The Virtual Institute of Microbial Stress and Survival

1
The Virtual Institute of Microbial Stress and
Survival
Environmental Monitoring
Pathway Models
Pathway Inference
Deduction of Stress Response Pathways in Metal
and Radionuclide Reducing Bacteria
Microbial Isolation
Cell/environmental models
Imaging
Comparative Genomics
sFTIR In situ physiology
Metabolomics
Proteomics
2
Participants
Adam Arkin, (UCB/LBNL, Director and Core Team
Co-Leader) Terry C. Hazen (LBNL, Co-Director and
Core Team Leader) Jay Keasling (UCB/LBNL, Core
Team Co-Leader) Eric Alm (LBNL, Core Team
Co-Leader) Inna Dubchak (JGI/LBNL) Matthew Fields
(Miami University) Hoi-Ying Holman (LBNL) Martin
Keller (Diversa Inc.) Aindrila Mukhopadhyay
(LBNL, Core Team Co-Leader) David Stahl
(University of Washington) Dorothea Thompson
(ORNL) Judy Wall (University of Missouri) Jizhong
Zhou (ORNL) and many others
National Laboratories Lawrence Berkeley Oak
Ridge Sandia JGI
Universities U.C. Berkeley University of
Washington U. Missouri, Columbia Miami U.
Industry Diversa, Inc
3
The Focus

• Established in August 2002. http//vimss.lbl.gov
• Experimentally elucidating and computationally
  modeling stress response pathways in metal and
  radionuclide reducing organisms
• Primary organism Desulfovibrio vulgaris
• Anaerobic, SRB, common in eutrophic environments,
  much less known about this organism
• Comparison organisms
• Shewanella oneidensis MR-1, microaerophile, more
  common in oligotrophic environments
• Geobacter metallireducens, anaerobe, more common
  in oligotrophic environments
• Stressors O2, metals, TEAs, PO4, NO2, NO3, pH

4
Why Sulfate Reducing Bacteria
S0
Global S Cycle
oxic
SO42-
H2S
anoxic
Assimilative sulfate reduction
Global C Cycle
Dissimilative sulfate and sulfur-reducing microbes
S0
Souring of oil reservoirs
Microbial-induced corrosion
5
Bioremediation
SEM micrograph courtesy of Hoi-Ying Holman (LBNL)
Reduction of metals to less toxic or less soluble
oxidation states in the environment depends on
how bacteria respond to environmental stressors
6
Themes

• Adaptability of cells to their environment
• Efficiency of metabolism within a cell and across
  its community
• Plasticity and Evolution of Genomes
• Network Inference and Interpretation of
  Functional Genomics

7
The VIMSS pipeline

• With a few results

8
The Virtual Institute for Microbial Stress and
Survival
Environmental Microbiology Core
Functional Genomics Core
Gene expression MAs
Environmental Stressors pH, O2, NO2-, NaCl, heat,

