Proofreading by single RNA polymerase molecules Joshua W' Shaevitz Elio A' Abbondanzieri Steven M' B

1
Fermilab Colloquium Wednesday October 13, 2004
Nature's Nanomachines The Physics of
Biomolecular Motors
Joshua W. Shaevitz Departments of IB, MCB, and
BioE Miller Institute for Basic Research in
Science Univ. of California, Berkeley
jshaevitz_at_berkeley.edu
2
Motion is the defining characteristic of life
Mitosis http//www.bio.unc.edu/faculty/salmon/lab
/mitosis/ Bacteria http//www.rowland.org/labs/b
acteria/index_movies.html http//cmgm.stanford.ed
u/theriot/movies.htm Insects
http//www.uku.fi/holopain/ento/digivideo.htm Ga
zelle http//tva.canoe.com/emissions/salutbonjour
we/salutptcom/ Human http//www.sip.uiuc.edu/t
dwg/baby/videos.html
3
Biological motion is produced by molecular motors
Molecular motors are enzymes that convert
chemical energy into mechanical work to produce
movement

• Myosin muscle contraction, cargo transport
• Kinesin cargo transport, chromosome separation
• Dynein cargo transport, sperm swimming
• Lambda exonuclease chews up a strand of DNA
• RNA polymerase copies DNA into RNA
• ATP synthase a rotary turbine that generates ATP
• and many others . . .

4
Molecular motors live in a nanoworld
Molecular motors . . .

• are small in size (1 10 nm)
• move in tiny steps (1 10 nm)
• produce tiny forces (1 100 pN)
• use a chemical energy source (e.g. by breaking a
  chemical bond)
• care about Brownian motion
• can be very efficient (50 100)

Molecular motors are true nanotechnology!
5
Biology presents a window into the nanoworld
P.N.A.S. 1981
Development of the ability to design protein
molecules will open a path to the fabrication of
devices to complex atomic specifications, thus
sidestepping obstacles facing conventional
microtechnology.
Foresight Institute
6
Relevant (biological) length scales
Red Blood Cells
Humans
1011 nanometers
109 nanometers
104 nanometers
E. Coli Bacteria
Viruses
Proteins
103 nanometers
102 nanometers
10 nanometers
7
Biological filaments are the cells highways
Microtubules Actin filaments DNA
25 nm
8 nm
9 nm
36 nm
2 nm
3.4 Å
3.4 nm
8
Myosin the motor that flexes your muscles
Sale and Sabbadini, UCSD

• movement is driven by ATP hydrolysis
• each myosin strokes by 5 nm and then lets go
• maximum force generation 10 pN

9
Kinesin the cells railroad engine

• is driven by ATP hydrolysis
• has two small catalytic feet (4 nm) that walk
  hand-over-hand on a microtubule
• walks in 8-nm steps
• moves at 100 steps/sec for 1 sec
• is 70 fuel efficient at maximum force (6 pN)

kinesin
Hirokawa, Science (1998)
Distance (nm)
Vale Milligan, Science (2000)
Time (s)
10
ATP synthase a molecular turbine
makes 10 ATP/s 1 ATP 10-19 J ? Power output
1 attowatt
Wang and Oster Nature (1998)
11
ATP synthase is a reversible engine!
If you give the turbine fuel (ATP), it spins
backwards
Yoshida et al. Nature Reviews Molecular Cell
Biology (2001)
12
Molecular motors require an interdisciplinary
approach

• Biochemistry (Molecular Reaction)
• Biostructure (Molecular Geometry)
• Biophysics (Molecular Mechanics)

Displacements and Forces in salty H2O at room
temp!
13
Tools for measuring displacements and forces
Microscopic markers
Atomic Force Microscopy
Fluid flow
Magnetic Force Microscopy
and many more
14
Optical traps are three-dimensional springs made
of light
Ashkin et al., Optics Letters (1986)
15
Small particles bend the trapping laser light
The trapping laser imparts a force onto the
particle directed towards the laser focus. The
magnitude of this force is P laser power n
particle refractive index c speed of light Q
trapping efficiency
Force
Objective
Reviewed in Svoboda and Block, Annu Rev Biophys
Biomol Struct (1994)
16
Tiny particles can be treated as dipoles (r ltlt ?)
The particle feels a Lorentz force
For a CW laser the Poynting vector is constant
Also a force from the light scattered by the
particle
17
Large objects require ray optics (r gtgt ?)
Refraction of each ray in the laser beam imparts
a momentum change to the sphere.
Integrate over all the rays in the beam (usually
a gaussian TEM00 mode)
18
Trapping force scales with particle size
nparticle 1.57 nmedium 1.33 ?n 800 nm
19
Our optical trapping apparatus
20
Optics Diagram
21
Acousto-optic deflectors allow the trap to be
moved quickly
Christoff Schmidt, Vrije University, Amsterdam

