Compartmentalisation in Droplets for

1
Compartmentalisation in Droplets for Directed
Evolution Andrew Griffiths Institut de
Science et dIngénierie Supramoléculaires
(ISIS), Strasbourg, France
2
Institut de Science et dIngénierie
Supramoléculaires (ISIS) - Strasbourg
Founded by Prof. J.-M. Lehn (Nobel Prize for
Chemistry, 1987) in 1997 Research Institute of
the Université Louis Pasteur. Financed by the
Ministère de la Recherche and the
CNRS Multidisciplinary research at the interface
between physics, chemistry and biology 140
people - 28 nationalities
Academic Laboratories Supramolecular chemistry -
J.-M. Lehn Organic and Bioorganic chemistry - N.
Winssinger Biophysical chemistry - M.
Karplus Nanochemistry - P. Samori Nanostructures
- T. Ebbesen Chemistry and recognition of
biomolecules - S. Ladame Chemical Biology - A.
Griffiths
Industrial Satellite Laboratories Present BASF Br
uker Biospin Alsachim OxyPlus
Past CAREX Prestwick Chemical AC
Immune FaustPharmaceuticals
3
Darwinian Evolutionary Systems Can evolutionary
systems be used in the laboratory provide insight
into biosilification
By cycles of Mutation Recombination Selection
4
Directed Evolution Using Darwinian principles to
evolve new molecules in the laboratory
Mutation
Selection

• Objectives
• To investigate the mechanism of evolution
• To evolve molecules of real practical utility

5
Requirements for Evolutionary Systems Linkage
between genotype and phenotype
It is essential to have linkage
between Genotype - a nucleic acid which can be
replicated and Phenotype - a selectable
activity (e.g. a binding or catalytic activity)
6
In NatureGenotype and phenotype linkage by
compartmentalisation in cells
2 µm
Bacteria (E. coli)
7
In Nature Genotype and phenotype linkage by
compartmentalisation in cells
2 µm
Bacteria (E. coli)

• Cells keep together
• Genes (DNA)
• The RNAs and proteins encoded by the genes
• The products of the catalytic activities of the
  RNAs and proteins

8
In Vivo Selection for Directed EvolutionGenotype
and phenotype linkage by compartmentalisation in
cells

• Advantage
• Can select for binding, catalysis and
• regulation

2 µm
Bacteria (E. coli)

• Cells keep together
• Genes (DNA)
• The RNAs and proteins encoded by the genes
• The products of the catalytic activities of the
  RNAs and proteins

9
In Vivo Selection for Directed EvolutionGenotype
and phenotype linkage by compartmentalisation in
cells

• Advantage
• Can select for binding, catalysis and
• regulation
• Disadvantage
• Limited range of activities selectable -
• e.g. complementing auxotrophic mutations
• Cells are very complicated genetically and
• biochemically - often contourn applied
• selection pressure

2 µm
Bacteria (E. coli)

• Cells keep together
• Genes (DNA)
• The RNAs and proteins encoded by the genes
• The products of the catalytic activities of the
  RNAs and proteins

10
In Vitro Compartmentalisation (IVC)Genotype and
phenotype linkage by compartmentalisation in
microdroplets in water-in-oil emulsions
2 µm
Emulsion
Bacteria (E. coli)

• Aqueous microdroplets have
• Average 2.6 µm diameter (same size as
  bacterium)
• 5 fl volume
• 50 µl dispersed into 1010 droplets

11
In Vitro Compartmentalisation (IVC)Genotype and
phenotype linkage by compartmentalisation in
microdroplets in water-in-oil emulsions

• Advantage
• Can select for binding, catalysis and
• regulation
• Wide range of activities selectable
• Microdroplets are very simple genetically
• and biochemically - easier to apply
• selection pressure

2 µm
Emulsion

• Aqueous microdroplets have
• Average 2.6 µm diameter (same size as
  bacterium)
• 5 fl volume
• 50 µl dispersed into 1010 droplets