Proteomics ICAT, DIGE, Comp., complexes
Biomass production
Pathway Predictions
Environmental sequencing
Data Analysis
Genome Annotation
Computational Core
9
WEEK 1-2
WEEK 3-4
Choose Perturbation
Growth and MIC
R4
R2
R1
WEEK 5-6
WEEK 6-6
yes
Design Time/Dose Series (S-FTIR)
Execute Time Series
d1
R6
R7
R3
E2
no
No, try again
No
Close Book
d2
R5
Close Book
Yes
E1
Store Backup samples
Ship to ORNL
For all reports there is an automated email sent
to the teams. Red means a data upload has to
occur. CC does analysis at this point. R1-
Start Stress Book- Record Name, Date, R2- Add
description of condition (protocol) and rationale
to stress book. R3- Add description of
experiment, upload biolog/growth, morphology
data, etc. complete phenomics, add conclusions
from experiment, note person, date, R4- Add
description of why decision was made R5- Upload
of S-FTIR time series R6- Upload experimental
design R7- Add comments, upload culture data,
assign sample IDs, delivery point, date shipped
added to ORNL samples. R8- Annotate that
shipment is received. Comments on conditions,
time to analysis, who is doing analysis,
protocols. R9- Upload analysis data. Annotate
with comments and decision.
R8
RNA Analysis
WEEK 7-8
E3
R9
No
Close Book
d3
Yes
Stage 2
D1- Do we believe we are seeing an interesting
response? D2- Did the time series work? (QA/QC
verification) D3- Is the RNA Good? (QA/QC
verification)
WEEK 9
Note AEMC can start new growth,MIC, and time
series design analyses as soon monthly. Nitrate,
nitrite, and salt have already been completed for
growth and MIC, FITC will be completed in the
next 2 months. Setup for time series can be done
and RNA analysis done to verify appropriateness
of samples for the rest of the team. Stored
samples can prioritized for the most interesting
responses.
10
WEEK 14-15
WEEK 10-14
WEEK 9-10
Metabolite Quality Report
Significant Responders Report
Metabolomics
R19
R11
WEEK 16
ICAT Proteome
R20
R12
Ship Samples
Stage 2
Significant Responders Report
R9
Proteomic Consistency/ Qual.Report
Diversa Proteome
R13
R21
Consistency/ Conclusions Report
Sandia DIGE
GO Report
R14
R22
Phenomics Summary Report
SEM
R23
Comparative Stress Report
R15
TEM
Yes
EPS Report
Significant Responders Report
R24
R16
EPS
Back to stage 1
d4
Lipidomics Report
No
R25
PLFA
R17
Operon/Regulon Report
Microarray Quality Report
Transcriptome
R10
R18
For all reports there is an automated email sent
to the teams. Red means a data upload has to
occur. CC does analysis at this point. R9- Add
shipping date to all other samples. R10-17 Add
receiving dates to samples. Upload Exp. Design,
protocols, contacts, predicted delivery time..
R18- Microarray data D4 Is microarray data ok?
Receive OK from CC before proceeding with
remaining Genomics R19-R25 Upload data, add
comments, Quality estimates, successful yes or
no, requests for special analysis, etc. Automate
analyses delivered back to stress book.Note If
method fails, there maybe should be a procedure
for requesting a backup sample. Can FGC do
analyses from different exp concomittantly to
increase throughput
GO Report
11
Choosing Perturbations

• Model Stressors, Field Conditions, and Phenotyping

12
Field Research Center, Oak Ridge TN

• S-3 Waste Disposal Ponds
• unlined
• Received liquid nitric acid and uranium-bearing
  wastes 320 million liters 1951-1983

13

• NABIR Field Research Center Uranium-contaminated
  site
• Sulfate reducers common

14
Principal Components Analysis 87 variance 107
compounds
15
Beginning to understand metal reduction/oxidation

• Following Dv in the Field

16
Reoxidation of Bioreduced Uranium
17
D. vulgaris (direct fluorescent antibody)
Changes in Cr(VI) and Total Cr Concentrations
after the HRC Injection
18
Learning about the bug

• What to do when youre not a model.

19
Not a model but also not unloved.

• A model sulfate reducing bacteria
• A venereal disease of industrial oil processing
• A unique life-cycle energetics that may adapt it
  to life in poor nutritive environments. (SRBs
  have been found in nearly monoculture in deep
  subsurface samples).
• A contentious theory about how the organism
  derives energy from sulfate reduction.

20
Biology of Sulfate Reduction
AMP
ATP
SO42 -
APS
HSO3 -
HS-
PPi
?
?
2e- 2H
6e- 6H
2 Pi
8e- 8H
21
Biology of Sulfate Reduction
AMP
ATP
SO42 -
APS
HSO3 -
HS-
PPi
?
?
2e- 2H
6e- 6H
2 Pi
8e- 8H
2 acetate 2 CO2
2 lactate
2 ATP
2 ADP
2 Pi
22
Biology of Sulfate Reduction
AMP
ATP
SO42 -
APS
HSO3 -
HS-
PPi
?
?
2e- 2H
6e- 6H
2 Pi
8e- 8H
ATP AMP 2Pi 2ADP 2Pi 2ATP 2ADP AMP
ATP
2 acetate 2 CO2
2 lactate
2 ATP
2 ADP
2 Pi
23
Hydrogen Cycling Model
8H 8e-
4H2
4H2
periplasm
Trans- membrane Complexes
Cytoplasmic Hydrogenase
cytoplasm
8e- 8H
2.7 ATP
2.7 ADP Pi
6e-
2e-
2 acetate 2 CO2
2 lactate
ATP
2 ATP
2 ADP
2 Pi
SO42 -
APS
HSO3 -
HS-
PPi
2H
6H
(Odom and Peck, 1981)
2 Pi
AMP
24
Molecular Basis for Hydrogen Cycling
8H 8e-
4H2
4H2
periplasm
Trans- membrane Complexes
Cytoplasmic Hydrogenase
cytoplasm
8e- 8H
2.7 ATP
2.7 ADP Pi
?
other redox proteins
?
?
Hmc
CO de-H2ase
Qmo
(Higuchi et al., 1987)
Ech
(Haveman et al., 2004)
Dhc
Nqr
Tmc
(Heidelberg, et al., 2004)
DsrMKJOP
Ohc
Desulfovibrio vulgaris
(Dahl, et al., 2005)
25
How do we find out what makes SRB special

