You need a circuit to stretch the pulse, since 2.5 microseconds is too short to perceive.
Here is something I knocked up quickly in LTSpice (a free simulator from Analog Devices/formerly Linear Tech). Here, the input pulse turns on Q1 which quickly charges the capacitor C1. The charged capacitor then turns on Q2 which drives the LED. The resistors R3 and R2 discharge the capacitor and ensures that Q2 turns off the LED when the voltage of C1 drops below a threshold voltage (otherwise the LED would slowly dim down). With this circuit, the LED stays lit for over 100 ms, with an input pulse of 2.5 us.