• Position accuracy 1 Å
• Can change position at speeds exceeding 100 kHz

22
We measure the angle that the light is bent
The position of the bead relative to the laser
focus is proportional to the bending angle. We
measure this angle by calculating the center of
intensity of the laser beam at the back focal
plane of the condenser with a position sensitive
detector.
Condenser
Objective
Pacific Silicon Sensors, Inc.
23
Two-dimensional position detection
24
Control experiment 1 Ångstrom stability and
resolution
Position (Å)
Time (s)
25
Measuring the spring constant
a 0.1 pN/nm
26
RNA Polymerase the DNA transcribing
machine with a Backspace key
27
The central dogma of molecular biology
Crick, Symp Soc Exp Biol (1958)
28
RNA polymerase (RNAP) carries out transcription

• 3 stages of transcription
• Initiation
• RNAP recognizes a start sequence in the DNA
  called the promoter
• Separates the two strands of DNA to expose the
  template strand
• Initiates RNA synthesis using free nucleotides
  from solution
• RNA Elongation
• Copies the sequence of DNA following the promoter
  into a new strand of RNA
• Termination
• RNAP recognizes a stop sequence in the DNA called
  a terminator
• Stops elongation and lets go of both the DNA and
  RNA

29
RNA polymerase as an enzyme
30
RNA polymerase as a motor!

• E. Coli RNAP
• moves along the DNA as it transcribes its a
  motor!
• is small 10x10x16 nm
• can synthesize up to 1,000,000 bases without
  letting go
• is driven by the free energy of PPi release, 50
  kJ/mol (80 pNnm)
• moves at 3 7 nm/s
• has a putative single-nucleotide step-size (3.3
  Å)
• can produce up to 25 pN of force
• is 10 energy efficient

Gelles and Landick, Cell (1998)
31
Dumbbell optical trap avoids noise drift
arising from motions of the microscope stage
32
Both the DNA and the optical traps have elastic
compliance
ktrap
ktrap
kDNA
33
DNA acts like an extensible worm-like chain
34
DNA acts like an extensible worm-like chain
35
DNA is a very non-linear spring
Contour Length 894 1 nm Persistence Length
41 2 nm
Wang et al., Biophys J (1997)
36
A force clamp allows for longer runs and
simplified analysis
track this bead
move this bead
Servo the position of the left bead in order to
maintain a constant force (position) of the right
bead
37
RNAP motion with the dumbbell assay
1 µm
60x speed 1 sec 1 min
38
Two states run and pause
1mM NTPs
Template Position (bp)
Time (s)
39
Force has a non-linear effect of run velocity
?Fstall? 28 2 pN
30
-30
15
-15
0
Force (pN)
Neuman et al., Cell (2003)
40
Pausing is important for RNAP function

• RNA polymerase pauses in order to
• Synchronize transcription and translation
• Signal a damage site in the template DNA (MfD)
• Perform termination

41
Pausing occurs on many timescales
Number
20 - 500
Pause Duration (s)
Neuman et al., Cell (2003) Adelman et al., PNAS
(2002)
42
Short pauses are everywhere and are unaffected by
force
We dont know what they are!
Neuman et al., Cell (2003)
43
Modulating temperature in a single molecule
experiment is difficult
A peripheral, not the central, portion of one of
the aggregates is illuminated by focusing an IR
laser beam. If the central portion is
illuminated, the aggregate is frequently blown
off or the surrounding medium gets boiled.
randomly shaped metal aggregate
- Kato et al., PNAS (1999)

• Potential Problems
• Boiling (!)
• Heats the microscope object, possibly ruining it
• Local temperature distribution is not well
  understood

