The Escherichia coli transcriptome mostly consists of independently regulated modules

Independent component analysis recovers consistent regulatory signals from disparate datasets

The availability of bacterial transcriptomes has dramatically increased in recent years. This data deluge could result in detailed inference of underlying regulatory networks, but the diversity of experimental platforms and protocols introduces critical biases that could hinder scalable analysis of existing data. Here, we show that the underlying structure of the E. coli transcriptome, as determined by Independent Component Analysis (ICA), is conserved across multiple independent datasets, including both RNA-seq and microarray datasets. We subsequently combined five transcriptomics datasets into a large compendium containing over 800 expression profiles and discovered that its underlying ICA-based structure was still comparable to that of the individual datasets. With this understanding, we expanded our analysis to over 3,000 E. coli expression profiles and predicted three high-impact regulons that respond to oxidative stress, anaerobiosis, and antibiotic treatment. ICA thus enables deep analysis of disparate data to uncover new insights that were not visible in the individual datasets.

Cells adapt to diverse environments by regulating gene expression. Genome-wide measurements of gene expression levels have exponentially increased in recent years, but successful integration and analysis of these datasets are limited. Recently, we showed that independent component analysis (ICA), a signal deconvolution algorithm, can separate a large bacterial gene expression dataset into groups of co-regulated genes. This previous study focused on data generated by a standardized pipeline and did not address whether ICA extracts the same quantitative co-expression signals across expression profiling platforms. In this study, we show that ICA finds similar co-regulation patterns underlying multiple gene expression datasets and can be used as a tool to integrate and interpret diverse datasets. Using a dataset containing over 3,000 expression profiles, we predicted three new regulons and characterized their activities. Since large, standardized expression datasets only exist for a few bacterial strains, these results broaden the possible applications of this tool to better understand transcriptional regulation across a wide range of microbes.

A palette of fluorophores that are differentially accumulated by wild-type and mutant strains of Escherichia coli: surrogate ligands for profiling bacterial membrane transporters

Our previous work demonstrated that two commonly used fluorescent dyes that were accumulated by wild-type Escherichia coli MG1655 were differentially transported in single-gene knockout strains, and also that they might be used as surrogates in flow cytometric transporter assays. We summarize the desirable properties of such stains, and here survey 143 candidate dyes. We eventually triage them (on the basis of signal, accumulation levels and cost) to a palette of 39 commercially available and affordable fluorophores that are accumulated significantly by wild-type cells of the 'Keio' strain BW25113, as measured flow cytometrically. Cheminformatic analyses indicate both their similarities and their (much more considerable) structural differences. We describe the effects of pH and of the efflux pump inhibitor chlorpromazine on the accumulation of the dyes. Even the 'wild-type' MG1655 and BW25113 strains can differ significantly in their ability to take up such dyes. We illustrate the highly differential uptake of our dyes into strains with particular lesions in, or overexpressed levels of, three particular transporters or transporter components (yhjV, yihN and tolC). The relatively small collection of dyes described offers a rapid, inexpensive, convenient and informative approach to the assessment of microbial physiology and phenotyping of membrane transporter function.

iModulonDB: a knowledgebase of microbial transcriptional regulation derived from machine learning

Independent component analysis (ICA) of bacterial transcriptomes has emerged as a powerful tool for obtaining co-regulated, independently-modulated gene sets (iModulons), inferring their activities across a range of conditions, and enabling their association to known genetic regulators. By grouping and analyzing genes based on observations from big data alone, iModulons can provide a novel perspective into how the composition of the transcriptome adapts to environmental conditions. Here, we present iModulonDB (imodulondb.org), a knowledgebase of prokaryotic transcriptional regulation computed from high-quality transcriptomic datasets using ICA. Users select an organism from the home page and then search or browse the curated iModulons that make up its transcriptome. Each iModulon and gene has its own interactive dashboard, featuring plots and tables with clickable, hoverable, and downloadable features. This site enhances research by presenting scientists of all backgrounds with co-expressed gene sets and their activity levels, which lead to improved understanding of regulator-gene relationships, discovery of transcription factors, and the elucidation of unexpected relationships between conditions and genetic regulatory activity. The current release of iModulonDB covers three organisms (Escherichia coli, Staphylococcus aureus and Bacillus subtilis) with 204 iModulons, and can be expanded to cover many additional organisms.

