Why Wavelet 1-D Denoising in Matlab can only happen on approximation coefficients?

I have been trying to get familiar with Wavelet toolbox in matlab for denoising purposes. I understand that wavelets assumes noise does not affect much low frequency bands which is why thresholding to denoise can only happen on details.

I also understand that details are for HPF and approximation are for the LPF part of the decomposition.

How can I apply thresholding on the approximation coefficients?

Also since my input to the Wavelet 1-D is discrete data without sampling frequency information how are the filter banks frequencies dictated when forming the decomposition?