1. MGH-PGA Genomic Analysis of Stress and Inflammation: Pseudomonas aeruginosa Infection Nicole T. Liberati, Dan G. Lee, Jacinto M. Villanueva and Frederick M. Ausubel Department of Molecular Biology Massachusetts General Hospital Carolyn Cannon, Fadie Coleman, Mike Kowalski Jeff Lyczak, Martin Lee, Gloria Meluleni, and Gerald Pier Channing Laboratory Brigham and Women’s Hospital

2. Ausubel/Pier PGA Project

3. Contents of Slide Show: Section I: Background Information on Multi Host Pathogenesis System Section II: Background Information on Screening Methodology and Rationale for Constructing the Uni-Gene Library Section III:Progress Report on Uni-Gene Library Construction and Detailed Methodology Section IV:Development of a Linux MySQL Uni-Gene Library Relational Database Section V:CF Mouse Oropharynx Colonization Model

4. Section I Background Information on Multi-Host Pathogenesis System

5. Pseudomonas aeruginosa Gram-negative rod Found throughout the environment in soil, water and plants Opportunistic human pathogen: - Nosocomial pulmonary infections -Immune compromised patients (chemotherapy/burns) -85% of adult CF patients suffer from chronic pulmonary P. aeruginosa infections

6. P. aeruginosa Multi-Host Pathogenesis System P. aeruginosa Strain PA14 Humans Plants Insects Nematodes Mice

7. % Nematodes Killed Hours of Feeding on P. aeruginosa strain PA14 P. aeruginosa Kills C. elegans and Colonizes the C. elegans Intestine 0 20 40 60 80 100 020406080 E. coli P. aeruginosa

8. P. aeruginosa Kills Galleria mellonella (Wax Moth Caterpillar) Larvae; LD 50 = 1

9. Section II Background Information on Screening Methodology and Rationale for Constructing the Uni-Gene Library

10. Random Transposon Mutagenesis PA14 Sequence Insertion Sites and Identify a Non-Redundant “Unigene” Set Screen Unigene Set for Mutants that Do Not Kill C. elegans or Wax Moths Test Mutants that Do Not Kill C. elegans in CF Mouse Model Identification of P. aeruginosa Virulence Factors by Screening “UniGene” Library for Mutants that do not Kill Wax Moth Caterpillars or Nematodes

11. Transposon: Kan r E. Coli PA14 6 Mb Transposase Generation of Transposon Insertion Mutations Select for insertions with Kanamycin

12. Unigene Library: A collection of P. aeruginosa strains containing a disruption in each non-essential open reading frame (ORF) in the P. aeruginosa genome Wild typeMutant #1Mutant #2

13. ~6 Mb 6 Mb genome 4800 non-essential genes 24,000 insertions 30,400 insertions ( S. cerevisiae ) 5 fold saturation Recovery failure Unigene Library Size

14. Selection of Unigene Library Mutants Approximately 5 hits per ORF: Choose the most 5’ disruption within the actual coding sequence ~6 Mb 30,400 insertions ~4800 catalogued Unigene mutants

15. Advantages of Unigene Library Screening 1)Mutation previously identified 2)Limited number of mutants to screen (4800) 3)Non-redundant mutations 4)Built-in confirmation of the involvement of known pathways. 5)Easy to confirm the importance of the mutated gene using other mutant alleles.

16. Section III Progress Report on Uni-Gene Library Construction and Detailed Methodology

17. Pick ~30,000 colonies with Qbot into bar-coded 96- well plates containing media + selective antibiotics Divide into 3 plates: 384-well (Master copy) 384-well (Duplicate copy) 96-well (Working copy) Grow overnight 25 ml for arbitrary (ARB) PCR reactions Add glycerol to 15% Generation of Unigene Library of Transposon Insertions in Non-Essential Genes

18. Current Status of the Unigene Library 1)48 x 96 (4608) mutants created. 60% of the insertion sites identified. 2)Insertion site identification protocol optimized. (1152 mutants created and identified in 2.5 weeks) 3) Accompanying database is operational. Quality assurance testing is in progress.

