Hi Christian. No matter how you configure your luciferase assays, there is always the potential that it will not completely recapitulate the normal behavior of a given promoter or enhancer. Many, many factors at work here such as

1) copy #, you have thousands of plasmids in a transfected cell compared to only two copies in a diploid genome. Depending on the abundance of nuclear factors that regulate transcription from these loci, dysregulation may occur with excessive copies of your enhancer element

2) chromatin context. While this will impact the normal accessibility, nucleosomal distribution and looping ability of endogenous enhancers, such contextual regulatory elements will be lacking on a plasmid

3) epigenetic modifications- again, histone modifications or DNA methylation will not be the same on plasmid and genomic sites

4) E. coli directed DNA methylation- this can sometimes change enhancer or promoter function in unintended ways

As such, tinkering with design of these reporters is likely to yield diminishing returns.

Nonetheless, my personal experience with intronic enhancers is that they work reasonably well (motif, and factor-dependent activity) if cloned upstream of a minimal promoter (I recommend pGL4.24 from Promega). Coupling these enhancers to strong basal promoters (CMV or SV40) or using consecutive repeats of your element entail certain risks of observing either dampened (former case) or non-physiologic (latter case) results. These approaches have been used successfully in the literature, but I don't recommend them unless the simpler approach of a single copy upstream of a minimal promoter fails. HTH.

Hi Christian. No matter how you configure your luciferase assays, there is always the potential that it will not completely recapitulate the normal behavior of a given promoter or enhancer. Many, many factors at work here such as

1) copy #, you have thousands of plasmids in a transfected cell compared to only two copies in a diploid genome. Depending on the abundance of nuclear factors that regulate transcription from these loci, dysregulation may occur with excessive copies of your enhancer element

2) chromatin context. While this will impact the normal accessibility, nucleosomal distribution and looping ability of endogenous enhancers, such contextual regulatory elements will be lacking on a plasmid

3) epigenetic modifications- again, histone modifications or DNA methylation will not be the same on plasmid and genomic sites

4) E. coli directed DNA methylation- this can sometimes change enhancer or promoter function in unintended ways

As such, tinkering with design of these reporters is likely to yield diminishing returns.

Nonetheless, my personal experience with intronic enhancers is that they work reasonably well (motif, and factor-dependent activity) if cloned upstream of a minimal promoter (I recommend pGL4.24 from Promega). Coupling these enhancers to strong basal promoters (CMV or SV40) or using consecutive repeats of your element entail certain risks of observing either dampened (former case) or non-physiologic (latter case) results. These approaches have been used successfully in the literature, but I don't recommend them unless the simpler approach of a single copy upstream of a minimal promoter fails. HTH.

By definition, an enhancer is an element that operates at a distance from a promoter regardless of position or orientation. The classical way of proving it to be an enhancer is to move it around once isolated. So, the short answer is that you can use any of the promoter-enhancing assays you describe. You ask about weak promoter versus strong promoter--I generally try to use something weak enough so that I can be confident that the additional effect added by the enhancer will be measureable. But I also want to know if the element might act as a repressor. So, I like to have some baseline activity in the absence of the enhancer. But the first thing to do is define your question and design the study accordingly.

Therefore, the design should be considered in the context of what you are asking. For instance, if I want to identify an enhancer working at a specific gene in a specific cell type, then I want too use the promoter from that gene as opposed to a minimal promoter from a different gene (although the latter can be done--it just provides a slightly different type of answer). Because of the specificity by which transcription factors/co-factors binding to enhancers interact with those at the promoters, I mostly am inclined to forget about the weak/strong promoter question and just use the relevant promoter (your real question then becomes 'what are the margins of my promoter'?). That's what's done when your question is 'does this enhancer work at that gene?'. In that case, I also want to compare activities in cell types in which endogenous gene activity varies from inactive to very active. However, there are instances in which you are globally looking for any gene regulatory elements out of the context of any single gene. In that case, you don't have a specific promoter. But you should then be aware that the results depend on the promoter used. See the comments from Daniel Cohen about other caveats of the assay.

