I am working with a protein of approx. 45kDa, I want to take spectra in far UV region, protein is highly prone to aggregation so I am getting very low conc. Please suggest what is minimum conc. needed to get a good spectra.
Minimum concentration of protein at which ellipticity reached -10 to -15 m degree at 222 or 208nm for helix rich proteins and at 218 nm for beta sheet rich proteins so that you can differentiate native protein signal and signal from random coil.
For typical measurements in a 0.1-cm cell: 0.05–0.2 mg/ml protein; for measurements in 0.01- to 0.02-cm cells: 0.2–1 mg/ml protein; and for 1-cm cells: 0.005–0.01 mg/ml protein.
As Md. Anzarul Haque said, the most important is the quality of the signal you get (which will depend on protein concentration, type of secondary structure content, cuvette path length, bandwidth, acquisition time and the voltage at the detector). For an alpha-helical protein I would trust a raw signal of ~ -5 mdeg (at 222 nm) provided independent repeats overlay.
In your specific case (and providing that your protein contains reasonable amounts of secondary structure) you could probably go down to 1 uM in a 4 mm path length (1 cm fluorescent cuvette turned side-ways). I would ensure sufficient scanning time per point (10-20 s) and you can increase the bandwidth to have more light reaching the detector (therefore reducing the noise) if needed.