Microbiology/Metabolomics Core John Cronan and Jonathan Sweedler Enzyme Function Initiative (EFI) Advisory Committee Meeting November 30, 2011
Feb 10, 2016
Microbiology/Metabolomics Core
John Cronan and Jonathan Sweedler
Enzyme Function Initiative (EFI)Advisory Committee Meeting
November 30, 2011
Outline
• Experimental scope
• Infrastructure
• Targets
• YidA from E. coli (HAD)
• YghU, YfcF, and YfcG from E. coli (GST)
• RuBisCO-like protein from R. rubrum (ENO)
• Future Directions
Experimental Scope
Bochner, B.R. (2003) New technologies to assess genotype-phenotype relationships, Nature Rev. Genetics. 4, 309-314.
Phenomics Transcriptomics Metabolomics
Conditions for target expression
(We now do qRT-PCR on each gene of interest)
Verification of hypothesized enzyme-catalyzed reaction
and/or evidence from relevant pathway
Infrastructure
Personnel
• John Cronan (Microbiology)• Jonathan Sweedler
(Metabolomics)
• Brad Evans (Metabolomics)• McKay Wood (Micro/Meta)• Kyuil Cho (Metabolomics)• Ritesh Kumar (Micro)• Amy Jones (Micro)
Instrumentation
• Microbiology– Biolog Omnilog phenotype
microarray plate reader/incubator
– Growth curve-ometer, BioscreenC
– E. coli single gene KO collection (Keio collection)
• Metabolomics– 11 Tesla LTQ-FT LC-MS– High resolution QTOF LC-MS– Custom XCMS LCMS data
analysis platform for untargeted metabolomics
Targets from around the EFI
• AHS:– E. coli
• SsnA, Php, TatD, YahJ, YjjV, HyuA, YcdX, Ade
– B. halodurans• LisM-RP
• ENO:– E. coli
• GudX, RspA, YcjG, YfaW– B. cereus
• NSAAR– S. enterica
• ManD-RP– A. tumefaciens
• 1RVK, 2NQL, GlucDRP, Atu0270, Atu4120, Atu3139, Atu4196…
• GST:– E. coli
• YfcG, YghU, YqjG, YliJ, YfcF, YncG, YibF, YecN
• HAD:– E. coli
• YidA, YigB, YbjI, NagD– P. fluorescens
• 3M9L
• IS:– A. tumefaciens
• IspB– C. glutamicum
• gi# 19551716– B. fragilis
• gi# 53711383
HAD SF: YidA from E. coli
Courtesy of D. Dunaway-Mariano
YidAkcat = 2 s-1
KM = 250 μMkcat/KM = 8 x 103 M-1s-1
dgoRdgoKdgoAdgoDdgoT yidA
Toxic if concentration
builds in the cell!
YidA (HAD): no effect after addition of galactonate
glycerol +galactonate
succinate + galactonate
glucose +galactonate
YidA(HAD): long lag when cells are resuspended in galactonate
YidA KOlikely mutated during lag
YidA (HAD): LCMS results for KDGP
Validated with standard from Hua Huang in the DDM Lab
YidA from E. coli (HAD): Results and Conclusions
• Phenomics is difficult with HAD SF members, as many are promiscuous housekeeping phosphatases
• An abrupt shift from a relatively poor carbon source to galactonate as sole carbon source causes the YidA KO to display a growth lag– The “abruptness” may be important for quickly building
levels of the toxic metabolite, KDGP– Growth of YidA following the lag may be due to mutation
• Metabolomics efforts so far do not support the connection between YidA KO lag with elevated KDGP levels
GST SF in E. coli: a role in oxidative stress response?
YfcF and YfcG (GST): NO sensitivity in null mutants
GST SF in E. coli: secreted to the periplasm?
