To be an effective countermeasure against crown gall disease, our CRISPR system targeting the Ti-plasmid must be capable of spreading efficiently in an Agrobacterium tumefaciens bacterial population. We investigated if conjugation, a natural bacterial process, could be potentially used to mobilize our plasmid carrying CRISPR/Cas9 between Agrobacterium cells.
Bacterial conjugation is the process by which a plasmid is spread from a donor cell directly into a recipient cell (Llosa et al. 2002). This DNA transfer occurs through a pilus-like structure on the exterior surface of the donor cell. This pilus inserts itself into the recipient when the two cells come into contact, forming a direct conduit. Conjugative transfer requires the donor cell to encode and express the conjugative machinery, including the pilus. Often, this machinery is carried on the plasmid it interacts with, allowing the plasmid to spread itself through a population of cells.
The Ti-plasmid itself uses conjugation to spread through the A. tumefaciens population in soil environments and inside crown gall tumour tissue. We considered exploiting this conjugative machinery for delivering our CRISPR system, but this idea was abandoned because the Ti-plasmid’s conjugation genes are strictly regulated. This regulation includes sensory systems that recognize small molecules produced by wounded plants and quorum-sensing systems that respond to high densities of Ti-plasmid in the cell population (Lang and Faure 2014, Subramoni et al. 2014, White and Winans 2007). Consequently, the Ti-plasmid’s conjugative machinery is adapted to conditions where plants are extremely susceptible to infection, conditions that are not ideal for an intervention designed to destroy Ti-plasmids before they can begin the disease process.
We identified an alternate conjugative plasmid called RP1 with several merits for our particular application. It belongs to a broad host range family of plasmids recognized for their capacity to readily propagate antibiotic resistance genes through diverse communities of bacteria. Their conjugation machinery is encoded on the plasmid, allowing it to mobilize itself from a variety of hosts, and this machinery is also not under stringent conditional regulation (Zatyka and Thomas 1998). Crucially, we found examples in the literature of RP1 and related plasmids conjugating between and within species of Agrobacterium (Quandt et al. 2004), making it an attractive candidate to test as a vehicle for mobilizing our Ti-plasmid-targeting CRISPR system.
We performed conjugative mating assays to characterize RP1’s capacity for conjugative transfer into Ti plasmid-containing A. tumefaciens recipient cells. We also characterized RP1’s ability to conjugate from A. tumefaciens donors to other A. tumefaciens recipients. We demonstrated that Ti plasmid-containing A. tumefaciens can readily receive RP1 by conjugation from E. coli donors, but failed to observe RP1 conjugation into an A. tumefaciens recipient from an A. tumefaciens donor. Future experiments will demonstrate if A. tumefaciens donors require longer conjugation incubations to fully express conjugation machinery due to their slower growth.
RP1 conjugation from E. coli to A. tumefaciens
To investigate the ability of Agrobacterium to act as a RP1 plasmid recipient, we transformed RP1 into E. coli host and performed a conjugative mating reaction with Agrobacterium tumefaciens strain GV3101. In this assay, a donor culture of E. coli containing RP1 was grown in LB supplemented with kanamycin (75 μg/ml), and a recipient culture of A. tumefaciens strain GV3101 was grown in LB supplemented with gentamicin (100 μg/ml). The cultures were harvested at OD 0.250, mixed together at various ratios, concentrated by centrifugation to facilitate cell-to-cell contact, and incubated to allow for conjugative transfer of RP1. Transconjugants were selectively grown by plating the mating reactions on LB agar plates supplemented with both kanamycin (75 μg/ml) and gentamicin (75 μg/ml). This allowed only for growth of A. tumefaciens cells where the kanamycin selection marker on RP1 had moved into recipient A. tumefaciens cells, which carry a gentamycin marker. To control for the possibility of false-positives from spontaneous kanamycin and gentamicin mutations, donor and recipient cultures were separately concentrated by centrifugation, incubated, and plated on selective medium to ensure the absence of growth.
