I would not make a random potentially concatameric integration cassette for the following reason. You will have all sorts of recombination going on, including transgene reduction to potentially the single copy. All of this recombination is not controllable even between different cells in the same organism, so every experiment is a different one and nothing can be compared. Let's say you have a concatamer of 10 copies. One cell can have reduction to one, while another one as none, some have reduction to 4, ... What if the concatamer has 50 copies? Concatameric recombination was used to their advantage in the Brainbow strategy, but you want to overexpress a cDNA with potentially a overexpression phenotypic outcome. Hence, my opinion is that this needs to be controlled as much as possible. You can screen your transgenic founders for a single copy insertion, use a transposon or lentivrus to move your transgene into mouse, include one FRT site into the transgene to reduce you concatamer to one copy, or do integrase mediated recombination using phic31, to avoid gene targeting. In addition, you can have repeat silencing. There are strong promoters available that should express at high levels even at single copy.

-Here is the paper related to phiC31:

Site-specific integrase-mediated transgenesis in mice via pronuclear injection.

Tasic B, Hippenmeyer S, Wang C, Gamboa M, Zong H, Chen-Tsai Y, Luo L.

Proc Natl Acad Sci U S A. 2011 May 10;108(19):7902-7. doi: 10.1073/pnas.1019507108. Epub 2011 Apr 4.

-Here is a paper related to transgene reduction by Cre, that can be extrapolated to Flp (it also discusses repeat silencing):

Repeat-induced gene silencing in mammals.

Garrick D, Fiering S, Martin DI, Whitelaw E.

Nat Genet. 1998 Jan;18(1):56-9.

-Lentiviral:

Gene delivery by lentivirus vectors.

Cockrell AS, Kafri T.

Mol Biotechnol. 2007 Jul;36(3):184-204. Review.

Lentiviral transgenesis.

Nakagawa T, Hoogenraad CC.

Methods Mol Biol. 2011;693:117-42. doi: 10.1007/978-1-60761-974-1_8.

Dolrudee, an easier route might be to order a vector or ES cells with a construct already in place and generate the mice yourself. I used to work for a company that makes these constructs: https://www.komp.org/. Check them out.

What is the purpose of an excisable reporter? Unless it is beta-gal, you are using up another fluor channel when imaging later on, as well as increasing the size of the cassette that you are inserting in.

You are correct that overexpression is highly dependent on where your bac lands. Depending on how picky your tissue is, you could either have an immense overexpression of your protein, or have your insertion silenced to the point of oblivion making the mouse useless.

I am a strong advocate of targeted knock-ins, as you can choose the site of recombination, such as the ROSA locus and save yourself this headache of potential silencing and multi-copy insertions, but I am aware of the increased difficulty of generating such a construct.

However, if you have not considered a targeted knockin approach, I would highly recommend you take a look at existing ROSA26 targeting vectors available on addgene, and see if they can be easily adapted for your purpose. The benefits are two fold: 1) it is easier than conventional targeting if you use a preexisting vector, and if it does homologously recombine, then you have a defined knock-in to a locus which is well expressed, not silenced, and works well in most if not all tissues. 2) if after screening for clones, you do not find *ANY* true homologous recombinants, you have all of the other clones which are basically bac insertions that you could use.

I am sorry to not have any suggestions on the question you actually asked, as we have very little experience with bac transgenics. They rarely work in my system (unfortunately) and so we have completely switched to making targeted knock-ins.

