Shengchang Su

Investigating the predictability of essential genes across distantly related organisms using an integrative approach

Rapid and accurate identification of new essential genes in under-studied microorganisms will significantly improve our understanding of how a cell works and the ability to re-engineer microorganisms. However, predicting essential genes across distantly related organisms remains a challenge. Here, we present a machine learning-based integrative approach that reliably transfers essential gene annotations between distantly related bacteria. We focused on four bacterial species that have well-characterized essential genes, and tested the transferability between three pairs among them. For each pair, we trained our classifier to learn traits associated with essential genes in one organism, and applied it to make predictions in the other. The predictions were then evaluated by examining the agreements with the known essential genes in the target organism. Ten-fold cross-validation in the same organism yielded AUC scores between 0.86 and 0.93. Cross-organism predictions yielded AUC scores between 0.69 and 0.89. The transferability is likely affected by growth conditions, quality of the training data set and the evolutionary distance. We are thus the first to report that gene essentiality can be reliably predicted using features trained and tested in a distantly related organism. Our approach proves more robust and portable than existing approaches, significantly extending our ability to predict essential genes beyond orthologs.

BdlA, DipA and Induced Dispersion Contribute to Acute Virulence and Chronic Persistence of Pseudomonas aeruginosa

The human pathogen Pseudomonas aeruginosa is capable of causing both acute and chronic infections. Differences in virulence are attributable to the mode of growth: bacteria growing planktonically cause acute infections, while bacteria growing in matrix-enclosed aggregates known as biofilms are associated with chronic, persistent infections. While the contribution of the planktonic and biofilm modes of growth to virulence is now widely accepted, little is known about the role of dispersion in virulence, the active process by which biofilm bacteria switch back to the planktonic mode of growth. Here, we demonstrate that P. aeruginosa dispersed cells display a virulence phenotype distinct from those of planktonic and biofilm cells. While the highest activity of cytotoxic and degradative enzymes capable of breaking down polymeric matrix components was detected in supernatants of planktonic cells, the enzymatic activity of dispersed cell supernatants was similar to that of biofilm supernatants. Supernatants of non-dispersing ΔbdlA biofilms were characterized by a lack of many of the degradative activities. Expression of genes contributing to the virulence of P. aeruginosa was nearly 30-fold reduced in biofilm cells relative to planktonic cells. Gene expression analysis indicated dispersed cells, while dispersing from a biofilm and returning to the single cell lifestyle, to be distinct from both biofilm and planktonic cells, with virulence transcript levels being reduced up to 150-fold compared to planktonic cells. In contrast, virulence gene transcript levels were significantly increased in non-dispersing ΔbdlA and ΔdipA biofilms compared to wild-type planktonic cells. Despite this, bdlA and dipA inactivation, resulting in an inability to disperse in vitro, correlated with reduced pathogenicity and competitiveness in cross-phylum acute virulence models. In contrast, bdlA inactivation rendered P. aeruginosa more persistent upon chronic colonization of the murine lung, overall indicating that dispersion may contribute to both acute and chronic infections.

Prediction and Analysis of the Protein Interactome in Pseudomonas aeruginosa to Enable Network-Based Drug Target Selection

Pseudomonas aeruginosa (PA) is a ubiquitous opportunistic pathogen that is capable of causing highly problematic, chronic infections in cystic fibrosis and chronic obstructive pulmonary disease patients. With the increased prevalence of multi-drug resistant PA, the conventional “one gene, one drug, one disease” paradigm is losing effectiveness. Network pharmacology, on the other hand, may hold the promise of discovering new drug targets to treat a variety of PA infections. However, given the urgent need for novel drug target discovery, a PA protein-protein interaction (PPI) network of high accuracy and coverage, has not yet been constructed. In this study, we predicted a genome-scale PPI network of PA by integrating various genomic features of PA proteins/genes by a machine learning-based approach. A total of 54,107 interactions covering 4,181 proteins in PA were predicted. A high-confidence network combining predicted high-confidence interactions, a reference set and verified interactions that consist of 3,343 proteins and 19,416 potential interactions was further assembled and analyzed. The predicted interactome network from this study is the first large-scale PPI network in PA with significant coverage and high accuracy. Subsequent analysis, including validations based on existing small-scale PPI data and the network structure comparison with other model organisms, shows the validity of the predicted PPI network. Potential drug targets were identified and prioritized based on their essentiality and topological importance in the high-confidence network. Host-pathogen protein interactions between human and PA were further extracted and analyzed. In addition, case studies were performed on protein interactions regarding anti-sigma factor MucA, negative periplasmic alginate regulator MucB, and the transcriptional regulator RhlR. A web server to access the predicted PPI dataset is available at http://research.cchmc.org/PPIdatabase/.

Anaerobic Pseudomonas aeruginosa and other obligately anaerobic bacterial biofilms growing in the thick airway mucus of chronically infected cystic fibrosis patients: an emerging paradigm or “Old Hat”?

