Hi Giulia, I'm not sure I understand your problem. To design an MLPA assay you need to know the sequence, since the two probes will have to lie adjacent to be ligated. So MLPA as far as I know is used to quantify, not to discover.

Secondly, for sure you need to use DNA on MLPA assays. Still, it can be genomic DNA (and you can quantify SNPs or ploidy) or cDNA (to quantify alternatively spliced isoforms). When you talk of deletion and break point I think at genomic DNA, but if you want to run a 3' RACE I think at cDNA.

If your problem is to demonstrate a deletion/alternative splicing in an mRNA, I'm not sure you need a RACE (rapid amplification of cDNA ends - this you use to find out where your mRNA ends at the 5' side or the far 3' side, in the UTRs, and you produce the cDNA with a specific primer in the 5' UTR in the first case or with modified oligo-dT priming in the second): since you already know the deletion (otherwise how did you design the MLPA probe?), just use primers on the two sides of it. Wt sequence will be the least problem, the deletion will provide a much shorter fragment that will prevail on the long wt. You can select against the wt by using very short elongation time, in case. Then sequence your product to check the exon junction. If you are on genomic DNA, the same is true concerning the PCR approach, and the sequence will show you the breakpoints.

Or am I missing something?

Hi Giulia, It's years since I've done this but I'm not sure RACE is the way to go. People generally use it on cDNA to discover new splice variants and fusion transcripts. From my own experience, the end of the 'known' sequence needs to be pretty close to the unknown sequence you're trying to identify. As you're doing MLPA, I guess your probes are in the exons of your gene, so the true breakpoint may be several kb away from the start of the first lost MLPA probe. This makes some type of RACE variant on genomic DNA technically tricky.

There are various alternative complicated PCR schemes to amplify from known sequence to unknown sequence but when I've tried to define genomic breakpoints in the past, the most successful method was designing several (5-10 at 5kb intervals) forward primers to the left of where you think the deletion started and several reverse primers to the right of where the deletion ended. You then do a long range PCR with all of the possible primer combinations. This approach uses a lot of genomic DNA and wouldn't be good if your DNA is fragmented eg. from FFPE. The plus side is that it's a quick and easy setup. The down side is that you also need a fairly good idea of where the breakpoints are to do this such as SNP array data.

I guess that, as your deletion extends past your final MLPA probe, you don't know where the deletion (or translocation or complex rearrangement or whatever it is) ends. This may not be a disaster if it's a know recurrent deletion from some type of constitutional syndrome. Some questions that may help us answer this better:

1) How precisely do you know the breakpoints.

2) Related - is there any other data about this deletion eg. array CGH

3) Is this a constitutional, cancer or other type of specimen?

4) Do you have cDNA from this patient (in which case, 3' RACE may be a more viable option).

Depending on the details of your project, you could look into Capture PCR orInverse PCR (and variations thereof (T. Kohda, DNA Research,2000, A simple and efficient method to determine the terminal sequences of restriction fragments containing known sequences.))

Dear Pietro Pilo Boyl, thank you for your answer.

I apologyze for any incomprensions: I have already carried out MLPA and I found the results, yet! Now I have to confirm the results by PCR.

I know where the deletion is located, but I don't know how wide is.

Dear Scott, thank you for your answer.

I give you some extra information: the affected gene causes a sindrome in an autosomal recessive manner. This gene has 11 exons and the deletion is from exon 8 to exon 11. My MLPA probes are both exonic and intronic and I have only DNA sample (no cells and no blood).

I am happy that you totally agree with my idea: many attempts by PCR drawing new primers is the best solution.

I have tried to amplify using the primer 11r but only one allele appears, is it sufficient that I design a reverse primer downstream of 11r?

you can design some more MLPA probe downstrean the gene (1 Kb, 5 kb or so)  to have an idea where the deletion ends. then is get much easiret know where to design your reverse primer.

Michela makes a great point. More MLPA after the gene might define the deletion and allow you to design reverse primers for validation.

Regarding the PCR you already tried: If you know that exon 11 is deleted, there's no point putting F primer in exon 10 (for example) and R in exon 11 - you'd only amplify the wild type allele.

One thing you could try is to look for similar deletions in Decipher and  ISCA (now called the ICCG)

https://decipher.sanger.ac.uk/

http://www.iccg.org/iccg-references-and-policies/iccg-signup-and-login/

Both of these have tracks on the UCSC genome browser too. Like I said before, some of these constitutional syndromes have recurrent deletion regions and that might help you guess where your patient's deletion ends.

Dear Michela Barbaro, thank you so much for your answer. It is a great idea!

Now I am testing this MLPA kit for this gene, but I will consider your solution actually.

Dear Scott, thank you so much for your answer. I have just checked when I started to work on this patient...unfortunately the deletion is unknown! I reckon that I will try by LONG-PCR (I found an antique reverse primer located downstream 11R) and then, if nothing change, I will design another MLPA probe.

The difficulty is the presence of a very homologous pseudogene which complicates the situation: it's hard to find gene specific regions (necessary to carry out a suitable PCR and sequence) which are not deleted in this patient.

happy to help you! do not esitate to contact me if you need help with MLPA. i've been designing a lot a kit myself.

Buona fortuna,

Michela

Hi Giulia,

I think that "LA PCR in vitro Cloning Kit" is a possible solution. That is similar to inverse PCR, but would be better.

Dear Akihiro, thank you for your answer.

I have already tried to carry out a cloning by a kit in order to separate alleles to study separately.

Unfortunately, it was a fail because of the low efficiency of transformation.

This is indeed challencing, 3' RACE works only on mRNA with a polyA tail.

Deletions are often mediated by ALU repeats, so you could look for the precense of such ALU sequences using this website: http://transpogene.tau.ac.il/

You can check for ALU sequences outsite your gene using UCSC genome browser.

Then you should design several primers around these ALU sequences and hope you obtain a long range PCR product.

You could also develop new MLPA probes at different points outside your gene to check for the absence/precense of the sequence.