Liberati et al. 10.1073/pnas.0511100103. |
Supporting Table 3
Supporting Table 4
Supporting Methods
Supporting Table 5
Supporting Table 6
Supporting Table 7
Supporting Figure 8
Supporting Figure 9
Supporting Figure 8
Fig. 8. Short genes are less likely than longer genes to be disrupted by transposon insertion and are overrepresented in the set of putative essential genes. (A) Theoretical likelihood (y axis) of gene disruption as a function of gene size (x axis), given a genome of 6.53 megabases (Mb) and a library of 60,000 mutants. The black line indicates 95% probability of insertion. (B) Relative abundances of genes (y axis) is represented as a function of gene length (x axis) for the set of putative essential genes (red) and for the set of all PA14 genes (blue). Genes were grouped in 50-bp bins by size, before totaling and dividing by the total number of genes, either for putative essential genes or for all genes. The black line indicates a gene of 327 bp (95% probability of insertion).
Supporting Figure 9
Fig. 9. Flow diagrams of PA14 Transposon Insertion Library and production schemes for the PA14 nonredundant (PA14NR) Set. Transposant colonies grown on LB agar with antibiotics were picked using a Qbot (Genetix, Boston, MA) to inoculate media in 96-well "working" (wor) plates. Mutant cultures were manipulated as shown. PA14NR Set mutants were streaked from frozen wor plate stocks onto LB agar with antibiotics before manual picking and inoculation of master blocks containing LB and antibiotics.
Table 3. Makeup of the parental PA14 transposon insertion library
Parental library | Total mutants | Processed mutants | RawSeqs | High-quality seqs | High-quality seqs with BLAST hits | Genes hit | Insertion sites |
PA14/MAR2xT7 | 34,176 | 30,336 | 30,321 | 25,035 | 24,089 | 4,469 | 20,530 |
PA14/TnphoA | 4,608 | 1,344 | 1,344 | 744 | 689 | 453 | 641 |
exoU psi3/MAR2xT7 | 9,216 | 2,304 | 2,304 | 1,656 | 1,497 | 927 | 1,402 |
exoUspcU /MAR2xT7 | 9,216 | 4,992 | 3,167 | 914 | 793 | 459 | 676 |
All backgrounds/All transposons | 57,216 | 38,976 | 37,136 | 28,349 | 27,068 | 4,598 | 22,881 |
Processed mutants
refers to the number of mutants subjected to arbitrary PCR and sequencing. RawSeqs indicates the number of sequences obtained from processed mutants. High-quality seqs indicates the number of sequences containing high-quality sequence as defined by the PHRED software application. High-quality seqs with BLAST hits indicates the number of high-quality sequences that aligned with PA14 and/or PA01 genomic sequences. For a description of the PA14NR Set selection scheme, see Supporting Methods.
Table 4. Makeup of the PA14 nonredundant (PA14NR) set
PA14NR set | Total |
PA14 MAR2xT7 mutants | 5,215 |
PA14 TnphoA mutants | 44 |
exoU (both backgrounds) MAR2xT7 mutants | 200 |
Total PA14NR set mutants | 5,459 |
Total genes represented | 4,596 |
Bacterial Strains. Transposon insertion mutants were generated in wild type Pseudomonas aeruginosa strain PA14 (1) and in two PA14 derivatives, DexoU and DexoUspcU. The rationale for using an exoU background was to decrease the cytotoxicity of PA14 (2). PA14 DexoU contains a 2-kb deletion of exoU, but the deletion is not in-frame, and a newly generated stop codon may alter the expression pattern of the immediate downstream gene spcU (2). An in-frame deletion encompassing the adjacent PA14 genes exoU and spcU, which are in the same operon, was constructed by replacing 4.29 kb of wild-type sequence with a 1.88-kb PCR-amplified fragment that contained a 2.41-kb deletion. The following oligonucleotide primers, based on P. aeruginosa strain PAO1 sequence data, were used to generate the exoUspcU deletion. ExoU111-SacI: 5'-CTGGAGCTCGAGTATCGACGCTGTAGCTAACG-3'; ExoU22-XbaI: 5'-TAACGCGCTGGTTCTAGATTGGATATGCATGTTCGCTC-3'; SpcU3-XbaI: 5'-CATATCCAATCTAGAACCAGCGCGTTAGTGTTCG-3'; SpcU4-HindIII: 5'-GTCAAGCTTATGGTGCAGACCGTCCAGGC-3'. The PCR-amplified fragment containing the deletion was subcloned into the SacI and HindIII sites of pEX18Ap (3) generating plasmid pEX18exoUspcU8, which was subsequently used to introduce the deletion into the wild-type PA14 genome by homologous recombination as described (3).
