Analysis of the evolution of resistance to multiple antibiotics enables prediction of the Escherichia coli phenotype-based fitness landscape.

The fitness landscape represents the complex relationship between genotype or phenotype and fitness under a given environment, the structure of which allows the explanation and prediction of evolutionary trajectories. Although previous studies have constructed fitness landscapes by comprehensively studying the mutations in specific genes, the high dimensionality of genotypic changes prevents us from developing a fitness landscape capable of predicting evolution for the whole cell. Herein, we address this problem by inferring the phenotype-based fitness landscape for antibiotic resistance evolution by quantifying the multidimensional phenotypic changes, i.e., time-series data of resistance for eight different drugs. We show that different peaks of the landscape correspond to different drug resistance mechanisms, thus supporting the validity of the inferred phenotype-fitness landscape. We further discuss how inferred phenotype-fitness landscapes could contribute to the prediction and control of evolution. This approach bridges the gap between phenotypic/genotypic changes and fitness while contributing to a better understanding of drug resistance evolution.

Experimental evidence for the correlation between RNA structural fluctuations and the frequency of beneficial mutations.

RNA has been used as a model molecule to understand the adaptive evolution process owing to the simple relationship between the structure (i.e., phenotype) and sequence (i.e., genotype). RNA usually forms multiple substructures with similar thermodynamic stabilities, called structural fluctuations. Ancel and Fontana theoretically proposed that structural fluctuation is directly related to the ease of change in structures by mutations and thus works as a source of adaptive evolution; however, experimental verification is limited. Here, we analyzed 76 RNA genotypes that appeared in our previous in vitro evolution to examine whether (i) RNA fluctuation decreases as adaptive evolution proceeds and (ii) RNAs that have larger fluctuations tend to have higher frequencies of beneficial mutations. We first computationally estimated the structural fluctuations of all RNAs and observed that they tended to decrease as their fitness increased. We next measured the frequency of beneficial mutations for 10 RNA genotypes and observed that the total number of beneficial mutations was correlated with the size of the structural fluctuations. These results consistently support the idea that the structural fluctuation of RNA, at least those evaluated in our study, works as a source of adaptive evolution.

Purifying selection enduringly acts on the sequence evolution of highly expressed proteins in Escherichia coli.

The evolutionary speed of a protein sequence is constrained by its expression level, with highly expressed proteins evolving relatively slowly. This negative correlation between expression levels and evolutionary rates (known as the E-R anticorrelation) has already been widely observed in past macroevolution between species from bacteria to animals. However, it remains unclear whether this seemingly general law also governs recent evolution, including past and de novo, within a species. However, the advent of genomic sequencing and high-throughput phenotyping, particularly for bacteria, has revealed fundamental gaps between the 2 evolutionary processes and has provided empirical data opposing the possible underlying mechanisms which are widely believed. These conflicts raise questions about the generalization of the E-R anticorrelation and the relevance of plausible mechanisms. To explore the ubiquitous impact of expression levels on molecular evolution and test the relevance of the possible underlying mechanisms, we analyzed the genome sequences of 99 strains of Escherichia coli for evolution within species in nature. We also analyzed genomic mutations accumulated under laboratory conditions as a model of de novo evolution within species. Here, we show that E-R anticorrelation is significant in both past and de novo evolution within species in E. coli. Our data also confirmed ongoing purifying selection on highly expressed genes. Ongoing selection included codon-level purifying selection, supporting the relevance of the underlying mechanisms. However, the impact of codon-level purifying selection on the constraints in evolution within species might be smaller than previously expected from evolution between species.

Development of a device that generates a temperature gradient in a microtiter plate for microbial culture.

