Вы находитесь на странице: 1из 3

A LOOP OF FAITH

One of the most puzzling problems in genome architecture may have a


simple solution. But no one can agree on what powers the process.
BY ELIE DOLGIN

L
eonid Mirny swivels in his office chair up like last year’s Christmas lights. explosive growth for research into the 3D
M. IMAKAEV/G. FUDENBERG/
N. NAUMOVA/J. DEKKER/L. MIRNY

and grabs the power cord for his laptop. He argues that DNA is constantly being structure of the genome. The models neatly
He practically bounces in his seat as he slipped through ring-like motor proteins to explain the data flowing from high-profile
threads the cable through his fingers, creating make loops. This process, called loop extru- projects on how different parts of the genome
a doughnut-sized loop. “It’s a dynamic process sion, helps to keep local regions of DNA interact physically — which is why they’ve
of motors constantly extruding loops!” says together, disentangling them from other parts garnered so much attention.
Mirny, a biophysicist here at the Massachusetts of the genome and even giving shape and But these simple explanations are not
Institute of Technology in Cambridge. structure to the chromosomes. without controversy. Although it has become
Mirny’s excitement isn’t about keeping Scientists have bandied about similar increasingly clear that genome looping regu-
computer accessories orderly. Rather, he’s hypotheses for decades, but Mirny’s model, lates gene expression, possibly contributing to
talking about a central organizing principle and a similar one championed by Erez cell development and diseases such as cancer,
of the genome — how roughly 2 metres of Lieberman Aiden, a geneticist at Baylor the predictions of the models go beyond what
DNA can be squeezed into nearly every cell College of Medicine in Houston, Texas, add anyone has ever seen experimentally.
of the human body without getting tangled a new level of molecular detail at a time of For one thing, the identity of the molecular

2 8 4 | NAT U R E | VO L 5 4 4 | 2 0 A P R I L 2 0 1 7
©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
FEATURE NEWS

DNA loops help to keep machine that forms scientists sequence bits of DNA that are close pair, why did loops form only locally, and not
local regions of the the loops remains a to one another and produce a map of each between distant sites?
genome together. mystery. If the lead- chromosome, usually depicted as a fractal- Mirny’s model assumes that CTCFs act as
ing protein candi- like chessboard. The darkest squares along stop signs for cohesin. If cohesin stops extrud-
date acted like a motor, as Mirny proposes, the main diagonal represent spots of closest ing DNA only when it hits CTCFs on each side
it would guzzle energy faster than it has ever interaction. of a growing loop, it will naturally bring the
been seen to do. “As a physicist friend of mine The Hi-C snapshots that Dekker and his proteins together.
tells me, ‘This is kind of the Higgs boson of collaborators had taken revealed distinct But singling out cohesin was “a big leap of
your field’,” says Mirny; it explains one of the compartmentalized loops, with interactions faith”, says biophysicist Geoff Fudenberg, who
deepest mysteries of genome biology, but happening in discrete blocks of DNA between did his PhD in Mirny’s lab and is now at the
could take years to prove. 200,000 and 1 million letters long5. University of California, San Francisco. “No
And although Mirny’s model is extremely These ‘topologically associating domains’, or one has seen these motors doing these things
similar to Lieberman Aiden’s — and the dif- TADs, are a bit like the carriages on a crowded in living cells or even in vitro,” he says. “But
ferences esoteric — sorting out which is right train. People can move about and bump into we see all of these different features of the
is more than a matter of tying up loose ends. If each other in the same carriage, but they can’t data that line up and can be unified under this
Mirny is correct, “it’s a complete revolution in interact with passengers in adjacent carriages principle.”
DNA enzymology”, says Kim Nasmyth, a lead- unless they slip between the end doors. The Experiments had shown, for example, that
ing chromosome researcher at the University human genome may be 3 billion nucleotides reducing the amount of cohesin in a cell results
of Oxford, UK. What’s actually powering the long, but most interactions happen locally, in the formation of fewer loops8. Overactive
loop formation, he adds, “has got to be the big- within TADs. cohesin creates so many loops that chromo-
gest problem in genome biology right now”. Mirny and his team had been labouring for somes smush up into structures that resemble
more than a year to explain TAD formation tiny worms9.
LOOP BACK using computer simulations. Then, as luck The authors of these studies had trouble
Geneticists have known for more than three making sense of their results. Then came
decades that the genome forms loops, bringing Mirny’s paper on bioRxiv. It was “the first time

