Try colony PCR i.e, directly from bacterial colony rather bacterial DNA as template.

Colony PCR usually works well. Be careful not to overload your PCR mix with bacterial biomass. It is sufficient to just pick the colony and transfer it to the reaction mix. If you add to much of bacteria, you end up with inhibition problems. Alternatively, add a colony to 0.5 ml PBS, boil for 5 min and spin down. DNA is in the supernatant and can be used as template for your PCR.

Make reaction mix as for a normal PCR including the primers for the target gene and add bacterial cells which will be your template (by just touching the bacterial colony with a microtip on the plate). Add an initial cell bursting step in the PCR steps that you follow, like 95 degree for 3 min.    

if your primers are specific for bacterial dna then they will amplify even in the presence of large amounts of human genomic dna so no further purification is needed. If amplification is weak then re amplification of a very dilute first round pcr product using nested primers will give large amounts of bacterial amplified product

The colony PCR is a good option. Pick the minimum quantity of bacteria that your eyes can appreciate. The final volume of each PCR can be 25 µl, but I've even done this kind of PCR in 12.5 µl when I've to check many colonies. The protocol is the same than a standard PCR, but you can boil the bacteria previously or you can extend the time in the first step of denaturing DNA, before doing the repeated cycles. E.g: 95ºC 5 min for E.coli or Lactococcus; 96ºC 7 min for Lactobacillus

I have repeated the PCR for three times: I have got nothing in Gell electrophoresis. 

I have grown bacteria on LB media and annealing temperature was 56 degree centigrade. I have got no band in Gell electrophoresis. 

16S RNA