Poulain – Biomonitoring: A Microbe’s Perspective – OGI Life Sciences and Mining Workshop 2014

from the environment to the genes and back

Biomonitoring: a microbe’s perspective

Alexandre Poulain Université d’Ottawa

Life Sciences and Mining Workshop Explore - Extract - Exchange

Vale for Living Lake Center, Sudbury, ON May 7th, 2014

?

Nutrients: C, H, O, N, S

Metals: Cu, Zn, Fe, Mn, U, Pb, Hg, As, Cr, Ag, Cd

Microbes are key players in biogeochemical cycles

Anoxic - Sulfidic - Oxic

Newman, 2010

Microbes are the oldest miners on

earth

Anbar, 2008

Microbes metabolic and

genetic diversity was shaped

through geological times

Falkowski, 2008

Organic chemicals

Inorganic chemicals

Chemotrophy Phototrophy

Chemicals Light

Energy Sources

(glucose, acetate, etc.) (H2, H2S, Fe2+, NH4+, etc.)

(glucose + O2 CO2 + H2O) (H2 + O2 H2O) (light)

Chemoorganotrophs Chemolithotrophs Phototrophs

Microbes exhibit incredible metabolic diversity eating and breathing minerals

Brock, 14th edition

Present over wide gradients of: - salinity - Temperature - pH - [O2] - [toxic metals]

Biomonitoring of microbes and microbial processes at several levels

Ribosomes

rRNAmRNA

DNA

Proteins

community FUNCTION

community STRUCTURE

feasibility

Functional relevance

Metabolites

A signature of the contamination can be observed in community structure

Mosher et al. 2012 Aqu. Microb. Ecol. Characterization of the Deltaproteobacteria in contaminated and uncontaminated stream sediments (Hg, U)

Microbial community function: what is active?

Metal resistance genes Cytochrome genes

Kang et al. 2013. FEMS Microb. Ecol. Functional gene array–based analysis of microbial communities in heavy metals-contaminated lake sediments

At the protein level

Mueller et al. 2011, Environ. Microb. Proteome changes in the initial bacterial colonist during ecological succession in an acid mine drainage biofilm community,

Mercury resistance operon

Boyd and Barkay, Frontiers in Microb., 2012

A B

[Mz+] total

insoluble

complexed with organic ligand

chlorocomplexes

Metal speciation matters

The total concentration of a metal is a poor predictor of its effect => what is/are the bioavailable form(s) of a metal?

• Bioanalytical tools applied to monitoring or to the discovery of microbe-metal interactions

!- Improving ecological risk

assessment of metal mining !

- Monitoring of effluents originating from metal mining activities

!- Monitoring of potential

remediation efforts

Biosensors

Sorensen, 2006

Broad range “lights on” biosensors

Luxilla biolamp, igem.org

Response/ Metal Specific biosensors

sensing element gene

reporter gene

promoter regionvector bacterial

chromosome

bacterium cell wall

(a) (b)

biosensor = reporting/sensing element + reporting gene

Response specific biosensor (e.g., oxidative stress)

H2O$+$O2$

Pro)oxidants$

superoxide$dismutase$(sodA)$

Pro)oxidants$

$ catalase$(katG)$

Promoters$of$genes$that$deal$with$H2O2$stress$(i.e.$katG)$

OxyR$

O2•)$ H2O2$

katGp)

Morin and Poulain, unpublished

Metal specific biosensors

Chemical or Stressor

Measurable signal

Sensing component

Reporter gene

Transcription

Reporter protein

Translation

merR o/p

R

lux C lux D lux A lux B

merR o/pR lux C lux D lux A lux B

merR R lux C lux D lux A lux BPHg

lux E

lux E

lux E

P

x

Arsenic Mercury

Harnessing metal detoxification pathways

Bacteria emit light in the presence of the metal

How does Hg bioavailability change while Hg and NOM are reaching equilibrium?