The Expanding Computational Toolbox for Engineering Microbial Phenotypes at the Genome Scale

Microbial strains are being engineered for an increasingly diverse array of applications, from chemical production to human health. While traditional engineering disciplines are driven by predictive design tools, these tools have been difficult to build for biological design due to the complexity of biological systems and many unknowns of their quantitative behavior. However, due to many recent advances, the gap between design in biology and other engineering fields is closing. In this work, we discuss promising areas of development of computational tools for engineering microbial strains. We define five frontiers of active research: (1) Constraint-based modeling and metabolic network reconstruction, (2) Kinetics and thermodynamic modeling, (3) Protein structure analysis, (4) Genome sequence analysis, and (5) Regulatory network analysis. Experimental and machine learning drivers have enabled these methods to improve by leaps and bounds in both scope and accuracy. Modern strain design projects will require these tools to be comprehensively applied to the entire cell and efficiently integrated within a single workflow. We expect that these frontiers, enabled by the ongoing revolution of big data science, will drive forward more advanced and powerful strain engineering strategies.

Machine learning uncovers independently regulated modules in the Bacillus subtilis transcriptome

The transcriptional regulatory network (TRN) of Bacillus subtilis coordinates cellular functions of fundamental interest, including metabolism, biofilm formation, and sporulation. Here, we use unsupervised machine learning to modularize the transcriptome and quantitatively describe regulatory activity under diverse conditions, creating an unbiased summary of gene expression. We obtain 83 independently modulated gene sets that explain most of the variance in expression and demonstrate that 76% of them represent the effects of known regulators. The TRN structure and its condition-dependent activity uncover putative or recently discovered roles for at least five regulons, such as a relationship between histidine utilization and quorum sensing. The TRN also facilitates quantification of population-level sporulation states. As this TRN covers the majority of the transcriptome and concisely characterizes the global expression state, it could inform research on nearly every aspect of transcriptional regulation in B. subtilis.

The systems-level regulatory structure underlying gene expression in bacteria can be inferred using machine learning algorithms. Here we show this structure for Bacillus subtilis, present five hypotheses gleaned from it, and analyse the process of sporulation from its perspective.

The biology of ergothioneine, an antioxidant nutraceutical

Ergothioneine (ERG) is an unusual thio-histidine betaine amino acid that has potent antioxidant activities. It is synthesised by a variety of microbes, especially fungi (including in mushroom fruiting bodies) and actinobacteria, but is not synthesised by plants and animals who acquire it via the soil and their diet, respectively. Animals have evolved a highly selective transporter for it, known as solute carrier family 22, member 4 (SLC22A4) in humans, signifying its importance, and ERG may even have the status of a vitamin. ERG accumulates differentially in various tissues, according to their expression of SLC22A4, favouring those such as erythrocytes that may be subject to oxidative stress. Mushroom or ERG consumption seems to provide significant prevention against oxidative stress in a large variety of systems. ERG seems to have strong cytoprotective status, and its concentration is lowered in a number of chronic inflammatory diseases. It has been passed as safe by regulatory agencies, and may have value as a nutraceutical and antioxidant more generally.

OxyR Is a Convergent Target for Mutations Acquired during Adaptation to Oxidative Stress-Prone Metabolic States

Oxidative stress is concomitant with aerobic metabolism. Thus, bacterial genomes encode elaborate mechanisms to achieve redox homeostasis. Here we report that the peroxide-sensing transcription factor, oxyR, is a common mutational target using bacterial species belonging to two genera, Escherichia coli and Vibrio natriegens, in separate growth conditions implemented during laboratory evolution. The mutations clustered in the redox active site, dimer interface, and flexible redox loop of the protein. These mutations favor the oxidized conformation of OxyR that results in constitutive expression of the genes it regulates. Independent component analysis of the transcriptome revealed that the constitutive activity of OxyR reduces DNA damage from reactive oxygen species, as inferred from the activity of the SOS response regulator LexA. This adaptation to peroxide stress came at a cost of lower growth, as revealed by calculations of proteome allocation using genome-scale models of metabolism and macromolecular expression. Further, identification of similar sequence changes in natural isolates of E. coli indicates that adaptation to oxidative stress through genetic changes in oxyR can be a common occurrence.

