Peter Adriaens

1

Microbial Sensing and Control for Bioremediation
and Water Quality A Genesis for Discussion

• Peter Adriaens
• (with Cyndee Gruden and Steven Skerlos)
• Professor, Civil and Environmental Engineering
• Director, College of Engineering Environmental
  Council
• The University of Michigan at Ann Arbor

2
Geographical Location
3
Overview

• Principles of Distributed Microbial Detection
  Quantification (DMDQ), or sensing, for
  Environmental Monitoring Applications
• Candidate Microbial Detection Quantification
  Technologies for Distributed Measurements
• Advantages and Disadvantages
• General Technological Challenges
• Current Approaches to Demonstrate Microbial
  Cause-and-Effect Relationships During Remediation
• - Bachman Road Site
• - Passaic River Superfund Site
• Distributed Microbial Sensing in Large Systems
• UM Micro Integrated Flow Cytometer (MIFC)

4
Distributed Microbial Detection and
Quantification Motivation

• Convergence of Two Approaches and Thought
  Processes

5
1. Control System Model from Manufacturing
6
2. Layered Information Systems Approach to Site
Characterization and Interpretation
PVA Polytopic Vector Analysis An Environmental
Forensics Tool for Source Apportionment
7
Have you ever noticed

• that in modeling applications there is a misfit
  between complex hydrogeochemistry algorithms, and
  Monod model?
• that environmental engineering consultants are
  queasy about collecting microbiological
  information?
• that there is an imbalance between the sampling
  frequency and availability of organic/inorganic
  geochemical data on the one hand, and microbial
  data on the other hand?
• that microbial contributions to natural or
  engineered remediation systems are usually
  indirectly derived from hydrogeological modeling?
• that we tend to treat the natural environment as
  a microbially static system?

8
Spatial and Temporal Microbial MonitoringObserva
tions Coupled hydrological/ecological/metabolic
activity interactions at field scale

• Recharge months
• Redox zonation months
• Community dynamics months
• Biodegradation kinetics - months to years
• Microbial evolution/emergence - years

Short-term changes in redox zonation and
microbial dynamics may affect long-term intrinsic
bioremediation
Weakest link Microbial Information
9
Keeping Perspective Technology Transfer Across
Socio-Economic Boundaries (Veracruz 02)
? Much higher return on investment by eliminating
or limiting contaminant source contributions
(stressors) when compared to remediation
investment
10
Motivation for DMDQ

• Faster, Better, Cheaper More Often / More
  Places
• Technological Advances MEMS / NanoTech
• High Profile Microbial (Pathogens) Contamination
  Events
• Invigorated Interest in Bioterrorism Prevention
• Bioremediation Ecological/metabolic mapping for
  RU siting
• Water Reclamation Strategies Optimization of
  Resource Allocation
• Control System Paradigm

11
Microbial Detection Principles

• Detection is the perception that the microbial
  state of the system has changed.
• Target Variables of State
• Total biomass
• Ecological composition of community
• Distribution of ecological composition
• Viability state of the community or members
• Variables Influencing the Perception of State
• Matrix state (Physical, Chemical, Biological)
• Instrument state
• Sampling approach

12
Principles of Validity Microbial Detection

• Accuracy
• Precision
• Specificity
• Selectivity
• False positive
• Missed positive
• Sensitivity
• Absolute and Ambient

• Likelihood
• Stability
• Modelability
• Range
• Robustness
• Manufacturing
• Operator, Materials
• Ambient

13
Microbial Detection Statistical Analysis is
Critical

• Outside of pure cultures or total counts, it is
  difficult to assess the state of microbial
  systems with absolute certainty.
• However, if we follow the principles of validity,
  we can provide evidence that a state has changed
  with statistical certainty.
• A confidence interval approach is essential

14
PNA Molecular Beacon Approach M. parafortuitum
Steven J. Skerlos / Mechanical Engineering The
University of Michigan at Ann Arbor
15
Detection of Mycobacterium w/ PNA MBs via Flow
Cytometry

Typical Control
Steven J. Skerlos / Mechanical Engineering The
University of Michigan at Ann Arbor
16
While a conceptual definition of microbial
detection is quite straight forward
Moral
an operational definition is technology,
experiment, and application specific and must be
derived from statistical concepts of certainty.
A definition of microbial detection that is
useful and general is not possible!
and the same must be said for microbial
quantification!
17
Distributed Microbial Detection Requirements

• The distributed network must be appropriately
  established in terms of maximal coverage with
  minimal cost.

18
Candidate Technologies for DMDQ
Prof. Steven J. Skerlos / Mechanical
Engineering The University of Michigan at Ann
Arbor
19
DMDQ Technology Categories

• Growth Methods
• Viability Methods
• Artifact Methods
• Nucleic Acid Methods

20
Common Growth-Based Approaches

• Plate Counts
• Alternative Plate Count Methods
• Most Probable Number
• Membrane Filtration Counts
• ATP Bioluminescence
• Impedance/Conductivity
• Radiometric
• Colorimetric CO2 Monitoring
• Manometry

21
Growth-Based Technology for DMDQ Networks

• Major Benefits
• Reliable No major sample preparation
• Full automation possible
• Good specificity achievable (routine)

• Major Challenges
• Speed (still 6-24 hours)
• Size, cost, input volumes
• Bias/specificity for unknown samples
• Regeneration

MicroStar (Benford, MA)

• Needed Advancements
• Microfluidics, pumps, and handling
• Miniaturized thermal management
• Micro-fluorimeters, colorimeters, turbidimers,
  manometers, etc.

