Competence beyond Genes: Filling in the Details of the Pneumococcal Competence Transcriptome by a Systems Approach

DNA uptake by natural competence is a central process underlying the genetic plasticity, biology, and virulence of the human respiratory opportunistic pathogen Streptococcus pneumoniae. A study reported in this issue (J.

ABSTRACT

DNA uptake by natural competence is a central process underlying the genetic plasticity, biology, and virulence of the human respiratory opportunistic pathogen Streptococcus pneumoniae. A study reported in this issue (J. Slager, R. Aprianto, and J.-W. Veening, J. Bacteriol. 201:e00780-18, https://doi.org/10.1128/JB.00780-18) combined deep-genome annotation and high-resolution transcriptome analyses to considerably extend the previous model of temporal regulation of competence at the operon and component gene levels. That extended study also provides a playbook for updating, refining, and extending genomic data sets and making them publicly available.

COMMENTARY

The ability of Streptococcus pneumoniae (the pneumococcus) to take up and incorporate homologous DNA into its genome by natural competence was first reported in landmark papers by Griffith (1) and by Avery et al. (2) and has been a central topic of study and a powerful tool for genetic manipulation ever since (3, 4). The regulation of pneumococcal competence is mediated by a secreted competence-stimulatory peptide (CSP), which has made competence a model for quorum sensing (5–7) and fratricide, whereby competent cells kill noncompetent cells, releasing their DNA and nutrients (8). The induction of competence depends on environmental conditions, such as pH and phosphate concentration (3), and can be induced by antibiotics and blocks to DNA replication, suggesting a role as a general stress response in a bacterium that lacks a LexA-mediated SOS response (9). Natural competence underlies the genetic plasticity and horizontal gene transfer that have resulted in the great genomic diversity of different S. pneumoniae isolates (10, 11) and has led to the development of antibiotic resistance through the assembly of mosaic genes encoding essential penicillin-binding proteins (12). Moreover, competence is required for optimal colonization of and virulence in its human host (3, 13).

Studies of the regulation of competence have centered on the mechanisms of competence induction and turnoff (4, 14–17) and on identifying the genes that encode the extraordinary molecular machines that carry out DNA uptake into cells, transportation across the cell membrane of this Gram-positive bacterium, and recombination into its chromosome (18). A number of years ago, several classical studies used microarray approaches to identify the temporal patterns of gene expression following induction of competence by addition of CSP to cultures (14–16). Results from these transcriptomic experiments were integrated with extensive genetic and biochemical characterization of the proteins that regulate competence to provide an outline of the competence regulatory cascade and to identify missing or unknown components (4). A report of a new transcriptome study in this issue by Slager et al. (3) refines, updates, and considerably extends the previous model of pneumococcal transcription regulation during the different stages of competence. Notably, besides being based on newer, more sensitive transcriptome sequencing (RNA-seq) approaches, that paper builds on a systems approach developed in two previous papers (19, 20).

In one paper, those authors provide a deep-genome annotation of S. pneumoniae serotype 2 strain D39 (20), whose genome had not been reannotated for over a decade since its sequence was first reported (21). This deep-genome annotation involved reprising the genome sequence using current short- and long-read DNA sequencing methods to verify the previous sequence, identify sites of DNA methylation, and provide information about invertible and repeated DNA sequences that play roles in S. pneumoniae antigenic variation (20). The resulting genome sequence was then fed into the latest automated and manual pipelines to identify and annotate gene functions throughout the genome. In addition, paired-end and “cappable” RNA-seq approaches and predictions of factor-independent transcription terminators were compiled to indicate the precise boundaries of single-gene and multigene operons and noncoding small RNAs (sRNAs) throughout the D39 genome. These analyses revealed a wealth of genomic features, including new sRNAs, whose number has since been expanded using a different bioinformatic approach (22); the presence of regulatory riboswitches in the 5′ untranslated regions (5′-UTRs) of numerous operons but the absence of 5′-UTRs upstream of others; a strong bias toward adenine (A) at transcription start sites; and the presence of likely regulatory motifs in most promoter regions. Importantly, this gold mine of information was assembled and made available to other researchers in an easy-to-use, intuitive website named PneumoBrowse (20).