Elucidation of Regulatory Modes for Five Two-Component Systems in Escherichia coli Reveals Novel Relationships

Escherichia coli uses two-component systems (TCSs) to respond to environmental signals. TCSs affect gene expression and are parts of E. coli's global transcriptional regulatory network (TRN). Here, we identified the regulons of five TCSs in E. coli MG1655: BaeSR and CpxAR, which were stimulated by ethanol stress; KdpDE and PhoRB, induced by limiting potassium and phosphate, respectively; and ZraSR, stimulated by zinc. We analyzed RNA-seq data using independent component analysis (ICA). ChIP-exo data were used to validate condition-specific target gene binding sites. Based on these data, we do the following: (i) identify the target genes for each TCS; (ii) show how the target genes are transcribed in response to stimulus; and (iii) reveal novel relationships between TCSs, which indicate noncognate inducers for various response regulators, such as BaeR to iron starvation, CpxR to phosphate limitation, and PhoB and ZraR to cell envelope stress. Our understanding of the TRN in E. coli is thus notably expanded.IMPORTANCE E. coli is a common commensal microbe found in the human gut microenvironment; however, some strains cause diseases like diarrhea, urinary tract infections, and meningitis. E. coli's two-component systems (TCSs) modulate target gene expression, especially related to virulence, pathogenesis, and antimicrobial peptides, in response to environmental stimuli. Thus, it is of utmost importance to understand the transcriptional regulation of TCSs to infer bacterial environmental adaptation and disease pathogenicity. Utilizing a combinatorial approach integrating RNA sequencing (RNA-seq), independent component analysis, chromatin immunoprecipitation coupled with exonuclease treatment (ChIP-exo), and data mining, we suggest five different modes of TCS transcriptional regulation. Our data further highlight noncognate inducers of TCSs, which emphasizes the cross-regulatory nature of TCSs in E. coli and suggests that TCSs may have a role beyond their cognate functionalities. In summary, these results can lead to an understanding of the metabolic capabilities of bacteria and correctly predict complex phenotype under diverse conditions, especially when further incorporated with genome-scale metabolic models.

The Bitome: digitized genomic features reveal fundamental genome organization

A genome contains the information underlying an organism's form and function. Yet, we lack formal framework to represent and study this information. Here, we introduce the Bitome, a matrix composed of binary digits (bits) representing the genomic positions of genomic features. We form a Bitome for the genome of Escherichia coli K-12 MG1655. We find that: (i) genomic features are encoded unevenly, both spatially and categorically; (ii) coding and intergenic features are recapitulated at high resolution; (iii) adaptive mutations are skewed towards genomic positions with fewer features; and (iv) the Bitome enhances prediction of adaptively mutated and essential genes. The Bitome is a formal representation of a genome and may be used to study its fundamental organizational properties.

Rewiring of growth-dependent transcription regulation by a point mutation in region 1.1 of the housekeeping σ factor

In bacteria, rapid adaptation to changing environmental conditions depends on the interplay between housekeeping and alternative σ factors, responsible for transcription of specific regulons by RNA polymerase (RNAP). In comparison with alternative σ factors, primary σs contain poorly conserved region 1.1, whose functions in transcription are only partially understood. We found that a single mutation in region 1.1 in Escherichia coli σ⁷⁰ rewires transcription regulation during cell growth resulting in profound phenotypic changes. Despite its destabilizing effect on promoter complexes, this mutation increases the activity of rRNA promoters and also decreases RNAP sensitivity to the major regulator of stringent response DksA. Using total RNA sequencing combined with single-cell analysis of gene expression we showed that changes in region 1.1 disrupt the balance between the "greed" and "fear" strategies thus making the cells more susceptible to environmental threats and antibiotics. Our results reveal an unexpected role of σ region 1.1 in growth-dependent transcription regulation and suggest that changes in this region may facilitate rapid switching of RNAP properties in evolving bacterial populations.

