Mathematical basis of stability analysis
),(
),(
yxgy
yxfx
=
=
&
&
system of two coupled differential equations
find nullclines and fixed point(s)
step 1
0),(0
0),(0
=→=
=→=
oo
oo
yxgy
yxfx
&
&
o
o
yyy
xxx
?≡
?≡
~
~
step 2
consider small deviation from fixed point
o
o
yyy
xxx
?≡
?≡
~
~
consider small deviation from fixed point
linearize around fixed point(s)
step 3
ydxc
y
g
y
x
g
xy
ybxa
y
f
y
x
f
xx
oo
oo
oo
oo
yx
yx
yx
yx
~~~~
~~~~
),(
),(
),(
),(
+≡
?
?
+
?
?
≈
+≡
?
?
+
?
?
≈
&
&
step 4
determine matrix A
?
?
?
?
?
?
=
dc
ba
A
?
?
?
?
?
?
=
dc
ba
A
determine matrix A
determine trace and determinant of A:
step 5
bcadA
daAtrace
?==?
+==
)det(
)(τ
determine stability of fixed point
step 6
only if τ < 0 and ? > 0, (x
o
,y
o
) is a stable fixed point
!!! be careful: only valid for 2 dimensional systems !!!
Last lectures: Genetic Switches
L3-4: Naturally occurring: lysis-lysogeny decision
L5-6: Engineered: genetic toggle switch
Switches are necessary for making ‘decisions’:
- development & differentiation (e.g. stem cells)
what to be ?
- metabolism
what to eat ?
- molecule synthesis (e.g amino acids)
what to produce ?
time scales for genetic regulation ~ 10 min - hours
Images removed to due copyright considerations.
What if faster response is needed ?
- finding food
- chasing bait
- signal transduction
Image removed due to copyright considerations.
genetics is too slow !
Protein switches (active/inactive states)
(total amount active + inactive is constant,
ignore gene expression)
timescales 1 ms - minutes
Introducing the H atom for signal transduction:
chemotaxis of Escherichia coli
Image removed due to copyright considerations.
See Alberts, Bruce, et al. Chapter 13 in Molecular biology of the cell.
4th ed. New York: Garland Science, 2002.
100, no. 23 (Nov 11, 2003): 13259-63.
Copyright (2003) National Academy of Sciences, U. S. A.
Figure 1A in Mittal, N., E. O. Budrene, M. P. Brenner, and A. Van Oudenaarden.
"Motility of Escherichia coli cells in clusters formed by chemotactic aggregation." Proc Natl Acad Sci U S A.
cell length ~ 1-2 μm, diameter ~ 0.5 μm
Images removed due to copyright considerations.
The Flagellum
Image removed due to copyright considerations.
Absence of chemical attractant
Image by MIT OCW.
Tumble
Run
Presence of chemical attractant
Image by MIT OCW.
Tumble
Chemical Gradient Sensed in a Temporal Manner
Run
Attractant
Image by MIT OCW. After figure 4 in Falke, J. J., R. B. Bass, S. L. Butler, S. A. Chervitz, and M. A. Danielson.
"The two-component signaling pathway of bacterial chemotaxis: a molecular view of signal transduction by
receptors, kinases, and adaptation enzymes." Annu Rev Cell Dev Biol 13 (1997): 457-512.
Chemotactic pathway in E. coli
Towards more complex system networks.
Image removed due to copyright considerations.
Proteins in the chemotactic network can be
modified in differents ways:
I Phosphorylation (CheA, CheY, CheB)
II Methylation (Tar receptor)
Image removed due to copyright considerations.
I Phosphorylation (CheA, CheY, CheB)
CheA (protein kinase), uses ATP to
phosphorylate one of its histidines.
ADP
p
CheAATPCheA +?+
CheA (CheA
p
)is bound to the Tar receptor
through an adapter protein CheW. CheW is
not known to have any enzymatic activity.
(these proteins are sometimes called
‘scaffolding protein’)
CheA
p
is unstable and transfers its
phosphoryl group to CheY (highly soluble,
diffuses through the cytoplasm
CheA
His48
Courtesy of Annual Review of Cell and Developmental Biology. Used with permission.
Falke, J. J., R. B. Bass, S. L. Butler, S. A. Chervitz, and M. A. Danielson.
"The two-component signaling pathway of bacterial chemotaxis: a molecular view of signal transduction by
receptors, kinases,
and adaptation enzymes." Annu Rev Cell Dev Biol 13 (1997): 457-512.
