2. Project Overview Landing Pads Divide-by two Circuit Modeling
General, useful tool for all synthetic biologists
A systematic, Biobrick-compatible
approach to put plasmid constructs onto the chromosome
Divide-by two Circuit
Genetic toggle switch switchable by one input
Essential computing component
Modeling
Characterize system
Determine robustness
and predict ways to
improve performance
Fabrication
Construct gene operons in plasmid form and assemble into the whole system
Insert DBT constructs
into landing pad and
cross into chromosome

3. Divide-By-Two Circuit
Idealized Divide-By-Two Circuit
The same input toggles the system between both states
(on and off)
Frequency of the output
is half of the frequency
of the input
Output can be any gene expression
input
Output (normalized)
Time

4. DBT Circuit: Single Input Toggle Switch
σ54
NR1
GFP
lacp
araB
lacl
gfp
glnG cI
glnKp
rpoN
LacI
araBp
RFP cl
nifHp cl
nifA
lacl
rfp
cIp
σ54
nifA

5. How It Works lacp lacl rpoN nifHp cl rfp NR1 GFP gfp glnG cI glnKp
araBp
nifHp cl
nifA
lacl
rfp
cIp

6. How It Works lacp lacl rpoN nifHp cl rfp σ54 NR1 GFP araB gfp glnG cI
glnKp
rpoN
araBp
nifHp cl
nifA
lacl
rfp
cIp
σ54

7. How It Works lacp lacl rpoN nifHp cl rfp σ54 NR1 GFP gfp glnG cI glnKp
LacI
araBp
nifHp cl
nifA
lacl
rfp
cIp
σ54

8. How It Works lacp lacl rpoN nifHp cl rfp σ54 gfp glnG cI glnKp LacI
araBp
nifHp cl
nifA
lacl
rfp
cIp
σ54

9. How It Works lacp lacl rpoN nifHp cl rfp gfp glnG cI glnKp araBp RFP
nifA
lacl
rfp
cIp
nifA

10. How It Works lacp lacl rpoN nifHp cl rfp σ54 araB gfp glnG cI glnKp
araBp
RFP
nifHp cl
nifA
lacl
rfp
cIp
σ54
nifA

11. How It Works lacp lacl rpoN nifHp cl rfp σ54 gfp glnG cI glnKp araBp
nifA
lacl
rfp
cIp
σ54
nifA

12. How It Works lacp lacl rpoN nifHp cl rfp σ54 gfp glnG cI glnKp araBp
nifA
lacl
rfp
cIp
σ54

13. How It Works lacp lacl rpoN nifHp cl rfp NR1 GFP gfp glnG cI glnKp
araBp
nifHp cl
nifA
lacl
rfp
cIp

14. Modeling

15. Basic ODE’s
Hill function and degradation
No Basal production

16. Steady State Analysis
Under which conditions will the system toggle and which conditions will it not?
If it can toggle, how “easy” is it to achieve this?
How robust is the system?

17. Reducing Variable Dimensions
lacp
gfp
LacI
lacl cI cl
araBp
rpoN
glnG
NR1
nifA
nifHp
glnKp
cIp
rfp
GFP
RFP
lacp
gfp
LacI
lacl cI cl
araBp
araB
rpoN
glnG
σ54
NR1
nifA
nifHp
glnKp
cIp
rfp
GFP
RFP

18. Steady State Equations
Two variables and 8 parameters
Generally the parameters are either unknown and/or vary over a range
Symmetric vs unsymmetric

19. Existence of Multiple Steady States
NO INPUT
3 steady states: 2 stable, 1 unstable
1 steady state: stable
Phase plane: three steady states
Phase plane: one steady state cI
lacI

20. Toggle System Needs Two Steady States
With INPUT
Two stable steady states: toggle possible
Single steady state: no toggle can happen
Input added: no toggle
Input added: toggle between two states
GFP cI
RFP
lacI

21. What Parameter Ranges Yield Toggle Activity?
Only a narrow range of parameter values give toggle behavior
How robust is the system?
Not very
This will be addressed later

22. So how do you toggle? Vary parameters Toggle No toggle possible
assuming…
Toggle
possible
No toggle
possible
How does input affect the system?
Is a toggle easily achieved?