Extreme Antagonism Arising from Gene-Environment Interactions

Antagonistic interactions in biological systems, which occur when one perturbation blunts the effect of another, are typically interpreted as evidence that the two perturbations impact the same cellular pathway or function. Yet, this interpretation ignores extreme antagonistic interactions wherein an otherwise deleterious perturbation compensates for the function lost because of a prior perturbation. Here, we report on gene-environment interactions involving genetic mutations that are deleterious in a permissive environment but beneficial in a specific environment that restricts growth. These extreme antagonistic interactions constitute gene-environment analogs of synthetic rescues previously observed for gene-gene interactions. Our approach uses two independent adaptive evolution steps to address the lack of experimental methods to systematically identify such extreme interactions. We apply the approach to Escherichia coli by successively adapting it to defined glucose media without and with the antibiotic rifampicin. The approach identified multiple mutations that are beneficial in the presence of rifampicin and deleterious in its absence. The analysis of transcription shows that the antagonistic adaptive mutations repress a stringent response-like transcriptional program, whereas nonantagonistic mutations have an opposite transcriptional profile. Our approach represents a step toward the systematic characterization of extreme antagonistic gene-drug interactions, which can be used to identify targets to select against antibiotic resistance.

Gene Dispensability in Escherichia coli Grown in Thirty Different Carbon Environments

While there has been much study of bacterial gene dispensability, there is a lack of comprehensive genome-scale examinations of the impact of gene deletion on growth in different carbon sources. In this context, a lot can be learned from such experiments in the model microbe Escherichia coli where much is already understood and there are existing tools for the investigation of carbon metabolism and physiology (1). Gene deletion studies have practical potential in the field of antibiotic drug discovery where there is emerging interest in bacterial central metabolism as a target for new antibiotics (2). Furthermore, some carbon utilization pathways have been shown to be critical for initiating and maintaining infection for certain pathogens and sites of infection (3–5). Here, with the use of high-throughput solid medium phenotyping methods, we have generated kinetic growth measurements for 3,796 genes under 30 different carbon source conditions. This data set provides a foundation for research that will improve our understanding of genes with unknown function, aid in predicting potential antibiotic targets, validate and advance metabolic models, and help to develop our understanding of E. coli metabolism.

Central metabolism is a topic that has been studied for decades, and yet, this process is still not fully understood in Escherichia coli, perhaps the most amenable and well-studied model organism in biology. To further our understanding, we used a high-throughput method to measure the growth kinetics of each of 3,796 E. coli single-gene deletion mutants in 30 different carbon sources. In total, there were 342 genes (9.01%) encompassing a breadth of biological functions that showed a growth phenotype on at least 1 carbon source, demonstrating that carbon metabolism is closely linked to a large number of processes in the cell. We identified 74 genes that showed low growth in 90% of conditions, defining a set of genes which are essential in nutrient-limited media, regardless of the carbon source. The data are compiled into a Web application, Carbon Phenotype Explorer (CarPE), to facilitate easy visualization of growth curves for each mutant strain in each carbon source. Our experimental data matched closely with the predictions from the EcoCyc metabolic model which uses flux balance analysis to predict growth phenotypes. From our comparisons to the model, we found that, unexpectedly, phosphoenolpyruvate carboxylase (ppc) was required for robust growth in most carbon sources other than most trichloroacetic acid (TCA) cycle intermediates. We also identified 51 poorly annotated genes that showed a low growth phenotype in at least 1 carbon source, which allowed us to form hypotheses about the functions of these genes. From this list, we further characterized the ydhC gene and demonstrated its role in adenosine efflux.

Correlated chromosomal periodicities according to the growth rate and gene expression

Linking genetic information to population fitness is crucial to understanding living organisms. Despite the abundant knowledge of the genetic contribution to growth, the overall patterns/features connecting genes, their expression, and growth remain unclear. To reveal the quantitative and direct connections, systematic growth assays of single-gene knockout Escherichia coli strains under both rich and poor nutritional conditions were performed; subsequently, the resultant growth rates were associated with the original expression levels of the knockout genes in the parental genome. Comparative analysis of growth and the transcriptome identified not only the nutritionally differentiated fitness cost genes but also a significant correlation between the growth rates of the single-gene knockout strains and the original expression levels of these knockout genes in the parental strain, regardless of the nutritional variation. In addition, the coordinated chromosomal periodicities of the wild-type transcriptome and the growth rates of the strains lacking the corresponding genes were observed. The common six-period periodicity was somehow attributed to the essential genes, although the underlying mechanism remains to be addressed. The correlated chromosomal periodicities associated with the gene expression-growth dataset were highly valuable for bacterial growth prediction and discovering the working principles governing minimal genetic information.