The second paper, by Aprianto et al. (19), is also linked to the paper on competence in this issue (3). In that earlier paper, the authors present a comprehensive, high-resolution RNA-seq transcriptome analysis of S. pneumoniae D39 grown in culture under 22 different medium conditions that approximate sugar sources and other compounds and temperatures relevant to pneumococcal colonization, infections, and interactions with human host epithelial cells (19). As part of that study, the authors included strain D39 cells grown in a standard laboratory medium (C+Y, pH 6.8) to which CSP was added for 3, 10, or 20 min to induce competence. The data, which form the basis for the transcript time courses reported in reference 3, were also incorporated into a weighted gene coexpression network analysis (WGCNA) that included the transcriptome data set for growth under all 22 conditions (19). The WGCNA provides a powerful tool to interrogate which genes are coregulated under all conditions and which genes display partial regulation, indicative of the presence of two or more regulatory mechanisms. Again, the authors developed an effective public interface, called PneumoExpress, for retrieval of relative transcript expression levels for the 22 growth conditions and to provide coexpression patterns of query genes (19).

The current paper (3) uses the deep-genome annotation data (20) and the high-resolution transcriptome data (19) to provide the most current, comprehensive summary of the transcription responses that occur 3 min, 10 min, or 20 min after induction of pneumococcal competence. The response entails a major alteration of overall transcription compared to exponential growth. The paper provides detailed, new information about the time-resolved expression of genes in the early-competence response mediated by the phosphorylated ComE (ComE∼P) response regulator and the late-competence response mediated by the ComX alternative sigma factor. The systems approach makes it possible to understand these responses at the operon and sRNA levels (3) instead of at the single-gene level of earlier studies (14–16). In addition, RNA-seq provides normalized transcript amounts (in units of transcripts per millions [TPM] of nucleotides), which allows comparisons of relative expression levels of different operons affected by competence. The coexpression matrix identifies new genes regulated during development of competence, such as briC, important for biofilm development and virulence (23), and helps to deconvolute transcription patterns that result from coregulation by ComE∼P and other regulators, such as the bacteriocin regulon controlled by the BlpR response regulator (3). Knowledge of operon structures distinguishes secondary regulation, caused by internal promoters and incomplete transcription termination within gene clusters, from direct regulation by transcription regulators, whose consensus binding sites are also refined (3). Finally, that paper provides additional support for late transcription of the gene encoding the housekeeping RpoD sigma A factor, which plays a role in turning off the elaborate competence regulatory response after a limited time (3, 24).

Besides the canonical early-competence and late-competence regulons, the study reported in reference 3 provides comparable integrated results for operon transcription, sRNA involvement, and regulator binding sites for delayed competence regulons controlled by the CiaR, VraR (LiaR), HrcA, and CtsR regulators (3, 15, 25–27). Because competence represents such a massive, concerted response, S. pneumoniae cells must deal with multiple cellular stresses at several levels. Induction of the two-component systems mediated by the CiaR and VraR (LiaR) response regulators, for example, is required to reduce peptidoglycan and cell wall stresses caused by assembly of the competence uptake apparatus and by DNA uptake (2, 25, 26). The large load of proteins induced by competence also requires induction of the HrcA and CtsR regulons, which encode protein refolding chaperones and regulated proteases (3). These delayed responses support the previous conclusion that competence induction in S. pneumoniae plays the dual roles of DNA uptake in response to a quorum sensing peptide (5, 6) and coping with additional types of stress responses (9). The new blueprint of the extended competence response emphasizes that the functions and regulation of many genes and operons remain to be determined in D39 serotype 2 and in strains of other serotypes of S. pneumoniae.

The views expressed in this article do not necessarily reflect the views of the journal or of ASM.