Synthetic Lethal Mutations

Two mutations are synthetically lethal if cells with either of the single mutations are viable but cells with both mutations are inviable. As with suppressor analysis, synthetic lethal mutations often indicate that the two mutations affect a single function or pathway. Synthetic lethal phenotypes are diagnostic of an interaction between the products of the two mutant genes in the cell. Combinations of synthetic lethals can also give information about what products are needed to complete a cellular process.

The following cartoon shows the concept of synthetic lethality [modified from Hartman et al. 2001].

Because synthetic lethal mutations are inviable, it is clearly not possible to directly isolate synthetic lethal mutations -- the desired mutants would be dead. This is where the "synthetic" part of the term synthetic lethal mutation comes from. To isolate synthetic lethal mutations requires a way to artificially construct (or synthesize) a strain with two mutations that have been characterized independently. Using conditional mutants in one of the genes makes it possible to construct a strain with the two alleles under permissive conditions, and then test the phenotype under the nonpermissive condition. Temperature sensitive or cold sensitive mutations are used most often. An alternative method is to build a conditional mutant by placing one the genes under control of an regulated promoter, such that the amount of the protein is controlled by addition or removal of an inducer.


Use of conditional mutations to construct synthetic lethal strains. An elegant example of how synthetic lethal mutations can be used to identify protein-protein interactions was described by Flower et al.. Secretion of proteins across the cytoplasmic membrane requires multiple membrane proteins. A variety of circumstantial evidence suggested that the secY (also called prlA) and secE (also called prlG) gene products must interact for proper protein secretion. Despite these inferences, there was no direct evidence for the interaction between prlA and prlG.

To determine how the mutant proteins interacted, various combinations of 23 different prlA alleles and 5 different prlG alleles (115 total combinations) were constructed. Complementation was tested by the ability (+) or inability (-) to export the LamB protein to the periplasm. Only 7 of the 115 combinations failed to complement. Five prlA alleles were not complemented by a single prlG allele affecting an amino acid located within the transmembrane domain. Five other alleles affected amino acids in other transmembrane regions. Other areas that failed to complement affected amino acids located in periplasmid domains.

The allele combinations were then checked by attempting to construct strains containing two mutant alleles. Phage P1 grown on a pool of Tn10 insertions closely linked to one of the prlG alleles was used to transduce a recipient with one of the prlA alleles. Then, the cotransduction frequencywas calculated to figure out whether the strains could be constructed -- inability to construct a strain was supported the conclusion that those alleles conferred a synthetic lethal phenotype. Because there are reasons other than synthetic lethality that strain construction could fail, the frequency of cotransduction of each of the prlG and Tn10 mutatiions into a wild-type strain was tested. In addition, a strain with a prlA allele that, though a strong suppressor, did not cause a synthetic lethal phenotype with any prlG alleles.

So how does this help determine protein-protein interactions? First, the mutants used were allele specific. Note that the authors used many different combinations of alleles to check for complementation and that only 7 out of 115 failed to complement. Furthermore, the alleles that failed to complement grouped in discrete regions of the gene, suggesting that the inability of these specific amino acid residues to interact was responsible for the secretion defect.

Depletion of a regulated gene product. Another example demonstrates how depletion of a gene product can be used to assay for synthetic lethality. This example involves the role of the smc and spoIIIE gene products in cell division of Bacillus subtilis.

SMC is required for proper chromosome partitioning in B. subtilis. smc null mutants have an abnormal nucleoid structure and some cells have no nucleoid. When grown in rich media, smc null mutants have a temperature sensitive (Ts) phenotype.

SpoIIIE aids in chromosome partitioning by transferring the chromosome through the septa during sporulation. spoIIIE mutants are not able to sporulate because their chromosomes are cut in half by the septum (2).

Britton and Grossman were unable to construct a double mutant of spoIIIE and smc, suggesting that spoIIIE smc double mutants are not viable. Because they could not construct a healthy strain with the haploid smc(Ts) mutation, the double mutant was constructed in a strain with the wild-type smc gene cloned behind an inducible promoter system on a complementing plasmid. The strain was grown in the presence of inducer, then the inducer was removed from the media. The strain grew for a few generations, but once the pre-made SMC was diluted the cells stopped dividing. Half of cells with a wild-type copy of spoIIIE survived after SMC concentrations plunged, but less than 15% of the double mutants survived. This suggests that the cell needs SMC for partitioning, but SpoIIIE can partially compensate for this function in mutants that have little or no SMC.


Return to Microbial Genetics supplement.

The examples were summarized by Stephanie French-Mischo for a Mcbio 490 project
at theDepartment of Microbiology, University of Illinois, Urbana

Please send comments, suggestions, or questions to
Last modified July 16, 2002