12
Selection of DNA methyltransferases by In Vitro
Compartmentalisation (IVC)
13
Selecting for Novel Sequence Specificity Directed
Evolution of M.HaeIII to Methylate AGCC
14
Selecting for Novel Sequence Specificity Directed
Evolution of M.HaeIII to Methylate AGCC
15
Selecting for Novel Sequence Specificity Directed
Evolution of M.HaeIII to Methylate AGCC
16
Selection for AGCC Methylation Methylation of
AGCC Increases With Selection
17
Catalytic Efficiency and Sequence
Specificity Mutant T29 methylates AGCC much more
efficiently than wt M.HaeIII
T29 has 700-fold improvement in catalytic
efficiency (kcat/KM) on AGCC Due to 5-fold
increase in kcat and 140-fold decrease in KM
18
Catalytic Efficiency and Sequence
Specificity Mutant T29 methylates AGCC much more
efficiently than wt M.HaeIII
19
Catalytic Efficiency and Sequence
Specificity Mutant T29 also methylates GGCC more
efficiently than wt M.HaeIII
20
Catalytic Efficiency and Sequence
Specificity Mutant T29 efficiently methylates
AGCC and has radically altered sequence
specificity
21
Catalytic Efficiency and Sequence
Specificity Mutant T29 efficiently methylates
AGCC and has radically altered sequence
specificity
T29 has 700-fold improvement in catalytic
efficiency (kcat/KM) on AGCC
22
Catalytic Efficiency and Sequence
Specificity Mutant T29 efficiently methylates
AGCC and has radically altered sequence
specificity
T29 has 700-fold improvement in catalytic
efficiency (kcat/KM) on AGCC Due to 5-fold
increase in kcat and 140-fold decrease in KM
23
DNA Recognition by M.HaeIII Base specific
contacts between M.HaeIII and DNA
24
DNA recognition by M.HaeIII Interaction with the
first base of the target GGCC
25
DNA recognition by M.HaeIII Mutations in T29
variant
26
Selection of Enzymes by In Vitro
Compartmentalisation (IVC)
27
Selection of Enzymes using IVC Early
demonstrations of selection of proteins for
catalysis
Selection of DNA-methyltransferases with novel
sequence specificities
28
Selection of Enzymes using IVC Early
demonstrations of selection of proteins for
catalysis
Selection of DNA-methyltransferases with novel
sequence specificities Selection of one of the
most efficient enzymes ever described - a
hydrolase 10 times faster than acetylcholinesteras
e
Insecticides
29
Selection of Enzymes using IVC Early
demonstrations of selection of proteins for
catalysis
Selection of DNA-methyltransferases with novel
sequence specificities Selection of one of the
most efficient enzymes ever described - a
hydrolase 10 times faster than acetylcholinesteras
e
Insecticides
Chemical Warfare Agents
30
Wild-type PTE Is already a very efficient enzyme
Very efficient enzymes are diffusion limited kcat
/KM reaches maximum of 3 x 108 M-1s-1
kcat
kcat /KM
(s-1 )
(M-1s-1 )
PTE-wild-type
2.3 x 103
9.9 x 107
31
PTE-h5 Has a kcat 60-fold higher than wild-type
PTE
Very efficient enzymes are diffusion limited kcat
/KM reaches maximum of 3 x 108 M-1s-1
kcat
kcat /KM
(s-1 )
(M-1s-1 )
PTE-wild-type
2.3 x 103
9.9 x 107
PTE-h5
1.4 x 105
1.8 x 108
32
Selection of RNAs using IVC Selecting ribozymes
catalysing the Diels-Alder cycloaddition
Diene
Dienophile
Cyclohexene Product
33
Selection of RNAs using IVC Selecting ribozymes
catalysing the Diels-Alder cycloaddition
Diene
Dienophile
Cyclohexene Product
Anthracene
Cyclic Product
Maleimide
34
Selection of Diels-Alderase Ribozyme Using
IVC Selections are for catalysis in trans
T7
RNA (Inactive)
Emulsion containing single gene per microdroplet
35
Selection of Diels-Alderase Ribozyme Using
IVC Selections are for catalysis in trans
T7
RNA (Inactive)
Emulsion containing single gene per microdroplet
36
Selection of Diels-Alderase Ribozyme Using
IVC Selections are for catalysis in trans
T7
RNA (Active Ribozyme)
Emulsion containing single gene per microdroplet
37
Selection of Diels-Alderase Ribozyme Using
IVC Selections can be under multiple-turnover
conditions
T7
RNA (Active Ribozyme)
Emulsion containing single gene per microdroplet
38
Diels-Alderase Ribozymes IVC yields ribozymes
which all catalyse multiple turnover reactions
in trans
39
Selection of Enzymes using Fluorogenic
Substrates Allows selection for a wide range of
reactions
40
Selection of Enzymes using Fluorogenic
Substrates Double emulsions for flow sorting
41
Selection of Enzymes using Fluorogenic
Substrates Double emulsions for flow sorting
42
Flow Sorting of Microdroplets Using fluorogenic
substrates to sort microdroplets
43
Evolving b-galactosidase (Ebg) In Vitro In vivo
only two mutations confer the ability to grow on
lactose
wt kcat/KM 0.0092
44
Evolving b-galactosidase (Ebg) In Vitro In vivo
only two mutations confer the ability to grow on
lactose
wt kcat/KM 0.0092
ebgA Asp-92 Asn kcat/KM 0.37
45
Evolving b-galactosidase (Ebg) In Vitro In vivo
only two mutations confer the ability to grow on
lactose
wt kcat/KM 0.0092
ebgA Asp-92 Asn kcat/KM 0.37
ebgA Trp-977 Cys kcat/KM 0.091
46
Evolving b-galactosidase (Ebg) In Vitro In vivo
only two mutations confer the ability to grow on
lactose
wt kcat/KM 0.0092
ebgA Asp-92 Asn kcat/KM 0.37
ebgA Trp-977 Cys kcat/KM 0.091
ebgA Asp-92 Asn, Trp-977 Cys kcat/KM
4.2
47
Evolving b-galactosidase (Ebg) In Vitro In vivo
only two mutations confer the ability to grow on
lactose
wt kcat/KM 0.0092
Many other mutations
ebgA Asp-92 Asn kcat/KM 0.37
ebgA Trp-977 Cys kcat/KM 0.091
