Download - Towards Scalable Synthetic Biology
![Page 1: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/1.jpg)
Towards Scalable
Synthetic Biology
http://genomics.lbl.gov
![Page 2: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/2.jpg)
The Enormous Potential of Engineered Microbes
Skerker, Lucks, Arkin Genome Biol. 2009
Drugs
Food Additives (Carotenoids)
Chemical Feedstocks
Fuels
Bioremediation
Anticancer Therapeutics
Adapted from Arkin Nature Biotech, 2008
Beyond the BioreactorChemical Factories
1) Engineered microbial solutions are being developed for energy, ecology and medicine.
2) The core of these solutions is gene expression.
Environmental Sensors
![Page 3: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/3.jpg)
Modern Metabolic
Engineering
Naturally useful cells
Naturally useful enzymesPipes
Programs
Discovery Design
Verifiable cells for Health, Agriculture, Environment
Current industry
De novo enzymes
Demo networks(toggle, oscillator, edge detector)
Verified Biological Design (VerBiD)
![Page 4: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/4.jpg)
Ci
Cellular Environment
Cellular Resource
Composability
Composability
Compatible Parts Families
Parts Registry
Tunable, Orthogonal Parts
Lucks, Qi, Whitaker, Arkin Curr. Op. Microbiol. 11, 2008
Goal: Develop Building Blocks that are Tunable, Orthogonal, Composable and
Physically Homogeneous
Scalable Genetic Engineering
![Page 6: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/6.jpg)
Control of Protein-Protein Interaction and Localization Nat Biotech, 2008
1’State 1
DNA Writing
1A A
A
A A
State 0
Control of Central Dogma
Julius Lucks
Lei Stanley Qi
Vivek Mutalik
John Deuber
Weston Whitaker
Jay Keasling
Michael Samoilov
Tim Ham
![Page 7: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/7.jpg)
Gene Expression Engineering
dmRNA
dtOC 1TAF Degm *mRNA
dP
dtTrans*mRNA Degp *P
Nat Biotech, 2008
![Page 8: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/8.jpg)
Protein decay
Quality
)(
)(
Protein :
mRNA :
ProteinmRNAmRNA:
mRNAPromoterPromoter:
dP
dR
s
t
k
k
k
k
Vastly simplified model
100~ConstProtein 100ConstProtein
Variance of protein levels
Same mean, Different variances
Mean protein levels Time to 90% Mean
![Page 9: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/9.jpg)
Three Stories
Orthogonality
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
Composability
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
Connectivity
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
Mechanism Molecule (RNA , Protein )
Many regulators to choose from
DNA
RNA
Transcription
Translation
Protein P
Regulatory Coding Sequence RegulatoryTranscription Factors
Self-Splicing Ribozymes
Transcription Terminators
RNA Binding Proteins
Translational Repressors
Self-Splicing Introns
Post-Translational Modification
Transcription Attenuators
Riboswitches
![Page 10: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/10.jpg)
PT181 Attenuator
![Page 11: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/11.jpg)
Transcriptional Attenuators: Families
Antisense RNA
Sense RNA
• It is possible to design
orthogonal mutant lock-key
pairs by introducing mutations
into recognition motifs
• Creating family of mutually
orthogonal lock-key pairs ready
to use
• 4096 possible pairs with this
version of the stem-loop.
![Page 12: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/12.jpg)
Engineering Dynamic Range and
Orthogonality
![Page 13: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/13.jpg)
Cracking the Code of Orthogonality
WT/WT MUT/MUT
WT/MUT MUT/WT
Orthogonality
Specificity - 8nt change!
Dynamic Range - 4nt change!