19. Library Construction: Mutagenesis/Plating TnPhoA: Kan r /Neo r E. Coli PA14 LB + Irgasan + Neomycin (3,000-5,000 colonies)

20. Library Construction: Colony Picking/Culture Inoculate 250 µL LB + Irgasan [50 µg/mL] Kanamycin [200 µg/mL]* Grow 40 hrs at 37°C (no shaking)

21. Working (wor) Master (mas) Duplicate (dup) Supernatant (sup) Culture (wor) (280 µL) 70 µL Add glycerol Mix and Seal -80°C Storage-20°C Storage Library Construction: Biomek-Automated Liquid Manipulation

22. Library Construction: Bar Coding A Side: Unique ID# B Side: Human Readable PA14_PhoA_100_xxx wor sup mas dup ar1 ar2 seq

23. Genomic DNA Transposon Transposon-specific Primer Arbitrary PCR Primers LEGEND 1st PCR Reaction 2nd PCR Reaction PCR Cleanup and Sequencing 1 1 2 2 3 Library Construction: Arbitrary PCR to Amplify Sequence Adjacent to Transposon Insertion

24. Supernatant (sup) Arb PCR 1 (ar1) Thaw, 99°C/6 min., pellet 3K/5 min. Arb PCR 1 Library Construction: Details of ARB1 PCR Temporary Storage -20°C 3µL supnt

25. ar1 Arb2 PCR (ar2) ARB2 PCR Temporary Storage -20°C 5µL Library Construction: Details of ARB2 PCR

26. Library Construction: PCR Cleanup: EXOSAP-IT

27. ar2 Sequencing plate (seq) + ExoSAP-IT 15’ at 37°C 15’ at 80°C Library Construction: PCR Cleanup Temporary Storage -20°C 7µL

28. Library Construction: Sequencing seq Add Sequencing primer to a [final] of 5 ng/µL and Seal Send to DNA Core for Sequencing (Store at 4°C)

29. Example of High Quality Sequence TnPhoA Sequence Length + Mixed + TnPhoA = Sequence Success Index

30. TnPhoA Example of Low Quality Sequence

31. Sequencing Success Index Optimization of [Taq] in Sequencing Reactions

32. Unigene Library Mutant Identification Optimized for: Taq Manufacturer Roche vs. Promega vs. Prepared Master Mixes Final Taq Concentration 1.25 U sufficient PCR Master Mix Preparations Fresh Master Mixes vs Stored (4°C) Master Mix Hybaid vs. MJ Research PCR Machines PCR Cleanup Protocol ExoSAP-IT vs. Clontech NucleoSpin

33. 1 1 2 2 3 3 2 2 3 Template-Specific Tn/Genomic Sequence No Sequence A Template-Independent Genomic Sequence Relevant Background Sequence: Template Independent Genomic Sequence

34. NNNNNNNNN ARB PRIMER Sequence High Quality Sequence (cont’d)

35. NNNNNNNNN ARB PRIMER Sequence Trouble Shooting: Buried ARB Sequence High Quality Sequence

36. 1 1 2 3 Relevant Background Sequence: Buried ARB Primer Sequence 1 1 2 3 2 2 1 2 1 2 +

37. Library Construction:Time Line for 4608 colonies (48 sup plates) Mutagenesis/growth on Qbot plate Qbot picking/growth in 96 well culture plate Biomek ARB1/ARB2 Reactions/PCR Cleanup/Seq prep Sequencing Total Time 3 days 2 days 1 day 10 days ? 16+ days For 7 sets of 48 plates: 114 days

38. P. aeruginosa PA14 Virulence-Related Factors Involved in Mammalian Pathogenesis Identified in Non-Vertebrate Hosts Category#Genes Regulators6 gacA, gacS, algU, plcR, ptsP, lasR Membrane Protein1 aefA Biosynthetic Enzymes3 phzB, hrpM, fabF Modifying Enzyme1 dsbA Multi-Drug Transport2 mexA, mtrR Type III Secretion1 pscD Helicases2 phoL, lhr Unknown Proteins16?

39. Section IV Development of a Linux MySQL Uni-Gene Library Relational Database

40. Unigene Library: Overview of Bioinformatics Catalog each sample in relational database Retrieve DNA sequence for each sample Process DNA sequence to remove low-quality and contaminant sequences (i.e. - vector) BLAST searches to distinguish: a.Pseudomonas sequence vs. contaminants. b.PA01 vs. PA14 sequences. BLAST searches to identify: a.Disrupted ORF. b.Coding sequence vs. putative promoter disruption.

41. What will the MySQL Database Do? 1)Store/catalog all of the data. 2)Process DNA sequences and perform BLAST searches. 3)Display the results and allow for user queries.

42. How will the Database Store the Data? The data will be stored in a relational database. Individual tables can be thought of as separate Excel spreadsheets with rows and columns. The tables are connected to each other via specified relationships.

43. How will the Database Store the Data? Tables will be populated (i.e. - individual cells in the table will be filled with entries) as plates, samples, and/or data are generated. Data entry into the Database will be “restricted” to parallel the creation of the physical library. – Order of different types of inputs is restricted. – Prevent duplicate entries.