RP1 conjugation between A. tumefaciens
To investigate the ability of RP1 to conjugate between Agrobacterium cells, we designed another conjugation similar to the one above. We grew a donor culture of A. tumefaciens strain GV3101 carrying RP1 plasmid (obtained from the mating assay above) in LB supplemented with kanamycin (75 μg/ml), and a recipient culture of A. tumefaciensstrain LBA4404 in LB supplemented with streptomycin (2000 μg/ml). Cultures were harvested at OD 0.250 and concentrated to OD 2.50 by centrifugation, mixed together in equal proportion, and incubated for 24 hours. Following the incubation, a ten-fold dilution series was prepared from the mating reaction, and aliquots of each dilution were plated on LB agar supplemented with kanamycin (75 μg/ml) and streptomycin (2000 μg/ml) for selective growth and CFU enumeration of transconjugants. In parallel, aliquots of the mating reaction dilutions were plated on LB agar supplemented with gentamicin (100 μg/ml) and streptomycin (2000 μg/ml) for selection and CFU enumeration of transconjugants arising from background Ti plasmid transfer. Control reactions of donor and recipient cultures were plated on the same conditions to ensure the absence of spontaneous kanamycin and streptomycin resistant mutants.
Conjugation of RP1 into A. tumefaciens GV3101 from E. coli
We obtained plasmid RP1 inside an E. coli host and performed a conjugative mating to move it into A. tumefaciens strain GV3101, which contains a Ti plasmid. Transconjugants were selected for by growth on kanamycin-gentamycin medium since a kanamycin selection marker is present on RP1 and a gentamicin selection marker is present in the recipient.
Different donor:recipient culture proportions and incubation durations were assayed to observe any qualitative differences in these conjugation conditions. The mating reactions were plated on dual-selective medium to select for transconjugants, and lawns of growth were observed for all conditions (Table 1). Furthermore, donor and recipient control reactions were plated on the same dual-selective medium, and no growth was observed, indicating the gentamycin resistant colonies from the mating reactions were true-positive transconjugants and not spontaneous resistant contaminants in the donor or recipient cultures (Table 1). Overall, these results demonstrate that A. tumefaciens GV3101 is a ready recipient of plasmid RP1 via conjugation, and that hosting a Ti plasmid does not prevent RP1 receipt.
Conjugation of RP1 between A. tumefaciens cells
One of the transconjugants from the previous assay was isolated as A. tumefaciens GV3101 + RP1. We next attempted to perform a conjugative transfer of RP1 from this strain into a different strain of A. tumefaciens called LBA4404. We modified the above assay by replacing gentamicin with streptomycin as the recipient selection marker, since LBA4404 has a streptomycin selection marker on its Ti plasmid. This introduced the possibility of kanamycin-streptomycin resistant colonies arising through background conjugation of the Ti plasmid instead of RP1, albeit at low frequency outside of inducing conditions specific for the Ti plasmid. To quantify the amount of any background Ti plasmid conjugation we plated the mating reactions in parallel on gentamicin-streptomycin to select for Ti plasmid transconjugants.
When these mating reactions were plated on kanamycin-streptomycin medium, we observed growth on the dual-selective medium. We did not initially consider these colonies to be transconjugants, however, because we also observed growth of the donor control reaction on the kanamycin-streptomycin and gentamicin-streptomycin medium (Table 2) This suggested that the donor culture was contaminated with spontaneous streptomycin-resistant cells, which were carried through the mating reaction and growing on the dual-selective plates.
We compared growth of the donor control reaction on dual-selective medium to growth on a positive control plate containing non-selective medium. The dual-selective plates had only isolated colonies whereas the positive growth control plate had a confluent lawn, despite the dual-selective plates being plated with 3.25-fold more volume of the donor control reaction. This indicated that only a sub-population of the donor culture was streptomycin-resistant.
We counted isolated colonies from the donor control reaction on both dual-selective conditions, and we counted isolated colonies from the mating reaction on both dual-selective conditions. Using these colony counts, we calculated the CFU/µl of kanamycin-streptomycin and gentamicin-streptomycin resistant cells in both the donor control reaction and mating reaction (Table 3). We observed comparable densities of cells with these phenotypes in both reactions, indicating that dual-selective colonies from the mating we likely spontaneous streptomycin-resistant mutants from the donor culture. Consequently, this assay provided no evidence of either RP1 or Ti plasmid conjugation from A. tumefaciens strain GV3101 to strain LBA4404.
