Genetic cargo and bacterial species set the rate of vesicle-mediated horizontal gene transfer

Abstract

Most bacteria release extracellular vesicles (EVs). Recent studies have found these vesicles are capable of gene delivery, however the consequences of vesicle-mediated transfer on the patterns and rates of gene flow within microbial communities remains unclear. Previous studies have not determined the impact of both the genetic cargo and the donor and recipient species on the rate of vesicle-mediated gene exchange. This report examines the potential for EVs as a mechanism of gene transfer within heterogeneous microbial populations. EVs were harvested from three species of Gram-negative microbes carrying different plasmids. The dynamics of gene transfer into recipient species was measured. This study demonstrates that vesicles enable gene exchange between five species of Gram-negative bacteria, and that the identity of the genetic cargo, donor strain, and recipient strain all influence gene transfer rates. Each species released and acquired vesicles containing genetic material to a variable degree, and the transfer rate did not correlate with the relatedness of the donor and recipient species. The results suggest that EVs may be a general mechanism to exchange non-specialized genetic cargo between bacterial species.

Figure 1

The impact of plasmid identity on the production and packaging of plasmid-containing vesicle. (A) Vesicle-mediated gene transfer. Donor cells load plasmid DNA into EV vesicles that can be acquired by a recipient cell. (B) Purification of EVs from liquid culture of bacterial cells through ultracentrifugation of cell-free supernatant. (C) 10% SDS-PAGE gel stained with Coomassie Blue showing concentration of outer membrane proteins, OmpA and OmpC/F, from EVs. (D) Distribution of EV diameters measured by dynamic light scattering. (E) Vesicle DNA content per pg of vesicle protein quantified by qPCR. P-value for all paired plasmid comparisons <0.001. Error bars signify standard deviation.

Figure 2

EVs facilitate HGT of multiple plasmids at a rate dependent on plasmid origin. (A) Purified EVs loaded with pLC291 were added to a recipient strain culture. Over time, the gain of antibiotic resistance was monitored by selective plating. The three curves show replicate experiments. The transfer time is the average time point at which resistant colonies were first observed. Inset shows recipient cells receiving either free plasmids (fp) or vesicles harvested from cells not containing the plasmid (p-) did not gain resistance by t = 60 h, whereas vesicle-mediated transfer (vm) occurred at 3.7 h. (B) Transfer time for EVs containing the plasmids pLC291, pUC19, and pZS2501. All pairs of transfer times have p-values < 0.05. (C) Gene transfer assay for three doses of vesicles containing pLC291. 1X corresponds to EV solution with 0.01 mg of characteristic outer-membrane protein (20% of total harvested). (D) Transfer rates for each plasmid were normalized by the number of plasmids per vesicle. Transfer rates have p-values < 0.001. Error bars signify standard deviation.

Figure 3

Variation of EVs and plasmid packaging is less dependent on donor species. (A) The production of EVs from A. veronii (Av), E. cloacae (Ecl), and E. coli (Eco), are measured using SDS-PAGE to quantify protein concentration. EV production is similar across species. (B) Packaging of pLC291 is confirmed using PCR amplification from purified vesicles. (C) Size distributions of EVs measured by dynamic light scattering shows similar diameters of vesicles harvested from all three species. (D) Vesicle DNA content quantified by qPCR was used to calculate plasmid copy number per pg of vesicle protein. P-values ≤ 0.0002. Error bars signify standard deviation.

Figure 4

Donor and recipient species influence rates of vesicle mediated gene transfer. (A) Relatedness of recipient species based on 16S rRNA sequence. Donor EVs packed with pLC291 from donor strains A. veronii (B), E. cloacae (C) and E. coli (D) were added to recipients strains A. veronii, C. violaceum, E. cloacae, E. coli, and P. aeruginosa. The time to plasmid transfer was measured by selective plating as in Fig. 2. (E) The transfer rate was not correlated with the relatedness of the donor and recipient strains, as measured by divergence of the 16S rRNA sequence. Error bars signify standard deviation from 4 replicates.