Transposons.
The vast majority of the PA14 mutants were created with MAR2xT7, an engineered derivative of the Himar1 transposon (4, 5). pMAR2xT7, a plasmid carrying MAR2xT7, was derived from pMFLGM.GB. pMFLGM.GB (obtained from R. Vance and J. Mekalanos, Harvard Medical School, Boston) was constructed by inserting the linker acgcgtgaagttcctatactttctagagaataggaacttcTCAAGCCAGCAAAGTGCGATACactagtacgctagctaccatggcagCcgcgggaagttcctatactttctagagaataggaacttcacgcgt (which contains FRT sites and SpeI, NheI, NcoI and SacII restriction sites) into the MluI site in pFAC (6). The gentamicin-resistance cassette from pBBR1MCS-5 (6) was excised with NcoI and SacII and cloned into the modified pFAC to create pMFLGM.GB.To create pMAR2xT7, the primers 5'-CATCGATCGTCTAGTAACAGGTTGGCTGATAAGTCCCCGGTCTCTAGACCCTATAGTGAGTCGTATTACGCGGCCGCCGCGGGGAC-3' and 5'-GTCCCCGCGGCGGCCGCGTAATACGACTCACTATAGGGTCTAGAGACCGGGGACTTATCAGCCAACCTGTTACTAGACGATCGATG-3' were annealed, digested with PvuI, extended with Taq polymerase, digested with SacII, and ligated to replace the SacII/PvuII fragment in pMFLGM.GB. The primers encode a T7 promoter and an adjacent inverted repeat (IR) sequence. The sequence encoding a T7 promoter and its adjacent IR sequence was amplified from pMycoMar (5) by using the following oligonucleotides: 5'-CATAGCTAGCCGCGGGACCGAGATAGGGTTGAGTG-3' and 5'-TCCCCCCGGGTCTGACGCTCAGTCGAACG-3'. The NheI-digested PCR fragment was inserted into the NheI site of the modified pMFLGM.GB to create pMAR2xT7, which was propagated in the pir+ Escherichia coli strain MC4100. A map and sequence of pMAR2xT7 are available at http://ausubellab.mgh.harvard.edu/cgi-bin/pa14/downloads.cgi.
Transposon Mutagenesis, Colony Selection, and Work Flow.
Tripartite matings were performed using saturated cultures of PA14 or PA14 derivatives, E. coli MC4100/pMAR2xT7, and the helper strain E. coli HB101/pRK2013 (7, 8). 1:2:2 mixtures of the cultures, respectively, were dropped on Kings’ B media plates (2% wt/vol peptone, 6.57 mM K2HP04, 6.08 mM MgSO4, 1% vol/vol glycerol) (8, 9) and incubated at 37°C for 2 h. Independent matings were spread on a single rectangular 20- ´ 20-cm LB agar plate containing 15 mg/ml gentamicin and 1 mg/ml Irgasan at a density of »1 cfu/cm2. After incubation at 37°C for 12-15 h, colonies were robotically picked with a Qbot (Genetix, Boston) and grown statically in 250 ml of LB containing 15 mg/ml gentamicin in 96-well microtiter plates for »40 h at 37°C. For TnphoA matings, saturated cultures of PA14 and E. coli SM10lpir carrying pRT731 were mixed in a 1:1 ratio, dropped onto King’s B media plates, and incubated for 6 h at 37°C (10). Independent matings were collected, resuspended in 100 mM MgSO4, and plated onto a rectangular LB agar plate as described above containing 200 mg/ml neomycin and 100 mg/ml Irgasan. Colonies were picked robotically as above into 250 ml of LB containing 200 mg/ml kanamycin and 50 mg/ml Irgasan and grown at 37°C statically for »40 h. Seventy microliters of the putative MAR2xT7 and TnphoA transformant cultures were transferred robotically to 96-well microtiter plates for arbitrary PCR and sequencing (see below). Glycerol was added to the cultures before transfer to long-term storage plates at a final concentration of 15%. Fig. 9 shows a flow diagram indicating how putative insertion mutant clones were cataloged and partitioned into various microtiter plates during the process of library construction.Transposon Insertion Site Identification.