VITEK (Hazelwood, MO)
22
Common Viability-Based Approaches

• Direct Epifluorescent Filter Technique
• Electron Transfer Techniques
• Redox Indicator Reduction
• Microcalorimetry
• Fluorescence Flow Cytometry
• Membrane Scanning Fluorescent Cytometry

23
Activity Determinations - Redox Dyes
CTC (bright red)
Dissolved redox dye (colorless)

Active oxidative enzymes (respiration)

Reduced Redox Dye (colored PPT)
cell wall
ALIVE Metabolically Competent
Must be used in conjunction with cell wall stains
24
Viability-Based Technology for DMDQ Networks

• Major Benefits
• High sensitivity
• Fast and universal
• Full automation possible

• Major Challenges
• Size, cost, input volumes
• Reagent costs and stability
• Sample prep may be significant
• Not necessarily specific

FacsCalibur (Franklin Lakes, NJ)

• Needed Advancements
• Microfluidics, pumps, and handling
• Miniaturized thermal management
• Micro-optical systems
• Fast DAQ with Large Storage (FCM)

MicroCal (NorthHampton, MA)
25
Artifact-Based Approaches

• Gram stain
• ELISA methods
• Whole Cell Immunosensors
• colorimetric, DNA-based, gravimetric,
  electrochemical
• Latex Agglutination
• Fluorescence Conjugation
• Biosensors
• Fatty Acid Profiling - GC Analysis
• MALDI-TOF Mass Spectrometry (also immuno)
• FTIR/Raman Spectroscopy

Immunosensor Methods
26
Immunosensors for DMDQ Networks
(Sensortek, Germany)
(Eugenii Katz, Israel)
27
Artifact Technology for DMDQ Networks

• Major Benefits
• High sensitivity, very small
• Very fast, disposable
• No sample preparation
• potentially very robust

• Major Challenges
• Cost matrix interactions
• Antibody stability library development
• Regeneration cost reduction

• MALDI MS-TOF

• Needed Advancements
• Microfluidics, pumps, and handling
• Miniaturized MS, GC, FTIR, Raman
• Micro-Integrated optoelectronics
• Method for sequential observations.

• SPR Sensor GaTech

28
Nucleic Acid-Based Approaches

• General Nucleic Acid
• Polymerase Chain Reaction (PCR)
• Hybridization Colorimetry
• Fluoresence In Situ Hybridization (FISH)
• Fluoresence Hybridization Flow Cytometry
• 16S rRNA Sequencing Techniques
• Ribotyping

29
General Nucleic Acid (DNA/RNA) Staining
Syto Staining of Activated Bread Yeast
Also useful for live/dead
Acridine Orange Staining of E. Coli
30
Fluorescence In Situ Hybridization (FISH)

• Fix cells to slide
• Permeabilize cells
• Add probe to Slide

31
Ribotyping
32
Nucleic Acid-Based Technology for DMDQ

• Major Benefits
• Speed sensitivity versatility size
• Minimal carry-over (FCM)
• Low reagent volume disposability

• Major Challenges
• Sample preparation activities
• Sample carry-over (PCR) strain detection
• Sensitivity (200 cells for PCR)
• Matrix interactions calibration
• Regeneration cost reduction

• Needed Advancements
• Microfluidics, pumps, and handling improved
  thermal management controlled surfaces
• Micro-Integrated optoelectronics and DSP
• Non-disposable platforms.

33
State-of-the-Art in Environmental Monitoring
Off-line Microbial Analysis
34
Case 1 Passaic River Superfund Site (NJ)
Sample locations of 1995 data set
35
Conceptual Model Dioxin Patterns and Fate
Mixed source pattern
SOURCES linear mixing

2378T/PCDD 0.3
FATE Dechlorination shifts initial ratios
Some Congeners increase others decrease
Sample
2378T/PCDD 0.6
? Source patterns are modified by (biotic and
abiotic) reactive processes
36
Dioxin Dechlorination Signature Loadings
Distribution

• Dechlorination is prevalent throughout, but it is
  especially important in the downstream half.

5.1
3.0
1.8

• A number of hot-spots occur both upstream and
  downstream.