Hopefully this helps, and I wish you the best of luck making your mice! I know well the pain involved in this process unfortunately :(

I have used the system with three loxP sites in ES cells and I needed to excise only NeoR cassette without excision of the floxed allele. In the first experiment everything bordered by loxP sites was cut out (I got a null allele, good for constitutive knock-out, but not our target as we wanted to get conditionaln knockout with a floxed allele). The problem of over-excision by Cre recombinase was solved by using a plasmid with Cre recombinase under a not so strong promoter as in the first experiment (the first promoter for Cre recombinase we used was a phosphoglycerate kinase 1 promoter, the second one was modified cytomegalovirus promoter) so its expression was much lower. Then, the Cre recombinase present in small amounts was able to excise only the shortest fragment between loxP sites (NeoR cassette as a fragment with allele was much longer). Therefore, regulating the concentration of Cre recombinase in cells You can tip the excision in the direction You prefer. If there is only a small amount of recombinase it will excise the fragment with the optimal size for excision with the highest efficiency, with high concentration of Cre recombinase it will cut out everything it can and with very very high concentration of Cre recombinase it can be even genotoxic as it will also recognize pseudo-Cre sites in genome.

Thank you everyone for your answers.

Brian, the purpose of the first reporter is to allow you to screen for transgene expression or activity prior to Cre excision. Since not all transgenic mice derived from pronuclear injection will express the transgene, this reporter would serve as a quick way to screen which founder to sac and which one to keep and further breed with a Cre line.

After Cre is introduced, the first reporter along with the 3X-STOP will be eliminated allowing the cDNA and the second reporter to be expressed. Therefore, there shouldn't be any problem using up both fluor channels later on. It will be only one or the other if Cre works efficiently.

I have looked at the targeted knock-in approach before, but I believe (please correct me if I'm wrong) that you can only gain, at best, moderate level of expression. Plus it takes more time and resources to make the mice.

So we chose the random integration approach for a higher possibility to achieve a vey high level of expression. Although it's not predictable what you will get, and of course, you need to screen many founders. But, if you get a few lines that overexpress your gene very well, you have a good model to work on.

I'm still looking for an answer to my original question. Please post it if you have any experience regarding the nature of Cre excision in transgenic mice (random integration).

Thank you!

It might be good to make a knock-in, rather than transgenic model in your case. Multi-site interation is inevitable for transgenic mice.

Hi Dolrudee, perhaps in addition to technical commments you could read Csikos et al. J. Immunol. Methods, 2010; 339:259 and M. Schmidt-Supprian and K. Rajewsky, Nat. Immunol. 2007; 8:665.

Good luck!

Hi Dolrudee

I would echo the insightful comments above about alternative vector design and/or exploiting existing Rosa 26 vectors and especially designing a knock-in rather than a random integration strategy. Define exactly what you want to do and why you need to do it that way and see if there are tools already out there.

If you do decide to go your own route, you should be aware of the incidence of cryptic loxP sites in the mouse genome. It is likely only to be an issue if you integrate close to one of these Cryptic loxP sites, thus creating a pair with one of the sites in your BAC, but it can also be a problem for BAC engineering when you introduce multiple loxP sites.

You might be interested to look at the following, which contains background information and a bioinformatic tool, FuzznucComparator, which can be used to search for cryptic loxP sites:

Semprini S, Troup TJ, Kotelevtseva N, King K, Davis JR, Mullins LJ, Chapman KE, Dunbar DR, Mullins JJ.

Cryptic loxP sites in mammalian genomes: genome-wide distribution and relevance for the efficiency of BAC/PAC recombineering techniques.

Nucleic Acids Res. 2007;35(5):1402-10

Hope that helps!

Steve Morley

I would not make a random potentially concatameric integration cassette for the following reason. You will have all sorts of recombination going on, including transgene reduction to potentially the single copy. All of this recombination is not controllable even between different cells in the same organism, so every experiment is a different one and nothing can be compared. Let's say you have a concatamer of 10 copies. One cell can have reduction to one, while another one as none, some have reduction to 4, ... What if the concatamer has 50 copies? Concatameric recombination was used to their advantage in the Brainbow strategy, but you want to overexpress a cDNA with potentially a overexpression phenotypic outcome. Hence, my opinion is that this needs to be controlled as much as possible. You can screen your transgenic founders for a single copy insertion, use a transposon or lentivrus to move your transgene into mouse, include one FRT site into the transgene to reduce you concatamer to one copy, or do integrase mediated recombination using phic31, to avoid gene targeting. In addition, you can have repeat silencing. There are strong promoters available that should express at high levels even at single copy.