ebgA Asp-92 Asn, Trp-977 Cys kcat/KM
4.2
48
Evolving b-galactosidase (Ebg) In Vitro In vivo
only two mutations confer the ability to grow on
lactose
wt kcat/KM 0.0092
Many other mutations
ebgA Asp-92 Asn kcat/KM 0.37
ebgA Trp-977 Cys kcat/KM 0.091
ebgA Asp-92 Asn, Trp-977 Cys kcat/KM
4.2
?
LacZ kcat/KM 124
49
Evolving b-galactosidase (Ebg) In Vitro In vivo
only two mutations confer the ability to grow on
lactose
wt kcat/KM 0.0092
Many other mutations
ebgA Asp-92 Asn kcat/KM 0.37
ebgA Trp-977 Cys kcat/KM 0.091
ebgA Asp-92 Asn, Trp-977 Cys kcat/KM
4.2
?
LacZ kcat/KM 124
?
Better than LacZ
50
Wide Dynamic Range Much greater than for in vivo
selection systems
51
Selection of Proteins and RNAs by IVC Selection
for catalysis, binding and regulation
Proteins Catalysis
DNA-methyltransferases Restriction
Endonucleases DNA polymerases
Phosphotriesterases Biotin protein ligase
b-galactosidases Thiolactonases Binding
Peptides Antibodies Regulation DNA-nuclease
inhibitors
RNAs Catalysis Diels-Alderases RNA
ligases RNA polymerase
52
Selection of Proteins and RNAs by IVC Selection
for catalysis, binding and regulation
Proteins Catalysis
DNA-methyltransferases Restriction
Endonucleases DNA polymerases
Phosphotriesterases Biotin protein ligase
b-galactosidases Thiolactonases Binding
Peptides Antibodies Regulation DNA-nuclease
inhibitors
RNAs Catalysis Diels-Alderases RNA
ligases RNA polymerase
DNA methlytransferase with a 670-fold increase in
catalytic efficiency on a novel target site
(AGCC) The new catalytic activity is greater
than the wild-type enzyme on the original target
site (GGCC)
53
Selection of Proteins and RNAs by IVC Selection
for catalysis, binding and regulation
Proteins Catalysis
DNA-methyltransferases Restriction
Endonucleases DNA polymerases
Phosphotriesterases Biotin protein ligase
b-galactosidases Thiolactonases Binding
Peptides Antibodies Regulation DNA-nuclease
inhibitors
RNAs Catalysis Diels-Alderases RNA
ligases RNA polymerase
A phosphotriesterase variant with a kcat 63
times higher than the already very efficient
wild-type enzyme The new enzyme is one of the
most efficient enzymes ever described kcat of 105
s-1 kcat/KM 108 M-1 s-1
54
Selection of Proteins and RNAs by IVC Selection
for catalysis, binding and regulation
Proteins Catalysis
DNA-methyltransferases Restriction
Endonucleases DNA polymerases
Phosphotriesterases Biotin protein ligase
b-galactosidases Thiolactonases Binding
Peptides Antibodies Regulation DNA-nuclease
inhibitors
RNAs Catalysis Diels-Alderases RNA
ligases RNA polymerase
Novel b-galactosidases with up to 1700-fold
improvements in catalytic efficiency Very
different spectrum of mutations compared to those
found in earlier in vivo selections in E. coli
55
Selection of Proteins and RNAs by IVC Selection
for catalysis, binding and regulation
Proteins Catalysis
DNA-methyltransferases Restriction
Endonucleases DNA polymerases
Phosphotriesterases Biotin protein ligase
b-galactosidases Thiolactonases Binding
Peptides Antibodies Regulation DNA-nuclease
inhibitors
RNAs Catalysis Diels-Alderases RNA
ligases RNA polymerase
Novel Diels-Alderase ribozymes All capable of
bimolecular, multiple turnover catalysis in
trans This is frequently not the case with
ribosomes selected by SELEX, which only selects
for single-turnover intra-molecular reactions
56
Emulsion PCR Massively parallel PCR in
emulsions
57
Emulsion PCR Massively parallel PCR in emulsions
Water
Oil
Millions of identical DNA fragments per bead
58
Sequencing By Synthesis
454 Picotiter Plate Sequencing Pyrosequencing
Repeated dNTP Flow Sequence
G
C
T
A
T
Anneal Primer
PPi
59
Sequencing By Synthesis
454 Picotiter Plate Sequencing Pyrosequencing
Repeated dNTP Flow Sequence
G
C
T
A
A A T C G G C A T G
C T A A A A G T C A
T
G
A
C
T
T
T
A
G
C
A
T
G
C
C
G
A
T
T
T
Anneal Primer
Process continues until user-defined number of
nucleotide flow cycles are completed.
Sepharose bead carrying millions of copies of a
single clonal fragment
60
Emulsion PCR on Beads Generates template DNA for
454/Roche Genome Sequencer 20 and
Agencourt/Applied Biosystems SOLiD systems
A whole bacterial genome can be sequenced on a
single plate in 4 hours
61
Emulsion PCR on Beads Generates template DNA for
454/Roche Genome Sequencer 20 and
Agencourt/Applied Biosystems SOLiD systems
A whole bacterial genome can be sequenced on a
single plate in 4 hours Used to sequence the
genome of Jim Watson! 2 million and 2 months
- compared to 3 billion and over 10 years for
the Human Genome Projects reference genome
Our IVC patents are licensed for the preparation
of template DNA for sequencing by 454 Life
Sciences (now Roche)
62
Current Applications of Emulsion PCR
DNA preparation for sequencing (on beads)
454 system - pyrosequencing (Roche, Genome
Sequencer 100) Church system - sequencing by
ligation (Agencourt/Applied Biosystems,
SOLiD) PCR amplification of complex gene
libraries Reducing bias towards short
fragment Reducing artefacts due to
recombination Quantitation of (rare) allelic
variants BEAMing Haplotyping Screening for
transcription factor targets
63
Polydispersity of EmulsionsCreates problems for
doing really quantitative experiments
64
Digital Microfluidics Droplet-based
Microfluidics
65
Making Emulsions by Microfluidics Generates
highly monodisperse emulsions (lt3 polydispersity)
Water

