From this OnSemi datasheet for the TIP42 you can find in a number of places where they spec its use as a switch using \$\frac{I_{_\text{C}}}{I_{_\text{B}}}=10\$. You may be able to get away with a higher value. But I'd play it safer and stick with that value. This means in your case that \$I_{_\text{B}}=150\:\text{mA}\$ to the PNP BJT. That's quite a bit.
The best way to cut that down, and it will work in this case, is to use the NPN BJT with it's full active mode \$\beta\$ value and to select one that guarantees \$\beta\ge 150\$. This means the NPN BJT will not be able to be used as a switch. It must remain in active mode.
This is the way to set that up:
simulate this circuit– Schematic created using CircuitLab
You will have to find an NPN that can guarantee \$\beta\ge 150\$ at the collector current of \$I_{_\text{C}}=150\:\text{mA}\$. Even the 2N2222A can probably get there. But you also have to handle \$300\:\text{mW}\$ dissipation in the NPN BJT. A TO-92 package is about \$\frac{200^\circ\text{C}}{\text{W}}\$ so that's already \$60^\circ\text{C}\$ rise over ambient. Very hot. So it would pay to find an NPN BJT in a better package than a TO-92 here.
I've added \$R_2\$ to pick up some of the dissipation load. It's not strictly necessary. But if you don't include it, then you will need to find a still bigger NPN package and maybe even provide heat-sinking for it. So the addition of \$R_2\$ allows you to off-load some of that heating to a resistor that may be cheaper/easier.
Note that \$R_1\$ itself needs to be half a watt all by itself. No avoiding that.
\$R_3\$ is a stopper resistor. It's optional here. But it may help avoid ringing/oscillation by dissipating some of transition energy when switching.
\$R_4\$ is there to just keep the circuit off if the input control is left floating. It's also not strictly necessary depending on circumstances. But I've added it as a footnote, just in case.
You can replace \$Q_2\$ with a Darlington arrangement. This helps guarantee enough \$\beta\$ and would likely allow cheaper, more readily available NPN BJTs while still guaranteeing high \$\beta\$.
\$Q_2\$ in the above schematic will still get hot. But perhaps that is okay because you are using cheap readily available NPN BJTs.
Bottom line is that you are pushing the limits given your need for output current and the limitation presented by the maximum drive current you want to support. It pushes things. But not beyond the breaking point. So the above shows a few ways you might consider BJT arrangements.
You might also wish to consider the use of a MOSFETs for a circuit like this. The recombination current goes away and this may make an appropriate circuit a little easier to achieve. MOSFETs commonly can tolerate gate-to-source voltages around the \$12\:\text{V}\$ range you need. So extra protection efforts won't be needed to complicate the final result. So that may be a very good approach here, if you are willing to consider MOSFETs. (They cost more, but they work well too.)