Should you first stack for focus or combine some of these images to an HDR?

I recently took some landscape images in very challenging conditions. I ended up shooting several photos which I want to stack for dynamic range (HDR) and focus.

My photos are like this:

1. Foreground Photo 1, Exposure 1 (the foreground/landscape is exposed well, the sky is overexposed)
2. Foreground Photo 2, Exposure 1 (I focussed a bit further away)
3. Foreground Photo 3, Exposure 1 (the horizon line is sharp, sky still overexposed)
4. Sky Photo 1, Exposure 2 (basically the same as 3., but darker – the landscape is underexposed)

So what would be the best workflow to combine these images into one? There are two possibilities:

1. Stack Photo 3 and 4 to an HDR, then combine it with 1 and 2
2. Stack photo 1, 2 and 3, then combine the new photo with photo 4

Which of these workflows is better?