This is fine as long as you realize that your hand has okay texture: you have the Td to start with, so it's a bit less likely ( e.g., compared to when you only have KQ763 ) that the BB wakes up with TT which is also a game-theoretic autopayoff hand postdraw. Now that you made a pair less than aces, you simply check and if the BB now bets, you have a call ( at least according to game theory and your likely range predraw; of course, many players in the BB won't be bluffing optimally or too often ). At the limit you were playing at, the BB must defend more than 7/15 of the time; otherwise, he can be exploitable predraw, so he really must defend against an aggressive / exploitive / optimal playing SB with 33 and a hand like AK suited. It's not altogether clear that he simply defended with his 55 because he noticed you had raised several times from the SB when it was his BB; statistically, these events are bound to happen, and he may simply never held a pair or playable draw on the previous five hands. The postdraw action is normal: 55 has significant showdown value since the SB should have 22 to 44 sometimes since the SB's bluffing range will normally be hands like AK/AQ.
How wide can the SB open raise in this spot? It's best to consider the one street "toy game" where the SB antes $0.10 and the BB, $0.25 and each of the players are "dealt" a real number which is a uniform random variable from the interval [0,1] which represents a poker hand ( 0 is the "nuts" and 1 has no chance ); assume these random variables are independent. The BB must defend with the legitimate range of [0,7/15); otherwise, the SB can profit by raising his entire range. For the BB's "hand" of 7/15, for him to be indifferent to calling/folding, since the pot now has $0.75 and his calling amount is $0.25, the SB's range must have 1 bluff for every 3 legitimate hands; therefore, the SB's bluffing range has a size ( "measure" is the technical mathematical term ) that is 1/3 that of his legitimate range of [0,7/15], so the best range includes [0, (7/15)(4/3) ) or [0, 28/45) . Now if you convert that 28/45 to actual five card draw hands, forgetting about the card removal effects after the players outside of the blinds have folded ( which tends to make the frequency of the ranks of cards biased towards the higher ranks ), C(52,5)(28/45) ~ 1617130.667, so subtracting off the combinations of hands that are a pair of deuces or better, there are about 32071066.67 combinations. The number of 8 -out to 15-out draws excluding a pair is 137100 ( a number I happen to remember offhand ), and there are 164x960 combinations of AK hands that aren't also a flush draw, which leaves about 26170.67 combinations. There are 25704 combinations of AQ(suited) hands, so we can get pretty close to 28/45 of all draw hands; however, I don't think against a nonexplotable and aggressive BB, the 28/45 frequency is correct - I think it's more likely that the optimal opening frequency is slightly lower than 28/45 and some hands to cut out that have already been mentioned are the weakest openended straight draws such as 8765 to 5432 and maybe a very weak flush draw. Also, if you started with a pair of deuces with a flush draw ( or a slightly higher pair with a 12-out or 15-out draw ), you would normally break the pair. It's also not clear what you would like to do with specifically AKQJx: if you draw one to this, then you probably need to reduce the combinations of ordinary flush/straight draws to open with; if you draw three to this, then you may not be able to bet out every 3-card draw that isn't at least 22 ( and BTW, with a hands that are already 22 with AK/AQ/AJ/KQ/KJ, it's probably better to draw three to the highest kickers if the BB just calls ).