Transposon insertion sites were identified by using a two-round arbitrary PCR protocol (11). Aliquots (70 ml) of the statically grown transformant cultures (see above) were incubated at 95°C for 10 min to lyse the cells and spun at 3,000 rpm for 5 min. Of the cleared lysate, 3 ml was used as template for the first round of arbitrary PCR (ARB 1). ARB 1 PCR mix contained 1´ Taq Buffer (Roche), 10% DMSO, 2.5 mM dNTPs, 1.25 units of Taq polymerase (Roche), and 1.0 ng of each primer per microliter of mix. For MAR2xT7 mutants, the transposon-specific primer, PMFLGM.GB-3a, (5'-TACAGTTTACGAACCGAACAGGC-3') was used. For TnphoA mutants, Tn5Ext (5'-GAACGTTACCATGTTAGGAGGTC-3') was used as the transposon-specific primer. Although different ARB primers were used at different times during library production to maximize the efficiency of obtaining useful sequence data, the ARB 1 primer used most frequently was ARB 1D (5'-GGCCAGGCCTGCAGATGATGNNNNNNNNNNGTAT-3'). Plates were sealed with adhesive foil (Alum-1000, Diversified Biotech). After initial denaturation at 95°C for 5 min, ARB 1 plates (Costar catalog no. 6511, Corning Incorporated) were cycled 30 times at 95°C for 30 s, 47°C for 45 s, and 72°C for 1 min. Plates were incubated 5 min at 72°C to allow for extension of PCR products.For the second round of arbitrary PCR, 5 ml of the ARB 1 reaction was used as template. ARB 2 reaction mix was the same as the ARB 1 reaction except that nested transposon-specific primers and a nested ARB 2 primer were used. The PMFLGM.GB-2a primer (5'-TGTCAACTGGGTTCGTGCCTTCATCCG-3') was used for MAR2xT7 mutants, and the Tn5Int2 primer (5'-GGAGGTCACATGGAAGTCAGATCCTGG-3') was used for TnphoA mutants, respectively, as transposon-specific primers. The ARB 2 primer, ARB 2A (5'-GGCCAGGCCTGCAGATGATG-3') was used for both MAR2xT7 and TnphoA mutants. Plates were sealed with adhesive foil. Reactions were cycled 40 times: 95°C for 30 s, 45°C for 30 s, and 72°C for 1 min with a 5-min extension at 72°C.
For PCR cleanup, 5 ml of ARB 2 reaction was mixed with 2 ml of EXOSAP-IT enzyme mix (catalog no. 78205, USB) in a new reaction plate (catalog no. AB-0800, Abgene). Plates were sealed with adhesive foil and incubated according to the manufacturer’s instructions. Sequencing primer was added directly to each PCR cleanup for a final concentration of 5 ng/ml. The PMFLGM.GB-4a primer (5'-GACCGAGATAGGGTTGAGTG-3') was used as the sequencing primer for MAR2xT7 mutants and the Tn5Int primer (CGGGAAAGGTTCCGTTCAGGACGC-3') was used for TnphoA mutants.