Out of 190nt:
![Page 14: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/14.jpg)
Composing Multiple Attenuators
Attenuator
Output Signal
(GFP)
Input Signal
(IPTG)
A
( )n=1,2,3,4
TAFn=1-(TAF)n
TA
Fn
Number of attenuators (n)
![Page 15: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/15.jpg)
Homogeneous and
combinatorial function
![Page 16: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/16.jpg)
Engineering a 2-Stage Cascade
A1
S1 S2
A2
Orthogonal Parts
S1
A1S2
A2
Stanley Qi, Arkin Lab
![Page 17: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/17.jpg)
S1
Engineering a 2-Stage Cascade
Stanley Qi, Arkin Lab
![Page 18: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/18.jpg)
S1
A1S2
Engineering a 2-Stage Cascade
Stanley Qi, Arkin Lab
![Page 19: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/19.jpg)
S1
A1S2
A2 A2
Engineering a 2-Stage Cascade
Stanley Qi, Arkin Lab
![Page 20: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/20.jpg)
Summary – Engineering Transcription Regulation
Orthogonality Composability Connectivity
0
20
40
60
80
1 2 3 4
% A
ttenuation
WT/WT WT/MUT MUT/WT MUT/MUT
![Page 21: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/21.jpg)
Translation engineering
• RNA IN/OUT
• >4096 possible pairs with
expanded specificity region
• Modular/composable
molecular function
Re
pre
ssio
n %
Designed
homogeneous
operation
Predictive design
Repre
ssio
n
%
G
(kcal/mol)
Part #
Designed
orthogonality
![Page 22: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/22.jpg)
Modular Control of Signal and Carbon Flux
Optimization of metabolism (Dueber/Keasling)
Control of Signal Transduction (Whitaker, Dueber, Arkin)
![Page 23: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/23.jpg)
DNA WriteTim Ham, Michael Samoilov
![Page 24: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/24.jpg)
Recombinases• Interesting operations
• Inversion
• Insertion
• Excision
• Nearly Boolean
• State accessible after death and
transmissible between cells
• Geometrically programmable?
• Possible huge state space for
logical machines encodable in
relatively little DNA
• Flexible acceptance of active
elements in internal regions. Invertases: the Fim system
![Page 25: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/25.jpg)
Invertase dynamics
Leakless
System can hold “state” after pulse
1A A
A
1’A AState 1
State 0
![Page 26: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/26.jpg)
1A A2 3
3’A A2’ 1’ 1A A2’ 3
1’A A2 3’
State 0
State 1State 2
State 3
Placement of two invertase sites.
N! input sequences
![Page 27: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/27.jpg)
B1A A2 3 B
B2’A A1’ 3 B B1A A3’ 2’ B
B2’A A3’ 1 B B3A A1’ 2’ B
State 0
State 1 State 2
State 3State 3
Placement of two invertase sites.
S(N ) k!C(N ,k)k1
N
Maximum number of states
available with N pairs of sites
![Page 28: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/28.jpg)
Log
10(C
or
S)
Number of Invertases
Number of devices and state space grows rapidly with N.
Configurations of 3 site pairs
a(0) 1; for n 0, a(n) (2n1)!! (2k1)!!a(n k)k1
n1
For n=N-1 there are:
configurations.
If the minimum length of a flippable region is
100 bp and inversion sites are 30bp then the
device is less than 2*30*N+100bp long at
minimum.
So for N= 10: S~ 1011 and Length~ 700 bp
![Page 29: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/29.jpg)
Cell Sends a Coded Message to
Recipient Cell
Addressable
riboregulators
![Page 30: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/30.jpg)
NAND Logic Gates
![Page 31: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/31.jpg)
Trainable Bacterial Networks
![Page 32: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/32.jpg)
Some numbers to contemplate106 microorganisms per gram of soil (1033 on earth)
109 in a ml of rich media
Divide every 20 minutes
Flip-operations-per-molecule-per-second-per cell ~ 0.01
100-200 plasmids per cell.
5-6 “machines” per plasmid
100-1000 “machines” per genome
Plasmids passed per cell per generation ~ 0.01
What is the computational capacity?
![Page 33: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/33.jpg)
Performance Assessment
General Considerations:
1. Modern computers have advanced to where solving a general P problem does not represent a compelling reason for developing a new technological platform beyond silicon in itself. Thus, solving NP-hard problems represents the main stimulus behind this and other approaches. And within those, NP-complete represents the most immediately interesting class.
2. At present, NP-complete problems have at best O(2N) algorithms, i.e. regardless of the implementation – in silico or in situ – finding a solution involves scanning through most of the solution space. Therefore, the relevant comparison between an in silico and in vivo or in vitro computer is best made within the context of an NP-complete problem solving application and its number of operations executed per unit time.
![Page 34: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/34.jpg)
A motivating example that doesn’t even use the full power of the system
1. Develop a synthetic biological platform for efficient generation of N-object permutation pool (N-OPP) uniformly distributed over states
2. Use biomolecular-scale implementation to generate a full realization of N-OPP state space
3. Engineer a sieving scheme to allow problem-specific searches of N-OPP state space by random sampling
4. As an application, engineer a scheme for NP-complete problem –start with Hamiltionan Path Problem (HPP)
5. Design and implement a scalable bioengineered device –potentially creating a practical scalable solver
![Page 35: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/35.jpg)
HPP Example
• Problem: find whether a path passing through each node of a graph of size N exists / is valid
• Biomolecular solution: 1. Generate state space of all paths on a graph that pass through each node
only once as N-OPP of N distinct elements on a plasmid (see Slide 2)
2. Check if any of those paths follow all valid edges (all neighboring node elements in a permutation have corresponding valid edges on the graph), i.e. (…,Ni,Nj,…) = Eij exists on the graph for all (i,j).
