This is an adapted method for identifying candidate insect colonization factors in a Burkholderia beneficial symbiont. The beetle host is infected with a random mutant library generated via transposon mutagenesis, and library complexity after colonization is compared to a control grown in vitro.
Inferring the function of genes by manipulating their activity is an essential tool for understanding the genetic underpinnings of most biological processes. Advances in molecular microbiology have seen the emergence of diverse mutagenesis techniques for the manipulation of genes. Among them, transposon-insertion sequencing (Tn-seq) is a valuable tool to simultaneously assess the functionality of many candidate genes in an untargeted way. The technique has been key to identify molecular mechanisms for the colonization of eukaryotic hosts in several pathogenic microbes and a few beneficial symbionts.
Here, Tn-seq is established as a method to identify colonization factors in a mutualistic Burkholderia gladioli symbiont of the beetle Lagria villosa. By conjugation, Tn5 transposon-mediated insertion of an antibiotic-resistance cassette is carried out at random genomic locations in B. gladioli. To identify the effect of gene disruptions on the ability of the bacteria to colonize the beetle host, the generated B. gladioli transposon-mutant library is inoculated on the beetle eggs, while a control is grown in vitro in a liquid culture medium. After allowing sufficient time for colonization, DNA is extracted from the in vivo and in vitro grown libraries. Following a DNA library preparation protocol, the DNA samples are prepared for transposon-insertion sequencing. DNA fragments that contain the transposon-insert edge and flanking bacterial DNA are selected, and the mutation sites are determined by sequencing away from the transposon-insert edge. Finally, by analyzing and comparing the frequencies of each mutant between the in vivo and in vitro libraries, the importance of specific symbiont genes during beetle colonization can be predicted.
Burkholderia gladioli can engage in a symbiotic association with Lagria villosa beetles, playing an important role in defense against microbial antagonists of the insect host4,5,6. Female beetles house several strains of B. gladioli in specialized glands accessory to the reproductive system. Upon egg-laying, females smear B. gladioli cells on the egg surface where antimicrobial compounds produced by B. gladioli inhibit infections by entomopathogenic fungi4,6. During late embryonic development or early after the larvae hatch, the bacteria colonize cuticular invaginations on the dorsal surface of the larvae. Despite this specialized localization and vertical transmission route of the symbionts, L. villosa can presumably also acquire B. gladioli horizontally from the environment4. Furthermore, at least three strains of B. gladioli have been found in association with L. villosa4,6. Among these, B. gladioli Lv-StA is the only one that is amenable to cultivation in vitro.
B. gladioli Lv-StA has a genome size of 8.56 Mb6 and contains 7,468 genes. Which of these genes are important for B. gladioli bacteria to colonize the beetle host? To answer this question, we used transposon-insertion sequencing (Tn-seq), an explorative method to identify conditionally essential microbial genes1,2,3. A mutant library of B. gladioli Lv-StA was created using a Tn5 transposon. Through conjugation from Escherichia coli donor cells to B. gladioli Lv-StA, a pRL27 plasmid carrying the Tn5 transposon and an antibiotic resistance cassette flanked by inverted repeats was transferred (Figure 1). Thereby, a set of mutants that individually carry disruptions of 3,736 symbiont genes was generated (Figure 2).
The mutant pool was infected onto beetle eggs to identify the colonization factors and, as a control, was also grown in vitro in King's B (KB) medium. After allowing sufficient time for colonization, hatched larvae were collected and pooled for DNA extraction. Fragments of DNA containing the transposon insert and the flanking genomic region of B. gladioli Lv-StA were selected using a modified DNA library preparation protocol for sequencing. Read quality processing followed by analysis with DESeq2 was carried out to identify specific genes crucial for B. gladioli Lv-StA to colonize L. villosa larvae when transmitted via the egg surface.