Oil
Oil
20,000 droplets per second
Started to collaborate with the group of David
Weitz at Harvard to develop droplet-based
Digital microfluidics systems for biology and
chemistry
66
Loading Module
Water

Oil
Oil
20,000 droplets per second
Each droplet functions like the well of a
microtitre plate but with a volume of 1 nl to 1
pl - a thousand to a million times smaller volume
than in a microtitre plate well
67
RainDance Technologies Guilford, CT
68
Founders and Scientific Advisory Board

• Founders
• Prof. Andrew D. Griffiths ISIS (Strasbourg,
  France)
• Prof. Jerome Bibette ESPCI (Paris, France)
• Darren R. Link, PhD RainDance Technologies, Inc.
• Prof. David A. Weitz Harvard University
• Jonathan M. Rothberg, PhD Founder 454, CuraGen
  Corp.
• Scientific Advisory Board
• Sir Aaron Klug Sir Aaron received the undivided
  Nobel Prize in Chemistry in 1982, for his
  development of crystallographic electron
  microscopy and the structural elucidation of
  protein-nucleic acid complexes of biological
  importance.
• Jean-Marie Lehn Prof. Lehn received the Nobel
  Prize in Chemistry in 1987 with Cram and Pedersen
  for his work on the chemical basis of molecular
  recognition.
• Richard J. Roberts Dr. Roberts shared the Nobel
  Prize in 1993 for the discovery of split genes.
  Dr. Roberts is Research Director, New England
  Biolabs, Inc.
• Sir Greg Winter Sir Greg is joint Head of the
  Division of Protein and Nucleic Acids Chemistry
  at the MRC, a recognized pioneer in antibody
  engineering and founder of Cambridge Antibody
  Technology and Domantis.