![Page 36: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/36.jpg)
N-OPP platform• Use all the same invertase sites.
• Starting with an initial state, the system stochastically samples configuration space to generate all possible permutations
![Page 37: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/37.jpg)
Generating Uniform N-OPP
• With the stochastic N-OPP generator, the state space is described by a distribution over element permutations
• The desired distribution is ~ uniform, since this insures that no state remains preferentially un-sampled
• With N-OPP represented by sequences of DNA elements on plasmids and recombinase acting on them (Slide 2), the (mixing) time required to reach a distribution exponentially close to uniform is
~O(N*Log(N))
• (Requires a long and involved derivation)
![Page 38: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/38.jpg)
HPP Sieve – in vitro
Sample first pass implementation:
1. Setup:
- Design nodes to have unique orthogonal tags at ends
- Design edges to be oligos flanked by two tags – each complementary to a distinct node
- Design edge interior to consist of have short homology to the inter-node region with a restriction site T on it
2. Inputs:
1. A tube w/ plasmids of uniformly distributed N-OPP over nodes
2. Resuspended oligos of the complementary graph (i.e. oligoscorrespond to missing edges).
3. Mix with high oligo excess, melt and re-anneal
N1
Orthogonal sequences
Node sequence
E1N3 N1
Node 3 & 1 ID sequences
Edge sequence
![Page 39: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/39.jpg)
HPP Sieve – in vitroSample first pass implementation:
4. Sieving correct solution- For any given plasmid containing any given permutation of nodes (a “test path”), if there is an
edge complementary to some pair of nodes => this is not a valid path (since edges are from the complementary graph and the path is traversing missing edges)
- Binding of an edge oligo to the test path (i.e. the test path is incorrect) generates dsDNA
- Restriction enzyme cuts site T of dsDNA, but not unpaired ssDNA, eliminating only those test paths that bound an oligo and leaving valid test paths intact
5. Presence of the path can now be detected by PCR
No complementary edge Complementary edgeRestriction enzyme
Exo 1
![Page 40: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/40.jpg)
Performance Assessment
Core Statistics: (First-pass implementation)
1. Scalability:
It is essentially infeasible to allocate or maintain > 10,000 CPUs for most computational tasks, but possible to do with > 100 L of culture
2. These numbers characterize the speed and efficiency of the biomolecular device in generating the N-OPP space, but not the sieving scheme for problem-specific searches
i5-680 CPU
in vivo in vitro
Speed (GHz/mL)
1.1 29.9 12.2
PowerRequired
73W Low 0
Cost ($ per GHz)
~$100 ~$0.1 ~$1
0
50
100
150
200
250
300
i5-680 in cella in vitro
Speed/Cost
![Page 41: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/41.jpg)
Towards a predictable control
layer in cellular engineering
Programmable sensing
Programmable transcription
Programmable translation
Programmable DNA
Goal: Make the programming the logic
of behavioral control “easy”
We need systems that are:
Orthogonal
Composable
Connectable
Homogeneous
Designable
Scalable
Robust to “Context”
![Page 42: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/42.jpg)
Orthogonality
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
Connectivity
Composability
0
.
1
5
0
.
1
5
0
.
1
5
0
.
1
5
The Larger Picture: Taking Design to the Next LevelSense Environmental
Signals
Process
Information and
Make Decisions
Carry Out Regulatory
Programs
Rival Complexity of Nature
Modular Signaling
Persistent state machines
![Page 43: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/43.jpg)
Conclusions
• We are at the edge of creating live organisms for critical human application.
• Therapeutic viruses and bacteria are being built from bottom up and require multiscale design and characterization from molecules to ecologies– and requires the building of the proper analytical and computational infrastructure.
• Using engineering principles is at least creating a network effect in providing parts, devices and systems for other researcher to create new artifacts
– Chassis is fundamental module for therapeutic bacteria.
• We ARE making progress in developing truly homogeneous designs for scalable circuit engineering in cells- but we are at the proof-of principle phase.
• We need a robust programming paradigm using our devices as a basis and scaling through population behaviors.
![Page 44: Towards Scalable Synthetic Biology](https://reader034.vdocument.in/reader034/viewer/2022042109/62570293b0dd8a597a6f3a66/html5/thumbnails/44.jpg)
Join up!
• Synthetic Biology Institute launches in
July with our Founding Partner Agilent
• BIOFAB biofab.org
• SYNBERC