“IT’S A COMPLETE
regulatory elements into close proximity with that a preprint has really changed the way peo-
genes that they control. But it was unclear how ple were thinking about stuff in this field”, says
these loops formed. Matthias Merkenschlager, a cell biologist at the
Several researchers have independently put MRC London Institute of Medical Sciences.
forward versions of loop extrusion over the
years. The first was Arthur Riggs, a geneticist
at the Beckman Research Institute of City of
REVOLUTION IN DNA (Mirny’s team eventually published the work
in May 2016, in Cell Reports6.)

MULTIPLE DISCOVERY?

ENZYMOLOGY.”
Hope in Duarte, California, who first proposed
what he called “DNA reeling” in an over- Lieberman Aiden says that the idea of loop
looked 1990 report1. Yet it’s Nasmyth who is extrusion first dawned on him during a con-
most commonly credited with originating the ference call in March 2015. He and his former
concept. mentor, geneticist Eric Lander of the Broad
As he tells it, the idea came to him in 2000, would have it, Mirny happened to attend a Institute in Cambridge, Massachusetts, had
after a day spent mountain climbing in the Ital- conference at which Marko spoke about his published some of the most detailed, high-
ian Alps. He and his colleagues had recently then-unpublished model of loop extrusion. resolution Hi-C maps of the human genome
discovered the ring-like shape of cohesin2, a (Marko coined the term, which remains in use available at the time10.
protein complex best known for helping to today.) It was the missing piece of Mirny’s puz- During his conference call, Lieberman
separate copies of chromosomes during cell zle. The researchers gave loop extrusion a try, Aiden was trying to explain a curious phenom-
division. As Nasmyth fiddled with his climb- and it worked. The physical act of forming the enon in his data. Almost all the CTCF landing
ing gear, it dawned on him that chromosomes loops kept the local domains well organized. sites that anchored loops had the same orienta-
might be actively threaded through cohesin, The model reproduced many of the finer-scale tion. What he realized was that CTCF, as a stop
or the related complex condensin, in much features of the Hi-C maps. sign for extrusion, had inherent directionality.
the same way as the ropes looped through his When Mirny and his colleagues posted their And just as motorists race through intersec-
carabiners. “It appeared to explain everything,” finished manuscript on the bioRxiv preprint tions with stop signs facing away from them,
he says. server in August 2015, they were careful to so a loop-extruding factor goes through CTCF
Nasmyth described the idea in a few para- describe the model in terms of a generic “loop- sites unless the stop sign is facing the right way.
graphs in a massive, 73-page review article3. extruding factor”. But the paper didn’t shy away His lab tested the model by systematically
“Nobody took notice whatsoever,” he says — from speculating as to its identity: cohesin was deleting and flipping CTCF-binding sites, and
not even John Marko, a biophysicist at North- the driving force behind the looping process remapping the chromosomes with Hi-C. Time
western University in Evanston, Illinois, who for cells not in the middle of dividing, when and again, the data fitted the model. The team
more than a decade later developed a math- chromosomes are loosely packed6. Condensin, sent its paper for review in July 2015 and pub-
ematical model that complemented Nasmyth’s they argued in a later paper, served this role lished the findings three months later11.
verbal argument4. during cell division, when the chromosomes Mirny’s August 2015 bioRxiv paper didn’t
Mirny joined this loop-modelling club are tightly wound7. have the same level of experimental valida-
around five years ago. He wanted to explain A key clue was the protein CTCF, which was tion, but it did include computer simulations
data sets compiled by biologist Job Dekker, known to interact with cohesin at the base of to explain the directional bias of CTCF. In fact,
a frequent collaborator at the University of each loop of uncondensed chromosomes. For a both models make essentially the same predic-
Massachusetts Medical School in Worcester. long time, researchers had assumed that loops tions, leading some onlookers to speculate on
Dekker had been looking at physical inter- form on DNA when these CTCF proteins whether Mirny seeded the idea. Lieberman
actions between different spots on chromo- bump into one another at random and lock Aiden insists that he came up with his model
somes using a technique called Hi-C, in which together. But if any two CTCF proteins could independently. “We submitted our paper