44
The answer Use the laser to heat the liquid
directly
45
Calibrating the temperature with two optical traps
46
Calibrating the temperature in an optical trap
?(T) viscosity of water at temperature T a trap
stiffness r bead radius
47
Run velocity changes with temperature, but short
pauses dont
?H-1111 pNnm (-77 kJ/mol) (-22 kcal/mol)
?H9111 pNnm (557 kJ/mol) (132 kcal/mol)
48
What happens when RNAP makes a mistake?
Shaevitz et al., Nature (2003)
49
What happens when RNAP makes a mistake?
incorrect base
DNA
RNA Polymerase
RNA
Shaevitz et al., Nature (2003)
50
Support for a proofreading mechanism

• Error rates in vitro 10-3- 10-4
• Error rates in vivo 10-5-10-6

Erie et al., Science (1993)
51
Error rates for different processes
perfection
chaos
52
Consensus model of RNA proofreading

• The paused RNAP can
• slide forward again, leaving the mistake
• cleave the RNA, correcting the mistake

Erie et al., Science (1993) Jeon and Agarwal,
PNAS (1996) Thomas et al., Cell (1998)
53
Why single molecules and optical traps?

• Misincorporation is a low probability, stochastic
  event
• Look at single molecule instead of bulk behavior
• Proofreading involves tiny motions of the
  polymerase relative to DNA
• Use optical tweezers with sub-nanometer resolution

54
Observations about the long pauses
1mM NTPs
Template Position (bp)
Time (s)
55
Observations about the long pauses

• Long pauses are sequence independent
• Long pauses happen about once every thousand
  bases
• RNAPs in vitro error rate is about one mistake
  per thousand bases

Are the long pauses associated with
misincorporation and proofreading?
56
Are long pauses associated with proofreading?

• Do the long pauses backtrack?
• Are the long pauses associated with
  misincorporation?
• Do the cleavage factors GreA and GreB relieve
  long pauses?
• Does the polymerase cleave the RNA at the end of
  long pauses in the presence of the GreA and GreB?

57
Long pauses start with a backtrack
Position (bp)
Time (s)
58
Backtracking accompanies long pauses
n56
elongation
backtracking (4.7 0.8 bp)
Position (bp)
elongation
slow recovery
Time (s)
59
Short pauses dont backtrack
Position (bp)
Time (s)
60
Backtracking pauses are force dependent
Characteristic Distance 4 1 bp (4.7 0.8 bp)
Pauses per kb
Unloaded Rate 1.2 10-4
Force (pN)
61
Evidence that long pauses are associated with
proofreading

• Do the long pauses backtrack?
• YES!
• Are the long pauses associated with
  misincorporation?
• Do the cleavage factors GreA and GreB relieve
  long pauses?
• Does the polymerase cleave the RNA at the end of
  long pauses in the presence of the GreA and GreB?

62
Inosine looks like a bad guanine
Inosine

• Inosine is a purine similar to guanine, but
  lacking an amine group
• RNAP selectively misincorporates an inosine at
  sites coding for guanine

Guanine
Prediction Addition of ITP should increase the
number of long pauses.
Thomas et al., Cell (1998)
63
Incorporation of ITP increases the number and
duration of the long pauses
Inosine
1mM NTPs
Template Position (bp)
Guanine
1mM NTPs 200 µM ITP
Time (s)
64
Incorporation of ITP increases the number and
duration of the long pauses
Inosine
Pauses per kb
Mean duration (s)
Guanine
200 µM ITP
1 mM NTPs
1 mM NTPs
200 µM ITP
65
Backtracking occurs at the beginning of
ITP-induced pauses
Inosine
1mM NTPs
Position (bp)
n56
Guanine
1mM NTPs 200 uM ITP
Position (bp)
n26
Time (s)
66
Evidence that long pauses are associated with
proofreading

• Do the long pauses backtrack?
• YES!
• Are the long pauses associated with
  misincorporation?
• YES!
• Do the cleavage factors GreA and GreB relieve
  long pauses?
• Does the polymerase cleave the RNA at the end of
  long pauses in the presence of the GreA and GreB?