I Phosphorylation (CheA, CheY, CheB)
ADP
p
CheAATPCheA +?+
autophosphorylation:
p
CheYCheACheY
p
CheA +?+
phosphoryltransfer:
CheYp binds to the motor (FliM),
motor rotates CW (= tumbles)
logic:
high levels of CheA -> high levels of CheYp
(lots of tumbles)
low levels of CheA -> low levels of CheYp
(straight swimming)
Image removed due to copyright considerations.
CheZ dephosphorylates CheY
p
(opposite function as CheA)
p
CheZCheYCheZ
p
CheY +?+
logic:
high levels of CheZ -> low levels of CheYp
(straight swimming)
II Methylation (tar receptor)
Image removed due to copyright considerations.
CheR adds methyl group
CheB
p
removes methyl group
phosphorylation state of CheB
is controlled by CheA
Methylation - Phosphorylation coupling
Image removed due to copyright considerations.
phosphorylation state of CheB
is controlled by CheA
Courtesy of Annual Review of Cell and Developmental Biology. Used with permission.
Falke, J. J., R. B. Bass, S. L. Butler, S. A. Chervitz, and M. A. Danielson.
"The two-component signaling pathway of bacterial chemotaxis: a molecular view of signal transduction by
receptors, kinases, and adaptation enzymes." Annu Rev Cell Dev Biol
13 (1997):
457-512.
Role of ligand binding
Image removed due to copyright considerations.
The rate of CheA phosphorylation is stimulated
by unoccupied receptors
Image by MIT OCW. After figure 4 in Falke, J. J., R. B. Bass, S. L. Butler, S. A. Chervitz, and M. A. Danielson. "The
two-component signaling pathway of bacterial chemotaxis: a molecular view of signal transduction by receptors, kinases,
and adaptation enzymes." Annu Rev Cell Dev Biol 13 (1997): 457-512.
why is this all so complex ?
Image by MIT OCW.
methylation is important for adaptation
(~ background subtraction)
E. coli can sense aspartate from 10 nM - 1 mM
and sense changes as small as 0.1%
1 2 3 4
tumbling
methylation
add
attractant
Correlation of Receptor Methylation with Behavioral Response
remove
attractant
add more
attractant
Before starting with the modeling,
first let’s look at some recent experiments
Alon et al. Nature 397,168 (1999)
Cluzel et al. Science 287, 1652 (2000)
Sourjik et al., PNAS 99, 123 (2002)
PNAS 99, 12669 (2002)
Nature 428, 439 (2004)
Remember scientists have been working on
E. coli chemotaxis for about 100 years now
Single cell chemotactic analysis
Image removed due to copyright considerations. See figure 1 in Cluzel, P., M. Surette, and S. Liebler.
Science 287, no. 5458 (Mar 3, 2000): 1652-5.
"An ultrasensitive bacterial motor revealed by monitoring signaling proteins in single cells."
correlation CW bias & CheY-P gene expression
cells have plasmids with CheY-GFP
under inducible promoter
assumption: all CheY is
phosphorylated
strain: CheY-, CheZ-, CheB-
Hill #: ~10
Image removed due to copyright considerations. See figure 1 in Cluzel, P., M. Surette, and S. Liebler.
Science
287, no. 5458 (Mar 3, 2000): 1652-5.
"An ultrasensitive bacterial motor revealed by monitoring signaling proteins in single cells."
low YFP/CFP:
unbound
high YFP/CFP:
bound
FRET (fluorescence resonant transfer)
Figures 1A, 1B in Sourjik, V., and H. C. Berg. "Binding of the Escherichia coli response regulator CheY to its
Proc Natl Acad Sci U S A 99, no. 20
Copyright (2002) National Academy of Sciences, U. S. A.
target measured in vivo by fluorescence resonance energy transfer."
(Oct 1, 2002): 12669-74.
CheY-YFP (yellow)
CheZ-CFP (blue
CheZ binds only to
CheYp !!
adding attractant
leads to immediate
lower concentration of
CheY
p
-CheZ complex,
lower [CheYp],
less tumbling
Figures 1A and 1B in
Proc Natl Acad Sci U S A 99, no. 1
(Jan 8, 2002): 123-7.
"Receptor sensitivity in bacterial chemotaxis."
Sourjik, V., and Berg HC.
Copyright (2002) National Academy of Sciences, U. S. A.
Hill # ~ 1
Copyright (2002) National Academy of Sciences, U. S. A.
Figure 2 in
Proc Natl Acad Sci U S A 99, no. 1
(Jan 8, 2002): 123-7.
"Receptor sensitivity in bacterial chemotaxis."
