I agree with Ikhtifar Rafi, you need at least 5 of them, but if you are doing preliminary analysis you can use one for positive, one for negative and one experimental group. After that you need to add more groups.
Also doses should be estimated based on previous similar researches or based on you own experience in experiments (or your supervisor).
Get a hold of a basic statistics text. You need to understand basic probabilities so that you know how many animals will be needed to give your experiment adequate power. You can't simply guess a number of animals, certainly your animal care committee will expect to see some sort of justification for your animal numbers.