Moving E. coli genes into Salmonella, by David Thaler

Salmonella and E. coli have almost the same gene order, and most of the genes in one have homologs in the other. However, the sequences of the two diverge enough to prevent homologous recombination in most cases.
Mismatch repair mutants lower the sequence identity required for homologous recombination (Rayssiguire,C,. D.S. Thaler, M. Radmann 1989 Nature 342:396-401; Shen, P. and H. Huang 1989 MGG 218:358-360). If the recipient is mutant for genes involved in mismatch repair (which involves recognizing non-Watson-Crick base pairs) then it can recombine more easily with sequences that have a difference every 5 or 10 bases (80-90% similar). One or two differences in 10 bases is enough to knock out most recombination in a wild type host. If two sequences diverge by more than about 20% the mismatch trick probably won't work; it functions only over a small window, but fortunately, that window is about how different E. coli and Salmonella are.

Theoretical problems, what to watch for
Sometimes this trick will work, sometimes it won't. Variables include the degree of homology for your gene and the possibility that "wrong" (i.e., non-homologous) regions of the chromosome may match up when sequence specificity is lowered.
The absolute rate is still 100-10,000 X lower than homolgous Salmonella x Salmonella recombination with P22 transduction. Conjugation gave a higher absolute rate (approaching homologous) but many of the product were unstable (see Table 3 in Rayssiguire et al. and the associated discussion). Transductive recombination appeared to be 99% simple allelic replacement (Table 4, Rayssiguire et al.). I rationalize this by thinking that the size of input DNA is less for transduction than for conjuagation so there is less opportunity for poor matches. In a smaller region the chances are that close sequence identity will occur only in the homologous position. The practical effect is that you'll need to use lots of phage to get any tranductants.
Mut alleles confer a high point mutation rate on the recipient (reviewed by E.G. Radman or Meselson). They probably confer a high rate of duplication formation as well (Thaler, unpublished).
These Mut alleles are also aassociated with a high rate of precise excision of Tn10 and Tn5 (Lundblad and Kleckner, 1985 Genetics 109:3-19). What this means is that the mutator alleles themselves, which are due to transposon insertions, will revert at a high frequency. It's a good idea to check the very culture that you use for any critical experiment to make sure that it has the mutator phenotype. An easy way to do this is to spread a lawn of the culture on NB plates (use top agar to get a nice even spread) and put a few crystals (a pinch) of rifampacin in the center of the plate. After incubation for two days what one sees is a very clear zone of inhibition about 3 cm in diameter, with a clear border between the killing zone and beyond where there is normal thick growth of the lawn. Inside the zone you'll see a few drug resistant papillae. For wild type strains(i. e., non-mutator) there will be zero to about 5 drug resistant papillae inside the zone. For the mutator alleles used here (mutS and mutL) there will typically be 50 or 100 papillae inside the zone inhibition.
Point mutations are also high in these backgounds (that's how the genes were originally identified). So use caution and get your mutations into a clean (non-mutator) background ASAP.

Strategy and Tactics:
The donor:
The object is to obtain a P22 tranducing lysate containing the E. coli gene. One source could be E. coli F's residing in Salmonella. If this is available it is probably the best source to use. A review of E. coli F's is : Low, K.B. 1973 Bact Rev 36:587-607. We have many of these already conjuagated into Salmonella in the lab collection. Tom Doak has a list of these, which he should probably put into the lab manual sometime.
The other way to get a donor is to induce P22 out of E. coli. Tom Elliot (was from this lab, now from the University of Alabama) has made a strain that is useful for this. This is TE1335 F' lac pro P22 dilysogen (one P22 is wild type, the other is the hft int3 version that we use for high frequency transduction) in an E. coli deleted fro lac-pro on the chromosome (what is the TT# and complete genotype of this strain?). F's will segregate, so check that your isolate is lac+. Construct the donor by moving your favorite mutation into this background, then induce the P22 prophage in the E. coli.
(A mitomycin C induction protocol:
Grow cells to mid log (dilute a fresh overnight 20 fold and grow 2 hours with good aeration). Add miomycin C to 2 mg/ml final. Shake overnight or until visible lysis. CHCl3 and give a slow (6K/10') spin to clear cell junk. Typical titers are about 2 x 1010 pfu. It's worth titering cause this method is more variable than simply growing P22 in Salmonella.)

Onto the Salmonella recipient:
Phage from E coli will be subject to restriction in wild type Salmonella, whereas the phage from an E. coli F' resident in Salmonella will not be subject to restriction. This could affect your choice of Salmonella recipient.
The best overall recipient woud be r-m+, mutL or mutS. Such as TT??
Incidentally, most r-m+ strains we have are also galE. This means that to absorb P22 the cells have to be grown in glucose plus galactose. Grow them in 0.2% glucose and 0.02% galactose. Feel free to play around with these conditons in order to get good absorption with P22 (whatever conditions allow for good plaque formation). A reference for the base strains is: Jun-ichi Ryu, Dept of Microbiology, Loma Linda University, Loma Linda CA 92350 Phone (714) 824-4480. Its complete genotype is:

JR501=r-SAkm+SA r-SBm+SB r-Tm+LT
hsdSA29 hsdB121 hsdL6 metA22 metE551 trpC2 ilv452
rspL120 xyl-404 galE719 H1-b H2-e,n,x, nml-(Fels2-)fla-66
(This is TT??)
I have made deriviatives of this strain that are mutS::Tn10d-cam and mutL::Tn10d-cam. The Tn10d-cam insertions were made by Chris Conner and will be given allele numbers by him. Please ask Chris for this stuff and then put my strains in the lab collections. (TT??)

Plaease keep me up on how this protocol works, any problems and successes. Best of luck! David Thaler/Apt 25F Faculty House/Rockefeller University/500 East 63rd St/ NY‚, NY 10021-6399. Phone: (212) 570-7809 Fax: (202)570-8651

Last Update: Thursday June 19 2014
This page has been viewed Failed to execute CGI : Win32 Error Code = 2
Eric Kofoid eckofoid at