Short Background

Enhancers, like all other types of transcription regulatory elements, are little more than a collection of binding sites for transcription factors. So the individual components of an enhancer are not much different than a distal control locus or a promoter except that one (the promoter) happens to be close to a transcription start site. Be aware that there may be elements at the margins of the enhancers that insulate their activity and that you may inadvertently include in your DNA fragment. So, when you do the transient enhancer assay, you are just identifying a collection of sites that constitutes a bona fide gene regulatory element. If active in the enhancer assay in a specific cell type, then you know that particular cell has the transcription factors capable of binding and activating the enhancer. The introduced DNA will not have the chromatin marks that also govern whether the gene is expressed. So, the transient enhancer assay will not tell you about those epigenetic considerations.

The main objectives of the assay is to show/detect that this particular sequence enhances transcription, while another sequence or its "corrupt sequence" as a control does not, using the same very plasmid& promoter. Before going into any in vitro studies to analyse the possible transcription factors, the efforts to screen a multiplex of cell lines and at different contions (e.g. temperature, O2, serum etc) in order to narraw the possible transcription factors involved from the specific cell line and at the specific condition.

I agree with Daniel and Fred useful remarks about using weak promoters.

I should add that a lot of enhancers found in (first) introns are actually alternate promoters. I stumbled upon such cryptic promoters twice in my carreer, for Pitx3 and c-fos genes.

So if you suspect transcription factor binding in your intron, you could also check if some ESTs are starting there, if promoter predictors in silico find something, if there is a TATAbox / initiator sequence, etc... In this case the enhancer quest may become a promoter quest, and it's a different story.

In that case, you also may want to clone your putative promoter in a promoterless reporter (and search for the corresponding RNA).

Plsamid-reporter assay is in vitro assay outside nuclear environment.. Gene regulation is a network with cis and trans factors. We need to think some better way to do intonic enhacer activity.

@Daniel: Thanks for the answer. We have been working with classical promoter elements so far, so cloning them into a luciferase vector (for example) an then do tests, mutate them and so on worked pretty fine. We now may have an intronic enhancer element and were discussing the best way.

@Fred: Thanks for the good points. I pretty aware, that this is an artificial system, especially when it come to the availability of factors and chromatin state. We know from our transcription factor work, that this can be crucial when you look at subsets of expressed genes between different cell types (lines). We are also setting up ChIP-seq to see wheter we can see binding in vivo.

@Raafat: Thats right, we want to identify the binding site to the exact location on the DNA (or show that its not binding, either way). I don't see how changing the conditions helps here. Can you explain this a bit further?

@Vincent: Good point, thank you. In our case we are pretty sure, that this is not the case, but I will put it on our to-do list anyway. Its only one cloning step and can help to clarify things.

@Qing-Rong: Thats why I was asking for other opinions. If there are any, please speak up.

There are always drawbacks to any scientific technique - nothing is 100% fool-proof, but we have to work with the tools that we have in hand. For this reason, I like to combine different approaches to prove the same hypothesis and to include a lot of controls to account for failures in the assay.

Luciferase Assay Controls: One thing that you should definitely do with your luciferase assays are include proper controls. For instance, if you are co-transfecting the transcription factor, also use a different transcription factor as a control. Or if you're not transfecting the factor, knock it down by siRNA/shRNA or use an inhibitor and see the effect on the luciferase expression. Alternatively, if you know the canonical motifs that your factor binds, use site directed mutagenesis to interrupt them.

Other assays: A good place to start is with bioinformatic prediction of how likely these sites are to transcription factor binding sites, and whether the motif for a given factor is present (I'm sure you've probably done this already to make your luciferase constructs). Another thing you could do is to combine the luciferase assays with something like ChIP. This can be as simple as ChIP-qPCR in cells that have no/low/moderate/high expression of your gene. I have found that the Cell Signalling Simple-ChIP kit is very easy to use and gives quite clean results.

Good luck.

There's a method called Chromosome conformation capture, it's a method to see if the enhancer motif can loop and bind the promoter. Shortly, you have to crosslink the DNA and proteins in vivo and then cut the DNA to short fragments, dilute so you have about 1 molecule/reaction and then ligate. In this case only if you have a protein complex which brings the promoter and enhancer close enough to each other you can get a ligation product. After that you just perform a PCR with a primer from the enhancer and a primer from the promoter and see if you have a product. You can find elaborated protocols online.