• And how can we begin to understand the basic
  genome organization of Dv?

26
A general comparative survey

• Comparative genomics pipeline
• MicrobesOnline
• Operon/regulon prediction
• Functional Genomic Database
• Environmental Sensing and HPKs
• Inference of the SRB Signature Genes

27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Comparative Genomics Reveals an Abnormal Number
of HPKs
32
Evolution of HPK Signaling Proteins
Identify HPK domain proteins
200
genomes
33
Evolution of HPK Signaling Proteins
HPK domains from different genomes
HPK
Other signaling domains
Identify HPK domain proteins
200
genomes
4000 HPKs
34
Evolution of HPK Signaling Proteins
Identify HPK domain proteins
Isolate HPK domains
200
genomes
4000 HPKs
35
Evolution of HPK Signaling Proteins
Identify HPK domain proteins
Isolate HPK domains
197
genomes
4000 HPKs
Cluster domains of high sequence similarity
(BLASTclust)
?
HPK1b
Lineage-specific expansions
Ancient gene family expansions
36
NJ tree of HPK1b, Cluster1 of 213 HPKs
Caulobacter
Gamma
D. Vulgaris
Vibrio
Bdellovibrio
Dechloromonas expansion
37
Pseudomonas
Desulfovibrio expansion
TM
HAMP
PAS
HPK
cheY
38
D. Vulgaris HPK, NJ 1000 bootstrap
39
?s
Models
40
There are diversity producer bacteria and
consumer bacteria
41
Comparative Approach to Understanding Sulfate
Reduction
Af
Dv
Desulfovibrio vulgaris Desulfovibrio alaskensis
G20 Desulfotalea psychrophila Archaeoglobus
fulgidis
Bacteria
Archaea
Dp
Da
What genes are shared among all SRB?
42
Comparative Approach to Understanding Sulfate
Reduction
Af
Dv
Desulfovibrio vulgaris Desulfovibrio alaskensis
G20 Desulfotalea psychrophila Archaeoglobus
fulgidis
Bacteria
Archaea
Dp
Da
What genes are shared among all SRB?

• Phylogenetic profiles
• Doesnt pick up subtle differences/shared traits
  among a group of proteins
• Some key proteins may be shared with similar
  organisms outside the set (S-reducing, or
  S-oxidizing bacteria)

43
Comparative Approach to Understanding Sulfate
Reduction
Af
Dv
Desulfovibrio vulgaris Desulfovibrio alaskensis
G20 Desulfotalea psychrophila Archaeoglobus
fulgidis
Bacteria
Archaea
Dp
Da
What genes are shared among all SRB?

• Phylogenetic profiles
• Doesnt pick up subtle differences/shared traits
  among a group of proteins
• Some key proteins may be shared with similar
  organisms outside the set (S-reducing, or
  S-oxidizing bacteria)
• Simple homology-based approach
• Find genes for which orthologs in the chosen set
  of bacteria are within top (ten) BLASTp hits for
  all 200 genomes

44
Molecular Basis for Hydrogen Cycling
8H 8e-
4H2
4H2
periplasm
Dsr
Cytoplasmic Hydrogenase
Qmo
MKJ OP
cytoplasm
8e- 8H
2.7 ATP
2.7 ADP Pi
Da
Fdh
6e-
2e-
ATP
SO42 -
APS
HSO3 -
HS-
Ech Coo
Dv
Aps
Dsr
PPi
Fdh F420
Dp
Ppa
2H
6H
2 Pi
AMP
45
New SRB Signature Genes