Although temperature is a fundamental parameter in biology, testing various temperature conditions simultaneously is often difficult. In the present study, we developed a device for generating a temperature gradient in arrays of wells on a microtiter plate. This device consists of a pair of Peltier elements and temperature sensors placed on both ends of a flat aluminum bar to generate a linear temperature gradient. The device loads a microtiter plate at the center of the aluminum bar and transfers the temperature gradient to the bottom of the wells in the plate. This device successfully maintained a temperature gradient of 38.2 to 43.1°C on the horizontal axis of a 96-well microtiter plate in an incubator at 31°C. Furthermore, using this device, we demonstrated a laboratory evolution experiment of Escherichia coli, which was selected on the basis of its ability to grow at high temperatures. The developed device also facilitates a two-dimensional assay to determine the effects of temperature and drug concentrations on cellular growth.

Mutational property of newly identified mutagen l-glutamic acid γ-hydrazide in Escherichia coli.

We previously found that an l-glutamine analog l-glutamic acid γ-hydrazide has high mutagenic activity through the high-throughput laboratory evolution of Escherichia coli. In this study, mutagenicity and mutational property of l-glutamic acid γ-hydrazide were examined by the Ames test and mutation accumulation experiments using E. coli. The Ames test revealed that l-glutamic acid γ-hydrazide showed higher mutagenic activity without metabolic activation than known mutagens 2-aminoanthracene, and cobalt(II) acetate tetrahydrate. This result indicates that l-glutamic acid γ-hydrazide does not require metabolic activation for mutagenic activity in E. coli. Mutation accumulation experiments and whole-genome sequencing analysis revealed the number and spectrum of the accumulated mutations with or without l-glutamic acid γ-hydrazide. In the presence of l-glutamic acid γ-hydrazide, MDS42 strain accumulated 392.3 ± 116.2 point mutations during 30 passages corresponding to 777 generations, while MDS42 strain accumulated 1.5 ± 2.5 point mutations without l-glutamic acid γ-hydrazide during 50 passages corresponding to 1341 generations. The mutational spectrum of l-glutamic acid γ-hydrazide was G/C to A/T transition (82.2 ± 4.3 %) and A/T to G/C transition (17.4 ± 4.3 %). These results indicated that l-glutamic acid γ-hydrazide has a strong mutagenic activity.

Laboratory evolution of Mycobacterium on agar plates for analysis of resistance acquisition and drug sensitivity profiles.

Drug-resistant tuberculosis (TB) is a growing public health problem. There is an urgent need for information regarding cross-resistance and collateral sensitivity relationships among drugs and the genetic determinants of anti-TB drug resistance for developing strategies to suppress the emergence of drug-resistant pathogens. To identify mutations that confer resistance to anti-TB drugs in Mycobacterium species, we performed the laboratory evolution of nonpathogenic Mycobacterium smegmatis, which is closely related to Mycobacterium tuberculosis, against ten anti-TB drugs. Next, we performed whole-genome sequencing and quantified the resistance profiles of each drug-resistant strain against 24 drugs. We identified the genes with novel meropenem (MP) and linezolid (LZD) resistance-conferring mutation, which also have orthologs, in M. tuberculosis H37Rv. Among the 240 possible drug combinations, we identified 24 pairs that confer cross-resistance and 18 pairs that confer collateral sensitivity. The acquisition of bedaquiline or linezolid resistance resulted in collateral sensitivity to several drugs, while the acquisition of MP resistance led to multidrug resistance. The MP-evolved strains showed cross-resistance to rifampicin and clarithromycin owing to the acquisition of a mutation in the intergenic region of the Rv2864c ortholog, which encodes a penicillin-binding protein, at an early stage. These results provide a new insight to tackle drug-resistant TB.

Morphological change of coiled bacterium Spirosoma linguale with acquisition of ß-lactam resistance.