Introduction: The cystic fibrosis (CF) airway mucus is an ideal niche in which many bacteria can develop antibiotic- and phagocyte-resistance in unique structures known as “mode II biofilms” where bacteria are embedded within the mucus, yet unattached to airway epithelial cells. Pseudomonas aeruginosa is the dominant CF pathogen, yet herein the authors provide burgeoning evidence that obligate anaerobic bacteria (e.g., Prevotella) actually thrive within the CF mucus, a paradigmatic shift that chronic CF is an “aerobic” disease. Interestingly, CF organisms repress virulence factor production (e.g., P. aeruginosa) while others (e.g., S. aureus) increase them under anaerobic conditions. Areas covered: The authors shed additional light on (i) the anoxic nature of the CF airway mucus, (ii) the relative commonality of anaerobic bacteria isolated from CF sputum, (iii) virulence factor production and cross-talk between obligate anaerobes and P. aeruginosa relative to disease progression/remission, (iv) the role of mucoidy in CF, and (v) the role of nitrosative stress in activation of bacteriophage and pyocins within biofilms. Expert opinion: The authors conclude with insight as to how we might treat some CF bacteria during mode II biofilm infections that utilizes a metabolite of bacterial anaerobic respiration and an aerobic oxidation product of airway-generated NO, acidified NO2 -.

Epistatic Roles for Pseudomonas aeruginosa MutS and DinB (DNA Pol IV) in Coping with Reactive Oxygen Species-Induced DNA Damage

Pseudomonas aeruginosa is especially adept at colonizing the airways of individuals afflicted with the autosomal recessive disease cystic fibrosis (CF). CF patients suffer from chronic airway inflammation, which contributes to lung deterioration. Once established in the airways, P. aeruginosa continuously adapts to the changing environment, in part through acquisition of beneficial mutations via a process termed pathoadaptation. MutS and DinB are proposed to play opposing roles in P. aeruginosa pathoadaptation: MutS acts in replication-coupled mismatch repair, which acts to limit spontaneous mutations; in contrast, DinB (DNA polymerase IV) catalyzes error-prone bypass of DNA lesions, contributing to mutations. As part of an ongoing effort to understand mechanisms underlying P. aeruginosa pathoadaptation, we characterized hydrogen peroxide (H2O2)-induced phenotypes of isogenic P. aeruginosa strains bearing different combinations of mutS and dinB alleles. Our results demonstrate an unexpected epistatic relationship between mutS and dinB with respect to H2O2-induced cell killing involving error-prone repair and/or tolerance of oxidized DNA lesions. In striking contrast to these error-prone roles, both MutS and DinB played largely accurate roles in coping with DNA lesions induced by ultraviolet light, mitomycin C, or 4-nitroquinilone 1-oxide. Models discussing roles for MutS and DinB functionality in DNA damage-induced mutagenesis, particularly during CF airway colonization and subsequent P. aeruginosa pathoadaptation are discussed.

A Statistical Framework for Improving Genomic Annotations of Prokaryotic Essential Genes

Large-scale systematic analysis of gene essentiality is an important step closer toward unraveling the complex relationship between genotypes and phenotypes. Such analysis cannot be accomplished without unbiased and accurate annotations of essential genes. In current genomic databases, most of the essential gene annotations are derived from whole-genome transposon mutagenesis (TM), the most frequently used experimental approach for determining essential genes in microorganisms under defined conditions. However, there are substantial systematic biases associated with TM experiments. In this study, we developed a novel Poisson model–based statistical framework to simulate the TM insertion process and subsequently correct the experimental biases. We first quantitatively assessed the effects of major factors that potentially influence the accuracy of TM and subsequently incorporated relevant factors into the framework. Through iteratively optimizing parameters, we inferred the actual insertion events occurred and described each gene’s essentiality on probability measure. Evaluated by the definite mapping of essential gene profile in Escherichia coli, our model significantly improved the accuracy of original TM datasets, resulting in more accurate annotations of essential genes. Our method also showed encouraging results in improving subsaturation level TM datasets. To test our model’s broad applicability to other bacteria, we applied it to Pseudomonas aeruginosa PAO1 and Francisella tularensis novicida TM datasets. We validated our predictions by literature as well as allelic exchange experiments in PAO1. Our model was correct on six of the seven tested genes. Remarkably, among all three cases that our predictions contradicted the TM assignments, experimental validations supported our predictions. In summary, our method will be a promising tool in improving genomic annotations of essential genes and enabling large-scale explorations of gene essentiality. Our contribution is timely considering the rapidly increasing essential gene sets. A Webserver has been set up to provide convenient access to this tool. All results and source codes are available for download upon publication at http://research.cchmc.org/essentialgene/.