1. Media and buffer preparation
2. Conjugation to generate the transposon mutant library
Figure 1: Conjugation protocol steps. The conjugation recipient Burkholderia gladioli Lv-StA (red) and donor Escherichia coli containing the pRL27 plasmid (pink) are grown in KB agar and LB, respectively, supplemented with kanamycin and DAP. After conjugative transfer of the plasmid for 12-18 h at 30 °C, the transconjugant B. gladioli cells are selected on KB containing kanamycin and pooled together. Abbreviations: DAP = 2,6-diaminopimelic acid; Kan = kanamycin. Please click here to view a larger version of this figure.
3. PCR and gel electrophoresis to confirm successful insertions in B. gladioli Lv-StA
4. Mutant pool infection on beetle eggs
5. Infected beetles and in vitro mutant library DNA extraction
NOTE: DNA extractions were performed using a DNA and RNA purification kit according to the manufacturer's protocol briefly outlined below.
6. Sequencing library preparation
NOTE: The protocol and reagents for DNA library preparation are adapted and modified from the instructions provided by the manufacturer of the DNA library preparation kit.
Figure 2: Schematic of the DNA library preparation steps. After shearing and adapter ligation, the modified protocol includes a streptavidin bead-selection step to enrich DNA fragments containing the insertion cassette. Please click here to view a larger version of this figure.
7. Sequencing and analysis
Host-associated bacteria can employ several factors to establish an association, including those mediating adhesion, motility, chemotaxis, stress responses, or specific transporters. While factors important for pathogen-host interactions have been reported for several bacteria13,14,15,16,17,18, including members of the genus Burkholderia19,20, fewer studies have explored the molecular mechanisms used by beneficial symbionts for colonization21,22,23. Using transposon insertion sequencing, the aim was to identify molecular factors that enable B. gladioli to colonize L. villosa beetles.
Transposon-mediated mutagenesis was performed using the pRL27 plasmid, which carries a Tn5 transposon and a kanamycin resistance cassette flanked by invert repeat sites. The plasmid was introduced into the target B. gladioli Lv-StA cells by conjugation with the plasmid donor E. coli WM3064 strain (as shown in Figure 1). After conjugation, the conjugation mix containing B. gladioli recipient and E. coli donor cells were plated on selective agar plates containing kanamycin. The absence of DAP on the plates eliminated the donor E. coli cells, and the presence of kanamycin selected for successful B. gladioli Lv-StA transconjugants. The pooled B. gladioli Lv-StA mutant library obtained from harvesting the 100,000 transconjugant colonies was prepared for sequencing using a modified DNA library preparation kit and custom primers. Figure 2 highlights the DNA library preparation steps. Sequencing yielded 4 Mio paired reads; 3,736 genes out of 7,468 genes in B. gladioli Lv-StA were disrupted.
To identify mutants that were colonization-defective in the host, the B. gladioli Lv-StA mutant library was infected on the beetle eggs and grown in vitro in KB medium as a control. The in vivo colonization bottleneck size was calculated before the experiment. A known number of B. gladioli Lv-StA cells was infected on beetle eggs, and the number of colonizing cells in freshly hatched first instar larvae was obtained by plating a suspension from each larva and counting colony-forming units per individual. These calculations were done to ensure that the number of colonizing cells is enough to assess all or a high percentage of the mutants in the library for their ability to colonize the host. Additionally, the growth time between in vitro and in vivo conditions was normalized based on the number of bacterial generations to make these samples comparable.
