Hi,

you could consider to use the conserved regions and then run a PCR through this. You might have a look into the attached paper as the primers work quite well. If you have very low amounts of bacterial DNA I would suggest to use a mastermix which does not carry bioburden to not bias your results. If you are interested to get more background please let me know

Sven

You could use the middle region to design primers that face "outwards" towards the 27F or 1492R primers. Check that you are using regions that appear unique to your bacteria of interest.

The problem with any PCR based approach is that it is practically impossible to design primers that are specific to any one 16S sequence. Your PCR reactions will always "work", but you will have no idea about product specificity. Even if you design primers that appear to specifically match your sequence of interest, you can't be sure that there are not many very similar sequences in the mixed population that you are going to try to amplify from. What happens when you BLAST your sequence of interest or look for it in a 16S database? Do you get hits that are identical or very similar? This should give you some idea about the difficulty of making a specific design. You might be able to achieve greater specificity if you could do PCRs with both primers specific to two different areas of your sequence. One wild idea would be to make 27f and 1492R primers with restriction site linkers on the 5' ends (choose a sticky end enzyme that does NOT cut 16S). Then PCR with these, cut with the enzyme and circularize the product by ligation. Then use outward facing primers designed with 3' ends specific to your sequence to amplify from this circularized template. This is essentially an "inverse PCR" approach. Alternatively, you could just design two sets of "specific" nested primers, one set facing toward the 27f primer and another set facing toward the 1492R primer. The problem with this approach is that you would probably also need nested 27f and 1492R primers in order to get clean products from your nested PCR. But this might be the easiest thing to try first and then move to the inverse PCR approach if it does not work. With less than 200 bp of sequence that you know, any of this may be impossible if the more specific nucleotide differences are not conveniently located!

Another wild idea would be to design a probe that is specific to the distinct portion of your 16S sequence of interest. Then dilute your sample to the point where only 1/3 to 1/4 of a series of parallel reactions with 27f + 1492R give a product. In that situation, you can be practically sure that each of these products arises from a single molecule. If you then test the positives for reaction with your probe (there are many ways this could be done based on what your equipment and budget are), you should find the 27f+1492r products that contain your sequence of interest The great advantage of this approach is that the PCR products that you eventually sequence will have been generated only with the 27f and 1492r primers, so that if they are identical to your sequence of interest, it will not be because you used primers containing those sequences to generate them--a very important consideration. The alternative would be to somehow make a clone library of 27f+1492R products and then screen with the probe. The dilution/PCR approach just avoids the cloning element.