23. System responds differently to varying input levels
Sufficient input
Low input
Excess input
Ideal!
Steady state 1
Response of lacI and cI: excess input
Response of lacI and cI: sufficient input
Response of lacI and cI: insufficient input
Transient
response
concentration
Quasi-steady state
Quasi-
Steady state
Steady state 1
Steady state 2
bifurcation
???
time

24. The Nature of Quasi-Steady State
Back to original
steady states
Original system:
Two stable steady states
ONE quasi-
Steady state
Output 1
Output 1
Output 2
High s54 level
Output 2
Excess input
added
S54 degrades

25. Quasi-Steady State Heavily Favors One Output
End: Output 1
Start: Output 2
Start: Output 1
NOT TOGGLE BEHAVIOR!!!
Output 1
Output 1
Output 2
High s54 level
Output 2
Excess input
added
S54 degrades

26. Dynamic Modeling of the DBT
State transition takes place
“Overthrowing” side:
Rapidly increases
Dips back down
Slowly rises to dominant level
lacI
Concentration (molecules) cI
Time (s) 26

27. Dynamic Modeling of the DBT
lacI
Concentration cI
Time 27

28. Potential Problem Spots
A: Input promoter must be very quiet
B: Cross-talk between the two values must be low
C: lacp and cIp side parameter values must be relatively close C B A 28

29. Concentration (molecules) Concentration (molecules)
Unbalanced, Gardner-based Simulations:
lacI: (blue)
cI: (gold)
rfp: (red)
gfp: (green)
Large difference in the parameters between sides
Device predicted to tend to have lac-dominant side more stable
Can switch into lac-dominant side, but can’t switch out
Concentration (molecules)
Time (s)
lacI: (blue)
cI: (gold)
rfp: (red)
gfp: (green)
Concentration (molecules)
Time (s) 29

30. Stochastic Modeling
Mass-Action Model and Gillespie Method 30

31. Tuning the Device with IPTG and Temperature
Attenuate Lac with IPTG
Attenuate cI with heat
Calibrate device by varying both IPTG and temperature
Current research focusing on if and where “sweet-spot” is located 31

32. Clamping the DBT
Regardless of input pulse length/decay rate, only one change-of-states occurs 32

33. Divide-By-Two Fabrication

34. Operons of the DBT Circuit
Initial construction requires five plasmids, each carrying one operon of the circuit
Our final design will have fewer plasmids, as we will place some on the chromosome and combine others

35. cIp-nifA-lacI-RFP
BBa_I720004
BBa_I720005 35

36. lacp-GFP-glnG-cI

37. lacp-GFP-glnG-cI (cont)
Sequencing and characterization failed
BioBrick had bad DNA (wrong lacp sequence)
Actually needed repressible, rather than constitutive, promoter
Sequencing results verified plasmid had all the genes, though
BBa_I720006

38. BBa_I720002
PCR: NCM 77
BBa_I720003
PCR: K. Pneumoniae genome 38

39. Also transformed rpoN alone: still have growth
BBa_I720000
BBa_I720001
Characterization:
transformed plasmid into rpoN mutants: growth regardless of arabinose input
Also transformed rpoN alone: still have growth
Too noisy; need a single copy
Back-up plan: tetO-rpoN 39

40. DBT Operons in Landing Pads
Synthetic Operon
Landing Pad Used
Leucine BioBrick Landing Pad
Arabinose BioBrick Landing Pad
glnK Landing Pad
BBa_I720007

41. DBT Operons in Landing Pads

42. BioBrick Landing Pad Built so that ANYONE can insert
ANY BioBrick onto the
Chromosome of E.
Coli!

43. BioBrick Landing Pad: Goals of Project
General Landing Pad Goal: Aid in insertion of constructs onto the chromosome
Develop a general method for constructing landing pads that:
Have BioBrick compatible restriction sites
Allow easy phenotypic screening
Limit noise
Allow nesting of sequential landing pads

44. BioBrick Landing Pad

45. BioBrick Landing Pad

46. BioBrick Landing Pad: Homologous Regions
Arabinose Homologous Regions
Homologous Recombination of Arabinose Landing Pad

47. BioBrick Landing Pad: Nesting Landing Pads
Benefits of Nested Landing Pads
Allow for more constructs to be inserted onto the chromosome at the same chromosomal location
Different drug resistance gene is used: Only one type of screening is required

48. BioBrick Landing Pad: Nesting Landing Pads
New Landing Pad Criteria
Homologous regions from previous drug resistance gene (ChlorR)
Different drug resistance gene (KanR)

49. BioBrick Landing Pad: Nested Landing Pads
Nested Product: Chloramphenicol Landing Pad nested in Arabinose Landing Pad

50. BioBrick Landing Pad: Fabrication Progress
* Nesting to be tried in the near future as well!

51. Questions? THANK YOU!!! Prof. Alex Ninfa Prof. Peter Woolf
Domitilla DelVecchio
Dong Eun Chang
Questions?

52. Unbalanced System: Bifurcation favors one state
In a more realistic unbalanced system, the quasi-SS leads to one of the two steady states
One steady state is favored over the other

53. Quasi-Steady State Bifurcates with Sigma54 level (Balanced System)
If the system is balanced, as sigma54 degrades past threshold both steady states are equally accessible
Stable SS
Unstable SS
Quasi SS

54. Low input: no sigma54 presence after response
Sigma54 concentration
Time