This can verify luciferase results very accurately.

Many nice points raised in this thread, but IMHO, but not enough consideration is given to how to interpret negative results. For example, if you place your intron within the luciferase cDNA, and you don't see enhancer activity, does this prove anything? Nope, no way to interpret why it failed. Only good if you get a positive result.

And how about 3C type ChIP approaches? Looping can be demonstrated in principle, but this assay is quite finicky (low reproducibility) and does not tell you which DNA-binding factors are important for looping, and has limited spatial resolution. If it works, then great, you have additional support for enhancer function, but still haven't demonstrated a functional requirement.

The ChIP-Seq experiment is critical if you want to make a link between between a specific TF or set of TFs and your putative enhancer. Beyond that, the best bang for your buck might be to try genome editing using the CRISPR system to delete or mutate your enhancer and assay for changes in gene expression. That is, assuming you are working in a diploid cell line that is amenable to transfection or transduction. This approach has some risks, too, on the technical side, but provided your mutagenesis works, there is no ambiguity about how to interpret a negative result.

@Shachar: Good point, thank you. I will keep this in mind.

@Daniel: You are touching the main point. All reporters used are cloned from cDNA, so there is no splicing here any more. If I want to add an enhancer element into luciferase (for example), then I need to transfer splicing sites as well. And make sure, that everything stay in frame. The other thing: Per definition an enhancer is a DNA element, which has reulatory function and is located in some distance to the regulated gene. How much of this "distance" would I add between the promoter and the intron element, when I clone it this way? I don't think there is an answer to this, so this will probably lead to a lot of technical problems (not to talk about comparability).

Additional to the luciferase experiments we will have to get into ChIP-seq (and hope we get a good antibody for our factor) to see if there is real binding.

CASPR is also a good idea, I will have to look into it (since this is not yet established here).

It is not hard to introduce short introns into a cDNA (I can suggest some that have worked for me), and I doubt you would have to worry about the exact spacing/distance since enhancers are generally position-independent. What would worry me the most is that you don't know where the regulatory elements that affect looping reside, and therefore you might run into artifacts related to exactly how you truncate your gene-specific promoter and intronic enhancer.

One other point- if you are working with human or mouse cells, you should check out the ENCODE datasets. They have some long-range DNA interaction data as well as a ton of information about factor-specific ChIP-seq, and epigenetic modifications that would be associated with an enhancer. These datasets don't translate universally across cell types but they offer some potential insights for designing your experiments.

I am no expert on the topic but I would like to propose an idea to study the enhancer effects of intronic regions.

Assuming that you have already selected which intron to study and therefore you would know its distance from the original promoter of the original gene the intron is isolated from. Since enhancers work by affecting the structure of chromatin hence distance is important (although different distances for different enhancers), I would suggest you clone/synthise the intron exactly the same distance from the promoter (a weak one so that distinction between control and test is obvious) in a reporter gene (be it luciferase or GFP or anything else), within the gene (without affecting the reading frame and anyway this would be sliced out) with a scrambled sequence of the intron (other than splice sites) as your control.

Again, I am not sure how feasible or cost-effective this is. Please feel free to mention any defects in the idea.

Cheers,

Hi Christian,

Very nice discussion going on.Only for curiosity how far is your intron from the promoter? I think that having an intron as an enhancer for the same gene may pose topological constrains on elongation. I understand that gene transcription in eukaryotes doesn’t proceed in a Gaussian way as in prokaryote as it has been shown. So, it is still possible. However, how can you be sure as Vicent said it is not an alternative promoter? It is imperative that in order for you to distinguish it from promoter to establish that it lay far away from it and that some kind of looping dos exist. I know that 3C, 4C or 5C techniques have their own drawback as Daniel has pointed out, but it is always worth trying. I know some people who have been trying for a year without success which brings an important point that Daniel referred to. Negative results will not discard your hypothesis because the experimental approach doesn't have an internal positive control even if you show that it does work for the nearby gene.