69
Miniaturisation in the Electronics Industry

• Miniaturization Enables Fast Processing Speeds
  (3.2 billion cycles per second)
• Integration of Multiple Functions on One Device
  (42 million transistors)
• Scaleable

70
Miniaturisation of Biology and ChemistryOnly
1000-fold reduction in volume
Test tube (ml) 1000s of assays per week
Microtitre plate (µl) 1000s of assays per hour
To reduce volumes below 1 µl requires completely
new technology
71
Screening in Microdroplets
Traditional Screening
Screening in Microdroplets
Robots 50 m long, weighing 200 tonnes
Microfluidic chips 10 cm long, weghing 10 g
Each assay is performed in a microtitre plate
well 1-200 µl
Each assay is performed in a microdroplet 1nl-1pl
72
Evolution of Fluid Handling

• Each chemical in a separate tube (ml)
• Hand manipulation
• Handwritten labeling
• 1000s of assays per week
• Each chemical in well of a plate (µl)
• Robotic handling
• Barcode labeling
• 1000s of assays per hour
• Each chemical in microscopic droplet (nl to pl)
• Microfluidic handling on chip
• Fluorescent labels (Barcodes)
• 1,000s of assays per second

Labels
Test tubes
Barcodes
Microtitre plates
Fluorescent Labels
NanoReactors
73
Chip Layout

• The standard chip interface includes
• 18 fluidic connections
• 2 coalesce connections
• 1 sort connection
• 1 non-volatile memory / serial number connection
  (in the carrier)
• 70mm x 75mm PDMS

74
Loading Module

• Less than 1.5 polydispersity
• Up to 20,000 per second
• 5 to 100 ?m diameter drops
• 1 nl to 1 pl volume drops

75
Loading Module

• Less than 1.5 polydispersity
• Up to 20,000 per second
• 5 to 100 ?m diameter drops
• 1 nl to 1 pl volume drops
• Each droplet is like a well of a microtitre plate

76
Loading Module

• Liquids
• Solutions
• Cells, Beads
• Gene libraries
• Each droplet can have a single molecule, bead or
  cell
• Precise, uniform
• Reproducible
• Up to 20,000 per second
• 5 to 100 ?m drops

20,000 droplets per second
77
Formulate Module

• Multiple components
• Dynamic control of input ratios
• drop-by-drop controlled formulation

78
Detection Module (Fluorescence)
5,000 NanoReactors per Second

• Up to 20,000 events per second
• Sensitive lt10,000 molecules
• Potential for lt1,000 molecules per droplet
• Quantitative

79
Splitting Module
80
Splitting Module

• Symmetrical
• Asymmetrical
• Precise
• Allows different assays on same sample

81
Combining Module - Charged Droplets

• Combine reagents
• Start or stop reactions
• Precise
• Loss-free
• Enable multi-step processes
• Up to 5,000 per second

400 Droplets per Second
82
Sequential Adder Module - Uncharged Droplets

• All pair-wise combinations
• Enable multi-step and combinatorial processes

83
Interdigitate and Coalesce Modules - Uncharged
Droplets

• Large droplets move slower than small droplets,
  hence droplet pairs
• Process Pre-Made Libraries
• Add Reagents to Uncharged Droplets

-
-


E
84
Interdigitate and Coalesce Modules - Uncharged
Droplets
-
-


E
85
Size Selection Module Removing unfused droplets
86
Detection Hardware Fluorescence Intensity (FI),
Fluorescence Polarization (FP)
Droplet Sample
87
Mixing Module

• Controlled rate
• Precise initiation
• Enables stop flow kinetics
• Repeatable

400 NanoReactors per Second
88
Delay Module Short Incubations

• Preservation of order
• Precise incubation
• ms to 1 hour
• Reaction kinetics

1,200 NanoReactors per Second
89
Store Module Long Incubations

• Emulsions can be stored for weeks at room
  temperature
• Or stored for months (probably years) frozen in
  liquid nitrogen
• There is no exchange of compounds between
  droplets due to the use of perfluorocarbon
  carrier fluids

90
Re-Loading Module

• Emulsions can be re-injected back onto a
  microfluidic device

2,500 Microdroplets per Second
91
No Visible Droplet Cross-Talk
Pure Dye Emulsion
Mixed Emulsion (11000)
After several days, the intensity does not
diminish even with just 1 in 1000 droplets
containing dye (Alexa 488).
92
Sorting Module

• Fast -up to 10,000 events per second
• Precise just the drops of interest

93
Building an Optical Sorter
Detect
Direct
Load

Goes beyond cell sorting with a standard FACS
machine Droplets allow beads, cells, chemicals,
reactions and their products to be sorted - as
individual packages
94
Flow Sorter - Sorting Using Electrical Fields

• No mechanical valves
• Allows sorting at up to 10,000 droplets s-1

95
Flow Sorter - Sorting Using Electrical Fields

• No mechanical valves
• Allows sorting at up to 10,000 droplets s-1

96
Flow Sorter - Sorting Using Electrical Fields

• No mechanical valves
• Allows sorting at up to 10,000 droplets s-1

97
Dielectrophoretic Sorting - Uncharged Droplets
Active Sorting
Population 1
Population 2
Initial droplet population

• In the absence of an electric field, droplets
  all enter the top channel
• When an electric field is applied droplets
  enter the bottom channel