37
H2 Diffusion in Sediments
N2out
N2in
Stop Cock
Septa
All-Pyrex Sediment Column w/Ball Joint
Bubbler
H2/N2
Stainless Steel Syringe Epoxied to Luer Lock
5 Gal Water-Filled Bucket
Tedlar Bag
Background
Luer Lock (Fixed to Column)
Septa
27 gau. needle
H2/N2
38
CTC Activity in Response to H2 (nM)error bars
represent standard deviations of the mean
39
Flow Cytometric Analysis
Density plot of green fluorescence (FL1) as a
function of internal complexity (SSC) at (a) 0 nM
H2 and (b) 25 nM H2.
R1 represents total sediment-eluted bacteria.
The population R3, which has more internal
complexity, was detected only at 25nM H2. R3
composed less than 10 of the total number of
microorganisms in the sample, however, it
accounted for 22 of the total CTC activity of
R1. R3 was 80 CTC active.
40
Dechlorination of Aromatic Compounds
aModified from Albrecht et al., 1999 bA negative
value indicates a net gain in 2,3,7,8-TCDD during
sequential dechlorination from OCDD or HpCDD.
41
Case 2 Bachman Road Residential Wells
Lake Huron
Est. annual total VOC flux 20-35 kg/y.
Halo- respiration
SEAR
Plume A Plume B
42
Plot Layout
GW flow
43
(No Transcript)
44
Project Timeline
45
Extraction well contaminant profiles
(bioaugmentation)

• Increase in cDCE at t0 resulting from aquifer
  reduction and mixing between 18 and 0 (plus
  residual from inoculum)
• Near-complete conversion of chloroethenes to
  ethene after 43 days
• Variability in control plot presumably due to
  aquifer mixing

46
Extraction well contaminant profiles
(biostimulation)

• Three-month lag phase prior to onset of
  dechlorination
• At that time, rates similar to bioaugmentation
• Approx. 80 conversion

47
Total Community Analysis T-RFLP Fragment Change
in Groundwater
48
Qualitative Analysis of Chlororespirers

• Control/biostimulation
• Dehalococcoides nondetect until after t44d.
• Desulfuromonas consistent detect after t15d.
  (sporadic occurrence before)
• Bioaugmentation
• Dehalococcoides immediate detect after t1d.
• Desulfuromonas immediate detect after t1d.

49
Quantitative Analysis of Chlororespirers
16S rRNA Gene Copies per gram of
Aquifer Material
Dehalococcoides spp. Desulfuromonas spp.
Inoculum 1.0E06 Dehalococcoides per mL ? 1.0E12
total cells in 200 L Post-inoculation 1.8E07
Dehalococcoides/ g soil ? 1.7E16 total cells in
plot therefore increase on the order of
1.0E04 cells in bioaugmentation plot!
50
Distributed Microbial Sensing A Case for
Micro-Integrated Flow Cytometry
51
Why Flow Cytometry?

• Allows direct optical detection avoids
  growth-based detection problems
• High speed multi-parametric data acquisition and
  multi-variate data analysis
• Fast and reliable enumeration, and determination
  of basic cell functions such as reproductive
  ability, metabolic activity and membrane
  integrity,
• Characterization of the physiological state or
  degree of viability of bacteria

52
Flow Cytometry Concept
53
FACSCalibur Flow Cytometer.
FL1 515-345 nm
SSC 483-493 nm
FL2 564-606 nm
FL3 670/L nm
FL4 564-606 nm
Focusing Lens
Beam of Light
FSC 488/10 nm
488nm Blue Laser
Red Diode Laser 635 nm
54
What Can a Flow Cytometer Tell Us About a Cell?

• Its relative size (Forward Scatter-FSC)
• Its relative granularity or internal complexity
  (Side Scatter-SSC)
• Its relative fluorescence intensity (FL1, FL2,
  FL3, and FL4) for
• Total an target organisms (pathogens, inoculum,)
• Viability indicators (membrane integrity,
  membrane potential, )
• Enzyme activity (hydrogenase, dehalogenase,
  other)

55
Motivation for Micro-Integrated Flow
CytometerClinical and industrial fluids (health
hazard applications)
56
Micro-Integrated Flow Cytometer (MIFC)

• Goal is to develop a hand-held flow cytometer
  with two-color excitation/detection capability.
• Miniaturized and low cost observation cell,
  excitation sources, and photodetectors have been
  proven in MIFC.
• Simultaneous two color detection of biological
  populations has been demonstrated (440nm and
  635nm excitation).
• Future work will include refining the optical
  system, adding on-device sample preparation, and
  miniaturizing the packaging and of the system
  components.

MIFC Schematic
Detection of Saccharomyces cerevisiae
Integrated Optics
57
Challenges to Micro Integrated Flow Cytometer
Observation Cell From Quartz to PDMS From
Water to Air
Opto-Electronics From Lasers to LEDs From
Vacuum Tubes to PINs
http//www.engin.umich.edu/news/flowcytometer/inde
x.html
58
Summary and Conclusions

• MDQ calls for close attention to validity and
  certainty by statistically accounting for
  instrument, ambient, sampling, and experimental
  variation.
• Distributed MDQ requires a well-designed network
  based upon technology that is low cost,
  automated, compact, fast, and versatile.
• Growth, viability, artifact, and nucleic acid
  methods all have application to DMDQ in potable
  water systems, but have not been applied in
  environmental monitoring network systems.
• Artifact and nucleic acid approaches are most
  amenable to DMDQ networks, but the true potential
  of DMDQ will require many years to achieve.