The PA14 Transposon Insertion Mutant Database (PATIMDB).
The PATIMDB was implemented using the MYSQL relational database managing system hosted on a multiprocessor Intel system running Red Hat (Raleigh, NC) LINUX. The data-entry application was written in JAVA and runs on WINDOWS 2000. This application implemented process tracking and an automated sequence analysis pipeline. Sequences in the form of ABI files were imported into the application, and base-calling was performed using PHRED. The resulting raw sequence was trimmed to select only high-quality sequence, which was then compared by BLAST (12) to the PA14 genome to identify the insertion site of the transposon and the identity of the disrupted ORF. The data-retrieval system was implemented using a PERL-based common gateway interface (CGI) web application hosted on a multiprocessor Intel system running Red Hat LINUX with APACHE. The functions of the data-retrieval web application include queries and the downloading of data from PATIMDB over the web, such as a list of mutants with identified insertion locations. PATIMDB is compatible with different genome sequences and is adaptable to library construction applications in other organisms. Extensive quality assurance testing was performed on PATIMDB, the data-input application, and the data-retrieval application, to ensure that the database accurately related and processed each mutant DNA sequence file. PATIMDB can be accessed at http://ausubellab.mgh.harvard.edu/cgi-bin/pa14/home.cgi.A map of all identified transposon insertions in the PA14 chromosome is also available at http://ausubellab.mgh.harvard.edu/cgi-bin/pa14/tnmap.cgi. The transposon insertion map is a web-based application written with PERL CGI and JAVASCRIPT and is deployed on the same server as the data-retrieval system. This database-driven application accesses PATIMDB via PERL database interface. To give users the best viewing experience, the application dynamically detects user screen resolution to decide the map size that can best fit the screen. Because of the large amount of data that has to be processed to generate the map, when a map window contains data for 50 kb or more, the application performance can be degraded. We used a server-side cache technique to pregenerate all large-window maps, so users can quickly retrieve maps of all sizes. Extensive quality assurance has been done to ensure that all maps display the correct data. A genome navigation bar (at the top of the screen) can locate any genome location with one click. An integrated search tool (below the map) helps users find specific transposon insertions. Two buttons on each side of the page allow users to scan nearby genome locations. With multiple viewing options and data filters, users can select data of interest. Links to information such as the genetic background of the mutant carrying each displayed insertion and gene and mutant reports are also incorporated into the transposon insertion map application.
Statistical Analysis of Transposon Insertion Distribution.
Gaps (in base pairs) between transposon insertions in the PA14 genome were identified by recording insertion locations along the 6,534,396 base pairs in the PA14 genome. Gap sizes were measured by counting the number of bases between insertion locations. The number of gaps of a given size, or in bins of a given size range, was then tabulated.A theoretical genome array the same length as the actual genome size was used in a Monte Carlo prediction of random insertion events. We created 26,534 simulated transposon insertions by using a random number generator, and insertion locations were recorded. Simulated gap sizes were then measured as with the actual data by counting the number of bases between insertion locations. Gap sizes were grouped in bins of 200 bp. The simulation was repeated 200 times and bin totals were averaged giving the theoretical distribution of gap sizes. Gaps larger than 2,500 bp were identified.
Probabilistic Calculation of Insertion likelihood as a function of Gene Size.
Assuming a random distribution of transposon insertions, the probability of getting at least 1 insertion in a gene of length l given a genome of size g and a library containing n mutants is:p
(one or more insertions given n mutants) = 1 – (1 – (l/g))nTherefore, a gene 327 bp in length has a 95% likelihood of insertion given a genome size of 6.5 megabases and library size of 60,000 insertions.
PA14/PAO1 Ortholog Assignment.
PA14/PAO1 orthologs were picked using an automated PERL script that performed reciprocal BLAST alignments of PA14 and PAO1 protein sequences. Orthologs were required to have the same amino acid length within 30% and to have at least 70% of the amino acid sequence length align with a minimum of 70% identity across the aligned sequence. In the case of multiple high-scoring hits, orthologs were assigned to maintain synteny between the PA14 and PAO1 genomes. Cases of redundancy due to gene duplication were resolved manually.PA14NR Set Selection and Production.
