Weak promoter activity in pGL3 reporter assay - (May/24/2013 )

I have recently cloned a 2.0 kB insert which I thought may be a promoter into pGL3 and have discovered it has no activity compared to pGL3-Basic (actually, it has about 0.3 x compared to pGL3 basic, but this is understandible because pGL3-Basic has low background activity in some cell lines).

The way I identified the insert as a putative promoter region was by identifying the location of the transcriptional start site (TSS) through 5' RACE. Based on this result, I cloned in the 2 Kb piece, 1.5 kB upstream of the TSS and 0.5 kB downstream.

I know fo four reasons why this could be a negative result:

Possibility 1: something wrong with plasmid identity, integrity or concentration used in the transfections to test for activity.
Possibility 2: something wrong with the sequence.
Possibility 3: ATG start sites that are downstream of the transcription start site but upstream of the ATG start site for luciferase may interefere with the reading of the ATG for luciferase.
Possibility 4: the 2.0 kB insert is missing crucial regions, either A.) part of the promoter, B.) missing an enhancer.

I have ruled out possibility 1 and 2 by doing restriction digests, agarose gels, and sequencing to suggest that the plasmid I transfeceted is correct. Notably, there is a great Kozak sequence at the ATG for luciferase.

For #3, there are 3 ATG sites between the ATG of luciferase and the TSS, but they have poor Kozak sequences.

For #4. I did clone the 2.0 kB putative promoter insert into the pGL3-enhancer vector, which is the same as pGL3 basic but includes the SV40 enhancer. This did not improve activity.

Questions:
1. Is there another possibility I am not thinking about?

2. Could a strong splice donor site between the TSS and the ATG for luciferase interfere?

3.Any other ideas on how I can identify the problem rapidly?

Thanks for any consideration.

1. Have you run a promoter prediction using online programs to see how likely the sequence contains a promoter? The best program for doing this is the BDGP: Neural Network Promoter Prediction
2. Does the sequence contain CpG islands which is a good indicator for the presence of promoter?
3. How big is the ORF for each ATG?
4. How is the activity of the 2kb insert compared to 1.5 kb insert?
5. How did you control your transfection efficiency?
6. If other technical issues have been ruled out, the low activity may have resulted from upstream silencers in the insert, you can do serial deletions to find out.

Thanks for the very thoughtful response, I will do (1) and (2) as you suggested.

For (3). The ATG's between the TSS and the Luc ATG have subsequent in-frame stop codons. For the most upstream one it is 56 codons (amino acids) between ATG and in-frame stop, and the other two it is 21 and 13 codons. All end before the ATG for luciferase.

For (4), I have only cloned a 2kB insert (that is inclusive of 1.5 downstream and 0.5 upstream of the TSS).

For (5), my promoter-pGL3, pGL3-Basic, and a positive control promoter-pGL3 are transfected during the same experiment, all samples with renilla. All have strong activity above background for Renilla, so there appears to be adequate amount of transfection.

For (6), thats a good idea and we can test your suggestion.

Thanks again.

>>For (3). The ATG's between the TSS and the Luc ATG have subsequent in-frame stop codons. For the most upstream one it is 56 codons (amino acids) between ATG and in-frame stop, and the other two it is 21 and 13 codons. All end before the ATG for luciferase.

The ORFs in front of luc is problematic to me. Try to include less sequence downstream of the TSS.

>>For (4), I have only cloned a 2kB insert (that is inclusive of 1.5 downstream and 0.5 upstream of the TSS).
Sorry for the misunderstanding. This is the same idea as point #6.

>>For (5), my promoter-pGL3, pGL3-Basic, and a positive control promoter-pGL3 are transfected during the same experiment, all samples with renilla. All have strong activity above background for Renilla, so there appears to be adequate amount of transfection.

I believe you obtain consistent renilla activity between pGL3 basic and pGL3 insert, right?

another question: how is the luc activity compared to renilla? If it is lower, can you increase the ratio of luc to renilla plasmid at transfection?

Yes, the level of renilla is about the same for inset-Luc and pGL3 basic-Luc samples.

Luc activity is lower than renilla, but this can be changed by lowering the amount of renilla plasmid in the transfection. The Luc activity for all samples is well above background. This includes pGL3-Basic...again, it does have activity above background, so I use enough lysate to see Luciferase activity in all samples.

I did try BDGP for promoter prediction as suggested with a postive control promoter where I know where the TSS is and it appears to have missed this control, but shows there could be a promoter in other locations in the test sequence (about 2000 bP). I will try other promoter prediction software. I also looked for GpC islands on the UCSC genome browser and there appears to be no islands in the region, so that may be a sign that I have missed the promoter.

Yes you are right. Promoter prediction programs behave poorly for CpG-less promoters.

I don't know which species you are looking at, if your species is covered by dbTSS, checking for TSS in the database can give you some idea whether and where transcription can be initiated from your sequence. the TSSs in the database are experimentally derived and reliable.