Genetic mechanisms of the resistance of Escherichia coli to amino acid antimetabolites. 2. Study of the frequency of induction and properties of glyphosate resistant mutants

The frequency of the induction by nitrosoguanidine of glyphosate resistant mutants was compared for recipient and donor, as well as for lysogenic and non-lysogenic E. coli cells. 11 was found that integration of viral genomes and also larger replicons such as F-factor into host chromosome increased the level of glyphosate resistance by the factor ranging from J.6 to 6. Mutants tolerating 0.2 rnM of the inhibitor were obtained one order of magnitude more frequently than mutants tolerating 1 mM of this inhibitor. One half of the mutants of every group were resistant not only to the analogue of glycine,but also to the analogue of lysine. An attempt to clone an insertion from a gene library of one of the mutants was attempted but failed. Study on the nature of this gene is in progress.

Introduction.Antimetabolites are extensively used as drugs and pesticides and are widely studied as «lead» compounds in «drug dcsign» programs [1 ] (for rev.see [2,3]).However in practice cellular drug re sistance developing in different taxonomic groups makes the use of such inhibitors less efficient.What is genetic and molecular basis of such a resistance?To investigate this the well-studied mechanisms of glyphosate inhibition may be used as a good ex perimental model.
The broad-spectrum non-selective herbicide gly phosate N-[phosphonomethyl 3-glycine specifically in hibits the 5-enolpyruvylshikimate-3-phosphate syn thase (EPSPS), which catalyses the sixth step in the shikimate biosynthetic pathway the occurence of which is restricted to bacteria, fungi and plants [4].Mutational changes rendering this enzyme insensitive to the inhibitor, as well as target site overproduction due to gene amplification (or specific promoter chan ges) have been shown to lead to the glyphosate resistance (for rev.see [5]).However these events as well as reduced inhibitor uptake, increase in its degradation are rare events [6][7][8], whereas the probability of glyphosate-resistant (gly r ) colonies ap pearance is 2 orders of magnitude higher in E. coli [6].We have recently shown that most of the gly r clones are single gene mutants which are mapped at 4 different loci of the E. coli chromosome map [9].We wondered if integration of different replicons into the host chromosome have an effect on the probability of the appropriate gene mutability.To answer this question the induction probability of gly r mutants in recipient and donor as well as lysogenic and nonlysogenic E. coli cells was measured.Also the nature of the mutants obtained was studied.
Materials and Methods-Chemicals.N-methyl-Nnitro-nitrosoguanidine (NG) and S-2-amino-ethyl-Lcysteine were supplied by «Sigma» (Germany).As a source of glyphosate the commercial herbicide Roun dup 00 with the isopropylamine salt of glyphosate as an active ingredient was used.
Bacterial Strains and Growth Media.The E. coli strains employed and their genotypes are listed in Table 1.As basal salt media, M9 and LB medium were used [10].
Mutagenesis and Genetic Methods.For muta genesis, nitrosoguanidine treatment was employed as described by Miller [10].Gly r mutants obtained in primary selection experiments were studied by a passage on increasing concentration of glyphosate alone or a mixture of glyphosate and AEC (200//g/ml).Construction of gly r mutant genomic DNA library and complementation.DNA isolated from this mutant was partially digested with Sau3A and fractionated on a 0.8 % agarose gel.Fragments smaller than 3 kb (A DNA HindiII markers) were recovered from the gel using the glassmilk procedure (BiolOl, Inc).Ana logously molecules of BamHl digested, dephosphorylated Bluescript SK + vector (from «Stratagene», USA) were recovered from the gel.The fragments from the two sources were ligated in a ratio of 1:1 (total 40 ng of DNA used) at 4 °С overnight [11 ] and were used to transform cells of competent E. coli strain DH5a (from Bethesda Res.Lab).Diluted aliquots of the transformed ceils were plated on LB agar containing ampicillin at 50 jug/ml and X-gal and IPTG thus allowing to determine the percent of plasmids having DNA inserts [11].
The remainder of the transformants were ino culated into 3 ml of LB supplemented with ampicillin and incubated overnight at 37 °С.A plasmid library was generated by harvesting the plasmids by alkaline lysis [11].The library obtained was used to trans form both DH5a and HB-101 cells made competent following the method of Hanahan 1121.The trans formants were plated on M9 medium containing different concentrations of glyphosate and all auxo trophic additions required.
Results and Discussion.The frequency of the induction of gly r mutants by NG for different E. coli cells represented by recipient -donor and lysogenicnonlysogenic cells are shown in Table 2.

Table 3 Number of mutants tolerating increasing doses of glyphosate and AEC
From the results shown in Table 2 it can be concluded that integration of different replicons into the host chromosome slightly increased the proba bility of a gly r mutation.The increase was more pronounced in the strain with both F-factor and A replicons integrated not far from each other (about 10 min on the E. coli chromosome map).As was shown earlier, the mutations obtained in this strain map at 4 different loci [9].
On the basis of their tolerance to increasing doses of the inhibitor, mutations obtained could be divided into 2 groups, i. e. those tolerating low doses and those tolerating higher doses.The frequency of the first group is one order of magnitude higher than the other.The frequency of mutants resistant to both glyphosate and AEC is evident from Table 3.
The results in Table 3 demonstrate that almost half of the mutants tolerating both low and higher doses of glyphosate also tolerated a toxic analogue lysine and thus were multiple-resistant.
Thus, gly r mutants appearing after NG treatment constitute 4 different groups tolerating different doses of glyphosate and cross-resistant to AEC.How many genes are involved in this phenotype and what is their nature?
Because mdr genes in eukaryotes are responsible for multidrugresistance based on drug efflux out of the cell (for rev.see [13]), and as such a gene should be dominant over its wild type allele the cloning of multiresistant gly r mutant DNA was tried.This mu tant DNA was partially digested with Sau3A (Fig. 1), fragments of ~2 kb recovered from the gel (Fig. 2) and ligated to the vector (Fig. 3) (see Materials and Methods).Thus, approximately 80 % of the transformants produced due to the ligation product were of white color on the indicator plates and the number of inserts obtained was about 10 000.Using this library, DH5a and HB-101 cells were transformed and after washing with M9 medium plated on selective medium containing 1 dose of glyphosate.A couple of white colonies were obtained in both cases but none of them were highly efficient in retransformation experiments and consequently the gene was not cloned.It this gene codes for a recessive trait another approach should be used [9] and this work is now in progress.
Acknowledgements.I appreciate the helpful dis cussions with professor Nikolaus Amrhein (ETH, Zurich).I thank professor Sydney Craig (UNC, Chapel Hill) for letting me use his lab to construct the gene library.The excellent technical assistance of Oxana Karpenko is also gratefully acknowledged.