I am not sure if chromatin signatures that identify enhancers are valid for all cell types but positive data may help support your hypothesis. I am also not sure if it is worth to invest in techniques such a ChIA Pet seq when you are looking for the regulation of a specific gene. But, you can lookup your gene in published dataset with this technique.

I think site specific mutation using CRASP is also a good alternative

Hi,

We had exactly this problem with the additional hiccup that the enhancer turned out to only work with its native promoter. We found that the luciferase and cell culture transfection model did not enable us to get much usable information so we made a construct from a BAC containing the whole gene of interest and engineered a GFP (or lacZ) into the coding sequence with a splice acceptor in such a way that it led to expression of the reporter instead of the native protein. This keeps the whole context of the genomic region intact and you can them use that in either for transfection in cell culture or pronuclear injection in mice to see if you get the endogenous expression pattern (we were looking at a developmentally expressed gene so this was quite quick and the preferred route for us). You can do all sorts of manipulations on your construct - deletions, insertions, site directed mutagenesis, recombination, etc.

I admit this is a very labour intensive way of going about it but this way you know you are looking at the system in something that closely approximates its "native" configuration.

For more details have a look at:

https://www.researchgate.net/publication/51735767_Elf5_regulation_in_the_trophectoderm

Good luck.

Article Elf5 regulation in the Trophectoderm

@Mitan: Interesting idea. It has only one drawback: The chromatin state of the plasmid will not be the same as of genomic DNA since the plasmid will be replicated in bacteria before transfection.

@Abdelhalim: Yes, I like this discsussion as well. We are pretty sure for a number of reasons: We have identified different alternative promoters (and already confirmed some of them biochemically and in-silico) at other places. This of course doesen't necessarily mean that there a not other sites which we haven't identified yet. But here are no transcripts available starting at this site (according to the databases) and we have some preliminary evidence that our factor in question really binds to the site(s).

@David: Thanks for the comment. I guess we will try both, since we have a BAC for amplifying from the region anyway. First a shorter construct containing the intron only, which should be easier to handle and to clone. And if this doesn't work, the whole sequence to the intron containing a part of the promoter as well. This will then be at least around 7kb, which is harder to make and also brings the usual expression plasmids to the limits in terms of maximum insert size.

Hi,

I have done lots of testing enhancers in vivo in zebrafish. Routinely we have been using a weak minimal promoter from zebrafish gata2 that had been working for enahncer trap before. In this paper: http://www.sciencedirect.com/science/article/pii/S0012160608013201

however, you can see what happens when you start using other minimal promoters. Compare expression when using the enhancer`s "own" promoter, human homolog`s promoter or neighboring gene`s promoter. We found, that promoters react to strong enhancers with the same expression pattern, with enahnced expression, but the level of expression somewhat reflects also the natural expression level of its own original gene. What was not published in the paper is usage of a strong Drosophila heatshock promoter where the expression was extreme.

We had the same experience also with intronic enahncers. Reagrding the orientation, I would keep the orientation as in the natural context as there were reports on both orientation independent and orientation sensitive enahncers. http://www.sciencedirect.com/science/article/pii/S001216060800242X

Hi Cristian,

If you have the BAC then I would suggest work on the BAC directly rather than subcloning into a plasmid. It is quite easy to insert a reported gene into your BAC via recombineering and you can use that for transfections or oocyte injection. Lower efficiency than with smaller plasmids but you will have kept your orientation, distances, etc intact. The recombineering is actually quite quick - the trargeting plasmid is the only tricky bit but you can get away with about 150-300bp flanking regions and those are quick to get via pcr.

@Pavla: Thanks for the comment.

@David: Its an idea. But I don't really like to handle BACs in the form of naked DNA, since they have a tendecy to break when not handled with a lot of care. On the other hand, amplifying 7kb of DNA is also a challenge. :-)

Hi Cristian,

have you ever considered to scale up your quest and use an approach such as STAR-seq? The idea is that an enhancer will promote it's own transcription if clone downstream a minimal promoter. Sequencing will pick up the amplified pieces of DNA and basically gave you a read-out of enhancer activity.

I you want to test an enhancer a weak promoter seems useful.