First of all, why spend over $400 million for a Shuttle mission to rescue a satellite that costs less than $300 million? Sure, government spending is often stupid but that would be absurd.

Next, I don't know the current orbital parameters but Atlas-Centaur missions typically launch into a Geosynchronous Transfer Orbit with a perigee down around 150-200 miles, an apogee out around 22,300 miles, and an inclination of 28.5 degrees. If they've been circularizing the orbit, the perigee would be higher. It'd take a lot of propellant to lower the orbit to a circular orbit at or below 350 miles where the Shuttle could reach it. Since a propellant malfunction is at the root of TDRS I's problems, it wouldn't surprise me if there isn't enough propellant to do that job.

Actually, here are the currently posted element sets for TDRS 9 with the element definitions defined here:

TDRS 9
1 27389U 02011A 02080.40361895 -.00000082 00000-0 10000-3 0 197

2 27389 21.3677 330.2800 6198498 183.1119 169.5508 2.06579675 306

It looks like the current inclination is 21.3677 degress, which is well below the minimum 28.5 degree inclination the Shuttle can reach. The eccentricity is still pretty high at 0.6198498 but it looks like they had been circularizing the orbit following launch. You can tell that by looking at the element sets for the Centaur rocket body. The eccentricity of that is 0.6848606.

ATLAS 2A CENTAUR R/B
1 27390U 02011B 02079.71169396 .00030668 00000-0 47133-2 0 45

2 27390 27.0103 328.6654 6849606 186.1785 156.1247 2.85254867 323

I wish I still had Satellite Toolkit loaded on my computer. I'd model those orbits and see if they're too far along to lower the satellite to rescue altitude. Since they have not lowered the inclination too much yet, there just might be enough fuel left. It still wouldn't make any economic sense to do a rescue mission, but when as ecomonic reality ever stopped the government from doing something?