HgII HgII

reactive Hg species

HgII HgIIhydrophobic Hg species

time required to reach equilibrium

t=0

t=24h

Chiasson-Gould et al. 2014, ES&T

Mercury specific biosensor (e.g., toxic metal)

Indu

cibl

e lig

ht p

rodu

ctio

n (C

PS)

102

103

104bioavailable HgII, t = 0h

95% Confidence Band

[DOC] mg.L-1

0 10 20 30 40

Cons

titut

ive

light

pro

duct

ion

(CPS

)

104

105

106

6

Constitutive light expression t = 0h

[DOC] mg.L-1

0 10 20 30 40

Chiasson-Gould et al. 2014, ES&T

Metals for which there exists a known microbial sensory regulatory network

harnessing the power of microbial experimental evolution

• Provide a biologically relevant perspective to metal(loid) speciation in the environment

!• Can be used for early detection !

• Have increased sensitivity !

• Can be tailored to be site-specific (e.g., by using bacteria hosts exhibiting natural adaptation to a wide range of pH); very robust

!• More cost-effective than current practices: most of the cost is to

develop the biosensor, the cost of its application is minimal

Advantages

Feasibility and HQP training

The NSERC CREATE Mine of Knowledge is an interdisciplinary training program, in a multi-institutional setup designed to train highly qualified individuals capable of fulfilling the demands of the mining industry in the field of technological innovation and environmental management.

outcomes. This will require integrating frameworks forunderstanding what is possible from mutations (howthey change functional properties of proteins and howthese alterations propagate to physiological traits underselection) with how the interplay of selection and driftact in populations to shape the distribution of observedoutcomes.

References

Atwood, K.C., Schneider, L.K., and Ryan, F.J. (1951) Periodicselection in Escherichia coli. Proc Natl Acad Sci USA 37:146–155.

Chou, H.H., Berthet, J., and Marx, C.J. (2009) Fast growthincreases the selective advantage of a mutation arisingrecurrently during evolution under metal limitation. PLoSGenet 5: e1000652.

Counago, R., Chen, S., and Shamoo, Y. (2006) In vivomolecular evolution reveals biophysical origins of organis-mal fitness. Mol Cell 22: 441–449.

Elena, S.F., and Lenski, R.E. (2003) Evolution experimentswith microorganisms: the dynamics and genetic bases ofadaptation. Nat Rev Genet 4: 457–469.

Novick, A., and Horiuchi, T. (1961) Hyper-production of beta-galactosidase by Escherichia coli bacteria. Cold SpringHarb Symp Quant Biol 26: 239–245.

Stanier, R.Y. (1970) Some aspects of the biology of cells andtheir possible evolutionary significance. In Organizationand Control in Prokaryotic and Eukaryotic Cells. Charles,H.P., and Knight, B.C.J.G. (eds). Cambridge, UK: Cam-bridge University Press, pp. 1–38.

Vishniac, W., and Santer, M. (1957) The thiobacilli. BacteriolRev 21: 195–213.

Woese, C.R. (1987) Bacterial evolution. Microbiol Rev 51:221–271.

Where reductionism meets complexity: a call forgrowth in the study of non-growth

Dianne K. Newman (Email: [email protected]) andMaureen L. Coleman (Email: [email protected]),Divisions of Biology and Geological & Planetary Sciences,California Institute of Technology, Pasadena, CA, USAWith the advent of metagenomics, we have unprec-edented access to the genetic blueprint of the microbialworld. Yet as metagenomic databases keep growing, ourability to interpret the information contained within themhas not kept up. This conundrum arises from the fact thatwe cannot assign functions to the vast majority of theirgenes. As Jo Handelsman pointed out in a Crystal Ballpiece two years ago, ‘the glory of the last 50 years ofmicrobiology is founded, in large part, on genetic analysis’(Handelsman, 2009). Amen. Yet as enticing as the pros-pect of environmental genetics or ‘metagenetics’ seems,how can we hope to interpret the unchartered world ofenvironmental metagenomes when after more than a half-century of rigorous genetic and biochemical analyses, the

functions of roughly a quarter of the genes in Escherichiacoli – arguably the most well-studied organism on theplanet – are still unknown (Karp et al., 2007)? Where havewe gone wrong? Perhaps it is time to re-examine ourassumptions about how to assign gene functions in lightof lessons from the field.