2 0 A P R I L 2 0 1 7 | VO L 5 4 4 | NAT U R E | 2 8 5
©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
NEWS FEATURE

before I ever saw their manuscript,” he School in Boston, Massachusetts, and his

NIK SPENCER/NATURE
says.
There are some tiny differences. The T H E TA M I N G O F T H E TA N G L E S colleagues made time-lapse Hi-C maps
of the bacterium Bacillus subtilis that
cartoons Mirny uses to describe his model A ‘loop extrusion’ model for genome organization helps reveal SMC zipping along the chromo-
to explain how certain regions of chromosomes stay
seem to suggest that one cohesin ring close together and why rings of the protein complex some and creating a loop at a rate of more
does the extruding, whereas Lieberman cohesin are often found in contact with CTCF proteins, than 50,000 DNA letters per minute. This
which bind to specific DNA sequences.
Aiden’s contains two rings, connected like tempo is on par with what researchers
a pair of handcuffs (see ‘The taming of the estimate would be necessary for Mirny’s
tangles’). Suzana Hadjur, a cell biologist model to work in human cells as well.
at University College London, calls this Packaged Models differ as to Rudner hasn’t yet proved that SMC
whether a single or
mechanistic nuance “absolutely funda- DNA
double loop of cohesin
uses ATP to make that happen. But,
mental” to determining cohesin’s role in is needed for extrusion. he says, he’s close — and he would be
the extrusion process. “shocked” if cohesin worked differently
Neither Lieberman Aiden nor Mirny Cohesin in human cells.
say they have a strong opinion on whether complex For now, the debate rages about what
the system uses one ring or two, but they cohesin is, or is not, doing inside the
do differ on cohesin’s central contribution cell — and many researchers, including
to loop formation. Mirny maintains that CTCF Doug Koshland, a cell biologist at the
As a loop is fed
the protein is the power source for looping, through cohesin, University of California, Berkeley, insist
whereas Lieberman Aiden summarily dis- local regions of the that a healthy dose of scepticism is still
chromosome are
misses this idea. Cohesin “is a big dough- kept close together. warranted when it comes to Mirny’s idea.
nut”, he says. It doesn’t do that much. “It “I am worried that the simplicity and
can open and close, but we are very, very elegance of the loop-extrusion model is
confident that cohesin itself is not a motor.” STOP already filling textbooks, coronated long
Instead, he suspects that some other fac- before its time,” he says.
tor is pushing cohesin around, and many in And although it may seem an academic
the field agree. Claire Wyman, a molecular A CTCF protein will dispute among specialists, Mirny notes
stop the extrusion of
biophysicist at Erasmus University Medi- the DNA it’s bound to,
that if it his model is correct, it will have
cal Centre in Rotterdam, the Netherlands, but only if it points in real-world implications. In cancer, for
points out that cohesin is only known to STOP the right direction. instance, cohesin is frequently mutated
consume small amounts of energy for and CTCF sites altered. Defective versions
clasping and releasing DNA, so it’s a stretch of cohesin have also been implicated in
to think of it motoring along the chromo- several rare human developmental disor-
some at the speeds required for Mirny’s ders. If the loop-extruding process is to
model to work. “I’m willing to concede that blame, says Mirny, then perhaps a better
