EcoGene

EcogeneTopic Page

TopicPage for Essential Genes of Escherichia coli K-12

Minimal Genome

Keio Collection

Description:

Several approaches have been taken to identify the Essential Genes of E. coli K-12. A rigorous approach is often taken when studying one gene at a time. This consists of getting a positive result by deleting the essential gene on the chromosome while having a good copy on a plasmid (or also on the chromosome) with a tightly regulated promoter. Removal of the inducer leads to cell death or growth cessation if the gene is essential under those growth conditions.

Details:

A relatively reliable set of genome-wide essentiality predictions come from the Keio collection, the product of a genome-wide targeted mutagenesis project: Baba (2006) made null mutations in 3985 E. coli genes, but were unable to do so for 303 candidate essential genes.

Three antitoxin genes (chpR, chpS, yefM) could not be deleted, presumably only because the adjacent toxin genes were still intact, illustrating the difficulty of providing a consistent, biologically meaningful definition of an essential gene (Baba, 2006). A fourth antitoxin gene yafN could only be deleted because the corresponding toxin gene yafO has a frameshift mutation inactivating it in the antitoxin deletion mutant strain, so yafN has been added to the list of essential genes (Singletary, 2009).

secM is a positive regulatory leader peptide normally required for translation initiation of the downstream essential secA gene that is classified as an essential gene, although a mutant with a suppressor presumably allowing constitutive expression of secA was isolated (Baba, 2006). This suppressed secM deletion strain is not included in the Keio collection.

The definition of an essential gene in EcoGene is evolving, but differs from that utilized by Baba (2006). The gene must be essential under all growth conditions tested. For example, if low/high temperature, oxygen, salt or nutrient supplements are found to allow growth of a haploid deletion mutant, the gene is considered as non-essential. Also in contrast to Baba (2006), a gene is still considered as essential for growth if null mutants do not grow in the absence of extragenic suppressor mutations.

One of the many complications of essential gene analysis is the definition of the standard essential growth conditions. A gene that is required for growth under one condition may not be required for growth under a different condition. In general, 37C, rich medium, and aeration are the minimum standard growth conditions to assess essentiality.

Baba (2006) categorize the gapA and dap genes as nonessential genes that appear to be essential because they expect they can be grown on different media, but failed to get mutants on the media used in the experiment. On this presumption that the dap deletions could be isolated using DAP-supplemented media and that gapA deletions will grow on a suitable medium, the genes gapA, dapA, dapB, dapD and dapE were omitted from this Essential Genes topic geneset as unlikely to be essential genes, even though they are on the list of the Baba (2006) 303 esssential genes.

Another complication is the definition of growth. Some deletions of non-essential genes allow growth, but at a greatly reduced growth rate. Second site suppressors of these growth defects (pseudorevertants) can arise, and suppressor mutations can allow growth even if the orginal mutation is lethal in the absence of suppressor mutations. For example, rpoE, requires a ydcQ suppressor mutation to survive (De Las Penas, 1997; Button, 2006). Baba (2006) did not isolate a suppressed rpoE deletion, therefore rpoE is in their list of 303 essential genes. Using the same technology, Egler (2005) were able to readily obtain suppressed deletions in rpoE.

Pre-existing duplications is another problem, particularly when global insertion mutagenesis is performed. A population of cells can contain large tandem duplications resulting from spontaneous recombination between rrn or rhs loci. As many as 1% of the cells can be duplicated for any particular gene. Thus insertions can be obtained in one copy, leaving one good copy.

Baba (2006) cite a personal communication from R. D'Ari and K. Nakahihashi that the essential glyS and ileS genes have intact gene copies present in the Keio disruption strains. glyS and ileS are included in the geneset for this Essential Genes topic. Another problem with global transposition experiments that are evalauted solely on the basis of negative results can be target size (small genes) and insertion cold spots.

Of the 303 essential gene candidates listed in Supplementary Table 6 of Baba (2006), six predicted gene candidates were omitted from this Essential Genes topic geneset as unlikely to be essential genes. They may represent recombination cold spots or other problems making the deletion constructs.

Three ORFs, JW5190, JW5193, and JW5379, have been dropped from GenoBase version 6 as spurious ORFs and were not annotated as ORFs in MG1655, in EcoGene or in GenBank.

1. JW5190 is a 73 codon ORF on the opposite strand and overlapping the narX promoter. There is no conservation or function evidence that it would encode a protein. Recombination may be problematic in this area of divergent narX_narG transcription, also near the terminus of replication. This ORF is annotated as a protein coding region in the genome sequences of several non-K12 strains of E. coli.

2. JW6379 is a 37 codon ORF that overlaps the start of accD, which is an essential gene, explaining why it was dropped from GenoBase and why it was erroneously listed by Baba (2006) as an essential gene candidate.

3. JW5193 is an ORF whose primer DNA sequence is not in MG1665.

Three apparent pseudogenes, JW0055 (yabP), JW0208 (yafF), and JW5647 (yibJ), that are unlikely to code for essential genes, are omitted from the linked geneset for this Essential Genes topic. Two of them are associated with Rhs RecA-dependent recombinational hotspsots which may have interfered with deletion recovery.

A review of the assignment of essentiality in EcoGene based upon the results of Baba (2006) is ongoing.

1. The rnc gene has been shown to be non-essential (Yu, 2000), and thus is omitted from the geneset for this Essential Genes topic.

2. The rpoH gene can be deleted, but mutants do not grow at temperatures above 20C (Zhou, 1988), explaining why there is no rpoH mutant in the Keio collection. The rpoH gene is considered as non-essential in EcoGene and is omitted from the geneset for this Essential Genes topic.

3. The folE gene is not essential, but grows poorly on LB plus thymidine (Klaus, 2005), explaining why Baba (2006) failed to isolate a mutant. folE is omitted from the geneset for this Essential Genes topic.

4. The lptB and ycdO Keio deletions are in genes reported to be essential by others. However, due to possible experimental differences, these genes are not yet added ot the essential gene set.

5. Bubunenko (2007) report additional discrepacies with the results of Baba (2006), further discussing problems associated with high-throughput mutagenesis. They report aditional false-negatives: no deletions recovered for non-essential genes: rnc (polarity); rpsI, rpsM, rpsQ (poor growth), and false-positives: strains with essential gene deletions can be isolated, but only in heterozygous merodiploids, e.g. rho and rpsU.
As a result of this analysis, rpsI, rpsM, and rpsQ are omitted from the geneset for this Essential Genes topic and rho and rpsU are added to the linked geneset for this Essential Genes topic.

6. Overexpression of yrbA, which is paralogous to the morphogene bolA, is toxic; an insertion in the upstream gene mlaB(yrbB) can be deleterious due to increased expression of yrbA, which is why the non-essential mlaB(yrbB) gene does not have a kan insertion in the Keio collection (Malinverni, 2009). mlaB(yrbB) is not listed as an essential gene in EcoGene.

Although the focus is on essential genes, essential functions can be considered. It is possible that a single essential function can be encoded by two or more genes, either homologous or of independent origins. These can be revealed by constructing synthetic lethal multiple mutants, e.g Yu (2006). However in this case, the growth defect of the sucABCD can be phenotypically suppressed by addition of succinyl-CoA to the growth medium, so we consider that to be a non-essential nutritional auxotrophy.

A high throughput global insertion mutagenesis project predicted 620 essential genes that failed to accumulate insertion mutations (Gerdes, 2003).
These results have been incorporated into EchoBase, but the results are not noted in EcoGene.