Synthetic cross-phyla gene replacement and evolutionary assimilation of major enzymes

The ability of DNA to produce a functional protein even after transfer to a foreign host is of fundamental importance in both evolutionary biology and biotechnology, enabling horizontal gene transfer in the wild and heterologous expression in the lab. However, the influence of genetic particulars on DNA functionality in a new host is poorly understood, as are the evolutionary mechanisms of assimilation and refinement. Here, we describe an automation-enabled large-scale experiment wherein Escherichia coli strains were evolved in parallel after replacement of the genes pgi or tpiA with orthologous DNA from donor species spanning all domains of life, from humans to hyperthermophilic archaea. Via analysis of hundreds of clones evolved for 50,000+ cumulative generations across dozens of independent lineages, we show that orthogene-upregulating mutations can completely mitigate fitness defects that result from initial non-functionality, with coding sequence changes unnecessary. Gene target, donor species and genomic location of the swap all influenced outcomes-both the nature of adaptive mutations (often synonymous) and the frequency with which strains successfully evolved to assimilate the foreign DNA. Additionally, time series DNA sequencing and replay evolution experiments revealed transient copy number expansions, the contingency of lineage outcome on first-step mutations and the ability for strains to escape from suboptimal local fitness maxima. Overall, this study establishes the influence of various DNA and protein features on cross-species genetic interchangeability and evolutionary outcomes, with implications for both horizontal gene transfer and rational strain design.

Independent component analysis of E. coli's transcriptome reveals the cellular processes that respond to heterologous gene expression

Achieving the predictable expression of heterologous genes in a production host has proven difficult. Each heterologous gene expressed in the same host seems to elicit a different host response governed by unknown mechanisms. Historically, most studies have approached this challenge by manipulating the properties of the heterologous gene through methods like codon optimization. Here we approach this challenge from the host side. We express a set of 45 heterologous genes in the same Escherichia coli strain, using the same expression system and culture conditions. We collect a comprehensive RNAseq set to characterize the host's transcriptional response. Independent Component Analysis of the RNAseq data set reveals independently modulated gene sets (iModulons) that characterize the host response to heterologous gene expression. We relate 55% of variation of the host response to: Fear vs Greed (16.5%), Metal Homeostasis (19.0%), Respiration (6.0%), Protein folding (4.5%), and Amino acid and nucleotide biosynthesis (9.0%). If these responses can be controlled, then the success rate with predicting heterologous gene expression should increase.

A quantitative method for proteome reallocation using minimal regulatory interventions

Engineering resource allocation in biological systems is an ongoing challenge. Organisms allocate resources for ensuring survival, reducing the productivity of synthetic biology functions. Here we present a new approach for engineering the resource allocation of Escherichia coli by rationally modifying its transcriptional regulatory network. Our method (ReProMin) identifies the minimal set of genetic interventions that maximizes the savings in cell resources. To this end, we categorized transcription factors according to the essentiality of its targets and we used proteomic data to rank them. We designed the combinatorial removal of transcription factors that maximize the release of resources. Our resulting strain containing only three mutations, theoretically releasing 0.5% of its proteome, had higher proteome budget, increased production of an engineered metabolic pathway and showed that the regulatory interventions are highly specific. This approach shows that combining proteomic and regulatory data is an effective way of optimizing strains using conventional molecular methods.

Revealing 29 sets of independently modulated genes in Staphylococcus aureus, their regulators, and role in key physiological response

Staphylococcus aureus infections impose an immense burden on the healthcare system. To establish a successful infection in a hostile host environment, S. aureus must coordinate its gene expression to respond to a wide array of challenges. This balancing act is largely orchestrated by the transcriptional regulatory network. Here, we present a model of 29 independently modulated sets of genes that form the basis for a segment of the transcriptional regulatory network in clinical USA300 strains of S. aureus. Using this model, we demonstrate the concerted role of various cellular systems (e.g., metabolism, virulence, and stress response) underlying key physiological responses, including response during blood infection.