98
Sorting Protocol PerformanceEnrichment factor of
2 x 104
Before Sorting
2 Bright Drops
288736 Dim Drops
3 bright in 1000 droplets
906 Bright Drops
1500 Dim Drops
One Second of Raw Data 1500 drops
Histogram of Data
After Sorting
2.5x106 droplets sorted
7965 Bright Drops
1 in 105 false positives
180 Dim Drops
One Second of Raw Data
Histogram of Data
High-throughput and low error sorting of droplets
enabled
99
Histograms of Sorted Populations
Waste
437,621 droplets
47 droplets (0.01)

• 10,000 fold enrichment
• Errors come from the startup and stop

Collected
13 droplets (0.02)
52,278 droplets
100
Breaking EmulsionsEmulsions can be broken on or
off-chip
On-chip electrocoalescence
101
Digital Microfluidic Modules
Load
Combine
Sequential Add
Formulate
Timing
Split
Mix
Heat/Cool
Detect
Direct
Store
Re-Loading

• In the semiconductor world, electronic gates
  control the flow of electrons
• Microfluidic modules control the flow, mixing,
  heating, cooling, identification, detection and
  storage of droplets

102
Digital Microfluidic Modules
Load
Combine
Sequential Add
Formulate
Load
Formulate
Combine
Sequential Add
Timing
Mix
Timing
Split
Mix
Heat/Cool
Split
Heat/Cool
Detect
Direct
Store
Re-Load
Detect
Direct
Store
Re-Loading
A wide range of modules enable many different
applications
103
Integrating Digital Microfluidic Modules on a Chip

Molded Silicone Chip
Chips have no moving parts - they simply control
the movement of droplets
104
Digital Microfluidic Modules on a Chip
Sequential Add
Protein Chip
Nucleic Acid Chip
Load
Mix
Heat/Cool
Combine
Direct
Chip Masters
Timing
Detect
Split
Store
Re-Loading
105
Chip Design
2 Piece Hard Shell Protection and
alignment
Memory Chip Serial number and
History
Molded PDMS Silicone Contains fluidic
interconnects, microchannels and electrical
contacts
Glass backing Provides rigidity and
good optical access to channels
106
The Personal Laboratory System (PLS)A flexible
platform for running the microfluidics chips

• Reference version ready to ship Q2 2007
• Full version ready to ship Q1 2008

107
The Personal Laboratory System (PLS)A flexible
platform for running the microfluidics chips

• Syringe pump banks
• Oil pumps (5)
• User reagents (4)
• Sort and store (2)

• Data acquisition and processing
• 10 channels 200,000 reads/second on each channel
• Real time processing of data

• Lasers and conditioning optics
• 3 lasers (375nm or 405nm, 488 nm and 555 nm)
• 10 detection PMTs (7 colors, 2 polarization
  states, backscatter)

• Chip dock
• Plug and play electrical and fluidic connections

108
The Personal Laboratory System (PLS)A flexible
platform for running the microfluidics chips
Reagent Inputs/ Outputs
Molded Silicone Chip
Graphical User Interface
Lasers (3 colors)
Chip Dock
Detectors
Personal Laboratory System (PLS) runs Digital
Microfluidics Chips
109
RainDance Technologies Business Plan

• Application Development collaborations
• Port over existing applications (kinase
  profiling, RNAi screening)
• Develop new applications (industrial enzyme
  evolution)
• Demonstrate unique capabilities in cancer
  diagnostics
• RDT Inside Partnerships
• Dedicated co-branded instruments (Genotyping,
  Expression)
• Dedicated front-ends to existing product lines
  (Sample Preparation)
• RDT Direct Sales of the Personal Laboratory
  System
• Suite of turnkey application kits
• Range of Chips Protocols available for broad
  range of applications
• Custom Chip design services and Application
  development support

110
Opportunities in Droplet-Based Biology

• Cells in Droplets
• Analysis of Large Libraries of Cells
• Analysis of Small Numbers of Cells
• Many Cell Types
• Bacteria
• Yeast
• Drosophila
• Mammalian
• Secreted Factor Analysis
• Cell Phenotype Analysis
• Cell BioMarker Analysis
• Proteins in Droplets
• Antibody-Antigen Interactions
• Protein-Protein Interactions
• Small Molecule-Protein Interactions
• in vitro Transcription-Translation
• Enzymatic Assays

