Testing and optimization

Initial note:

When using probes that have been previously described, it is important to always recheck the probes against an up-to-date rRNA database such as SILVA for current specificity and group coverage!


This test can be done using e.g. the online-tool probeCheck.

1. in silico Probe specificity

Ideally, the probe of interest shows at least a single base mismatch to all non-target microorganisms. For the design of new probes, it is important to keep these discriminatory positions central since a mismatch between probe and non-target rRNA at the 3'- or 5'- end of the oligonucleotide is only weakly destabilizing.


Tip: Competitor oligonucleotides have been shown to strongly enhance single mismatch discrimination by supression of unspecific probe binding to a particular one-mismatch sequence. Competitors are unlabeled oligonucleotides which are fully complementary to the mismatch-containing non-target sequence. Give it a try, it really works!

2. Empirical probe testing

If possible, new probes should be tested by FISH of isolates which have none, one, and more mismatches to the oligonucleotide. A series of hybridizations is performed at increasing stringency by either increasing the temperature of hybridization or by increasing the concentrations of a denaturing agent like formamide in the hybridization buffer (often referred to as "melting curve"). The changes in the fluorescence intensities of individual cells can be quantified by computer-assisted image analysis (Neef et al., 1996) or by flow cytometry (Fuchs et al., 1998).


The most desirable hybridization stringency often occurs at the point immediately before the target cell fluorescence begins to decrease ("drop-off point"). At this formamide concentration, hybridization to the non-target organism should be low or absent. As a rule of thumb, an 18mer oligonucleotide with a G+C content between 50 and 60% will start to dissociate from its fully complementary rRNA-target at a formamide concentration of approximately 30–40% in our standard buffer at 46°C.


Frequently, new probes are designed to target yet uncultured microorganisms which are only known from their rDNA sequences. In this case, it is not possible to test the probes using isolates. Two strategies are available to optimize the hybridization conditions:


  • (i) The cloned 16S rRNA gene of interest can be transcribed to RNA in vitro which is then blotted on a nylon membrane and hybridized with a labeled oligonucleotide at increasing levels of formamide (e.g., Pernthaler et al., 1998). This quite laborious method is based on the assumption that the temperature of dissociation from isolated rRNA is the same as from rRNA in fixed cells.
  • (ii) 16S rRNA clones carrying the target sequence of a new probe are used for adjusting the hybridization conditions (CLONE-FISH). Firstly, clones are grown with chloramphenicol and IPTG. If a vector with an inducible promoter upstream of the multiple cloning site is chosen, then this treatment leads to an in vivo transcription of the cloned 16S rDNA and accumulation of 16S rRNA of the uncultured organism inside the E. coli cell. After standard fixation of these clones they can be used as analogs to cultured organisms for determining the melting point of probes (Schramm et al., 2002).