Modeling/docking by Backy Chen, Computation Core
GST SF in E. coli: protein localization via gene fusionyghU-phoA yqjG-phoAyfcG-phoA empty vector treA-phoA gapA-phoA
yghU-lacZ yqjG-lacZyfcG-lacZ empty vector treA-lacZ gapA-lacZ
Cyt
opla
smPe
ripla
sm
YghU (GST): protein localization via proteomics
YfcF(GST): culture labeling and metabolite extraction
YfcF (GST): differential labeling provides higher accuracy
Ions from WT Ions from mutant
YfcF (GST): contaminant peaks remain unlabeled
YfcF (GST): affect of nitric oxide on metabolites
GST SF: results and conclusions
• YfcF and YfcG are implicated in reduction of nitric oxide– NO sensitivity phenotype identified– YfcF metabolomics with cutting-edge labeling protocol
allows measurement of small changes in metabolites
• Cellular localization is an important aspect of enzyme function– YghU and YfcG appear to remain in the cytoplasm
RuBisCO-like protein, RLP, from R. rubrum (ENO)
Canonical methionine salvage pathway (e.g. B. subtilis)
Seemingly incomplete MSP (R. rubrum)
?RLP
Work with Tobias Erb, Gerlt Lab
RLP: evidence for novel fate of methionine sulfur
Work with Tobias Erb, Gerlt Lab (ENO)
RLP: whole cell untargeted metabolomics
Work with Tobias Erb, Gerlt Lab (ENO)
RLP: whole cell untargeted metabolomics
Work with Tobias Erb, Gerlt Lab (ENO)
Data Processing
Preprocessing(XCMS)
Data Quality Control
Data Normalization
Peak detection/alignment
Retention time correction
Noise filtering
Retention time filter Adducts/Salt filter Missing value
imputation
Time-wise, condition specific
Mean-, Z-value …
Peak Grouping Formula Prediction
Pathway Activity Profiling
Isotope Pattern Analysis Mass check Retention time check Intensity ratio check
Peak Grouping
Deisotoping
Formula modeling
Theoretical Isotope Pattern Modeling
Perturbation Exp.LC-MS Analysis
Primary Peaks Isotope pattern ≥ 20% intensity change Secondary Peaks Isotope pattern < 20% intensity change
Monoisotopic peaks
Primary peaks used first Round Robin Recursive Backtracking
First order Markov Forward Trellis
Bayesian Statistics Isotope pattern
comparison experimental v.s
theoretical
Heuristics Prior prob. for C, N, S 6 Golden rules
Top 3 hits
DB Search
Seed Metabolites
Pathway Analysis
Activity Profiling
Active Pathways
2ppm mass tolerance Top hits formula
Isotope pattern High intensity change Exist in current DB
Seed metabolites info. DB Hits mono. peaks Shared pathways detection
Sort detected peaks upon fold change
p-values by MSEA
Pathways: p < 0.05
Potential Target Peaks Highly up- or down-
regulated, but not yet annotated peaks
Further experiments are needed
MTRu-1P01
2
MTA08
16
MTR-1P04
8
0min
10m
in20
min
0min
10m
in20
min
Met
abol
ite in
tens
ity (
x 10
6 )
Control +MTA
met-salvagepathway
p-value= 1.2 x 10-3
Isoprenoidpathway
p-value= 0.048
Glutathionemetabolism
p-value= 7.3 x 10-4
Purinemetabolism
Butanoatemetabolism
DXP
CDP-MEP
c-MEPP
00.51.0
0 1 2
0 4 8
0min
10m
in20
min
0min
10m
in20
min
Met
abol
ite in
tens
ity (
x 10
6 )
Control +MTA
up-regulated pathway
down-regulated pathway
pathway showing no big difference
metabolite
p-value= 4.8 x 10-4
p-value= 0.02
RLP: whole cell untargeted metabolomics
Work with Tobias Erb, Gerlt Lab (ENO)
RuBisCO-like protein from R. rubrum
RLPCupin
Work with Tobias Erb, Gerlt Lab (ENO)
RuBisCO-like protein (ENO): Results and Conclusions
• Perfect starting point for Micro./Metabolomics Core– Collaboration with ENO bridging project– Phenotype was known– High profile project (Ashida, et.al. Science, 2003)
• Genome context and measured thiol release suggested novel fate of MTA– Key enzymes in known MSP missing from genome– Cell extracts mixed with MTA produced methanethiol
• LC-MS-based metabolomics uncovered connection between MTA feeding and isoprenoid biosynthesis– Untargeted metabolite profiling of R. rubrum uncovered:
• Predicted MTA degradation products• Unexpected isoprenoid biosynthesis intermediates
Taking advantage of existing samples…Noncovalent Protein: Ligand Interactions Measured by Native ESI-MS
(from test cases to EFI samples…)
Microbiology/Protein/Structure Core Collaboration
Future work will use the samples stored in the Protein / Structure Core
Micro./Metabolomics Core: future directions
• Application of Biolog and custom phenotype microarrays to null mutants of targets from additional organisms
• Transcriptional analysis coupled to growth condition screens to gain complementary evidence for when target genes are expressed
• Further improvements in XCMS software to better detect metabolites of low abundance
• Application of differential labeling and multiple chromatographies for each metabolomics experiment to increase accuracy
• Continued and increasing collaboration with the BPs and Cores