Genetic analysis provides a powerful way to learn whatgenes are required for a phenotype of interest. Inventedby physicists, it is steeped in reductionism, permittingclear insights into biological phenomena through theapplication of simple logical rules. If we want to knowwhich genes are involved in a specific process for anorganism that is genetically tractable, we make mutantsand then design a screen or a selection that will permit usto assign a ‘yes’ or ‘no’ (or sometimes ‘partial’) level ofinvolvement to any given one. Thus, a key question wemust answer at the beginning is: what phenotype(s) do wecare about? What conditions are most relevant for ourfavourite model organism or favourite uncultured micro-bial community in the environment? Clearly, there is noone answer. Even the concept of ‘the environment’ ismisleading, because organisms reside in a dynamicworld, with changing physical, chemical and biologicalparameters. Given this complexity, is it even reasonableto think that reductionist approaches can be of value?Absolutely.

So where to begin? Recent work performed in CarolGross’ laboratory at UCSF provides an example. Theseinvestigators took a high-throughput approach to growinga collection of E. coli mutants under a battery of stressfulconditions to assign roles to genes whose functions wereunknown (Nichols et al., 2010). The idea was simple: ifmany conditions were tested, some of the unknown geneswere bound to be involved in growth on some of them.The results bore this out, and, comfortingly, strains con-taining mutations in genes that had previously beenshown to be involved in the response to particular stres-sors performed as expected. This type of approach is astart, and one can imagine doing this with any modelorganism for which a collection of mutant strains exist. Nodoubt such approaches will significantly reduce thenumber of genes of unknown functions in pure cultures.Yet, will they be enough to bring this number down tozero? Almost certainly not, as it is difficult, if not impos-sible, to capture every environmental variable in responseto which genes have evolved. Moreover, some genes maynot be useful at all. When we pluck an isolate from anatural community, we are capturing not an evolutionaryend-point but a work-in-progress, a snapshot of ongoinggene gain and loss complete with pieces that have not yetbeen, and may never be, integrated into the networks ofthe cell.

Even if we could wave a magic wand and somehowcapture all the relevant parameters and test them in

14 Crystal ball

© 2011 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology Reports, 3, 1–26

outcomes. This will require integrating frameworks forunderstanding what is possible from mutations (howthey change functional properties of proteins and howthese alterations propagate to physiological traits underselection) with how the interplay of selection and driftact in populations to shape the distribution of observedoutcomes.

References

Atwood, K.C., Schneider, L.K., and Ryan, F.J. (1951) Periodicselection in Escherichia coli. Proc Natl Acad Sci USA 37:146–155.

Chou, H.H., Berthet, J., and Marx, C.J. (2009) Fast growthincreases the selective advantage of a mutation arisingrecurrently during evolution under metal limitation. PLoSGenet 5: e1000652.

Counago, R., Chen, S., and Shamoo, Y. (2006) In vivomolecular evolution reveals biophysical origins of organis-mal fitness. Mol Cell 22: 441–449.

Elena, S.F., and Lenski, R.E. (2003) Evolution experimentswith microorganisms: the dynamics and genetic bases ofadaptation. Nat Rev Genet 4: 457–469.

Novick, A., and Horiuchi, T. (1961) Hyper-production of beta-galactosidase by Escherichia coli bacteria. Cold SpringHarb Symp Quant Biol 26: 239–245.