After the eggs hatched, 1,296 larvae were collected in 13 pools. The corresponding in vitro mutant cultures were grown and stored as glycerol stocks. DNA of the in vivo and in vitro grown mutant libraries was extracted and fragmented in an ultrasonicator. Figure 3 shows the size distribution of the sheared DNA, where the majority of the fragments span between 100 and 400 bp, as expected. This step was followed by the modified DNA library preparation protocol for sequencing. At each step of the protocol, the concentration of remaining DNA was checked to ensure that the steps were performed correctly and to track losses of DNA. A quality check (see the Table of Materials) before sequencing revealed that the DNA libraries contained unexpectedly large (>800 bp) DNA fragments, and this was more pronounced in the in vivo libraries. Given the difficulty in optimizing the clustering of fragments in the sequencing lanes, it was necessary to increase the sequencing depth to 10 Mio paired reads in the in vivo libraries to attain the desired number of reads. The analysis of the sequencing results revealed that an average of 4 Mio reads in the in vivo libraries and 3.1 Mio reads in the in vitro libraries contained the Transposon edge in the 5' end of Read-1 (Table 9), which was satisfactory for this experiment. The distribution of the 24,224 unique insertions across the B. gladioli genome in the original library is shown in Figure 4. An analysis carried out using DESeq2 revealed that the abundances of 271 mutants were significantly different between the in vivo and in vitro conditions.
Figure 3: Agarose gels of a mutant and DNA libraries. (A) Agarose gel with unsheared DNA of a mutant in lane x and a 1 kbp ladder for scale. (B) Gel with sheared DNA library. The band sizes of the ladder in the first lane are indicated on the left side. The first three lanes a, b, and c contain sheared DNA fragments of the in vivo libraries. Lanes d, e, f, and g contain sheared DNA fragments of the in vitro libraries. Please click here to view a larger version of this figure.
Figure 4: Location of unique insertion sites in the original library across the four replicons in the Burkholderia gladioli Lv-StA genome. Each bar along the x-axis is located at a site of insertion. The height of a bar along the y-axis corresponds to the number of reads associated with that site. Note that the two chromosomes and two plasmids are shown in full length and thus have different scales on the x-axis. Please click here to view a larger version of this figure.
King’s B medium/ agar | |
Peptone (soybean) | 20 g/L |
K2HPO4 | 1.5 g/L |
MgSO4.7H2O | 1.5 g/L |
Agar | 15 g/L |
Dissolved in distilled water | |
LB medium/agar | |
Tryptone | 10 g/L |
Yeast extract | 5 g/L |
NaCl | 10 g/L |
Dissolved in distilled water |
Table 1: Media components.
No. | Primers | Sequence | PCR annealing temp. (°C) | |
1 | tpnRL17–1RC | 5’-CGTTACATCCCTGGCTTGTT-3’ | 58.2 | |
2 | tpnRL13–2RC | 5’-TCGTGAAGAAGGTGTTGCTG-3’ |
Table 2: Primers to confirm the success of conjugation.
Component | Volume (μL) |
HPLC-purified water | 4.92 |
10x Buffer S (high specificity) | 1 |
MgCl2 (25 mM) | 0.2 |
dNTPs (2 mM) | 1.2 |
Primer 1 (10 pmol/µL) | 0.8 |
Primer 2 (10 pmol/µL) | 0.8 |
Taq (5 U/µL) | 0.08 |
Mastermix total | 9 |
Template | 1 |
Table 3: PCR master mix to confirm the success of conjugation. Abbreviations: HPLC = high-performance liquid chromatography; dNTPs = deoxynucleoside triphosphate.
Steps | Temperature °C | Time | Cycles |
Initial Denaturation | 95 | 3 min | 1 |
Denaturation | 95 | 40 s | |
Annealing | 58.2 | 40 s | 30 to 35 |
Extension | 72 | 1-2 min | |
Final Extension | 72 | 4 min | 1 |
Hold | 4 | ∞ |
Table 4: PCR conditions to confirm the success of conjugation.