• Known (or suspected) SO42- reduction pathway (
  genes)
• ABC transporters
• Gln (sulfate?) transporter
• Enzymes
• CarAB (carbamoyl-P)
• PflA
• CobB (Dsr operon)
• Molybdopterin biosynthesis
• DapF
• Redox (14 genes)
• Ferredoxins (3)
• Glutaredoxin/ ferredoxin-thioredoxin reductase
  operon (2)
• Rubredoxin operon (3)
• Ferritin
• Other (5)
• Unknown (5)
• Many new targets for genetic characterization!
• Are these genes co-regulated with known SO42-
  reduction genes?

46
Co-expression of Signature Genes
up-regulated
O2
-
NaCl
log
NO32 /NO2-
heat
cold
DsrMKJO complex ABC transporter
Difference mainly in nitrate/nitrite response
Expression level
Aps, Qmo, DsrAB, PpaC ferredoxin II,
glutaredoxin, ferritin, 2 unknown
down-regulated
47
Functional and Comparative Genomics Links
Signature Genes
White Operon/regulon links Blue?Red Increasing
microarray similarity
48
Genomic Plasticity
per-H2ase
e-
Niche-specific organic e- donors
H
H2
cyt-H2ase
conserved SO42- pathway
Genome-specific TM complexes/ e- acceptors
specialized e- acceptors
49
Comparative Genomics Summary

• There is clearly a very coherent sulfate
  reduction module of function in SRBs and Dv
• Dv and SRB are highly sensitive to its
  environment and rapidly evolve to new
  environments as evidenced by
• Large number of HPKs
• Status as HPK generating strains and the lack
  of correlation of HPK expression in Dv
• Plasticity of the electron donor/accept ends
  of the proposed SRB pathway.
• Maybe the hydrogen cycling process needs to be
  revisited?
• Functional genomics strong aids interpretation of
  comparative genomic analyses.

50
Characterizing environmental sensitivity in the
laboratory

• The VIMSS functional genomics/genetics pipeline

51
Goals of the pipeline

• Comprehensive cell and molecular biological
  picture of the global cellular responses to
  field stressors
• Reconstruction of cellular regulatory networks
  from a compendium of such data (still too early
  for comprehensive reconstruction)

52
WEEK 1-2
WEEK 3-4
Choose Perturbation
Growth and MIC
R4
R2
R1
WEEK 5-6
WEEK 6-6
yes
Design Time/Dose Series (S-FTIR)
Execute Time Series
d1
R6
R7
R3
E2
no
No, try again
No
Close Book
d2
R5
Close Book
Yes
E1
Store Backup samples
Ship to ORNL
For all reports there is an automated email sent
to the teams. Red means a data upload has to
occur. CC does analysis at this point. R1-
Start Stress Book- Record Name, Date, R2- Add
description of condition (protocol) and rationale
to stress book. R3- Add description of
experiment, upload biolog/growth, morphology
data, etc. complete phenomics, add conclusions
from experiment, note person, date, R4- Add
description of why decision was made R5- Upload
of S-FTIR time series R6- Upload experimental
design R7- Add comments, upload culture data,
assign sample IDs, delivery point, date shipped
added to ORNL samples. R8- Annotate that
shipment is received. Comments on conditions,
time to analysis, who is doing analysis,
protocols. R9- Upload analysis data. Annotate
with comments and decision.
R8
RNA Analysis
WEEK 7-8
E3
R9
No
Close Book
d3
Yes
Stage 2
D1- Do we believe we are seeing an interesting
response? D2- Did the time series work? (QA/QC
verification) D3- Is the RNA Good? (QA/QC
verification)
WEEK 9
Note AEMC can start new growth,MIC, and time
series design analyses as soon monthly. Nitrate,
nitrite, and salt have already been completed for
growth and MIC, FITC will be completed in the
next 2 months. Setup for time series can be done
and RNA analysis done to verify appropriateness
of samples for the rest of the team. Stored
samples can prioritized for the most interesting
responses.
53
Minimum Inhibitory Concentration

• Definition that concentration that causes a
  doubling of the generation time without
  significant mortality

