How to implement an HP filter with non-inverting amplifier - doing it right?

[Verdefluox]

Hi folks,
as title says, consider the following: I'm using a NE5534 op-amp in a non-inverting x10 configuration (choice of the op-amp is because of its low-noise characteristics). I'd like the possibility to add AC coupling (so, some kind of switch has to be implemented at the input, APART FROM THE INPUT SWITCH). The circuit is meant to be used as a general purpose amplifier, so I cannot know in advance the input impedance of the source (usually low so to test ripple in power supplies, but I'd like to go up a bit if possible so to use it in various circuits I may work on). Hence the non-inverting configuration. Some thoughts and characteristics:

- Since the x10 is supposed to be a precision one (resistors are 0.1%), initially I thought about not touching at all the non-inverting loop (since a switch has some resistance, and the low-noise requirement already implied using low values resistors). MIGHT CHANGE LATER (read text)

- System is powered by +-5V.

- BW of the overall system would be 10 kHz minimum

- In AC configuration, I'd like to pass > 1 Hz, 10 Hz accepted in case the problem reveals itself hard for whatever reason

- The low-value resistances are chosen to keep the noise level as low as possible

Initially, I thought about using a single HP filter at the input, but I'm having some doubts on how to proceed (refer to the attachment): for instance, the NE5534 has an input resistance of ~300kOhm per datasheet (NOTE: I don't know if this value is what I need to look at), hence I may choose an high value resistor (which circuitally should be in parallel with the input resistance). From this, recover the capacitor C1 from the frequency equation.
First question: if the input resistor goes in parallel with the op-amp internal resistance (hence Req ~= Ropamp), can I completely remove R1 so to simplify things, and hence recover C1 directly from the op-amp internal resistance? Or am I missing something?
Second: In case I cannot remove R1 for whatever reason, can I go "arbitrarly high" when choosing the values of R1 (eventually) and C1 or do I need to consider something particular? Aka. I'd just like to pass the signal as untouched as possible, but the response time isn't particularly important here.

Furthermore, as stated before, an alternative might be adding the capacitor in the non-inverting loop (in series with R3), but the main problem here lies in choosing a switch. In would have to be VERY low resistance not to interfere with R3 (again, resistors are ideally 0.1%). What are my options here (like: multiple switches in parallel so to reduce resistances, MOSFETS, etc..)?

Note that I may go with another op-amp (like a JFET one for higher input impedance in general), but ideally I'd like something < 5nV/sqrt(Hz) and with very low offset (ideally 10's of uV max), and I'm not finding anything particular. IMPORTANT: Regarding the offset, note that the NE5534 is already corrected externally by using another precision opamp, so no problems there. In this question we can simply ignore it.