Primers | Sequence | Tm °C | Use | Source | |
Transposon-specific biotinylated primer | 5’-Biotin-ACAGGAACACTTAACGGCTGACATG -3’ |
63.5 | 6.7.1. PCR I | Custom | |
Modified Universal PCR primer | 5’- AATGATACGGCGACCACCGAGATC TACACTCTTTCCCTACACGACGCTC TTCCGATCTGAATTCATCGATGAT GGTTGAGATGTGT – 3’ |
62 | 6.10.1. PCR II | Custom | |
Index primer | Refer to the manufacturer’s manual | 6.7.1. PCR I & 6.10.1. PCR II | NEBNext Multiplex Oligos for Illumina (Index primers set 1) | ||
Adapter | Refer to the manufacturer’s manual | 6.5. Adapter ligation | NEBNext Ultra II DNA library prep kit for Illumina |
Table 5: Primers and adapter for PCR I and II during DNA library preparation.
PCR mix | (µL) |
Adapter-ligated DNA fragments | 15 |
NEBNext Ultra II Q5 master mix | 25 |
Index primer (10 pmol/ µL) | 5 |
Transposon specific biotinylated primer (10 pmol/ µL) | 5 |
Total volume | 50 |
Table 6: DNA library preparation-PCR I master mix.
Steps | Temperature | Time | Cycles |
Initial Denaturation | 98 °C | 30 s | 1 |
Denaturation | 98 °C | 10 s | 6 to 12 |
Annealing | 65 °C | 30 s | |
Extension | 72 °C | 30 s | |
Final Extension | 72 °C | 2 min | 1 |
Hold | 16 °C | ∞ |
Table 7: DNA library preparation-PCR I and II conditions.
PCR mix | (µL) |
Bead-selected DNA | 15 |
NEBNext Ultra II Q5 master mix | 25 |
Index primer | 5 |
Modified universal PCR primer | 5 |
Total volume | 50 |
Table 8: DNA library preparation-PCR II master mix.
Libraries | Invivo-1 | Invivo-2 | Invivo-3 | Invitro-1 | Invitro-2 | Invitro-3 | Original library | |
No. of reads (PE) | 56,57,710 | 39,19,051 | 30,65,849 | 35,73,494 | 28,83,440 | 36,61,956 | 46,09,410 | |
No. of reads containing Tn – edge on 5’ end of Read-1 | 54,15,880 | 37,31,169 | 29,36,247 | 33,00,499 | 27,35,705 | 33,50,402 | 41,53,270 | |
Bowtie2 overall alignment rate (%) (Read-1 only) | 95.53% | 83.71% | 89.87% | 80.79% | 78.00% | 73.06% | 74.92% | |
Number of unique insertions | 8,539 | 4,134 | 7,183 | 18,930 | 18,421 | 20,438 | 24,224 | |
Number of genes hit | 1575 | 993 | 1450 | 2793 | 2597 | 3037 | 3736 |
Table 9: Summary of sequencing output and transposon insertion frequency per library. Abbreviation: PE = paired-end.
A B. gladioli transposon mutant library was generated to identify important host colonization factors in the symbiotic interaction between L. villosa beetles and B. gladioli bacteria. The major steps in the protocol were conjugation, host-infection, DNA library preparation, and sequencing.
As many strains of Burkholderia are amenable to genetic modification by conjugation24,25, the plasmid carrying the transposon and antibiotic insertion cassette was conjugated successfully into the target B. gladioli Lv-StA strain from E. coli. Previous attempts of transformation by electroporation yielded very low to almost no B. gladioli transformants. It is advisable to optimize the transformation technique for the target organism to efficiently yield a large number of transformants.
One round of conjugation and 40 conjugation spots disrupted 3,736 genes in B. gladioli Lv-StA. In hindsight, multiple rounds of conjugation would be necessary to disrupt most of the 7,468 genes and obtain a saturated library. Notably, the incubation time during conjugation was not allowed to exceed 12-18 h, which is the end of the exponential growth phase of B. gladioli. Allowing conjugation beyond the exponential growth phase of bacterial cells reduces the chances of success of obtaining transconjugants26. Therefore, the conjugation period should be adjusted according to the growth of the bacterial species.