-Here is the paper related to phiC31:

Site-specific integrase-mediated transgenesis in mice via pronuclear injection.

Tasic B, Hippenmeyer S, Wang C, Gamboa M, Zong H, Chen-Tsai Y, Luo L.

Proc Natl Acad Sci U S A. 2011 May 10;108(19):7902-7. doi: 10.1073/pnas.1019507108. Epub 2011 Apr 4.

-Here is a paper related to transgene reduction by Cre, that can be extrapolated to Flp (it also discusses repeat silencing):

Repeat-induced gene silencing in mammals.

Garrick D, Fiering S, Martin DI, Whitelaw E.

Nat Genet. 1998 Jan;18(1):56-9.

-Lentiviral:

Gene delivery by lentivirus vectors.

Cockrell AS, Kafri T.

Mol Biotechnol. 2007 Jul;36(3):184-204. Review.

Lentiviral transgenesis.

Nakagawa T, Hoogenraad CC.

Methods Mol Biol. 2011;693:117-42. doi: 10.1007/978-1-60761-974-1_8.

@Dolrudee Interesting, your reasoning for using the excisable reporter does offer advantages during founder screening in bac transgenics.

Other than using a promoter of lesser strength as Patrycja Koszałka suggested, I would suggest perhaps a CreER system, so you give a single dose of tamoxifen to induce the excision, but not under constitutive expression. It would be much less efficiency than in the previous case, but perhaps the lower amounts of activity/concentration of Cre would ameliorate the problem.

However, regardless of which Cre driver you end up using, you will need to breed it out once you verify excision, and with the large number of integrations, I would forsee that most of the progeny would be mosaic in cuts, so perhaps multiple rounds of breeding would be necessary?

While a knock-in strategy does take more time, I would argue that its stability and specificity are much more important. With respect to stability, because a targeted knock-in is present as a single copy in a define locus, such as rosa26, you avoid several issues.

1)the current concern that you have of concatenating and joining genomic dna

2)you are no longer working with multi-copy insertions and thus will not have variations of expression in each downstream mouse since each will have a varying number of insertions. (An issue we have had happen to us unfortunately)

3) avoid issues of extreme overexpression and/or silencing due to the epigenetic landscape of where your cassette landed

4) finally, you are much more certain that your targeting has not interfered with any gene or locus structure that may be causing a phenotype in whatever model you are studying.

In my opinion, a knock-in would most likely not keep you up at night with questions such as these after you have already gone through the trouble of generating, breeding a transgenic, and using it in your studies. *However* as someone who has spent a few years making knock-ins, I understand an aversion to doing it, since it's a pain to make them regardless of whose protocol or whose lab you are in. If I have not swayed you into making a knock-in, I think that the excisions across larger genomic areas would be highly inefficient, and using an inducible creER or a lower power driver should do the trick.

Good luck in your mouse making endeavours!

Thank you very much everyone for your very comments!

Hi Koen,

Thank you very much for your insightful comment, suggestion and references. That's exactly the answer I'm looking for. So, different kinds of Cre recombination will happen in random integration transgenics, and it cannot be predicted or controlled.

If this is the case then I prefer not to go through that route and then end up not able to interpret my data or to defend the data that derive from experiments that cannot be controlled.

I'm checking the PhiC31 integrase knock-in approach (TARGATT mice) that you suggested and so far it looks very promising in terms of efficiency and time it takes to make the mice. Have you used this approach before? If so, can you comment on its efficiency? They claimed that it's a lot more efficient than ROSA26 knockin approach, but since it's a relatively new technology, I just want to explore its pros and cons before committing in doing it.

Thank you again for your answer and the references. They are incredibly useful!

Hi Brian,

Thank you very much for your suggestion! I think you have already swayed me into making a targeted knock-in instead of a random integration transgenics. And yes, this question about uncontrollable Cre excision has been bothering me and keeping me up at night. That's how I ended up posting this question in ResearchGate community!

