcgas from mgmd and mghd possess the residues 146-506. the red boxes represent the identically overlapped regions. the elements above the sequences show the secondary structures of hcgas and mcgas. the medium (α) and small squiggles (η) describe the α-helices and 310-helices, respectively. arrows represent β-strands and tt letters show the β-turn. the “x” letters represent the cym residue connected to one zinc ion. the green box is the zn-thumb domain.

File Name : figure s2.jpg

Caption : figure s2. the treatment of the hdna and mdna sequences. left side is the overlapped structure between the hcgas-dna complex and the mcgas-dna complex, with the structure of the full dna length being modelled from the crystal structure of pdb ids 6ct9 and 4o6a, respectively. right side is the alignment of the two dna sequences according to their positions in the overlapped structure in the left. the colors of cgas and dna are the same as the description of figure 1 in the main text. after structural alignment between hghd and mgmd, we found that two identical dna structures did not completely overlap, but shifted by one base relative to each other. to achieve a one-to-one correspondence between the residue-nucleotide interactions in hcgas-dna and mcgas-dna complexes using cgas structure as the reference, one terminal base pair of the original dna sequence in the two structures (t18-a19 in the hcgas-dna complex and t1-a36 in the mcgas-dna complex) was removed, which gave the hdna of the sequence d(tttcgtcttcggcaatt) and the mdna of the sequence d(ttcgtcttcggcaattt).

File Name : figure s3.jpg

Caption : figure s3. the pka values for the residues lys395 and asp213 in mcgas as predicted by h++, propka, and delphipka tools. most residues are predicted to possess the default protonation states by all the three methods above, while inconsistent protonation states are assigned for two residues: mcgas lys395 and asp213. as shown, the pka values of mcgas lys395 and asp213 predicted by propka are 6.41 and 7.11, respectively, while they are considered as the default protonation states by h++ and delphipka predictions. therefore, we choose to treat the hcgas and mcgas residues arg and lys in a protonation state and his, glu, asp in a deprotonation state. 

File Name : figure s4.jpg

Caption : figure s4. a snapshot of the zinc-thumb domain in hcgas during 50 ns simulation without restraints. the four residues of hid390, cym396, cym397, and cym404 are shown as sticks, where c-, n-, o-, h-, and s-atoms are displayed as pink, blue, red, white, and yellow, respectively. hcgas protein is shown as green cartoon and zinc ion is shown as grey ball.

File Name : figure s5.jpg

Caption : figure s5. (a) left: superposition of apo hcgas between md structure (yellow) and crystal one (pdb id: 4km5, pink). right: superposition of apo mcgas between md structure (limon) and crystal one (pdb id: 4k8v, blue). note that wild-type lys185 and leu195 in the hcgas crystal structure (pdb id: 4km5) were mutated to asn185 and arg195, respectively, in order to compare with mutant hcgas-dna complex (pdb id: 6ct9). (b) superposition of apo hcgas (yellow) and the complex hghd (pdb id: 6ct9, green protein and blue&purple strands) obtained from md trajectory. (c) superposition of apo mcgas (limon) and the complex mgmd (pdb id: 4o6a, magenta protein and cyan&orange strands) obtained from md trajectory. the close-up pictures of the activation loop close to its catalytic center and the kink region are shown in (b) and (c) for hcgas and mcgas, respectively. the predicted important residues were shown as sticks. 

File Name : figure s6.jpg

Caption : figure s6. double stranded dna sequences in hcgas-dna (pdb code: 6ct9, top) and mcgas-dna (pdb code: 4o6a, bottom) complexes. the nucleotides resolved in the pdb structures are in black, while the missing nucleotides are in either red or blue. the segment in the black box is used to align the modelled b-form dna using avogadro software to the one in the crystal structure.