I have a theory that I reckon is reasonably accurate, but I've no way of proving it other than trying to do so through technical specifications, so I'll be happy if anyone can conclusively prove me wrong on this. Anyway...
The Astra satellites at 19E & 28E largely follow, at least for services on European-wide & European-based spot beams, a 27 MHz wide transponder bandplan for the FSS & Telecom Ku Band frequencies (between 10.70 - 11.70 & 12.50 - 12.75 GHz) - this dates back to implementing such a plan when they originally launched the DTH satellites in the late 1980's & early 1990's, as opposed to the more common 36 or 72 MHz transponders in use by most other satellite operators. A narrower bandwidth back in these analogue television days allowed Astra to squeeze in a few more channels in a given 250 Mhz block of bandwidth at the slight expense of picture quality, though also slightly better marginal receiving properties.
As DVB-S was about to come into commercial deployment in the mid-90's, with the upcoming launches of Astra 1E & 1F these satellites were to use the DBS frequencies between 11.70 - 12.50 GHz at 19E. Since this was virgin broadcasting territory for Astra at the time, they decided to reserve this spectrum space for digital broadcasts and as such make the transponders on these satellites 33 MHz wide (though there was a analogue Scandinavian sports channel available on 11.720 GHz for a short time here during the 90's) to be somewhat more in common with other satellite companies.
However, when it came to using DVB-S in the FSS & Telecom bands, it was in Astra's interest to maintain compatibility with their original bandplan. Therefore the 33 MHz wide transponders had a default symbol rate of 27500 while the 27 MHz wide transponders had a symbol rate of 22000 - as a rule for DTH broadcasting in Europe, Astra don't do low symbol rate SCPC split transmissions on their transponders.
When Astra started broadcasts for Sky Digital at 28E in the late 90's, they originally used the DBS frequencies (11.7 - 12.5 GHz) (though both Astra 2A & 2B were both capable of transmitting in the FSS & Telecom bands if needed). However, when Astra 2D was to be brought into service, it was decided to dedicate its transmissions in the low part of the FSS band (10.7 to 10.95 GHz). As 28E at this time was now an all-digital orbital slot, in theory a 33/36 MHz transponder bandplan similar to that used in the DBS band could be operated here especially as the satellite was intended to provide UK & Ireland spot-beam coverage rather than a broader European footprint. However since the FSS bandplan on Astra at 19E was a mix of analogue & digital at the time, were there a case to swap satellites between these two positions then to have the minimum of headaches it would be best for a common bandplan to be kept to these two orbital slots, at least by SES (whom operate the "Astra" satellites).
Now you might say "but Fisty, I'm looking at the Astra 2E/F/G page on Lyngsat/Flysay/KingOfSat and there are DVB-S transponders in the FSS band that have a symbol rate of 27500! Look! You're wrong! Ha!" - there is a slightly complex exception to this rule. Prior to the launch of Sky Digital, there was a dispute between SES and Eutelsat about whom had the rights to operate in the Ku Band frequencies at approx. 28E to 28.5E, which was originally home to a German Kopernikus satellite. An eventual settlement at the time allowed Eutelsat to operate in the 11.2 to 11.45 GHz segment of the FSS band as well as the 12.5 - 12.75 GHz Telecom band, with SES allowed to operate in the remaining spectrum. Eutelsat took this opportunity to launch "Eurobird 1" (Later Eutelsat 28A) to reside at 28.5E, close to the neighbouring Astra satellites at 28.2E but visible to viewers as one combined position in the sky. As you can guess, Eutelsat didn't abide by Astra's own bandplan and instead used the more traditional 36 & 72 MHz wide transponders for Eurobird 1, though still using a DVB-S symbol rate of 27500 to maintain compatibility with Sky Digital's receivers**. In addition, since this was "virgin" territory for Eutelsat they decided to have their separately polarised transponders on top of each other on the same (or very near) frequencies from other another - this potentially allowed even more of the available bandwidth to be used in a particular Ku-Band segment, whereas in the old analogue days, signals of different polarities were spaced to be at the guard band gaps between transponders of the opposite polarity to minimise cross-polarity interference (it seems DVB-S/S2 is a good bit more tolerant to such interference than analogue TV broadcasts were, with more modern LNBs better at cross-polarity rejection as well). However, the Astra frequency bandplan to this day maintains the old "guard band" arrangement, even in the DBS frequency segment, despite no more analogue satellite TV broadcasts existing in Europe...
...however, back in the early to mid 2010's with the original satellites at 28E now getting on in old age and SES planning to launch Astra 2E, F & G in their place (with no public sign of Eutelsat replacing Eurobird 1/Eutelsat 28A), the original agreement between SES & Eutelsat was due to expire and went back to court in Germany, whom determined that SES now had the rightful claim to all Ku-Band spectrum at 28.2/.5E. An agreement between the two operators saw Eutelsat agree to release transponder space on the new Astra satellites (what Eutelsat now term as Eutelsat 28E, 28F & 28G). With SES now able to access the 11.2 - 11.45 GHz & 12.5 to 12.75 GHz bits of Ku-Band spectrum at 28E, it now had a headache in looking to see how the present transmissions in these frequencies could fit into the Astra bandplan. What did they do? Essentially, they didn't do much at all! The new Astra 2x satellites were now technologically capable enough to configure its own transponder bandwidths & polarities and in the end they decided to largely keep the old Eutelsat bandplan in the 11.2 - 11.45 GHz segment - there was talk at the time as to how this might have effected Freesat receivers that were designed to see a "home" transponder and as to wherever certain receivers could be amended with an OTA software upgrade to change this, in the end nothing was changed. As for the Telecom 12.5 - 12.75 GHz band, SES decided to clear this space of all European DTH broadcasts and instead it is used in Europe now for satellite feeds (mostly outside broadcasts).
Hopefully the above explains how spectrum & symbol rate usage has evolved on the 28E orbital slot over the past 25-30 years without leaving you too confused - essentially it's down to Astra's analogue legacy bandplan, even though Astra, with one exception, never used 28E for analogue DTH broadcasting. That one exception was back in the early days of Sky Digital when a lone transponder carrying a BSkyB testcard was radiated presumably to help Sky installers set up customers for digital while having only an analogue meter to help them align the Sky minidish. This analogue transponder didn't last any more than two years IIRC.
Also, it was a nice co-incidence that on DVB-S a symbol rate of 27500 with an FEC of 2/3 gave the exact same net bitrate as a symbol rate of 22000 and an FEC of 5/6 - the latter needing a stronger signal however. With the current 2E/F/G satellites in place being more powerful than their predecessors, the transponder operators have been able to change the FEC (and modulation in the case of DVB-S2 transponders) to allow a higher net bitrate for their transmission whilst giving a similar "rain fade" margin compared to that from the original 2A/2B/EB1 satellites.
** Original Sky Digital DVB-S receivers had for their "Other Channels" feature to tune into services not on the Sky EPG a limit on their symbol rate between a choice of 22000 & 27500 - I'm not sure if channels on the EPG themselves could have been displayed from transponders with a different SR (I don't think we've ever found out as it's never happened IIRC), but for channels that might want to be accessible to Sky viewers without EPG access then the transponder SR needed to be one of these two values.