Mutant selection criteria. Selection of mutants for the nonredundant set was automated by creating a hierarchical set of priorities. Mutants were prioritized as follows. (i) The mutants corresponding to a particular gene were divided into two groups: those with BLAST bit scores less than or equal to 80 and those with scores of >80. Priority was given to those with BLAST scores of >80. (ii) Each of these two groups was further subdivided by strain background (PA14; exoU; exoUspcU) and transposon (MAR2xT7; TnphoA), with priority given to PA14/MAR2xT7 over PA14/TnPhoA and PA14/TnPhoA over exoU/MAR2xT7 or exoUspcU/MAR2xT7. (iii) Each of these six groups was ordered by distance of the transposon insertion site from the start of the gene, with priority given to more 5' insertions. A 200-bp handicap was given to sequences that exhibited evidence of contamination based on the sequencing. (iv) Finally, the mutants were ordered by BLAST bit score such that if all other criteria were equal the mutant with the higher BLAST score would have priority. In cases where there was a mutant available in the PA14 background but a more 5' mutant was available in either the exoU or exoUspcU background, both mutants were included in the PA14NR Set.Colony purification.
Mutants that were selected for inclusion in the PA14NR Set were picked manually from frozen 96-well "working plates" (see Fig. 9) on dry ice and streaked onto LB agar containing 15 mg/ml gentamicin. Plates were incubated at 37°C for 14–16 h, cooled to room temperature, and stored at 4°C for no more than 8 days before picking. Colonies were examined under a dissecting scope. A single colony with a morphology representative of the majority of colonies for a given mutant was used to inoculate 600 ml of LB containing 15 mg/ml gentamicin in a 96-well deep-well block (USA Scientific, Ocala, FL). Small colony variants were avoided except in cases where nonvariants were not available (<20 mutants).Culture conditions.
Master deep-well blocks were incubated in a HiGro incubator/shaker (Genomic Solutions, Ann Arbor, MI) set to 37°C with O2 injection for 2 s every 5 min of incubation for 14–16 h. Plates were set at room temperature for 0–6 h before transferring an aliquot of each culture to master blocks (see Fig. 9).Biomek transfer to master plates.
A Biomek FX (Beckman Coulter) liquid-handling robot was used to transfer cultures between microtiter plates. We found that unless glycerol (15%) is added to PA14 cultures before transfer, both manual and robotic culture transfer is prone to cross-contamination, possibly because of aerosolization of the cultures. Even with the inclusion of glycerol in transferred cultures, we were unable to define a methodology for transfer to 384-well plates that did not result in cross-contamination of adjacent wells. For this reason, the PA14NR Set has been arranged in a 96-well format. Exhaustive testing of the following protocol demonstrated that during test runs, no cross-well contamination was observed. We transferred 200 ml of sterile 60% glycerol from a reservoir to each deep-well master block and mixed each three times by pipeting 200 ml each time up and down. Tips were touched to the side of the wells to remove liquid clinging to the end of the tips and then discarded. Fresh tips were used to transfer 40 ml of the mixed cultures to each of 10 96-well master plates containing 160 ml LB with 15 mg/ml gentamicin and 15% glycerol. The Biomek FX was programmed to pipet culture 1 mm from the top of the culture for each transfer to minimize the surface area of the transfer tip coated with culture. Transferred cultures were dispensed by dropping the culture slowly into the recipient wells without blowing air out of the tips and without mixing. Tips were dipped into the culture/media mix in the recipient wells, touched to the side of the well and removed. Master blocks and plates were sealed (Alumina Seals, Diversified Biotech, Boston) and stored at –80°C immediately after a set of 10 plates were transferred and sealed.Quality Control.