54
Different Osmotic Challenges and EM
55
Proteomics (ICAT) and Metabolite Profiling
Glycine betaine ABC transporter Up-regulated in
salt by ICAT Mass-spec
Changes in selected metabolites under salt stress
56
PFLA profiling and Synchrotron FTIR Physiology
PFLA profiling shows large changes in fatty acid
composition of membrane and its resultant
fluidity.
sFTIR shows increase in total lipid content
57
Microarray data at MicrobesOnline
58
(No Transcript)
59
MicrobesOnline Analysis
60
Models for stress response
61
Salt Stress

• Accumulation of the neutral, polar
  osmoprotectant, glycine betaine, is the primary
  mechanism used by D. vulgaris to counter stress
  in a hyper-ionic environment.
• Ectoine, another well documented osmoprotectant
  was also found to be effective in alleviating
  salt stress.
• The bacterium also appears to employ a variety of
  efflux systems to counter excess ions.
• A dramatic increase in ATP synthesis pathways was
  also documented.
• Expression of the iron uptake regulon (FUR)
  increased significantly, even though the iron
  concentration did not appear to affect
  sensitivity to salt.
• The salt stress response included upregulation of
  chemotaxis, helicase genes and a change in the
  composition of phospholipid fatty acids.
• Downregulated systems included the flagellar
  biosynthesis pathway, lactate uptake permeases,
  and ABC transport systems.
• The extensive NaCl stress analysis was also
  compared with microarray data from KCl stress and
  surprising similarity between the two stresses
  was observed, setting D. vulgaris apart from many
  other bacteria that have notable differences in
  the two salt stresses.

62
Another quick stress Nitrite Stress Response
Prevention of oil reservoir souring
NO3- reducing bacteria
NO2-
SO42- reducing bacteria
Bioremediation in high NO3- soils
63
Approach

• Genome annotation
• Identify regulons (and DNA motifs) associated
  with nitrite and other pathways
• Crp/Fnr family
• Fur/Per/Zur family
• Follow pathway behavior through gene expression
  studies (nitrite vs. unstressed cells) from VIMSS
  collaborators
• Construct whole-cell model of nitrite response

64
Nitrite Stress Up-Regulated Pathways
Fe2
Fe3
Data Analysis
Modeling
Zhou Lab (ORNL)
FURFe(II) represses
?
NO2-
PerRFe(II) represses
NH3OH
Nitrite reductase
HcpR
Hcp
NH3
Aspartate
Asparagine
Glutamate
Glutamine
Key up-regulated genes require Fe
65
H2
60 min
Up-regulated
Down-regulated
NH3
Periplasm
NO2-
Fe only Hase
NrfHA
Formate dehydrogenase
2H
2e-
NiFe Hase-2 HynAB-2
NiFeSe Hase
2H
2e-
NiFe Hase-1 HynAB-1
Fdh-associated Cytochrome C
H2
C3,208021
HCOOH
C3
Coo MKLXUHF
DsrMKJOP Triheme
Hmc
ATP Synthase
QmoABC
Ldh
Lactate ADP Pi
Pyruvate
2e-
6e-
HCOOH
PorAB
e-
AcetylCoA HCOO- H
SO42- ATP
Acetate CO2 2e- 2HATP
Sat
PPi APS
Cytoplasm
PpaC
ApsAB
2Pi
HSO3-AMP
DsrAB
GlnA
Fumarate
Succinate
glutamate
glutamine
HS-
Rnf ABCDEFG Decaheme
FrdABC
Ech ABCDEF
Periplasm
NH3
66
Nitrite Fe Requirement
67
Comparative Genomic Survey Provides Insight to
Functional Data
HcpR
TTGTgAnnnnnnTcACAA
SO42- down / Hcp and ferredoxin Strongly
up-regulated
68
Comparative Genomic Survey Provides Insight to
Functional Data
HcpR
TTGTgAnnnnnnTcACAA
SO42- down / Hcp and ferredoxin Strongly
up-regulated
FUR
wTGAAAatnatTTTCAw
Strongly up-regulated
69
Comparative Genomic Survey Provides Insight to
Functional Data
HcpR
TTGTgAnnnnnnTcACAA
SO42- down / Hcp and ferredoxin Strongly
up-regulated
FUR
wTGAAAatnatTTTCAw
Strongly up-regulated
PerR
CAGTAAnnnTTACTG
Only genes with predicted PerR sites change
significantly
70
Why Does Fur Respond to Nitrite?