Isolation of a streptomycin-sensitive A. tumefaciens strain GV3101 + RP1 host
Having shown that our RP1 donor culture contained a sub-population of streptomycin-resistant cells, we decided to isolate donor cells with a streptomycin-sensitive phenotype for future conjugation assays. This was accomplished by streaking our GV3101 + RP1 host for isolated colonies, and picking these colonies in parallel onto medium with and without streptomycin.Thirty-two colonies were picked, and we identified three colonies with the appropriate streptomycin-sensitive phenotype (Figure 2).
We subsequently inoculated these colonies into liquid medium with and without streptomycin to confirm their phenotype, and in all cases, there was no growth in the streptomycin-supplemented medium (Figure 2 and Table 4). This indicates we successfully isolated a streptomycin-sensitive RP1 donor for future conjugation assays into an A. tumefaciens strain LBA4404 recipient.
With these results, we were able to begin characterizing RP1 as a vehicle for mobilizing our Ti plasmid-targeting CRISPR system in a population of A. tumefaciens. We demonstrated that Ti plasmid-containing strains of A. tumefaciens can receive RP1 via conjugation from E.coli, express its kanamycin selection marker, and maintain this plasmid. These features support its use as a mobilizing vehicle for the Ti plasmid-targeting CRISPR system. We were not able to show RP1 moving from A. tumefaciens into other cells, however. If the conjugation had occurred with high frequency in our assay, we should have observed some kanamycin-streptomycin transconjugants after accounting for the spontaneous streptomycin-resistant mutants in the donor culture. This suggested to us that either A. tumefaciens was incapable of donating RP1, contrary to the literature (Quandt et al. 2014), or that our conjugation reaction conditions were not appropriate. A. tumefaciens capacity to receive RP1 readily, express genes on the plasmid, and maintain it in culture indicate that it is a suitable RP1, as indicated in the literature. Instead, we speculate that our conjugation incubation of 4 hours may have been insufficient to allow a significant number of conjugation events. We chose this duration because our mating with an E. coli donor gave confluent transconjugant growth after a 6 hour incubation. In hindsight, E. coli much faster growth rate likely allows for the quicker expression of RP1’s conjugation machinery and more efficient transfer from this host. Now, having isolated a streptomycin-sensitive A. tumefaciens RP1 host, future conjugation assays should be conducted with longer incubations.
Adamczyk M, Jagura-Burdzy G. Spread and survival of promiscuous IncP-1 plasmids. Acta Biochim Pol. 2003;50(2):425-53.
Cook DM, Farrand SK. The oriT region of the Agrobacterium tumefaciens Ti plasmid pTiC58 shares DNA sequence identity with the transfer origins of RSF1010 and RK2/RP4 and with T-region borders. J Bacteriol. 1992 Oct;174(19):6238-46.
Lang J, Faure D. Functions and regulation of quorum-sensing in Agrobacterium tumefaciens. Front Plant Sci. 2014 Jan 31;5:14. doi: 10.3389/fpls.2014.00014. eCollection 2014.
Llosa M, Gomis-Rüth FX, Coll M, de la Cruz Fd F. Bacterial conjugation: a two-step mechanism for DNA transport. Mol Microbiol. 2002 Jul;45(1):1-8.
Quandt J, Clark RG, Venter AP, Clark SR, Twelker S, Hynes MF. Modified RP4 and Tn5-Mob derivatives for facilitated manipulation of large plasmids in Gram-negative bacteria. Plasmid. 2004 Jul;52(1):1-12.
Subramoni S, Nathoo N, Klimov E, Yuan ZC. Agrobacterium tumefaciens responses to plant-derived signaling molecules. Front Plant Sci. 2014 Jul 8;5:322. doi: 10.3389/fpls.2014.00322. eCollection 2014.
White CE, Winans SC. Cell-cell communication in the plant pathogen Agrobacterium tumefaciens. Philos Trans R Soc Lond B Biol Sci. 2007 Jul 29;362(1483):1135-48.
Zatyka M, Thomas C. Control of genes for conjugative transfer of plasmids and other mobile elements. FEMS Microbiol Lett, Feb 1998;21(4): 291–319. Doi: 10.1111/j.1574-6976.1998.tb00355