This is the first shadowed image I have seen of rSP-D (original by oft cited Crouch et al) which has what I think is an Ntermini ring side by side configuration, leaving a dark hole in the center of the multimer. This has been seen with various AFM images which is where I first got the idea that the center might not be end to end Ntermini connections, but side to side.  Normalized to a diameter of 100nm (which is arbitrary since no ones measurements have been consistent) the central ring which might be the Ntermini is about 7nm in thickness, the white dotted circle is where the first and highest peak (in AFM images) would occur and is pretty clearly visible here at about 34nm diameter.  A comparison with the place for the 1st LUT peak in the AFM images would be necessary but it looks like it will coincide nicely. There is obvious disadvantage of background noise in shadowed images, but this particular image fits the concept that at least some of the time the Ntermini are side to side.