it’s possible,” she says. “But the Magic 8-Ball STOP understanding of the motor could help fix
would say that, ‘All signs point to no’.” A finished loop the problem.
will have two
One group of proteins that might be CTCF proteins But his main interest remains more
doing the pushing is the RNA polymer- STOP
and cohesin at fundamental. He just wants to under-
ases, the enzymes that create RNA from a its base. stand why DNA is configured in the way
DNA template. In a study online in Nature it is. And although his model assumes a
this week12, Jan-Michael Peters, a chromo- lot of things about cohesin, Mirny says,
some biologist at the Research Institute of “The problem is that I don’t know any
Molecular Pathology in Vienna, and his other way to explain the formation of
colleagues show that RNA polymerases these loops.” ■
can move cohesin over long distances on
the genome as they transcribe genes into Elie Dolgin is a science writer in
RNA. “RNA polymerases are one type of Somerville, Massachusetts.
motor that could contribute to loop extrusion,” extrusion in action in a test tube. He throws
1. Riggs, A. D. Phil. Trans. R. Soc. Lond. B 326,
Peters says. But, he adds, the data indicate that in just three ingredients: DNA, some ATP to 285–297 (1990).
it cannot be the only force at play. provide energy, and the bacterial equivalent 2. Haering, C. H., Löwe, J., Hochwagen, A. &
Frank Uhlmann, a biochemist at the Francis of cohesin and condensin, a protein complex Nasmyth, K. Mol. Cell 9, 773–788 (2002).
Crick Institute in London, offers an alternative known as SMC. 3. Nasmyth, K. Annu. Rev. Genet. 35, 673–745 (2001).
4. Alipour, E. & Marko, J. F. Nucleic Acids Res. 40,
that doesn’t require a motor protein at all. In his “We see evidence of DNA being compacted 11202–11212 (2012).
view, a cohesin complex might slide along DNA into these kinds of flowers with loops,” says 5. Nora, E. P. et al. Nature 485, 381–385 (2012).
randomly until it hits a CTCF site and creates a Noy, who is collaborating with Mirny on the 6. Fudenberg, G. et al. Cell Rep. 15, 2038–2049
(2016).
loop. This model requires only nearby strands project. That suggests that SMC — and by 7. Goloborodko, A., Imakaev, M. V., Marko, J. F. &
of DNA to interact randomly — which is much extension cohesin — might have a motor func- Mirny, L. eLife 5, e14864 (2016).
more probable, Uhlmann says. “We do not tion. But then again, it might not. “The truth is 8. Zuin, J. et al. Proc. Natl Acad. Sci. USA 111,
need to make any assumptions about activities that we just don’t know at this point,” Noy says. 996–1001 (2014).
9. Tedeschi, A. et al. Nature 501, 564–568 (2013).
that we don’t have experimental evidence for.” 10. Rao, S. S. P. et al. Cell 159, 1665–1680 (2014).
Researchers are trying to gather experi- BACTERIAL BATTERY 11. Sanborn, A. L. et al. Proc. Natl Acad. Sci. USA 112,
mental evidence for one model or another. The experiment that perhaps comes the E6456–E6465 (2015).
12. Busslinger, G. A. et al. Nature http://dx.doi.
At the Lawrence Livermore National Labora- closest to showing cohesin acting as a motor org/10.1038/nature22063 (2017).
tory in California, for example, biophysicist was published in February13. David Rudner, 13. Wang, X., Brandão, H. B., Le, T. B. K., Laub, M. T. &
Aleksandr Noy is attempting to watch loop a bacterial cell biologist at Harvard Medical Rudner, D. Z. Science 355, 524–527 (2017).

2 8 6 | NAT U R E | VO L 5 4 4 | 2 0 A P R I L 2 0 1 7
©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.

Похожие интересы