Elsevier

Metabolic Engineering

Volume 7, Issue 4, July 2005, Pages 318-327
Metabolic Engineering

Functional analysis of 1440 Escherichia coli genes using the combination of knock-out library and phenotype microarrays

https://doi.org/10.1016/j.ymben.2005.06.004Get rights and content

Abstract

Escherichia coli is one of the best elucidated organisms. However, about 40% of E. coli genes have not been assigned to their function yet. We analyzed 1440 single gene knock-out mutants using the GN2-MicroPlate, which permits assay of 95 carbon-source utilizations simultaneously. In the knock-out library there are 1044 of so called y-genes with no apparent function. The raw dataset was analyzed and genes were interrelated by the clustering method of the GeneSpring™ software. In the resulted dendrogram of genes, a group of genes with known and related function tended to be assembled into a cluster. Our clustering method would be useful for functional assignment of so called y-genes with no apparent function, since the resulted dendrogram could connect y-genes to phenotype and function of well-studied genes.

Introduction

Recent advance in the field of microbial genomics seems to open the way to comprehensive elucidation of the life of microbes. However, we have been still confronting functionally unknown genes. Even in smaller genomes, genes were reported as unknown to the extents of 16%, 26%, 12%, and 14% for Wigglesworthia glossinidia (698 kbp), Tropheryma whipplei TW08/27 (926 kbp), Blochmannia floridanus (706 kbp), and Buchnera sp. APS (641 kbp), respectively (Akman et al., 2002; Bentley et al., 2003; Gil et al., 2003; Shigenobu et al., 2000). Though Mycoplasma genitalium G-37 is thought to be the self-replicating organism with the smallest genome of 580 kbp, 20% of CDSs were reported to have no matches to protein sequences from the database (Fraser et al., 1995). We can still recognize unknown genes in the genome of M. genitalium G-37 almost to the same extent even in the case that we use the ERGO database constructed with the recent knowledge (Overbeek et al., 2003). Last year genome sequence of Mycoplasma hyopneumoniae strain 232 was reported (Minion et al., 2004). The 893-kbp genome of the bacterium was found to include functionally unknown genes up to 56% of its CDSs. It is still difficult to comprehensively understand whole content of genes in a bacterium even in the case of one with the genome smaller than 1 Mbp. Furthermore about 40% of Escherichia coli genes have not been assigned to their function yet, though E. coli is one of the best elucidated organisms. It is still difficult to elucidate function of so-called y-genes having no assigned function.

A new technology for phenotypic analysis was developed by Bochner and colleagues (Bochner et al., 2001) and called as Phenotype MicroArrays (PM). PM technology was designed to test a large number of cellular phenotypes simultaneously. Nearly 2000 phenotypic tests can be performed in wells of 20 96-well microplates. Each well contained a particular chemical and tetrazolium violet and develops a purple color quantitatively according to the respiration of inoculated cells. Panoramic view of the color change was reported to be quantitative and highly reproductive. PM technology was already applied to analyze E. coli mutants with deletions of all two-component systems (Zhou et al., 2003). In the analysis, many phenotypes were shown to be as expected, and several new phenotypes were also revealed. PM technology would be useful for phenotypic analysis of knock-out mutants. On the other hand, Blattner and his coworkers have been constructing a set of mutant strains of MG1655, designated as Sce-poson mutants, and analyzing them with PM plates. Current number of mutants for individual CDSs is about 2000 (Kang et al., 2004; http://www.genome.wisc.edu/functional/tnmutagenesis.htm). Part of the dataset of PM analysis was used to assist the construction of in silico E. coli strain (Covert et al., 2004). However, clustering of mutant phenotypes has not been studied.

Independent from the Sce-poson mutants, another library of gene knock-out mutants has been established in Japan and designated as the KO library (Mori, 2004). It consists of more than 4000 gene knock-outs. We have been analyzing phenotypic changes of the KO library using the GN2-MicroPlate™ (GN2) that is a part of PM plates (Biolog Inc., CA). The GN2 plate can qualify 95 carbon-source utilizations of a particular mutant simultaneously. First we selected 1440 strains, 48% of which are knock-outs of poorly characterized genes in the COGs database. The raw dataset of GN2 analysis was processed by the GeneSpring™ software (Silicon Genetics, CA) in order to find clusters of interrelated genes that showed similar phenotypic changes when they were deleted. In the resulting dendrogram of genes, genes of known and related function tended to be assembled into a cluster. This result indicated that our clustering method would be useful for functional assignment of y-genes, since y-genes could be connected to phenotype and function of well-known genes in the same cluster.