To successfully carry out an experiment involving the infection of mutant libraries in a host, it is important to assess the bacterial population bottleneck size during colonization and the diversity of mutants in the library before infection1,2,27. In preparation for the experiment, we estimated the minimum number of beetles that must be infected to have a high chance that each mutant in the library is sampled and allowed to colonize. The approximate in vivo bacterial generation time and the number of generations for the duration of the experiment were also calculated. The in vitro culture was then grown up to a comparable number of generations by adjusting the incubation time. For a similar infection experiment in other non-model hosts, the ability to maintain a laboratory culture and a constant source of the host organisms is desirable.
Following the growth of the mutant library in vivo and in vitro and sample collection, a modified DNA library preparation protocol for transposon insertion sequencing was carried out. The modification in the protocol involved designing custom PCR primers and adding PCR steps to select for DNA fragments containing the insertion cassette. Because the protocol was customized, additional PCR cycles in the protocol increased the risk of overamplification and obtaining hybridized adapter-adapter fragments in the end libraries. Hence, a final cleanup step (without size selection) after the two PCRs is recommended, as it helps in removing these fragments. The size distribution of the DNA libraries was still broader than expected. However, increasing the sequencing depth provided sufficient data that were filtered during bioinformatics analysis, obtaining satisfactory results.
As transposon-mediated mutagenesis generates thousands of random insertions in a single experiment, it is possible to generate a saturated library of mutants that contains all except those mutants where genes essential for bacterial growth have been disrupted. We most likely did not work with a saturated mutant library, given the estimations of essential genes in other studies on Burkholderia sp.28,29. A non-saturated library nevertheless helps in exploring various candidate genes for further studies using targeted mutagenesis. Before the experiments, it is also important to remember that some transposons have specific insertion target sites that increase the abundance of mutants at certain loci in the genome30. Mariner transposons are known to target AT sites for insertion31, and Tn5 transposons have a GC bias32,33. Including steps during bioinformatics analysis to recognize hotspots for transposon insertions will help in assessing any distribution bias.
Although prone to setbacks, a well-designed transposon insertion sequencing experiment can be a powerful tool to identify many conditionally important genes in bacteria within a single experiment. For example, a dozen genes in Burkholderia seminalis important for the suppression of orchid leaf necrosis were identified by combining transposon mutagenesis and genomics34. Beyond Burkholderia, several adhesion and motility genes and transporters have been identified as important colonization factors in Snodgrassella alvi symbionts of Apis mellifera (Honeybee)22, and in the Vibrio fischerii symbionts of Euprymna scolopes (Hawaiian bobtail squid)23 using the transposon-insertion mutagenesis approach.
As an alternative approach, transposon mutagenesis may be followed by screening for individual mutants using selective media instead of sequencing. Phenotypic screening or bioassays to identify deficiencies, such as motility, production of bioactive secondary metabolites, or specific auxotrophies, are feasible. For example, screening of a Burkholderia insecticola (reassigned to genus Caballeronia35) transposon mutant library has been key in identifying that the symbionts employ motility genes for colonizing Riptortus pedestris, their insect host36. Furthermore, using transposon mutagenesis and phenotypic screening, the biosynthetic gene cluster for the bioactive secondary metabolite caryoynencin was identified in Burkholderia caryophylli37. An auxotrophic mutant of Burkholderia pseudomallei was identified following transposon mutagenesis and screening and is a possible attenuated vaccine candidate against melioidosis, a dangerous disease in humans and animals38. Thus, transposon mutagenesis and sequencing is a valuable approach in studying the molecular traits of bacteria that are important for the interactions with their respective hosts in pathogenic or mutualistic associations.
The authors have nothing to disclose.
