We don't use the Talairached version of the MNI 2009c dataset, but we do have the MNI 2009c in MNI space. The Desai atlases provide a maximum probability map of Freesurfer parcellations from a group of 75 subjects, affinely aligned to the N27 template. There are a few options:

1. Move the Desai TT_N27 atlas to MNI space. While the transformation from the TT_N27 space to MNI_N27 is exact by a 12-piece affine transformation, the transformation to MNI 2009c and other MNI variants is not. One could use an approximation like the two-piece Brett transform or a simple precomputed affine transformation, but a nonlinear transformation will be much more accurate.
2. Move the 2009c template to the Talairach space. That can be done by manually Talairaching the dataset with the stereotaxic procedure. That's still available inside AFNI in a hidden menu by right-clicking on the DataDir button. You can also just apply the same 12-piece transformation from the TT_N27 dataset to the MNI_2009c template. That won't be quite the same thing, but it will be close.
3. Use the FreeSurfer parcellation of the MNI_2009c template. This is not from a particular subject or from a set of subjects, but it does give a general parcellation with similar regions as the Desai atlas.
4. Use another atlas with the MNI 2009c template. The HCP Glasser and Brainnetome atlases are provided with AFNI.
5. Continue with the Desai atlases on the TT_N27 template. These are fine with the same resolution as the MNI 2009c 1mm3 templates, and the Talairached size is closer to most human brain subjects.