Transcript GST SF in E. coli - Institute for Genomic Biology

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
Phenomics
Transcriptomics
Conditions for target
expression
Metabolomics
Verification of hypothesized
enzyme-catalyzed reaction
and/or evidence from
relevant pathway
(We now do qRT-PCR on each
gene of interest)
Bochner, B.R. (2003) New technologies to assess genotype-phenotype relationships, Nature Rev. Genetics. 4,
309-314.
Infrastructure
Personnel
Instrumentation
• John Cronan (Microbiology)
• Jonathan Sweedler
(Metabolomics)
• Microbiology
•
•
•
•
•
Brad Evans (Metabolomics)
McKay Wood (Micro/Meta)
Kyuil Cho (Metabolomics)
Ritesh Kumar (Micro)
Amy Jones (Micro)
– 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
• GST:
– E. coli
• SsnA, Php, TatD, YahJ,
YjjV, HyuA, YcdX, Ade
– B. halodurans
• LisM-RP
• ENO:
– E. coli
• YfcG, YghU, YqjG, YliJ,
YfcF, YncG, YibF, YecN
• HAD:
– E. coli
• YidA, YigB, YbjI, NagD
– P. fluorescens
• GudX, RspA, YcjG, YfaW
– B. cereus
• NSAAR
– S. enterica
• ManD-RP
– A. tumefaciens
• 1RVK, 2NQL, GlucDRP,
Atu0270, Atu4120,
Atu3139, Atu4196…
• 3M9L
• IS:
– A. tumefaciens
• IspB
– C. glutamicum
• gi# 19551716
– B. fragilis
• gi# 53711383
HAD SF: YidA from E. coli
dgoT
dgoD dgoA dgoK dgoR
yidA
YidA
Courtesy of D. Dunaway-Mariano
kcat = 2 s-1
KM = 250 μM
kcat/KM = 8 x 103 M-1s-1
Toxic if
concentration
builds in the cell!
YidA (HAD): no effect after addition of galactonate
glycerol +
galactonate
succinate +
galactonate
glucose +
galactonate
YidA KO
likely mutated during lag
YidA(HAD): long lag when cells are resuspended in galactonate
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 fusion
yfcG-phoA
empty vector yqjG-phoA
treA-phoA
gapA-phoA
yghU-lacZ
yfcG-lacZ
empty vector yqjG-lacZ
treA-lacZ
gapA-lacZ
Cytoplasm
Periplasm
yghU-phoA
YghU (GST): protein localization via proteomics
YfcF(GST): culture labeling and metabolite extraction
Ions from WT
Ions from mutant
YfcF (GST): differential labeling provides higher accuracy
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
Data Processing
Perturbation Exp.
LC-MS Analysis
Preprocessing
(XCMS)
 Peak
detection/alignment
 Retention time
correction
 Noise filtering
Data Quality Control
Peak Grouping
Isotope Pattern
Analysis
 Mass check
 Retention time check
 Intensity ratio check
Peak Grouping
 Primary Peaks
 Isotope pattern
 ≥ 20% intensity change
 Secondary Peaks
 Isotope pattern
 < 20% intensity change
 Retention time filter
 Adducts/Salt filter
 Missing value
imputation
Deisotoping
Data
Normalization
 Monoisotopic
peaks
 Time-wise, condition
specific
 Mean-, Z-value …
Formula
Prediction
Formula modeling
 Primary peaks used first
 Round Robin
 Recursive Backtracking
Theoretical Isotope
Pattern Modeling
 First order Markov
 Forward Trellis
Bayesian Statistics
 Isotope pattern
comparison
 experimental v.s
theoretical
Heuristics
Pathway Activity
Profiling
DB Search
 2ppm mass tolerance
 Top hits formula
Seed Metabolites
 Isotope pattern
 High intensity change
 Exist in current DB
Pathway Analysis
 Seed metabolites info.
 DB Hits mono. peaks
 Shared pathways detection
Activity Profiling
 Sort detected peaks
upon fold change
 p-values by MSEA
Active Pathways
 Pathways: p < 0.05
 Prior prob. for C, N, S
 6 Golden rules
Potential Target Peaks
Top 3 hits
 Highly up- or downregulated, but not yet
annotated peaks
 Further experiments are
needed
Work with Tobias Erb, Gerlt Lab (ENO)
RLP: whole cell untargeted metabolomics
p-value
= 0.02
p-value
= 4.8 x 10-4
met-salvage
pathway
p-value
= 7.3 x 10-4
Purine
metabolism
+MTA
Glutathione
metabolism
MTA
8
4
0
MTR-1P
2
1
0
MTRu-1P
0min
10min
20min
1.0
0.5
0
DXP
p-value
= 1.2 x 10-3
2
1
0
CDP-MEP
8
4
0
Isoprenoid
pathway
up-regulated pathway
down-regulated pathway
pathway showing no big difference
metabolite
0min
10min
20min
c-MEPP
0min
10min
20min
Metabolite intensity ( x 106 )
Control
p-value
= 0.048
16
8
0
+MTA
0min
10min
20min
Butanoate
metabolism
Metabolite intensity ( x 106 )
Control
Work with Tobias Erb, Gerlt Lab (ENO)
RuBisCO-like protein from R. rubrum
RLP
Cupin
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…)
Future work will use the samples stored in the Protein / Structure Core
Microbiology/Protein/Structure Core Collaboration
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