Stanier, R.Y. (1970) Some aspects of the biology of cells andtheir possible evolutionary significance. In Organizationand Control in Prokaryotic and Eukaryotic Cells. Charles,H.P., and Knight, B.C.J.G. (eds). Cambridge, UK: Cam-bridge University Press, pp. 1–38.

Vishniac, W., and Santer, M. (1957) The thiobacilli. BacteriolRev 21: 195–213.

Woese, C.R. (1987) Bacterial evolution. Microbiol Rev 51:221–271.

Where reductionism meets complexity: a call forgrowth in the study of non-growth

Dianne K. Newman (Email: [email protected]) andMaureen L. Coleman (Email: [email protected]),Divisions of Biology and Geological & Planetary Sciences,California Institute of Technology, Pasadena, CA, USAWith the advent of metagenomics, we have unprec-edented access to the genetic blueprint of the microbialworld. Yet as metagenomic databases keep growing, ourability to interpret the information contained within themhas not kept up. This conundrum arises from the fact thatwe cannot assign functions to the vast majority of theirgenes. As Jo Handelsman pointed out in a Crystal Ballpiece two years ago, ‘the glory of the last 50 years ofmicrobiology is founded, in large part, on genetic analysis’(Handelsman, 2009). Amen. Yet as enticing as the pros-pect of environmental genetics or ‘metagenetics’ seems,how can we hope to interpret the unchartered world ofenvironmental metagenomes when after more than a half-century of rigorous genetic and biochemical analyses, the

functions of roughly a quarter of the genes in Escherichiacoli – arguably the most well-studied organism on theplanet – are still unknown (Karp et al., 2007)? Where havewe gone wrong? Perhaps it is time to re-examine ourassumptions about how to assign gene functions in lightof lessons from the field.

Genetic analysis provides a powerful way to learn whatgenes are required for a phenotype of interest. Inventedby physicists, it is steeped in reductionism, permittingclear insights into biological phenomena through theapplication of simple logical rules. If we want to knowwhich genes are involved in a specific process for anorganism that is genetically tractable, we make mutantsand then design a screen or a selection that will permit usto assign a ‘yes’ or ‘no’ (or sometimes ‘partial’) level ofinvolvement to any given one. Thus, a key question wemust answer at the beginning is: what phenotype(s) do wecare about? What conditions are most relevant for ourfavourite model organism or favourite uncultured micro-bial community in the environment? Clearly, there is noone answer. Even the concept of ‘the environment’ ismisleading, because organisms reside in a dynamicworld, with changing physical, chemical and biologicalparameters. Given this complexity, is it even reasonableto think that reductionist approaches can be of value?Absolutely.

So where to begin? Recent work performed in CarolGross’ laboratory at UCSF provides an example. Theseinvestigators took a high-throughput approach to growinga collection of E. coli mutants under a battery of stressfulconditions to assign roles to genes whose functions wereunknown (Nichols et al., 2010). The idea was simple: ifmany conditions were tested, some of the unknown geneswere bound to be involved in growth on some of them.The results bore this out, and, comfortingly, strains con-taining mutations in genes that had previously beenshown to be involved in the response to particular stres-sors performed as expected. This type of approach is astart, and one can imagine doing this with any modelorganism for which a collection of mutant strains exist. Nodoubt such approaches will significantly reduce thenumber of genes of unknown functions in pure cultures.Yet, will they be enough to bring this number down tozero? Almost certainly not, as it is difficult, if not impos-sible, to capture every environmental variable in responseto which genes have evolved. Moreover, some genes maynot be useful at all. When we pluck an isolate from anatural community, we are capturing not an evolutionaryend-point but a work-in-progress, a snapshot of ongoinggene gain and loss complete with pieces that have not yetbeen, and may never be, integrated into the networks ofthe cell.

Even if we could wave a magic wand and somehowcapture all the relevant parameters and test them in

14 Crystal ball

© 2011 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology Reports, 3, 1–26

Acknowledgements