Horizontally Acquired Homologs of Xenogeneic Silencers: Modulators of Gene Expression Encoded by Plasmids, Phages and Genomic Islands

Abstract

Acquisition of mobile elements by horizontal gene transfer can play a major role in bacterial adaptation and genome evolution by providing traits that contribute to bacterial fitness. However, gaining foreign DNA can also impose significant fitness costs to the host bacteria and can even produce detrimental effects. The efficiency of horizontal acquisition of DNA is thought to be improved by the activity of xenogeneic silencers. These molecules are a functionally related group of proteins that possess affinity for the acquired DNA. Binding of xenogeneic silencers suppresses the otherwise uncontrolled expression of genes from the newly acquired nucleic acid, facilitating their integration to the bacterial regulatory networks. Even when the genes encoding for xenogeneic silencers are part of the core genome, homologs encoded by horizontally acquired elements have also been identified and studied. In this article, we discuss the current knowledge about horizontally acquired xenogeneic silencer homologs, focusing on those encoded by genomic islands, highlighting their distribution and the major traits that allow these proteins to become part of the host regulatory networks.

Keywords: H-NS; Lsr2; MvaT; Rok; genomic islands; horizontal gene transfer; transcriptional network; xenogeneic silencers.

Figure 1

Multiple sequence alignment of the H-NS protein from Salmonella ser. Typhimurium LT2 and selected full-length (A) and short (B) homologs found in EARL genomic islands. The secondary structure elements of H-NS are represented with red cylinders (α-helices) and green arrows (β-strands). The blue, yellow, and green backgrounds represent the dimerization/oligomerization, linker, and DNA-binding domains, respectively. Numbers indicate the corresponding residue of the H-NS protein. The triangles under the alignment indicate the amino acid residues required for function of H-NS (A) or H-NST (B) that are also conserved in the homologs encoded by the EARL GIs. Aligned proteins correspond to H-NS from Salmonella ser. Typhimurium LT2 (STmLT2; WP_001287383.1), full-length homologs encoded in the EARL islands from Salmonella ser. Enteritidis P125109 (ROD21; WP_001287371.1), Klebsiella pneumoniae ST258 NJST258-2 (ICEKp258.2; WP_001588074.1), Salmonella ser. Typhi PM016/13 (StyPM01613_GI; WP_045354302.1), Escherichia coli SF-088 (EcoSF088_GI; WP_000005143.1), Serratia marcescens SM39 (SmaSM39_GI; WP_041035854.1), Pectobacterium atrosepticum SCRI1043 (HAI13; WP_011094424.1), and the short homologs encoded in the EARL islands from E. coli O127:H6 E2348/69 (EcoE2348_IE3; WP_000564595.1), K. michiganensis M1(KmiM1_GI; WP_038424693.1), Samonella ser. Sloterdijk ATCC 15791 (SsloATCC_GI; WP_023201852.1), S. marcescens UMH3 (SmaUMH3_GI; WP_089187391.1), Yersinia intermedia Y228 (YinY228_GI; WP_042569548.1), and P. parmentieri RNS08.42.1A (PpaRNS_GI; WP_033071994.1). The multiple alignment was made in MEGA X [116] using MUSCLE, and the graphic representation was made using ESPript3 [117]. The complete alignment of all 34 full-length and 20 short homologs found in EARL islands is provided as Supplementary Figures S1 and S2.

Figure 2

The EARL genomic islands possess a significantly low GC content. (A) Violin plots of the GC content of the genomic islands found in Enterobacteriaceae in the Islander database (GIs), their host chromosomes (Chroms. with GIs), the EARL genomic islands (EARL GIs) and the chromosomes that harbor the EARL GIs (Chroms. with EARL GIs). Kruskall–Wallis tests followed by Dunn’s multiple comparisons were used to assess differences between medians (α = 0.05, **** p < 0.0001). (B) Comparison of the GC content of genomic islands and the host chromosome. The dotted line represents a 1:1 correspondence. Data was obtained from the Islander database (15/10/19) [123] and the supplementary information from Piña-Iturbe et al. (2018) [50]. The 597 Enterobacteriaceae genomic islands stored in Islander were manually filtered to eliminate possible false positives (indicated in the database), putative prophages (>20% overlap with a PHAST call), duplicated genomic islands (islands with length ≥300 pb in the same species, the same integration site, and GC difference <1%), and islands found in plasmids, resulting in 244 genomic islands present in 103 host chromosomes. All the 56 EARL islands identified in [50] were used.

Figure 3

Xenogeneic silencer homologs encoded by genomic islands. The entire Islander database (4065 genomic islands) was interrogated for the presence of xenogeneic silencer homologs using tBLASTn (E-value cutoff = 10) using the amino acid sequences of H-NS (WP_001287383.1), MvaT (WP_003093888.1), Lsr2 (WP_003419513.1), and Rok (WP_003232378.1). The genomic islands corresponding to the BLAST hits were manually filtered to exclude possible false positives, putative prophages, duplicated islands, and islands in plasmids using the same criteria as in Figure 2. Then, the individual hits were manually examined to exclude those corresponding to pseudogenes, alignments outside coding sequences, or alignments in a reading frame different from the annotated coding sequence.