For MS3 we have tested exactly what you are describing with Phil Remes of Thermo, and it worked. I am saying that Skyline essentially ignores the MS2-level isolation information. MS3 has 2 levels of isolation information, the MS1 level - the peptide precursor m/z, and the MS2 level - the fragment ion or ions that get isolated from the first time in the collision cell. These are recorded in the raw data along with the final resulting spectrum. Skyline will extract product ions from essentially any level of spectrum MS2, MS3, etc. as long as its MS1 precursor m/z matches, which works great as long as all of the spectra collected in your raw data for a specific precursor m/z are for the same precursor target. It would not work so well if you were collecting both MS2 and MS3 spectra for the same precursor m/z or you were collecting MS3 spectra for the same precursor m/z value which were differentiated by their MS2 isolation. Both of these turn out to be relatively infrequent, and neither occurred in Phil's test data.
As for the mystery of the missing TMT ions, that is due to the Transition Settings - Instrument - Min m/z setting being 250. Essentially, you have told Skyline that your instrument is not capable of measuring these ions. If you lower this value to 100 m/z, then the TMT ions will be accessible to you. You must have a lower value in the document you created with Skyline 4.2. If you were trying to use this minimum to limit the size of the y- and b-ions that you will accept, then you should instead increase your ion range to something like "ion 3" (i.e. excluding y1, y2, b1, b2) instead of using "ion 1" which allows even y1.
One setting you have which has been a longtime confusion since Andrew Stergachis in our lab used it for an early paper, and that is limiting your ion range to "last ion - 1". Andrew mistook this to mean what he was calling in his paper y(n - 1), where n was the length of the peptide. However, it actually means y(n - 2). The option "last ion" is not referring to the precursor itself. That is, when you have a 10 amino acid peptid, its "last ion" will be y9 or b9. Whereas, its "last ion - 1" will be y8 or b8. This has been confusing enough for Skyline users that I am strongly considering removing the "last ion - n" options and warning if a user tries to click the OK button in a Transition Settings form where a value like this has been brought forward through backward compatibility with an old document.
Were you really intending to exclude the longest possible y- and b-ions and if so, why exactly?
Hope this gets you past your immediate issue, which is not a version related issue. Thanks for supplying your document and working with us on the Skyline support board.