Sourjik, V., and Berg HC.
amplification between receptors and CheYp: ~35
amplification between CheYp and motor: ~10
total amplification ~ 350
our models should reproduce this (hint: receptor clustering)
Copyright (2002) National Academy of Sciences, U. S. A.
Figure 2 in
Proc Natl Acad Sci U S A 99, no. 1 (Jan 8, 2002): 123-7.
"Receptor sensitivity in bacterial chemotaxis." Sourjik, V., and Berg HC.
adaptation
slow
perfect
adaptation
Models should also reproduce
qualitative properties such as perfect adaptation
excitation(fast)
Figures 1a and 1 b in
Proc Natl Acad Sci U S A 99, no. 1
(Jan 8,
2002): 123-7.
"Receptor sensitivity in bacterial chemotaxis."
Sourjik, V., and Berg HC.
Copyright (2002) National Academy of Sciences,
U. S. A.
Images removed due to copyright considerations. See Figure 1 in Alon, U., M. G. Surette, N. Barkai,
"Robustness in bacterial chemotaxis." Nature 397, no. 6715 (Jan 14, 1999): 168-71.and S. Leibler.
Perfect adaptation is robust against changes
in Che-protein concentrations
not all parameters are robust !
Images removed due to copyright considerations. See Figure 2 in Alon, U., M. G. Surette, N. Barkai,
"Robustness in bacterial chemotaxis." Nature 397, no. 6715 (Jan 14, 1999): 168-71.and S. Leibler.
Goal of next lecture is develop models that qualitatively
and quantitative reproduce these phenomena, such as:
huge gain
sensitivity
perfect adaptation
All these effects are ubiquitous in signal
transduction pathways in general.
‘Fine tuned model for perfect adaptation’
Spiro et al. PNAS 94, 7263-7268 (1997)
A model of excitation and adaptation in
bacterial chemotaxis
key player: Tar-CheA-CheW complex
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 1 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
assumptions:
1. Tar is only receptor type, CheW and CheA always
bound to Tar
2. Methylation occurs in specific order
3. Consider only 3 highest methylation states
4. Only CheB
p
demethylates
5. Phoshorylation of CheA does not affect ligand (un)binding
6. Tar-CheR binding does not affect ligand un(binding)
and phosphorylation of CheA
7. CheZ is not regulated
8. Phosphotransfer from complex to CheY or CheB is not
affected by occupancy or methylation state.
fast
slow
i
n
t
e
r
m
e
d
i
a
t
e
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 2 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
Ligand bound states generally have lower
autophosphoryalation rates
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 2 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
CheR methylates ligand-bound states more
rapidly
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 2 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
Consider step in aspartate concentration
time ~ 1 ms, increase in ligand bound complex
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 2 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
time ~ 5 s, total # of phosphorylated complexes
decreases gradually because ligand bound
complexes do not autophoshorylate very well
also: CheB
p
decreases
low CheA
p
, low CheY
p
,
tumble suppression
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 2 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
time ~ 50 s, slowly the unbound complex
methylate. Note that demethylation is switched
because of low levels of CheA
p
(low CheB
p
).
also low CheA
p
, low CheY
p
,
tumble suppression
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 4 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
Higher methylation states autophosphorylate
easier, so slowly CheA
p
adapts to its initial
level
high CheA
p
, high CheY
p
,
tumbling
Copyright (1997) National Academy of Sciences, U. S. A.
Figure 4 of Spiro P. A., J. S. Parkinson, and H. G. Othmer. "A model of excitation and adaptation in bacterial
chemotaxis." Proc Natl Acad Sci U S A 94, no. 14 (Jul 8, 1997): 7263-8.
CheY
p
CheB
p
Spiro, P. A., J. S. Parkinson, and H. G. Othmer. Figures 1, 2, and 4 in "
Proc Natl Acad Sci U S A 94, no. 14 (July 8, 1997): 7263-8.
A model of excitation and adaptation in
bacterial chemotaxis."
Copyright (1997) National Academy of Sciences, U. S. A.
Image removed due to copyright considerations. See Figures 2 and 3 in Barkai, N., S. Leibler.
"Robustness in simple biochemical networks." Nature 387, no. 6636 (Jun 26, 1997): 913-7.
Image removed due to copyright considerations. See Figures 2 and 3 in Barkai, N., S. Leibler.
"Robustness in simple biochemical networks." Nature 387, no. 6636 (Jun 26, 1997): 913-7.
Image removed due to copyright considerations. See Figures 2 and 3 in Barkai, N., S. Leibler.
"Robustness in simple biochemical networks." Nature 387, no. 6636 (Jun 26, 1997): 913-7.