Spirosoma linguale is a gram-negative, coiled bacterium belonging to the family Cytophagaceae. Its coiled morphology is unique in contrast to closely related bacteria belonging to the genus Spirosoma, which have a short, rod-shaped morphology. The mechanisms that generate unique cell morphology are still enigmatic. In this study, using the Spirosoma linguale ATCC33905 strain, we isolated ß-lactam (cefoperazone and amoxicillin)-resistant clones. These clones showed two different cell morphological changes: relatively loosely curved cells or small, horseshoe-shaped cells. Whole-genome resequencing analysis revealed the genetic determinants of ß-lactam resistance and changes in cell morphology. The loose-curved clones commonly had mutations in Slin_5958 genes encoding glutamyl-tRNA amidotransferase B subunit, whereas the small, horseshoe-shaped clones commonly had mutations in either Slin_5165 or Slin_5509 encoding pyruvate dehydrogenase (PDH) components. Two clones, CFP1ESL11 and CFL5ESL4, which carried only one mutation in Slin_5958, showed almost perfectly straight, rod-shaped cells in the presence of amoxicillin. This result suggests that penicillin-binding proteins targeted by amoxicillin play an important role in the formation of a coiled morphology in this bacterium. In contrast, supplementation with acetate did not rescue the growth defect and abnormal cell size of the CFP5ESL9 strain, which carried only one mutation in Slin_5509. These results suggest that PDH is involved in cell-size maintenance in this bacterium.

High-throughput laboratory evolution reveals evolutionary constraints in Escherichia coli.

Understanding the constraints that shape the evolution of antibiotic resistance is critical for predicting and controlling drug resistance. Despite its importance, however, a systematic investigation of evolutionary constraints is lacking. Here, we perform a high-throughput laboratory evolution of Escherichia coli under the addition of 95 antibacterial chemicals and quantified the transcriptome, resistance, and genomic profiles for the evolved strains. Utilizing machine learning techniques, we analyze the phenotype-genotype data and identified low dimensional phenotypic states among the evolved strains. Further analysis reveals the underlying biological processes responsible for these distinct states, leading to the identification of trade-off relationships associated with drug resistance. We also report a decelerated evolution of ß-lactam resistance, a phenomenon experienced by certain strains under various stresses resulting in higher acquired resistance to ß-lactams compared to strains directly selected by ß-lactams. These findings bridge the genotypic, gene expression, and drug resistance gap, while contributing to a better understanding of evolutionary constraints for antibiotic resistance.

Suppression of antibiotic resistance evolution by single-gene deletion.

Antibiotic treatment generally results in the selection of resistant bacterial strains, and the dynamics of resistance evolution is dependent on complex interactions between cellular components. To better characterize the mechanisms of antibiotic resistance and evaluate its dependence on gene regulatory networks, we performed systematic laboratory evolution of Escherichia coli strains with single-gene deletions of 173 transcription factors under three different antibiotics. This resulted in the identification of several genes whose deletion significantly suppressed resistance evolution, including arcA and gutM. Analysis of double-gene deletion strains suggested that the suppression of resistance evolution caused by arcA and gutM deletion was not caused by epistatic interactions with mutations known to confer drug resistance. These results provide a methodological basis for combinatorial drug treatments that may help to suppress the emergence of resistant pathogens by inhibiting resistance evolution.

Complete Genome Sequence of a Radioresistant Bacterial Strain, Deinococcus grandis ATCC 43672.

Deinococcus grandis is a radioresistant bacterial species isolated from freshwater fish. In this article, we report the complete genome sequence of D. grandis strain ATCC 43672. This sequence is useful for comparative genomics to understand the traits of Deinococcus species and can be used as a reference in experimental genetics.

Complete Genome Sequences of Three Star-Shaped Bacteria, Stella humosa, Stella vacuolata, and Stella Species ATCC 35155.

Stella species are unique star-shaped alphaproteobacteria found in various environments. We report the complete genome sequences of three Stella strains, Stella humosa ATCC 43930, Stella vacuolata ATCC 43931, and Stella species ATCC 35155. These are the first complete genome sequences of members of the genus Stella.

Prediction of Cross-resistance and Collateral Sensitivity by Gene Expression profiles and Genomic Mutations.