Nucleic Acids in Droplets
111
Digital Microfluidics Applications Cell-Sorting
for Protein Engineering and Directed Evolution
112
Directed Evolution of Enzymes
Molecular biology methods allow for generation of
large libraries of mutant enzymes
Primary screening methods are low-throughput, low
accuracy, and have high reagent costs
Colony Halo Primary Assay
Microtitre Dish Secondary Assay
An accurate high-throughput assay is needed
113
Microfluidic Devices for Directed
EvolutionHigh-throughput screening of enzymes
expressed in vivo by cells
114
Cells in Microdroplets

• The carrier fluid is a perfluorocarbon
• Perfluorocarbons can dissolve more than 20 times
  the amount of O2, and three times the amount of
  CO2, than water.
• A mouse can survive undamaged for more than an
  hour immersed in oxygenated perfluorocarbon
• They are used clinically both for liquid
  ventilation and as blood substitutes
• They facilitate respiratory gas-delivery to both
  prokaryotic and eukaryotic cells in culture

Cells remain viable for days in droplets in
perfluorocarbon carrier fluid
115
Cells in MicrodropletsCells can grow in droplets
Yeast growing over 24 hours
116
Cells in NanoReactorsCells can grow in droplets
Yeast growing in droplets
117
Cells in NanoReactorsCells can grow in droplets

• Yeast cell proliferation in 5 hours

118
Cells in Microdroplets
Jurkat and HEK2932T cells remain fully viable for
3 days in droplets in perfluorocarbon carrier
fluid Cells can be recovered and cultivated
after emulsions broken gt95 viability
119
Cells in Microdroplets
Cells in droplets follow a Poisson distribution
120
Cells in Microdroplets Cells in 660 pl droplets
in FC40 containing 0.5 AEH19 (FSH-DMP)
Jurkat and HEK2932T cells remain fully viable for
gt3 days in 660 pl droplets in perfluorocarbon
carrier fluid
121
Cells in Microdroplets Cells in 660 pl droplets
in FC40 containing 0.5 AEH19 (FSH-DMP)
Jurkat and HEK2932T cells remain fully viable for
gt3 days in 660 pl droplets in perfluorocarbon
carrier fluid
122
Multicellular organisms in MicrodropletsC.
elegans can undergo a complete life cycle in
droplets
Adult worms 5 days after encapsulation of eggs
C. elegans in 660 nl droplets
123
Sorting Cells Secreting a Protease
Droplet Generation
Re-load Droplets
Sort Droplets
Incubation
Filter(10um)
Oil
Sorting Electrode
Oil
Collect
Keep
bacteria
Droplets
Re-inject Droplets
substrate
Waste
Laser focus
Oil
Oil
Example of trace showing sorting voltage trigger
Keep Drop
Voltage Applied
Waste Drops
Threshold-based sorting enabled
124
Sorting Cells Secreting a Protease
Biological Growth
Single GFP-expressing Bacteria
Green Filter Shows Discrete Bacteria
Replication
Overnight
Example Overlay of Green and Red Images
(37oC)
Fluorescent Protease Assay
Red Filter Shows Secreted Protease Assay
Droplet environment allows for bacterial growth
and secreted protease assay
125
Sorting Cells Secreting a Protease
Cell-Generated Protease Assay
Sort/recover an enriched population of
protease-secreting bacteria
Step 1
Step 2
Substrate
Cell Library
Incubate Reaction
Generate Emulsion
Detect Reaction
Re-inject Emulsion
a) 2 member
Sort/Recover
b) 3-5 member
c) 10x member
Chip Designs
For generation and sorting of protease-secreting
bacterial emulsions
Oil
Oil
Waste
Waste
Laser Focus
Filter(10um)
Cells
Emulsion Re-injection
Sorted Positives
Collect
Substrate
Sorting Electrode
Emulsion
Oil
Oil
Sorting on red signal gives 2500-fold enrichment
for Protease Positive cells
Proprietary and Confidential
126
Sorting Cells Secreting a Protease
Combining modules creates a droplet-based optical
sorter
Reload
Detect
Sort
RainDance provides a microfluidic solution for
sorting
127
Bacterial Growth and Protease Secretion in
Droplets
Green Filter Shows Discrete Bacteria
Green Filter Shows Discrete Bacteria
Replication
Overnight
(37oC)
Red Filter Shows Protease Assay

• Single GFP Protease-positive cell
  encapsulated/droplet
• limited dilution encapsulation
• Replication to approximately 20-30 cells
  overnight
• Cell viability and motility observed over 6 days

128
Bacterial Growth and Protease Secretion in
Droplets
Overlay of Green/Red Images
40x Objective on Zeiss Fluorescence Microscope
129
Mixtures of Protease Positive and Negative Cells