So, I'm looking into making a knock-in using TARGATT mice (PhiC31 integrase) as Koen suggested above. Have you used or do you know anyone who has used this approach before? I'm more inclined to use this method rather than the ROSA26 knock-in because of higher efficiency and shorter time to generate the mice.

I see that you have extensive experience in making knock-in constructs for ROSA26 locus. How long did it take you (on average) from the beginning to when you get the mouse you want?

Thank you again for your comments and suggestions!

@Koen: Thanks very much for bringing this technique up, I had only briefly heard about it and had never really looked into it. It definitely looks like a promising technique. I wish I had stumbled across it earlier!

@Dolrudee: I am pretty optimistic of the phic31 technique, especially for the efficiency of the approach. I work with fairly large inserted cassettes and so efficiency is a huge roadblock. Specifically, check this paper out:

--

Chen, C., Krohn, J., Bhattacharya, S. & Davies, B. A comparison of exogenous promoter activity at the ROSA26 locus using a ΦiC31 integrase mediated cassette exchange approach in mouse ES cells. PloS one 6, e23376 (2011).

--

They show an extremely impressive efficiency of targeting into the rosa locus using the integrase approach. I haven't had time to really read this or the previously cited papers in-depth, but if you are going to do the targeting and selection yourself, I would agree with you that it's probably a good idea to try.

Unfortunately, our lab uses perhaps 10-15 strains of rosa26 knock-ins, drivers and reporters--I have not directly tried to target that locus. We are interested in adult neurogenesis and I have been targeting the hes1 and ascl1 loci. The time it has taken us to make mice is a rather confounding answer, as I am applying several new techniques in both the vector construction and the actual cassette design. To actually answer your question though, the last one I did took one week to generate the targeting vector. We use a core facility to do about 3 weeks of ES cell work, 2 weeks of screening, and then perhaps half a year of breeding to get homozygote offspring. The last 6 months are the worst. It helps to have a bunch of projects that don't require mice so you aren't just twiddling your thumbs all the time.

The majority of our labs work is with in-vivo genetic manipulations, so while making a knock-in mouse is very time consuming, the results are extremely satisfying...or so I've heard.... :-P

Good luck with your project!

Sorry Brian and Dolrudee, I am a fruit fly guy, so I can not give efficiency and alike for mouse experiments. However, I am interested in technology and methodology in general, and developing new ways for gene manipulation and alike in fruit flies. In context of that, I have used the phiC31 system in flies a lot. It is very efficient and the way to go. I am very familiar with it.

We made a complete overlapping duplication kit for the entire X chromosome in one specified site on chromosome 3. This project was totaling about 500 large transgenes and subsequent transgenics. This was a pretty crazy project.

-A molecularly defined duplication set for the X chromosome of Drosophila melanogaster:

Venken KJ, Popodi E, Holtzman SL, Schulze KL, Park S, Carlson JW, Hoskins RA, Bellen HJ, Kaufman TC. Genetics. 2010 Dec;186(4):1111-25. doi: 10.1534/genetics.110.121285. Epub 2010 Sep 27.

For this we made genome-wide phiC31 libraries first:

-Versatile P[acman] BAC libraries for transgenesis studies in Drosophila melanogaster.

Venken KJ, Carlson JW, Schulze KL, Pan H, He Y, Spokony R, Wan KH, Koriabine M, de Jong PJ, White KP, Bellen HJ, Hoskins RA.

Nat Methods. 2009 Jun;6(6):431-4. doi: 10.1038/nmeth.1331.

The biggest published thing we integrated was 146 kb:

-P[acman]: a BAC transgenic platform for targeted insertion of large DNA fragments in D. melanogaster.

Venken KJ, He Y, Hoskins RA, Bellen HJ.

Science. 2006 Dec 15;314(5806):1747-51. Epub 2006 Nov 30.

