I usually use the strain (containing the pKD3 vector) as the template to avoid the false positives. There are still some false positives, I think the fragment might integrated at other site of the chromosome.

Are you getting a single band with the primers flanking the gene or are you getting two bands from the same colony, one that would correspond to the length of the wild-type gene and one that would correspond to the length of your mutant with the kan cassette replacing the gene? If that's the case then the region probably just got duplicated and you just need to restreak your mutant a couple times on kanamycin until recA fixes it.

Xue Liu's answer is definitely a possibility, BLASTing pKD4 against the BL21(DE3) genome (genbank# CP001509.3) produces some multiple regions of homology, which all seem to map to the rrnB terminators on pKD4 (which makes sense considering those terminators came from E. coli in the first place.) If that's the problem you could also try gel extracting the PCR fragment which would eliminate all (or at least most) template plasmid DNA.

It's also possible that you're getting spontaneous kanamycin resistant mutants which don't actually contain the kanR cassette. The way to verify that would be to order primers that specifically amplify a fragment of the kanR cassette and see if you get a band. I've had to do this a lot with MDR strains of bugs that tend to become spontaneously resistant to everything to verify that they do actually contain the intended construct.

https://academic.oup.com/femsle/article/325/2/140/566480

Alexandra Johnson , Xue Liu Xue Liu, and Hiba Riyadh Al-abodi thank you very much for your answers. I'm still struggling with the obtention of the mutant. I recently obtained a new pair of primers to extend the homology region up to 80bp for each terminal end. I also obtained 2 different gel-excised, purified cassettes (one with pKD3 and the other with pKD4) using these new primers. Then I electroporated the BL21(DE3)-pKD46 cells (formerly induced with 10mM and 100mM of L-arabinose) with up to 3 ug of each cassette. Then, cells were incubated 3h at 37ºC in 1 mL of LB. Half the volume was scattered on each corresponding plate LB+Cm (25ug/ul) and LB+Kan (100ug/ul). The remainig volume was kept at room temperature and finally scattered on each plate the next day.

I did not obtain a single colony with the pKD3 linear cassette. On behalf of the transformation with the pkD4 product, I obtained several tiny colonies, but all of them contained the WT gene. Also, when re-streaking them on a new LB+Kan plate, they don't grow. I am using 2 pairs of primers for the colony screening, 1 of them binds to the internal part of the gene that I expect to delete (should render no product in the mutant) and the other one flanking the cassette that I have introduced.

The gene that I want to delete is the aceA from BL21(DE3):

5'ATGAAAACCCGTACACAACAAATTGAAGAATTACAGAAAGAGTGGACTCAACCGCGTTGGGAAGGCATTACTCGCCCATACAGTGCGGAAGATGTGGTGAAATTACGCGGTTCAGTCAATCCTGAATGCACGCTGGCGCAACTGGGCGCAGCGAAAATGTGGCGTCTGCTGCACGGTGAGTCGAAAAAAGGCTACATCAACAGCCTCGGCGCACTGACTGGCGGTCAGGCGCTGCAACAGGCGAAAGCGGGTATTGAAGCAGTCTATCTGTCGGGATGGCAGGTAGCGGCGGACGCTAACCTGGCGGCCAGCATGTATCCGGATC........AAGCACTACGTTGAGAAAGTGCAGCAGCCGGAATTTGCCGCCGCGAAAGATGGCTATACCTTCGTATCTCACCAGCAGGAAGTGGGTACAGGTTACTTCGATAAAGTGACGACTATTATTCAGGGCGGCACGTCGTCGGTTACAGCGCTGACCGGCTCCACCGAAGAATCGCAGTTCTAAGCAACAACAACCGTTGCTGACT3'

Primers in bold:

Forward:

5'GAAAACCCGTACACAACAAATTGAAGAATTACAGAAAGAGTGGACTCAACCGCGTTGGGAAGGCATTACTCGCC 3'

Reverse:

5'CGTCACTTTATCGAAGTAACCTGTACCCACTTCCTGCTGGTGAGATACGAAGGTATAGCC 3'

If you have any idea to improve this process i would really appreciate it! Thanks!

