PNP and NPN transistor are diametrically different: NPN transistor conducts current / switched on when saturated (specified current applied to its base) while PNP transistor inhibits current / switched off when
saturated current applied to its base (cut-off condition).
The NPN transistors on your circuit diagram is arranged in Darlington pair, a simple voltage polarity swap will NOT work on this configuration.
Anyway, you can use PNP transistors by changing the position of the resistors (see: PNP Darlington Pair) and put an inverter (you can built it with a transistor and two resistors, or simply use inverter IC such 7404) right after the opAmp output. But why would you do it (adding complexity and more components to your circuit), while the NPN transistor is as available as it's PNP counterpart?