67
GreA and GreB stimulate RNA cleavage
RNAP

• RNAP has an intrinsic cleavage ability at the
  polymerization active site that is stimulated by
  Gre proteins or high pH
• Functional analogs of GreA and GreB have been
  found in over 60 organisms, including TFIIS in
  eukaryotes

DNA
RNA
GreB
Prediction Addition of GreA and GreB should
reduce the duration, but not number of long pauses
Opalka et al., Cell 2003
68
Cleavage factors reduce the duration of long
pauses
Inosine
Pauses per kb
RNAP
DNA
Duration (s)
RNA
1 mM NTPs, 200 µM ITP
2 µM GreA, 1 µM GreB
69
Evidence that long pauses are associated with
proofreading

• Do the long pauses backtrack?
• YES!
• Are the long pauses associated with
  misincorporation?
• YES!
• Do the cleavage factors GreA and GreB relieve
  long pauses?
• YES!
• Does the polymerase cleave the RNA at the end of
  long pauses in the presence of the GreA and GreB?

70
GreA and GreB remove inosine from the transcript
radiolabeled ITP
GreA
GreB
full-length transcript
log (RNA Transcript Size)
cleavage products
single nucleotides
71
Recovery of motion is abrupt upon cleavage
Inosine
NTPs ITP
Position (bp)
sliding
n26
RNAP
NTPS ITP GreA,GreB
DNA
Position (bp)
cleavage
n22
RNA
Time (s)
72
The model of a molecular backspace key

• RNAP backtracks by 5 bp during long pauses.
• These pauses occur about once every thousand
  bases, and are likely due to misincorporation.
• Without GreA and GreB, the polymerase escapes
  from pause via slow forward sliding does not
  remove the error.
• With GreA and GreB, the polymerase escapes from
  pause by cleaving the newest RNA removes the
  error.

73
Biology presents a window into the nanoworld
74
Conclusions

• Molecular motors are fascinating machines that
  are important for both Biology and Nanotechnology
• Single molecule techniques offer a lot to these
  studies
• Optical traps are particularly useful for
  studying molecular motors displacements, force,
  temperature
• RNA polymerase is a VERY sophisticated enzyme
  able to correct errors

75
Acknowledgments
Steven Block Elio Abbondanzieri Keir
Neuman Charles Asbury Matt Lang Tom Perkins Megan
Valentine
Michael Woodside Ravi Dalal Adrian Fehr Polly
Fordyce Will Greenleaf Nicholas Guydosh Kristina
Herbert
Will Parks Viviana Risca Becky Wong
Bob Landick, U. Wisc.
http//www.stanford.edu/group/blocklab
76
(No Transcript)
77
Motors are everywhere in the cell
78
Nanobotsto the rescue!
sources Scientific American, Apr 96
Foresight online Nanomedicine Art Gallery
79
Popular Science, March 1989 work of R. Muller, UC
Berkeley
80
Optical traps are tools used to study individual
motor proteins
81
More temp figs
82
The DNA polymerase proofreading mechanism
http//ww2.mcgill.ca/biology/ undergra/c200a/sec2-
4.htm

• When a mismatched base is incorporated, the
  polymerase stalls
• The 3 end of the nascent strand is moved to a
  second active site
• An exonuclease activity removes the mismatched
  base
• Polymerization resumes after editing

Is something similar happening in RNA polymerase?
83
Short pause average (durations lt 5 seconds)
84
GreA and GreB promote cleavage of different
lengths of RNA
GreA
RNAP
DNA
Position (bp)
n3
RNA
GreB
GreB
Position (bp)
n8
Time (s)
85
GreA and GreB reduce the number of long pauses
RNAP
DNA
Pauses per kb
RNA
GreB
Duration (s)
2 µM GreA
1 mM NTP
1 µM GreB
86
GreB edits out XX of the mistakes
Length of transcribed region 262 Number of G
sites 66 Hot I conc 0.15 uM Cold G conc 0.50
uM I/(IG) ratio 23 GreB edits out 96 GreA
edits out 46 Denaturing Sequencing gel 10
Polyacrylimide RXns ran out long, 3 hours
87
Force tilts the energy landscape
F
elongation state
pause state
kp
k-p
Two parameters Pause Frequency (pauses per
kb) ? kp Pause Lifetime (seconds) 1 / k-p
free energy
reaction coordinate
d1
position
d2
88
(No Transcript)