• Can differentiate Protease Positive/Negative
  Cells

130
Droplet Generation - Single Cell per Droplet
Limited dilution of bacteria used to generate
clonal droplets
131
After Overnight Incubation - GFP image
Droplets provide a good environment for bacterial
growth
132
After Overnight Incubation - Assay
Secreted protease activity fills the droplet
volume
133
After Overnight Incubation - Overlay
Discrete bacteria, secreted protease fills
droplet volume
134
After Sorting - Overlay
50 droplets in view
Every sorted droplet contains bacteria and strong
assay signal
135
After Sorting - Overlay
200 droplets in view
Every sorted droplet contains bacteria and strong
assay signal
136
After Sorting - Overlay
800 droplets in view
Every sorted droplet contains bacteria and strong
assay signal
137
Digital Microfluidics Applications Other
Applications of Cells in Droplets
138
Cells in DropletsDigital microfluidics - more
than simply a FACS

• Cells can be analysed, sorted and recovered
• Digital microfluidics allows
• Analysis of Large Libraries of Cells
• Analysis of Small Numbers of Cells
• Many Cell Types
• Bacteria
• Yeast
• Drosophila
• Mammalian
• Secreted Factor Analysis
• Cell Phenotype Analysis
• Cell BioMarker Analysis

Not limited to sorting based on cell-surface
markers or on fluorogenic substrates which do
not leak from cells
139
BioMarker Detection and Cell Sorting
Cell
Add biotinylated antibody
Wash and bind SA-BGal
Load
Timing
Re-load
Collect Statistics/Sort
(Enzymatic Amplification)
Enzymatic signal amplification for low copy
number biomarkers Provides for
identification sorting of rare cells
140
BioMarker Detection and Cell Sorting
Biology in Droplets
Empty Droplets
Jurkat Cells in Droplets
Assay in Droplets

• Expected limits of detection 5-50 cell surface
  molecules

Opportunity for discovery of relevant cell
surface biomarkers
141
Digital Microfluidics Applications In Vitro
Systems for Protein Engineering and Directed
Evolution
142
Measuring Enzyme ActivityA Beta-Galactosidase
Assay
Fluorescence Detection
FDG fluorescein di-beta-D-galactoside
FMG fluorescein mono-beta-D-galactoside
143
Beta-Galactosidase Assay - Chip Design
Average Droplet Frequency 1kHz
144
Beta-Galactosidase Assay - Results Quantitative
and reproducible
4 S
12 units/mL
25 units/mL
50 units/mL
9 S
12 units/mL
25 units/mL
50 units/mL
15 S
12 units/mL
25 units/mL
50 units/mL

• Establishing activity in emulsion at four
  concentrations (0, 12, 25, 50 units/mL).

145
Beta-Galactosidase Assay - Results Sensitive

• Can detect 10 pM Beta-Galactosidase
• 10pM in a 30um droplet 85 molecules

146
Microfluidic Devices for Directed
EvolutionEnzymes expressed in vitro using
cell-free translation systems
Aqueous phase
Aqueous phase - in vitro translation system
- ?-galactosidase genes -
Fluorescein Di b-D-galactopyranoside
Oil
147
Microfluidic Devices for Directed
EvolutionEnzymes expressed in vitro using
cell-free translation systems
Aqueous phase
Aqueous phase - in vitro translation system
- ?-galactosidase genes -
Fluorescein Di b-D-galactopyranoside
Oil
148
Microfluidic Devices for Directed
EvolutionEnzymes expressed in vitro using
cell-free translation systems
Aqueous phase
Aqueous phase - in vitro translation system
- ?-galactosidase genes -
Fluorescein Di b-D-galactopyranoside
Oil
149
Cell-free Transcription and Translation in
Droplets
LacZ plasmid DNA concentration 20 fM, 2hr
incubation (1 gene in 40 droplets) 10 µm drops
made at 10 kHz per second
Transcription followed by translation from
single genes encapsulated in droplets
50 um
150
Microfluidic Devices for Directed
EvolutionEnzymes expressed in vitro using
cell-free translation systems
Aqueous phase
Aqueous phase - in vitro translation system
- ?-galactosidase genes -
Fluorescein Di b-D-galactopyranoside
Oil
Signal of Ebg class IV
151
Microfluidic Devices for Directed
EvolutionQuantitative systems
FACS of double emulsions
Aqueous phase
Oil
Microfluidics
152
Microfluidic Devices for Directed
EvolutionUncoupling translation and catalysis
Translation
Catalysis
153
Microfluidic Devices for Directed
EvolutionSimulating protocell growth and
division using microdroplets
154
Microfluidic Devices for Directed
EvolutionMassively parallel serial dilution
experiments