• Accidental response triggered by oxidation of
  Fe(II)
• Iron-dependent genes required for nitrite
  response (i.e. Hcp/Ferredoxin)
• Damage to iron containing enzymes - need to be
  replaced

71
This is one of many functional stories

• We have characterized and inferred networks for
• Oxygen response
• new physiology of oxygen tolerance
• Prediction of oxygen response regulons
• Confirmation by microarray
• Integrated prediction of membrane lipid changes
• Confirmation by lipidomics and imaging
• Also Heat shock, cold shock, pH, salt
  adaptation, strontium, exponential vs. stationary
  growth, co-culture with archea and more
• AND Same conditions for Shewanella oneidensis
• Now using genetics to parse out interactions in
  pathways.

72
Phenotypic Microarray

• Omnilog System - 2000 assays,
• 50 - 96-well plates at one time
• gt750 metabolic assays
• 239 inhibition/sensitivity assays

73
Genomic context of the two HK knockouts shown in
the MicrobesOnline Genome Browser (Top) and the
OperonBrowser (bottom).
74
Protein Complex Isolation and Identification
Using genetically-tagged bait proteins, protein
complexes were affinity isolated, separated by
gel electrophoresis and identified by nanoLC-
mass spectrometry
Pull down and Identification of proteins
associated with dnaK protein
Associated Proteins
D. Vulgaris mutant with Strep-tagged Dnak protein
was created in Judy Walls Laboratory at U of
Missouri, Columbia
75
All centralized and served from the VIMSS EDR
Janet Jacobsen and Keith Keller
76
What have we accomplished?

• Weve created an efficient, organized facility
  for rapid ecogenomics and phenomics for microbes.
• Weve demonstrated its utility in rapidly
  characterizing environmentally important stress
  response pathways in at least two microorganisms.
• And to do so weve developed a number of key
  technologies and had a number of spin-off and
  central scientific results.
• Weve developed a community resource in
  MicrobesOnline which is in fairly widespread use.

77
What else?

• We have developed a new theory of operon
  evolution.
• We are beginning to understand the different
  dynamics of creating and acquiring genes like
  HPKs for evolutionary adaptation to the
  environment.
• We have identified the critical signature genes
  for sulfate reduction and have begun to link
  their expression dynamics with different field
  stressors.
• We have developed fairly deep data on Dv and So
  stress responses in diverse conditions.
• We have the foundational infrastructure to
  support our future genetic and molecular complex
  studies and for our comparative functional
  genomics work to understand the evolution of
  regulatory networks.

78
Acknowledgements

• The Virtual Institute of Microbial Stress and
  Survival (VIMSS) Deduction of Stress Response
  Pathways in Metal/Radionuclide Reducing Microbes
• Carl Abulencia4,Eric Alm1,Gary Anderson1,Adam
  Arkin1 (APArkin_at_lbl.gov), Kelly Bender5, Sharon
  Borglin1,Eoin Brodie1,Swapnil Chhabra3, Steve van
  Dien6, Inna Dubchak1, Matthew Fields7, Sara
  Gaucher3, Jil Geller1, Masood Hadi3, Terry
  Hazen1, Qiang He2, Zhili He2, Hoi-Ying Holman1,
  Katherine Huang1, Rick Huang1, Janet Jacobsen1,
  Dominique Joyner1, Jay Keasling1, Keith Keller1,
  Martin Keller4, Aindrila Mukhopadhyay1, Morgan
  Price1, Joseph A. Ringbauer, Jr.5, Anup Singh3,
  David Stahl6, Sergey Stolyar6, Jun Sun4, Dorothea
  Thompson2, Christopher Walker6,Judy Wall5, Jing
  Wei4, Denise Wolf1, Denise Wyborski4, Huei-che
  Yen5, Grant Zane5, Jizhong Zhou2, Beto Zuniga6
• 1. E. O. Lawrence Berkeley National Laboratory,
  2. Oak Ridge National Laboratory, 3. Sandia
  National Laboratory, 4. Diversa, Inc., 5.
  University of Missouri, Columbia, 6. University
  of Washington, 7. Miami University
• Department of Energy, Office of Science.