44. How will the Database Store the Data? The Database will store organizational information: – Date created. – Created by. – Storage locations. – Bacterial strain. – Mutagen/Transposon used. The Database will store experimental data: – DNA sequences obtained by PCR. – Location of insertion with respect to PAO1 genome. – Identity of PAO1 ORF disrupted. – Phenotypic data?

45. Plate PlateID PlateType ProcessPlateLink ExecutionID PlateID InOrOut ProcessExecution ExecutionID ProcessType PlateIDPlateType 5PCR1 6PCR2 ExecutionIDProcess Type 12ARB2 PCR ExecutionIDPlateIDInOrOut 125In 126Out

46. Plate PlateID PlateType ProcessPlateLink ExecutionID PlateID InOrOut ProcessExecution ExecutionID ProcessType Protocol PlateSample SampleID PlateID Well Position MutantID Mutant MutID Library RawSequence RawSeqID MutantID ChromatPath Sequence

48. How will the Database Analyze the Data? ChromatogramChromatogram Phred Raw SequenceQuality ScoresTrimmed Sequence

49. How will the Database Analyze the Data? Trimmed Sequence Remove transposon, vector, and/or other contaminant sequences. Processed Sequence BLAST PAO1 genome BLAST PAO1 annotated ORFs

50. How will the Database Analyze the Data? Other BLAST searches that can be performed in the future: – Internal BLAST against the contents of the database to identify siblings vs. adjacent independent insertions. – BLAST against other public databases to determine gene identity of ORFs not found in PAO1.

51. How will we Retrieve/View the Contents of the Database? Current status: – A web-accessible table viewer can allow us to examine the contents of each table in the database. – To organize and search the contents, the html file can be opened in Excel and then sorted. Future goals: – A web-accessible browser with multiple query and view options.

52. How will we Retrieve/View the Contents of the Database? Types of queries: – Insertions in a given gene. – Insertions upstream of a given gene. – Insertions near a given gene. – Insertions near a given physical location. – Insertions in PAO1 non-coding sequences. – Insertions in sequences NOT found in PAO1. – Insertions in genes of a particular pathway/family. – Insertions in PAO1 ORFs of known function. – Insertions in putative PAO1 ORFs of unknown function. – Multiple queries.

53. How will we Retrieve/View the Contents of the Database? Options for Viewing Database Contents: – Table view in alphabetical order. – Table view in linear order. – Graphical view with ORF orientation and transposon orientation (zoom in/out/, click on ORF or transposon, etc).

54. Future Steps Select members for unigene (non-redundant) library. Physically pick members for unigene library. Store, duplicate and disseminate unigene library. Incorporate non-PAO1 sequences into unigene set. “Completing” the unigene set (targeted deletions, inducible antisense?).

55. Trouble Shooting Test cases - designed to test either the entire system (i.e. - start to finish) or a particular module. Can be designed to test “ideal” inputs or “incorrect” inputs. Example: – Input 96 (unknown) chromatogram files that contain a few known sequences at defined well positions. – Determine if the expected BLAST hits are associated with the appropriate well position. – This tests: Ability to process the chromatogram files. Ability to correctly perform BLAST searches. Ability to correctly map the resulting BLAST search onto the correct well-position.

56. Trouble Shooting: Example Test Case After chromatograms are retrieved, is each well- position mapped correctly as the sequences are processed and BLASTed? ChromatogramChromatogram Phred Raw Sequence Quality Scores Trimmed Sequence Remove transposon, vector, and/or other contaminant sequences. Processed Sequence BLAST PAO1 genomeBLAST PAO1 ORFs

57. A: aroE B: braB C: coxA D: dnaA E: exoT F: fabF G: galE H: hmgA Trouble Shooting: Example Test Case

58. Plate 1 A: aroE B: braB C: coxA D: dnaA E: exoT F: fabF G: galE H: hmgA Plate 2Plate 3 Trouble Shooting: Example Test Case