We are thankful to Junbeom Lee for providing the E. coli WM3064+pRL27 strain for conjugation and guidance in the procedure, Kathrin Hüffmeier for helping with troubleshooting during mutant library generation, and Prof. André Rodrigues for supporting insect collection and permit acquisition. We also thank Rebekka Janke and Dagmar Klebsch for support in the collection and rearing of the insects. We acknowledge the Brazilian authorities for granting the following permits for access, collection, and export of insect specimens: SISBIO authorization Nr. 45742-1, 45742-7 and 45742-10, CNPq process nº 01300.004320/2014-21 and 01300.0013848/2017-33, IBAMA Nr. 14BR016151DF and 20BR035212/DF). This research was supported by funding from the German Science Foundation (DFG) Research Grants FL1051/1-1 and KA2846/6-1.
2,6- Diaminopimelic Acid | Alfa Aesar | B22391 | For E.coli WM3064+ pRL27 |
Agar – Agar | Roth | 5210 | |
Agarose | Biozym | 840004 | |
AMPure beads XP (magentic beads + polyethylene glycol + salts) | Beckman Coulter | A63880 | Size selection in step 6.6 |
Bleach (NaOCl) 12% | Roth | 9062 | |
Bowtie2 v.2.4.2 | Bioinfromatic tool for read mapping. Reference 10 in main manuscript. | ||
Buffer-S | Peqlab | PEQL01-1020 | For PCRs |
Cell scraper | Sarstedt | 83.1830 | |
Cutadapt v.2.10 | Bioinformatic tool for removing specific adapter sequences from the reads. Reference 8 in main manuscript. | ||
DESeq2 | RStudio package for assessing differential mutant abundance. Usually used for RNAseq analysis. Reference 12 in main manuscript. | ||
DNA ladder 100 bp | Roth | T834.1 | |
dNTPs | Life Technology | R0182 | PCR for confirming success of conjugation |
EDTA, Di-Sodium salt | Roth | 8043 | |
Epicentre MasterPure Complete DNA and RNA Purification Kit | Lucigen | MC85200 | |
Ethidium bromide | Roth | 2218.1 | |
FastQC v.0.11.8 | Bioinformatic tool for assessing the quality of sequencing data. Reference 7 in main manuscript. | ||
FeatureCounts v.2.0.1 | Bioinformatic tool to obtain read counts per genomic feature. Reference 11 in main manuscript. | ||
Glycerol | Roth | 7530 | |
K2HPO4 | Roth | P749 | |
Kanamycin sulfate | Serva | 26899 | |
KCl | Merck | 4936 | |
KH2PO4 | Roth | 3904 | |
MgSO4.7H2O | Roth | PO27 | |
Na2HPO4 | Roth | P030 | |
NaCl | Merck | 6404 | |
NEBNext Multiplex Oligos for Illumina (Index primers set 1) | New England Biolabs | E7335S | |
NEBNext Ultra II DNA library prep kit for Illumina | New England Biolabs | E7645S | |
Peptone (soybean) | Roth | 2365 | For Burkholderia gladioli Lv-StA KB-medium |
peqGOLD 'Hot' Taq- DNA Polymerase | VWR | PEQL01-1020 | PCR for confirming success of conjugation |
Petri plates – 145 x 20 mm | Roth | XH90.1 | For selecting transconjugants |
Petri plates – 90 x 16 mm | Roth | N221.2 | |
Qiaxcel (StarSEQ GmbH, Germany) | Quality check after DNA library preparation | ||
Streptavidin beads | Roth | HP57.1 | |
Taq DNA polymerase | VWR | 01-1020 | |
Trimmomatic v.0.36 | Bioinformatic tool for trimming low quality reads and also adapter sequences. Reference 9 in main manuscript. | ||
Tris -HCl | Roth | 9090.1 | |
Tryptone | Roth | 2366 | For Escherichia coli WM3064+pRL27 LB medium |
Ultrasonicator | Bandelin | GM 70 HD | For shearing |
USER enzyme (uracil DNA glycosylase + DNA glycosylase- lyase Endonuclease VIII) | New England Biolabs | E7645S | Ligation step 6.5.2 |
Yeast extract | Roth | 2363 |