Now we have integrated something larger than 200 kb. We are writing that up as we speak.

Most recently we also made a phiC31 based transposon for RMCE:

MiMIC: a highly versatile transposon insertion resource for engineering Drosophila melanogaster genes.

Venken KJ, Schulze KL, Haelterman NA, Pan H, He Y, Evans-Holm M, Carlson JW, Levis RW, Spradling AC, Hoskins RA, Bellen HJ.

Nat Methods. 2011 Sep;8(9):737-43.

PhiC31 has revolutionized the fruit fly community. Other people made RNAi libraries, promoter bashing GAL4 libraries, ... I bet down the road it will revolutionize the mouse field as well.

The TARGATT technique was developed by Liqun Luo and colleagues. Liqun is a fruit fly/mouse person and developed also the MARCM/MADM technologies for mitotic labeling of mutant cells. I do not think it was used yet to integrate BAC sized material of >100 kb in mice. So far regular plasmids only I believe, unless I missed it. Actually it is good we have this discussion. I will scout for it later on Pubmed.

Thank you very much, Koen, for your comment! Fruit fly people rock! We, the mouse people, always follow you.

The more I read about this PhiC31 integrase approach, the more inclined I am to try it!

I like the fact that it takes much shorter time than the traditional ROSA26 knock-in, and the "up to 40% efficiency" sounds really attractive. Plus, you don't have to deal with screening ES cells and breeding for germline transmission, only pronuclear injection into TARGATT embryos will do. Time is so precious as we all know. I just hope that the transgene expression will be robust. So, I guess I just have to try it.

My construct is actually not a BAC transgene (my apology for not making myself clear in the earlier posts). So, the cloning part into TARGATT vector should be straightforward and not time-consuming. Charles River Lab now carries the TARGATT mice, but only in FVB strain though. I plan to make my transgenic in C57BL/6 strain and they said that the C57BL/6 strain will be available soon.

Thank you very much for the references. You're so productive! Please keep the discussion alive if you discover anything new about this approach. This discussion is very interesting and super useful!

Cheers!

Hi Brian,

Thanks so much for the comment and the reference you pointed me to. As Koen mentioned above, it seems like this PhiC31 approach might be adopted widely in the future based on its advantages over the random integration and the traditional ROSA knock-in.

I'm wondering if anyone has made transgenic mice using this method yet. I'm curious to know the expression level of the transgene.

Anyways, I'm exploring the details of this approach and probably will try it.

Fingers crossed!

Indeed, I am always jealous of the fly and worm systems. I briefly dabbled in both during lab rotations years ago and within a few months of joining my current lab I already missed how fast and sophisticated genetic modifications were. Ah well!

I agree the phic31 approach has a lot of advantages. I need to take some time and look into it and see if people have found a work-around to knock-in to other loci.

This is kind of a side note, but man do I wish CRISPR had a better off-target rate, otherwise I would be swimming in transgenic mice already :p Koen, I know of people working on crispr in flies, but do you know if they are having similar issues to the mammalian use where there are tons of off-target effects?

Thanks both to you Koen for bringing this tech to our discussion and Dolrudee for asking the question in the first place!

Good luck in both of your studies!

Hi, I think much of the advice above is sound. Overall the CAG-Cre is a fairly well studied one. I would add that it is important to be careful in the Cre that you choose, if you decide to go that route. For instance EIIa-Cre is known for uneven excision efficiency across different organ systems. Jackson Labs, www.jax.org, has a variety of additional information and research resources on various Cre systems. I would encourage you to check out their website.

Another approach depending on your target gene expression level is exchanging the 3'-UTR for a different one as shown in: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0012909;jsessionid=ECB1E3197B1AABDD9333B0AEE7C2E4D8 . Gene expression can be both increased or decreased depending on how stable the mRNA is in the gene where you get the 3'UTR. Good luck!

@Brian.

I am following the Crispr pubs with much caution. At this point, the risks out way the benefits to me.

Good luck with your studies too.