Section snippets

Strain

E. coli BW25113 [lacIq rrnBT14 ΔlacZWJ16 hsdR514 ΔaraBADAH33 ΔrhaBADLD78] was used as the parent for constructing in-frame null mutants for 1440 open reading frames (Datsenko and Wanner, 2000). Gene disruptants were constructed systematically by replacing a target gene by PCR-amplified kanamycin resistance gene with short homologous regions to the target at both ends (Datsenko and Wanner, 2000; Mori, 2004).

Phenotype analysis using the GN2-MicroPlate

The GN2-MicroPlate (Biolog Inc., CA) was used for the phenotypic analysis of 1440 mutants

Used gene knock-out mutants

We used 1440 knock-out mutants that are the part of the KO library. In the library, the target gene was replaced by kanamycin resistance gene (kan) to make null mutants (Mori, 2004). Content of the deleted genes are summarized in Table 1 according to their function described in the COGs database (http://www.ncbi.nlm.nih.gov/COG/new/) and the GenoBase (http://ecoli.aist-nara.ac.jp/GB5/search.jsp/). These mutants included genes of energy production process, regulation, and especially poorly

Acknowledgments

We thank Natsuka Shimodate for her excellent technical assistance. This study was carried out as a part of The Project for Development of a Technological Infrastructure for Industrial Bioprocesses on R&D of New Industrial Science and Technology Frontiers by Ministry of Economy, Trade & Industry (METI), and supported by New Energy and Industrial Technology Development Organization (NEDO).

References (33)

  • B.R. Bochner et al.

    Phenotype microarrays for high-throughput phenotypic testing and assay of gene function

    Genome Res.

    (2001)
  • M.W. Covert et al.

    Integrating high-throughput and computational data elucidates bacterial networks

    Nature

    (2004)
  • K.A. Datsenko et al.

    One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products

    Proc. Natl. Acad. Sci.

    (2000)
  • C.M. Fraser et al.

    The minimal gene complement of Mycoplasma genitalium

    Science

    (1995)
  • R. Gil et al.

    The genome sequence of Blochmannia floridanus: comparative analysis of reduced genomes

    Proc. Natl. Acad. Sci.

    (2003)
  • P. Golby et al.

    Identification and characterization of a two-component sensor-kinase and response-regulator system (DcuS-DcuR) controlling gene expression in response to C4-dicarboxylates in Escherichia coli

    J. Bacteriol.

    (1999)
  • Cited by (38)

    • Genome-wide identification of the targets for genetic manipulation to improve l-lactate production by Saccharomyces cerevisiae by using a single-gene deletion strain collection

      2013, Journal of Biotechnology
      Citation Excerpt :

      Gene deletion strain collections are available for some microorganisms such as S. cerevisiae (Giaever et al., 2002), Schizosaccharomyces pombe (Kim et al., 2010; Spirek et al., 2010), Escherichia coli (Baba et al., 2006; Yamamoto et al., 2009), and Bacillus subtilis (Kobayashi et al., 2003). These deletion strain collections are powerful tools not only for comprehensive phenotypic analyses but also for the identification of genes related to specific cellular behaviors; for example, identification of the genes required for growth under stress conditions (Ando et al., 2006, 2007; Auesukaree et al., 2009; Fujita et al., 2006; Kawahata et al., 2006; Mira et al., 2010; Teixeira et al., 2009; van Voorst et al., 2006; Yoshikawa et al., 2009), analysis of pleiotropy of genes (Dudley et al., 2005), identification of the genes responsible for swarming activity (Inoue et al., 2007), identification of the genes responsible for the sensitivity to antibiotics (Tamae et al., 2008), analysis of amino acid production (Cooper et al., 2010) and analysis of comprehensive metabolic activity using phenotypic arrays (Ito et al., 2005; Tohsato and Mori, 2008). Moreover, these deletion strain collections can be used in the comprehensive identification of target genes to increase the industrial production of target substance(s).

    • The diversity of microbial aldo/keto reductases from Escherichia coli K12

      2013, Chemico-Biological Interactions
      Citation Excerpt :

      Related sequences are found in a wide range of bacteria from enterobacteria, α-, β- and χ proteobacteria and high GC gram positive bacteria. Of all the AKRs described the ΔydbC knockout is the only one which has an observed phenotype, in that it is not able to respire when grown on acetate [42]. Structurally, YdbC shows some similarity to AKR11 proteins with a long loop 1 and 3, however it has a short loop B as is seen in the 25DKG reductases and no significant C-terminal loop.

    View all citing articles on Scopus
    View full text