59. Example Test Case: Expected BLAST Results 1A1 : aroE 1A2 : --- 1A3 : --- 1A4 : --- 1A5 : --- 1A6 : --- 1A7 : --- 1A8 : --- 1A9 : --- 1A10 : --- 1A11 : --- 1A12 : --- 1B1 : braB 1B2 : --- 1B3 : --- 1B4 : --- 1B5 : --- 1B6 : --- 1B7 : --- 1B8 : --- 1B9 : --- 1B10 : --- 1B11 : --- 1B12 : --- 1C1 : coxA 1C2 : --- 1C3 : --- 1C4 : --- 1C5 : --- 1C6 : --- 1C7 : --- 1C8 : --- 1C9 : --- 1C10 : --- 1C11 : --- 1C12 : --- 1D1 : dnaA 1D2 : --- 1D3 : --- 1D4 : --- 1D5 : --- 1D6 : --- 1D7 : --- 1D8 : --- 1D9 : --- 1D10 : --- 1D11 : --- 1D12 : ---

60. “System Requirements” Programming language / database platform. – Microsoft Access vs. MySQL. Backups and database restore. Storage issues. – Each plate generates ~30MB of chromatograms (I.e. - 3 X 96 chromatograms on a zip disk). – Each chromatogram spawns several types of data: a raw sequence, a quality score for each nucleotide of the raw score (~2.67 MB for an average 96-well plate), a processed sequence, and blast results. Documentation of database development and test cases.

61. Database Summary Data storage is mostly complete. Needs some testing. Sequence analysis is currently being tested. – Once it’s operational, sequence analysis will have to be updated to include more complex scenarios (i.e. - sequences not found in PAO1). Data retrieval/viewing is currently undeveloped. Non-redundant (unigene) library is undeveloped.

62. Section V CF Mouse Oropharynx Colonization Model

63. Utility of transgenic CF mice for identifying novel P. aeruginosa virulence factors To date, no apparent phenotype relevant to acquisition and establishment of chronic P. aeruginosa infection has been found in a variety of transgenic CF mouse lines CF mice given acute P. aeruginosa infections manifest increased inflammation and pathology but do not get chronic infections

64. Aspects of chronic oropharyngeal colonization in mice Maintain mice on water with antibiotic to prevent P. aeruginosa growth in water-0.1 mg gentamicin/ml Treat mice for 5-7 days with 250 ug levofloxacin/ml in drinking water – Eliminates oropharyngeal colonization by a mucoid Enterobacter that grows on Pseudomonas isolation media and interferes with P. aeruginosa colonization – Remove 48 hrs prior to introduction of P. aeruginosa Our approach: try to recapitulate method of natural acquisition of P. aeruginosa by CF patients

65. Aspects of chronic oropharyngeal colonization in mice Colonize mice by placing 10 7 CFU P. aeruginosa /ml drinking water for 5 days Remove contaminated water, culture mouse throats, give sterile water for 1 week followed by water containing 0.1 mg gentamicin/ml to keep bacteria from growing in it Monitor mice by throat culture every 1-2 weeks.

66. CWP= contaminated water/Pseudomonas P. aeruginosa strain PA14 chronically colonizes oropharynx of CF, but not wild-type, C57Bl/6 mice Mouse Oropharyngeal Colonization Model 0 20 40 60 80 100 1368101215171921232629 Time (Weeks) After Infection C57sBl/6 CF mice CWP Percent Positive Throat Cultures

67. Wild-type PA14 mucD mutant mucD mutant complemented in trans algD mutant mucDalgD double mutant 0 20 40 60 80 100 P. aeruginosa s train Percent that develop sepsis P=.06 P <.001 A contribution of algD to pathogenesis is shown in a mouse thermal injury model--the double mucD/algD mutant is more attenuated for virulence than the mucD mutant alone From: Yorgey P, Rahme LG, Tan MW, Ausubel FM. The roles of mucD and alginate in the virulence of Pseudomonas aeruginosa in plants, nematodes and mice. Mol Microbiol. 2001 Sep;41(5):1063-1076.

68. The algD mutant of P. aeruginosa PA14 fails to chronically colonize the oropharynx of CF mice CWP= contaminated water/Pseudomonas Time (Weeks) After Infection C57Bl/6 WT-mice CF mice 0 20 40 60 80 100 1368101215171921232629 CWP PA14 DalgD in transgenic CF mice WT-PA14 in transgenic CF mice 0 20 40 60 80 100 137111517192123 Percentage colonized

69. Another mutant of P. aeruginosa PA14, with an interruption in the gacA (global accessory regulator) gene, previously shown to have reduced virulence in the multi-host pathogen system, also has a reduced ability to chronically colonize the oropharynx of CF mice Time (Weeks) After Infection 0 20 40 60 80 100 135711151921232527 C57Bl/6 WT-mice CF mice Percentage colonized

70. Summary of CF Mouse Model A model of chronic P. aeruginosa oropharyngeal colonization in CF mice has been developed and tested for applicability for confirming the role of P. aeruginosa multi-host virulence factors.