Initially, Southern blotting was used to confirm a single insertion in 10 independent MAR2xT7 mutants. Further confirmation of single MAR2xT7 insertions was obtained from the sequence derived from the ARB PCR product of each mutant. The vast majority of mutants displayed clear, easily readable sequences rather than mixed sequences that would be expected if MAR2xT7 were prone to multiple insertion events.Extensive quality controls were put in place to prevent cross-contamination of colony-purified cultures and to avoid propagation of variant subpopulations during construction of the PA14NR Set. A full description of the techniques used and a "users manual" for the PA14NR Set are available at http://ausubellab.mgh.harvard.edu/cgi-bin/pa14/home.cgi.
From the PA14NR Set, 120 random clones were resubjected to arbitrary PCR and sequencing to check whether expected mutants were located in the appropriate well positions. Readable sequence was isolated from 109 clones. Of these, all but three sequences (2.8%) were the same as the sequences assigned by PATIMDB to the corresponding well positions.
PVC Attachment Screen.
Testing the PA14NR Set for attachment to polyvinylchloride (PVC) plastic was carried out essentially as described in ref. 13. Briefly, a sterilized, metal, 96-pin head was used to inoculate LB with stored, frozen cultures from the PA14NR Set. After 16 h at 37°C with shaking at 250 rpm (Innova4400, New Brunswick Scientific), cultures were diluted 1:10 in M63 media containing 1% casamino acids, 0.3% glucose, 0.5 mM MgSO4, and 0.025% vitamin B1. These 1:10 dilutions were diluted again in the same medium in PVC plates for a final dilution of 1:100. Diluted cultures (1:10 and 1:100 ) were grown statically for 8 h at 37°C. Media was removed from the PVC plates by pipeting, and the plates were stained with 1% crystal violet for 10 min. After several destaining rinses in water, stained rings at the air–liquid interface were scored by eye. A subset of rings was dissolved in 95% ethanol, and absorbance at 550 nm was recorded. After 8 h of static growth at 37°C, the optical density of cultures grown from the 1:10 dilutions were read at a wavelength of 600 nm using a Spectra Max Plus plate reader (Molecular Devices, Sunnyvale, CA).1. Rahme, L. G., Stevens, E. J., Wolfort, S. F., Shao, J., Tompkins, R. G. & Ausubel, F. M. (1995) Science 268, 1899–1902.
2. Miyata, S., Casey, M., Frank, D. W., Ausubel, F. M. & Drenkard, E. (2003) Infect. Immun. 71, 2404–2413.
3. Hoang, T. T., Karkhoff-Schweizer, R. R., Kutchma, A. J. & Schweizer, H. P. (1998) Gene 212, 77–86.
4. Lampe, D. J., Grant, T. E. & Robertson, H. M. (1998) Genetics 149, 179–187.
5. Rubin, E. J., Akerley, B. J., Novik, V. N., Lampe, D. J., Husson, R. N. & Mekalanos, J. J. (1999) Proc. Natl. Acad. Sci. USA 96, 1645–1650.
6. Wong, S. M. & Mekalanos, J. J. (2000) Proc. Natl. Acad. Sci. USA 97, 10191–10196.
7. Darzins, A. & Chakrabarty, A. M. (1984) J. Bacteriol. 159, 9–18.
8. Figurski, D. H. & Helinski, D. R. (1979) Proc. Natl. Acad. Sci. USA 76, 1648–1652.
9. Jones, D. L. & King, S. (1954) Am. J. Clin. Pathol. 24, 1316–1317.
10. Rahme, L. G., Tan, M. W., Le, L., Wong, S. M., Tompkins, R. G., Calderwood, S. B. & Ausubel, F. M. (1997) Proc. Natl. Acad. Sci. USA 94, 13245–13250.
11. Caetano-Anolles, G. & Bassam, B. J. (1993) Appl. Biochem. Biotechnol. 42, 189–200.
12. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. J. (1997) Nucleic Acids Res. 25, 3389–3402.
13. O’Toole, G. A. & Kolter, R. (1998) Mol. Microbiol. 30, 295–304.