TL;DR - You probably played it fine, but more information and detail about the hand, your opponents and your thought process would allow people to give better answers.
So now for the long version. ;)
Answering your question
To start by trying to answer your specific question "What have I done wrong in the hand?", maybe nothing, maybe a lot. We're missing a fair bit of information here about our opponents (especially the SB) and so we have to make quite a few assumptions and treat everyone as an ABC player, which is unlikely to be the case and so makes any analysis a bit less useful.
We also assume that the turn and river don't bring a backdoor flush (you state the flop was rainbow but give no information about the suits on the turn and river), although this shouldn't make a significant difference to our analysis of the hand.
Furthermore, we have to make assumptions about the other stacks at the table (I have assumed that everyone has you covered as you're relatively short stacked for a cash game with 50BBs), which is another factor which can affect analysis of a situation.
I would suggest you think about presenting hand history type questions more clearly, to make it easier for people to read and understand them and therefore to provide good answers (I've thrown together a very simple - but clear - summary of my understanding of the hand below).
As is often the case, answering impartially is made more difficult by the inclusion of results - in my opinion we would be likely to get much more helpful discussion if we don't know that the SB has TT. For what it's worth, I would have left the description as:
SB posts $1
BB posts $2
UTG+1 calls $2
MP calls $2
CO raises to $10
Hero  ($105) calls $10
SB calls $10
UTG+1 calls $8
MP calls $8
Flop (5 players, $52) - 4 6 J
SB bets $15
MP calls $15
Hero ($95) calls $15
Turn (3 players, $97) - (4 6 J) T
Hero ($80) checks
River (3 players, $97) - (4 6 J T) 9
SB bets $35
and invited people to answer as to what action they would take on the river facing the bet from the SB (and why), as well as any tangential advice on how they might have played prior streets differently to you.
With regards to what to do on the river, this boils down to assigning a range to the SB, assessing our hand against that range and then looking at our pot odds to see if we can justify a call or not, perhaps additionally considering raising (which would require some further consideration of what we expect Villain to do with each part of his range against a raise).
Villain's Range, street by street
We know nothing about Villain except his position and actions in this hand. We don't even know if this is a Live 1-2 game or an online game (I suspect the former), so we have very little to go on.
Based on the little information we do have though, I would suggest that to make the preflop call, SB could have a range like any pair (perhaps excluding some of the highest pairs which we'd expect them to 3bet with), any two broadway cards (perhaps excluding AK and AQ that we might expect them to 3bet with), any connectors and maybe some gapped connectors, perhaps even wider.
Once SB leads the flop though, we have to narrow this range significantly - they've donked out (albeit on a pretty dry flop) into 5 players so I think we have to give them credit for either holding a strong pair or having connected with the board in some way almost all of the time here. I would suggest we can narrow their range to something like this (a mixture of sets, pocket pairs, top pair hands and straight draws):
44, 66-JJ, [QQ, KK, AA], 75, 53, 87, AJ, KJ, QJ [JT, T9]
(less likely hands in square brackets)
As a point of interest, Hero's
99 is about a 2-1 dog against this range on this flop, when run hot and cold. But we are getting over 5-1 on a call here with the action in front, so I don't believe we can fold.
When SB then checks on the turn, I think we can further narrow their range to:
77-TT, 75, 53, 87, JT
If they were willing to donk with a flopped set, it seems unlikely that they would shut down with a flopped set on this pretty safe turn card. The overpairs now seem all but impossible for the same reason - you'd think with these hands they'd be looking to get money in the pot and if they weren't going to checkraise the flop (where they could reasonably expect to have got a bet out of at least one of the 4 other players in the pot), I can't see why they'd go for a checkraise now on the turn unless they're just exhibiting FPS and trying to be "deceptive".
All of the straight draws still make sense - they took a stab on the flop as some kind of semi-bluff and now they haven't hit a straight (although their gutshot draw to the nuts has improved to a double belly buster), so they're looking to slow down and try to get a cheap river card to see if they hit. 77-99 make sense for the same reason and both TT and JT are still in the mix because they may be following the classic line of betting without a hand and then slowplaying once they hit it because they're scared of losing customers.
Against this range, Hero's 99 is over 70%, so I probably would've just jammed the turn here. Even if we give MP the same range as the SB (which is not entirely accurate probably, but just to illustrate), Hero is still almost 50% to win the whole pot against both of them if they call. Factoring in fold equity as well, I think this would be a profitable move in the long run. Some more detailed analysis of what we'd expect Sb and MP to do with different parts of their range might yield some more concrete guidance here.
When the river comes and they lead out for ~35% of the pot, I would say their range converges to about:
77, 88, TT, 87, JT
99 is removed because it's no longer possible for them to hold it based on card elimination. 77 and 88 are possible because SB believes the ehcks of MP and Hero on the turn signified weakness and he can turn these hands into bluffs successfully, getting someone to fold a 9 perhaps. TT, 78 and JT all make sense as value bets.
Against this range on the river, Hero's 99 is about a coinflip, but getting almost 4-1 makes this an easy call for me. It may even be potentially profitable to jam your remaining $45 as well if Villain might call with worse (like 77, 88 or most likely JT) - you could do a quick EV calculation to figure this out given some assumptions about what Villain would call the shove with.