Your kanamycin concentration seems a bit high for a single insertion in E. coli (think my old lab used 25 µg/mL) but I don't really see anything obviously wrong. Can you give the sequences of all the primers you're using in this process?

Also at this point you might want to try to knockout another gene in the same strain just to make sure everything is working. I believe BL21(DE3) is lacZ positive so wild type should turn blue in the presence of X-Gal + IPTG and a lacZ mutant would be white, just as a suggestion for an easy phenotype to test.

Xavier Carol Morera , definitely I can understand your feeling. Around 2 months ago, I started to knockout two genes in BL21(DE3) using pKD46 systems, but almost all failed. The only success after one month hard working was achieved by using kanamycin resistant cassette (from pKD4) with 500 bp homologous arms (>200 ng, 2.5 kv, bio-rad cuvette with 0.2 cm gap). Only got one colony for each gene and luckily they're correct.

Then I think something might be wrong. I checked internet and found people saying BL21(DE3) is always troublesome in terms of recombineering using pKD46. pSIM series plasmids from Dr. Donald L. Court (https://redrecombineering.ncifcrf.gov/strains--plasmids.html) is a better choice. I requested and used pSIM6. It works amazingly in BL21(DE3). Just follow their published protocol (Thomason, L. C., et al. (2014). "Recombineering: genetic engineering in bacteria using homologous recombination." Curr Protoc Mol Biol 106: 1 16 11-39.) and I could obtain mutants consistently (I could even get an "essential gene" knockout mutant). I strongly recommended you give it a try if you're still bothered by your mutation work.

If you need to learn some principle, here is one (Datta, S., et al. (2006). "A set of recombineering plasmids for gram-negative bacteria." Gene 379: 109-115.).

Hi Xavier Carol Morera

Im trying to make gene knockouts in Shigella using the same lambda red recombination system and Im facing the same issues with etting false positives on kanamycin plates. Did you end up getting any solutions for this?? did pkd3 worked better ?

Hi Tanuka Sen ! I finally obtained the knock-out strain. I have a detailed protocol that might fit with your needs. Among different approaches tested (using pKD3 instead of pKD4, testing different arabinose concentrations, increasing homology of the cassette towards the target gene,....) the turning point for the successful integration was to increase the pureness of the DNA sample. That is to say, to gel excise the PCR product obtained using pKD3/4 as template in order to rule out: DpnI-escaping plasmid template, and (the most important part) primer-dimer side-products, which somehow triggered the integration of the linear DNA. By doing so I successfully obtained the k.o. strain. Below I am attaching the detailed protocol. If you need more help do not hesitate to contact me!

Hi, transformation of linear DNA obtained from pKD3

I have been trying to delete the gene sifA from Salmonella Typhimurium and keep failing. I have successfully transformed the pkd46 plasmid to my target strain but I keep failing to transform the linear DNA which I have amplified from pKD3. when I do gel purification for the PCR product obtained from pKD3 I'm getting very low levels of DNA concentrations (less than 20ng). I have induced my target strain containing pKD46 with 10mM arabinose. For transformation, I have applied an electric field from 2.0kV-2.5kV. Can you please help me to solve this problem. Thank you.

Hi Chamith Hewawaduge, I had the same limitation when trying to obtain the pure DNA sample (free of the template). I would recommend preparing several preparative PCRs and gel-excise the PCR product. Following this, pass all gel-excised bands through the same purification column and elute using pre-warmed DNAasa-free milli-Q water with the minimum volume possible. By doing so, I reached concentrations above (400ng/ul). Remember that DNA recovery yield is generally reduced when the elution volume is little. If you have a SpeedVac, I would suggest passing each gel-excised band through a different column, and then elute each column with an appropriate volume. Following this, I would join together all samples and concentrate the DNA using SpeedVac.

;)