The ability of Staphylococcus aureus to infect many different tissue sites is enabled, in part, by its transcriptional regulatory network (TRN) that coordinates its gene expression to respond to different environments. We elucidated the organization and activity of this TRN by applying independent component analysis to a compendium of 108 RNA-sequencing expression profiles from two S. aureus clinical strains (TCH1516 and LAC). ICA decomposed the S. aureus transcriptome into 29 independently modulated sets of genes (i-modulons) that revealed: 1) High confidence associations between 21 i-modulons and known regulators; 2) an association between an i-modulon and σS, whose regulatory role was previously undefined; 3) the regulatory organization of 65 virulence factors in the form of three i-modulons associated with AgrR, SaeR, and Vim-3; 4) the roles of three key transcription factors (CodY, Fur, and CcpA) in coordinating the metabolic and regulatory networks; and 5) a low-dimensional representation, involving the function of few transcription factors of changes in gene expression between two laboratory media (RPMI, cation adjust Mueller Hinton broth) and two physiological media (blood and serum). This representation of the TRN covers 842 genes representing 76% of the variance in gene expression that provides a quantitative reconstruction of transcriptional modules in S. aureus, and a platform enabling its full elucidation.

Identification of Modules With Similar Gene Regulation and Metabolic Functions Based on Co-expression Data

Biological systems respond to environmental perturbations and to a large diversity of compounds through gene interactions, and these genetic factors comprise complex networks. In particular, a wide variety of gene co-expression networks have been constructed in recent years thanks to the dramatic increase of experimental information obtained with techniques, such as microarrays and RNA sequencing. These networks allow the identification of groups of co-expressed genes that can function in the same process and, in turn, these networks may be related to biological functions of industrial, medical and academic interest. In this study, gene co-expression networks for 17 bacterial organisms from the COLOMBOS database were analyzed via weighted gene co-expression network analysis and clustered into modules of genes with similar expression patterns for each species. These networks were analyzed to determine relevant modules through a hypergeometric approach based on a set of transcription factors and enzymes for each genome. The richest modules were characterized using PFAM families and KEGG metabolic maps. Additionally, we conducted a Gene Ontology analysis for enrichment of biological functions. Finally, we identified modules that shared similarity through all the studied organisms by using comparative genomics.

Adaptive evolution reveals a tradeoff between growth rate and oxidative stress during naphthoquinone-based aerobic respiration

Evolution fine-tunes biological pathways to achieve a robust cellular physiology. Two and a half billion years ago, rapidly rising levels of oxygen as a byproduct of blooming cyanobacterial photosynthesis resulted in a redox upshift in microbial energetics. The appearance of higher-redox-potential respiratory quinone, ubiquinone (UQ), is believed to be an adaptive response to this environmental transition. However, the majority of bacterial species are still dependent on the ancient respiratory quinone, naphthoquinone (NQ). Gammaproteobacteria can biosynthesize both of these respiratory quinones, where UQ has been associated with aerobic lifestyle and NQ with anaerobic lifestyle. We engineered an obligate NQ-dependent γ-proteobacterium, Escherichia coli ΔubiC, and performed adaptive laboratory evolution to understand the selection against the use of NQ in an oxic environment and also the adaptation required to support the NQ-driven aerobic electron transport chain. A comparative systems-level analysis of pre- and postevolved NQ-dependent strains revealed a clear shift from fermentative to oxidative metabolism enabled by higher periplasmic superoxide defense. This metabolic shift was driven by the concerted activity of 3 transcriptional regulators (PdhR, RpoS, and Fur). Analysis of these findings using a genome-scale model suggested that resource allocation to reactive oxygen species (ROS) mitigation results in lower growth rates. These results provide a direct elucidation of a resource allocation tradeoff between growth rate and ROS mitigation costs associated with NQ usage under oxygen-replete condition.

Adaptive laboratory evolution of Escherichia coli under acid stress

The ability of Escherichia coli to tolerate acid stress is important for its survival and colonization in the human digestive tract. Here, we performed adaptive laboratory evolution of the laboratory strain E. coli K-12 MG1655 at pH 5.5 in glucose minimal medium. After 800 generations, six independent populations under evolution had reached 18.0 % higher growth rates than their starting strain at pH 5.5, while maintaining comparable growth rates to the starting strain at pH 7. We characterized the evolved strains and found that: (1) whole genome sequencing of isolated clones from each evolved population revealed mutations in rpoC appearing in five of six sequenced clones; and (2) gene expression profiles revealed different strategies to mitigate acid stress, which are related to amino acid metabolism and energy production and conversion. Thus, a combination of adaptive laboratory evolution, genome resequencing and expression profiling revealed, on a genome scale, the strategies that E. coli uses to mitigate acid stress.