In adaptive evolution, an increase in fitness to an environment is frequently accompanied by changes in fitness to other environmental conditions, called cross-resistance and sensitivity. Although the networks between fitness changes affect the course of evolution substantially, the mechanisms underlying such fitness changes are yet to be fully elucidated. Herein, we performed high-throughput laboratory evolution of Escherichia coli under various stress conditions using an automated culture system, and quantified how the acquisition of resistance to one stressor alters the resistance to other stressors. We demonstrated that resistance changes could be quantitatively predicted based on changes in the transcriptome of the resistant strains. We also identified several genes and gene functions, for which mutations were commonly fixed in the strains resistant to the same stress, which could partially explain the observed cross-resistance and collateral sensitivity. The integration of transcriptome and genome data enabled us to clarify the bacterial stress resistance mechanisms.

Improvement of isopropanol tolerance of Escherichia coli using adaptive laboratory evolution and omics technologies.

Isopropanol (IPA) is the secondary alcohol that can be dehydrated to yield propylene. To produce IPA using microorganisms, a significant issue is that the toxicity of IPA causes retardation or inhibition of cell growth, decreasing the yield. One possible strategy to overcome this problem is to improve IPA tolerance of production organisms. For the understanding of tolerance to IPA, we performed parallel adaptive laboratory evolution (ALE) of Escherichia coli under IPA stress. To identify the genotypic change during ALE, we performed genome re-sequencing analyses of obtained tolerant strains. To verify which mutations were contributed to IPA tolerance, we constructed the mutant strains and quantify the IPA tolerance of the constructed mutants. From these analyses, we found that five mutations (relA, marC, proQ, yfgO, and rraA) provided the increase of IPA tolerance. To understand the phenotypic change during ALE, we performed transcriptome analysis of tolerant strains. From transcriptome analysis, we found that expression levels of genes related to biosynthetic pathways of amino acids, iron ion homeostasis, and energy metabolisms were changed in the tolerant strains. Results from these experiments provide fundamental bases for designing IPA tolerant strains for industrial purposes.

Roles for the Yb body components Armitage and Yb in primary piRNA biogenesis in Drosophila.

PIWI-interacting RNAs (piRNAs) protect genome integrity from transposons. In Drosophila ovarian somas, primary piRNAs are produced and loaded onto Piwi. Here, we describe roles for the cytoplasmic Yb body components Armitage and Yb in somatic primary piRNA biogenesis. Armitage binds to Piwi and is required for localizing Piwi into Yb bodies. Without Armitage or Yb, Piwi is freed from the piRNAs and does not enter the nucleus. Thus, piRNA loading is required for Piwi nuclear entry. We propose that a functional Piwi-piRNA complex is formed and inspected in Yb bodies before its nuclear entry to exert transposon silencing.

A regulatory circuit for piwi by the large Maf gene traffic jam in Drosophila.

PIWI-interacting RNAs (piRNAs) silence retrotransposons in Drosophila germ lines by associating with the PIWI proteins Argonaute 3 (AGO3), Aubergine (Aub) and Piwi. piRNAs in Drosophila are produced from intergenic repetitive genes and piRNA clusters by two systems: the primary processing pathway and the amplification loop. The amplification loop occurs in a Dicer-independent, PIWI-Slicer-dependent manner. However, primary piRNA processing remains elusive. Here we analysed piRNA processing in a Drosophila ovarian somatic cell line where Piwi, but not Aub or AGO3, is expressed; thus, only the primary piRNAs exist. In addition to flamenco, a Piwi-specific piRNA cluster, traffic jam (tj), a large Maf gene, was determined as a new piRNA cluster. piRNAs arising from tj correspond to the untranslated regions of tj messenger RNA and are sense-oriented. piRNA loading on to Piwi may occur in the cytoplasm. zucchini, a gene encoding a putative cytoplasmic nuclease, is required for tj-derived piRNA production. In tj and piwi mutant ovaries, somatic cells fail to intermingle with germ cells and Fasciclin III is overexpressed. Loss of tj abolishes Piwi expression in gonadal somatic cells. Thus, in gonadal somatic cells, tj gives rise simultaneously to two different molecules: the TJ protein, which activates Piwi expression, and piRNAs, which define the Piwi targets for silencing.