troubleshooting: gene deletions by method of Wanner and Datsenko - (Jul/07/2009 )

I'm trying to delete an operon in E. coli using the homologous recombination method of Wanner and Datsenko and encountering difficulties. I got this method to work once the first time for a single gene knockout (~1700-bp). Since then I've tried deleting two operons (4500 and 3400-bp) and have had identical problems with both. I'm looking for others who might recognize these problems!

This is what is done initially:
- PCR amplifying the linear recombination product off pKD13, no problem. My primers have 40 (or sometimes 45-bp) homologous regions. The PCR product is DpnI treated and gel extracted and purified prior to use.
- Make K-12/pKD46 electrocompetent: start a 250 mL culture from an overnight culture grown in LB + 100 ug/mL Amp at OD 0.05 and grow in LB + 1 mM arabinose + 100 ug/mL Amp at 30C to OD 0.4-0.5. After this a typical prep with 10% glycerol washes (I do this all the time for many different strains and get high transformation efficiencies)
- Transform 0.1 ug of the linear PCR product into K-12/pKD46. (Time constants are ordinarily ~5.5 ms) Resuspend in LB.
- Outgrow 1 hour at 37C
- Plate 0.5 mL onto both 25 ug/mL kanamycin and 10 ug/mL kanamycin. Grow at either 37C or 30C (tried all combinations).

What is seen the next day:
- Colonies are slow to come up, as expected. Less slow on 10 ug/mL kan. By 24 hours after plating, there are a lot of colonies. There are always less colonies on 25 ug/mL kan vs. 10 ug/mL kan.
- Pick colonies and streak them onto kan plates.
- Run colony PCR on each with appropriate primers
- *Every single time*, colony PCR indicates no integration of the kan cassette in the desired region, just a band indicating the original gene is still there.
- And additionally, the colonies struck on the new plate either do not grow at all, or come up extremely sickly and slowly. There is significant variation between each colony.

My only guess is there's some sort of misintegration occuring preferentially over having the kan cassette recombine where it should. I've repeated these attempts over and over again for these two operons and don't know where to start with troubleshooting. I thought the first problem operon was just a bad place to do recombination, maybe lots of bound proteins, and had eventually given up - but seeing this a second time with a different operon is worrying me. Any tips about what's happening and how to overcome it would be greatly appreciated!

Is the operon you are knocking out important? I would debug this with a test construct, removing a useless gene. You could be strongly selecting against what you want to have happen.

phage434 on Jul 7 2009, 08:21 PM said:

They're completely non-essential genes that others have knocked out in the past. One of them is already gone in the background strain in the entire Keio collection! I eventually found a group I could request a strain with the antibiotic casette from.

The one I'm trying to do now is involved in acetate assimilation and both genes or one or the other have been deleted before. I have the deletion (without antibiotic cassette) in another strain and it does increase lag a bit but after a few extra hours it grows at the same rate and reaches the same final OD.

So any idea what's happening if these genes are absolutely non-essential?