Due to recurrence of requests, I describe here a protol to create gfp fusion constructs ready for injection within one day. The protocol entails a simple fusion PCR, which fuses two primary PCR products with a set of nested primers, schematically outlined in the Figure. In two separate PCR reactions, the promoter (or the complete gene) and the gfp coding sequence + unc-54UTR are amplified; the former from genomic DNA (or a cosmid prep), the latter from the standard Fire vector pPD95.75. The 3 primer for the promoter/gene, termed B, has a 24nt overhang to the gfp vector pPD95.75 (see Fig for details). In older versions of the protocol the 5primer for the gfp coding sequence, termed C, also had an overlap to the promoter sequence (thus making it the reverse complement of primer B), which increased the total overlap of the 2 primary PCR products by a factor of two; however, this was found not to be necessary; a 24nt overlap of the 2 PCR products created by the B primer is entirely sufficient to give a fused PCR product. The obvious advantage of this is that the C primer is no longer promoter/gene specific and can thus be used for different reactions.
An important, though enigmatic trick is NOT to purify (by whatever gel-elution method) the 2 PCR products. Just run them on a gel, eye-ball the concentration, dilute an aliquot of the PCR reaction with water to roughly 10-50 ng/ul of each product (in case the yield of the PCR product is low, it can also be used undiluted; I have encountered cases where the first PCR product was invisible on a gel and nevertheless got a fusion product) and then use 1 ul of each diluted PCR reaction in the fused PCR reaction. For the fusion reaction, nested primers must be used (A* and D*; see Figure).
Although in most cases one will get a single band from the fusion PCR reaction, one can occasionally see another band, possibly some sort of a GFP-dimer; sometimes this additional band may even be much stronger than the fusion PCR product. It can be ignored and considered as some sort of carrier DNA for the injection. The concentration of the fusion-PCR is eye-balled on a gel and the DNA, again NON-purified, injected into worms at a final concentration of rougly 20-50ng/ul. As injection marker, rol-6 in N2 and pBX into pha-1 has been used successfully. Note that neither of these injection markers has any sort of sequence overlap to the co-injected PCR product, yet co-segration of the injected DNAs has virtually always been found.
We typically use 4 kb of promoter; other people used this protocol successfully with pieces of >10 kb. For the PCR reaction, I had best luck with Boehringers Expand Long Template PCR system using their buffer #2 and a PCR program that they recommend in the datasheet (which adds extension time for every cycle).
Provided that PCR #1 worked (in those few cases it did not work initially, shifting the A and/or A* primer usually eliminated the problem), the fusion PCR never failed. In my exprierence with more than 50 constructs so far, I also never had problems getting lines and in >90% of cases got clearly discernably gfp expression.
A slightly less optimized version of the protocol is hidden in the Method section of Hobert et al., 1999, J.Cell Biol.144, 45-57.
Thanks to Stephen Nurrish for an initial inspiration.
Primers:
A= 5 upstream, approx.20-25 nt
A* = nested to A (3-10 bp away from A
B = spanning 20-24 nt of end of gene to fuse + 24 nt of gfp-vector pPD95.75 (= sequence of the PLUS strand: 5-AGCTTGCATGCCTGCAGGTCGACT-3). Example: If the 3end of the gene/promoter has the plus strand sequence 5-agagagagagagagagagagag-3 , the whole primer B would be: 5- AGTCGACCTGCAGGCATGCAAGCTctctctctctctctctctctct-3
C = polylinker beginning of pPD95.75: 5-AGCTTGCATGCCTGCAGGTCGACT-3
D=at the end of unc-54 3UTR: 5AAGGGCCCGTACGGCCGACTAGTAGG-3
D*= immediately nested to D: 5-GGAAACAGTTATGTTTGGTATATTGGG-3