A Putative ABC Transporter Permease Is Necessary for Resistance to Acidified Nitrite and EDTA in Pseudomonas aeruginosa under Aerobic and Anaerobic Planktonic and Biofilm Conditions

Pseudomonas aeruginosa (PA) is an important airway pathogen of cystic fibrosis and chronic obstructive disease patients. Multiply drug resistant PA is becoming increasing prevalent and new strategies are needed to combat such insidious organisms. We have previously shown that a mucoid, mucA22 mutant PA is exquisitely sensitive to acidified nitrite (A-NO2−, pH 6.5) at concentrations that are well tolerated in humans. Here, we used a transposon mutagenesis approach to identify PA mutants that are hypersensitive to A-NO2−. Among greater than 10,000 mutants screened, we focused on PA4455, in which the transposon was found to disrupt the production of a putative cytoplasmic membrane-spanning ABC transporter permease. The PA4455 mutant was not only highly sensitive to A-NO2−, but also the membrane perturbing agent, EDTA and the antibiotics doxycycline, tigecycline, colistin, and chloramphenicol, respectively. Treatment of bacteria with A-NO2− plus EDTA, however, had the most dramatic and synergistic effect, with virtually all bacteria killed by 10 mM A-NO2−, and EDTA (1 mM, aerobic, anaerobic). Most importantly, the PA4455 mutant was also sensitive to A-NO2− in biofilms. A-NO2− sensitivity and an anaerobic growth defect was also noted in two mutants (rmlC and wbpM) that are defective in B-band LPS synthesis, potentially indicating a membrane defect in the PA4455 mutant. Finally, this study describes a gene, PA4455, that when mutated, allows for dramatic sensitivity to the potential therapeutic agent, A-NO2− as well as EDTA. Furthermore, the synergy between the two compounds could offer future benefits against antibiotic resistant PA strains.

Construction and characterization of stable, constitutively expressed, chromosomal green and red fluorescent transcriptional fusions in the select agents, Bacillus anthracis, Yersinia pestis, Burkholderia mallei, and Burkholderia pseudomallei

Here, we constructed stable, chromosomal, constitutively expressed, green and red fluorescent protein (GFP and RFP) as reporters in the select agents, Bacillus anthracis, Yersinia pestis, Burkholderia mallei, and Burkholderia pseudomallei. Using bioinformatic approaches and other experimental analyses, we identified P0253 and P1 as potent promoters that drive the optimal expression of fluorescent reporters in single copy in B. anthracis and Burkholderia spp. as well as their surrogate strains, respectively. In comparison, Y. pestis and its surrogate strain need two chromosomal copies of cysZK promoter (P2cysZK) for optimal fluorescence. The P0253‐, P2cysZK‐, and P1‐driven GFP and RFP fusions were first cloned into the vectors pRP1028, pUC18R6KT‐mini‐Tn7T‐Km, pmini‐Tn7‐gat, or their derivatives. The resultant constructs were delivered into the respective surrogates and subsequently into the select agent strains. The chromosomal GFP‐ and RFP‐tagged strains exhibited bright fluorescence at an exposure time of less than 200 msec and displayed the same virulence traits as their wild‐type parental strains. The utility of the tagged strains was proven by the macrophage infection assays and lactate dehydrogenase release analysis. Such strains will be extremely useful in high‐throughput screens for novel compounds that could either kill these organisms, or interfere with critical virulence processes in these important bioweapon agents and during infection of alveolar macrophages.

Characterization of stable, constitutively expressed, chromosomal green and red fluorescent transcriptional fusions in the select agent bacterium, Francisella tularensis Schu S4 and the surrogate type B live vaccine strain (LVS)

Here, we constructed stable, constitutively expressed, chromosomal green (GFP) and red fluorescent (RFP) reporters in the genome of the surrogate strain, Francisella tularensis spp. holarctica LVS (herein LVS), and the select agent, F. tularensis Schu S4. A bioinformatic approach was used to identify constitutively expressed genes. Two promoter regions upstream of the FTT1794 and rpsF(FTT1062) genes were selected and fused with GFP and RFP reporter genes in pMP815, respectively. While the LVS strains with chromosomally integrated reporter fusions exhibited fluorescence, we were unable to deliver the same fusions into Schu S4. Neither a temperature-sensitive Francisella replicon nor a pBBR replicon in the modified pMP815 derivatives facilitated integration. However, a mini-Tn7 integration system was successful at integrating the reporter fusions into the Schu S4 genome. Finally, fluorescent F. tularensis LVS and a mutant lacking MglA were assessed for growth in monocyte-derived macrophages (MDMs). As expected, when compared to wild-type bacteria, replication of an mglA mutant was significantly diminished, and the overall level of fluorescence dramatically decreased with infection time. The utility of the fluorescent Schu S4 strain was also examined within infected MDMs treated with clarithromycin and enrofloxacin. Taken together, this study describes the development of an important reagent for F. tularensis research, especially since the likelihood of engineered antibiotic resistant strains will emerge with time. Such strains will be extremely useful in high-throughput screens for novel compounds that could interfere with critical virulence processes in this important bioweapons agent and during infection of alveolar macrophages.