NEWS & VIEWS RESEARCH
of A and B subunits. But further mutations protein Vma3 and one of its paralogue Vma16. processes might drive a system towards com-
might result in each subunit losing the ability In fungi such as budding yeast, by contrast, one plexity is more than a decade old8–10. But the
to bind others of its own type, A to A or B to B. of the five Vma3 proteins is replaced by Vma11, study by Thornton and colleagues1 may provide
However, as long as A still binds B and B still which is a fungus-specific paralogue of Vma3. the most compelling experimental evidence to
binds A, functional rings can form. The abilities of these subunits to bind to one date. Of course, one can never prove that some
As A and B continue to co-evolve through another7 are such that Vma11 must lie between subtle, unidentified selective advantage was
the acquisition of further neutral mutations, the Vma16 and Vma3 in the ring, and Vma3 cannot not involved in the evolution of the V-ATPase
likelihood of reversing back along the muta- form the interface with the side of Vma16 that protein ring, but neutrality would seem the
tional path to the initial single-gene homo- Vma11 can. This represents a sort of functional most justifiable default hypothesis. Thus, a
oligomeric complex becomes vanishingly degeneration, because the last common ances- neutral theory of molecular evolution, normally
small. We can thus see the homo-oligomer to tor of Vma3 and Vma11 must have been able invoked for nucleotide substitutions, may also
hetero-oligomer transition as an inexorable to interface with either side of Vma16, as Vma3 apply to certain higher-order structures such
evolutionary ratchet. And even if there were must still do in organisms other than fungi. as multi-subunit protein rings. How general
some slight selective disadvantage to using Thornton and colleagues 1 synthesized such neutral mutational drives to complexity
a hetero-oligomeric complex (for example, proteins that matched the phylogenetically might be is one of evolutionary theory’s deeper
complex formation could become slower), this inferred ancestor of Vma3 and Vma11, as well unanswered questions. ■
could be overcome in small populations — in as that of Vma16 and other intermediates in
which the efficiency of selection is reduced the predicted evolutionary pathway. They W. Ford Doolittle is in the Department
and mildly harmful mutations can be fixed demonstrate that Anc.3-11, the inferred of Biochemistry and Molecular Biology,
by chance2. ancestor of fungal Vma3 and Vma11, does Dalhousie University, Halifax, Nova Scotia
Some chaperonins — molecular machines indeed interface with either side of Vma16. B3H 4R2, Canada.
that assist other proteins to fold into functional The authors tested the functionality of these e-mail: ford@dal.ca
configurations — have until recently pro- and other inferred evolutionary ancestors and
1. Finnigan, G. C., Hanson-Smith, V., Stevens, T. H. &
vided the best evidence for such a scenario. intermediates by expressing them in yeast Thornton, J. W. Nature 481, 360–364 (2012).
Microorganisms known as archaea have mutants that lacked current versions of 2. Lynch, M. Proc. Natl Acad. Sci. USA 104, 8597–8604
one, two or three paralogous genes encoding one or more of the three genes. According (2007).
3. Archibald, J. M., Logsdon, J. M. Jr & Doolittle, W. F.
chaperonin subunits that form rings: homo- to this picture, when Anc.3-11 underwent Curr. Biol. 9, 1053–1056 (1999).
oligomeric eight-membered rings when duplication, one paralogue (which was to 4. Ruano-Rubio, V. & Fares, M. A. Mol. Biol. Evol. 24,
1384–1396 (2007).
there is a single gene, or hetero-oligomeric become modern fungal Vma3) lost the ability 5. Fares, M. A. & Wolfe, K. H. Mol. Biol. Evol. 20,
eight- or nine-membered rings when there to interact with the ‘anticlockwise’ side of 1588–1597 (2003).
are two or three genes. Such paralogy seems to Vma16. The other paralogue (which became 6. Ohno, S. Evolution by Gene Duplication (Springer,
1970).
reoccur frequently in archaea, and a neutral Vma11) lost the ability to interact with 7. Wang, Y., Cipriano, D. J. & Forgac, M. J. Biol. Chem.
co-evolutionary picture of the sort described the ‘clockwise’ side of Vma16 and/or the 282, 34058–34065 (2007).
above has been suggested3 to explain it. anticlockwise side of Vma3. 8. Stoltzfus, A. J. Mol. Evol. 49, 169–181 (1999).
9. Force, A. et al. Genetics 151, 1531–1545 (1999).
Comparative bioinformatic studies4 sup- A general theory, sometimes called construc- 10. Gray, M. W., Lukeš, J., Archibald, J. M., Keeling, P. J.
port the idea that hetero-oligomeric archaeal tive neutral evolution, to explain how neutral & Doolittle, W. F. Science 330, 920–921 (2011).
chaperonins consisting of two different sub-
units are not functionally distinct from the
homo-oligomers, and thus did arise neutrally. GA L A XY FOR M AT I ON
Adaptive changes can follow later, however.
Indeed, there is direct evidence for specialized
roles for the eight distinct chaperonin subunits
of eukaryotes (organisms such as plants,
Distant dwarfs
animals and fungi)5.
So a more realistic picture might be as Astronomers value the smallest galaxies for the clues they can provide about how
follows. Rarely, gene duplication is quickly galaxies form. But only those nearest to us are easily detectable. A neat technique
followed by functional differentiation (as pro- has allowed one such object to be found at a large distance. See Letter p.341
posed in early formulations of the hypothesis
of evolution by gene duplication6). Selection
against loss of either duplicate would then be R O B E R T W. S C H M I D T a satellite in the vicinity of a galaxy located at a
ensured early on. Perhaps more often, though, cosmological distance from Earth that allows
S
the paralogous condition might first be locked ome objects do not shine brightly enough them to test the simulation predictions for gal-
in by the above-described co-evolutionary to be seen easily. Examples of this are the axies beyond those of the Local Group.
ratchet, allowing functional differences (if small and extremely faint galaxies called The object studied by Vegetti et al. is a distant
any) to be acquired later (if ever). That is what dwarf or satellite galaxies. Only when they are elliptical galaxy that acts as a gravitational lens:
Thornton and colleagues1 argue has happened near to our Galaxy might we be lucky enough light from a background source is deflected
in the case of a six-membered hetero-oligomeric to detect the light of their stars. Satellite galax- from its otherwise straight path by a massive
ring that forms part of the enzymes known as ies are of great interest because simulations of foreground object, the lens. When compared
vacuolar H+-ATPases (V-ATPases) in fungi. the process of galaxy formation predict that with camera lenses, gravitational lenses have
Their argument is especially compelling there are many more satellites in the Local rather dismal optical properties because they
because it rests not only on bioinformatics, but Group, which includes the Milky Way and can lead to large distortions and multiple
also on experiments, including the synthesis other galaxies in its vicinity, than have yet been images of the background object. But from an
and testing of proteins that the authors propose found1–3. This begs the question of whether astronomer’s point of view this is a good thing:
are ancestral to the V-ATPase subunits. there is something wrong with our under- the mass of the gravitational lens can be meas-
In most eukaryotes, the V-ATPase six- standing of how galaxies form. On page 341 of ured by reconstructing the background object
membered ring is made up of five copies of the this issue, Vegetti et al.4 report the discovery of from the distorted images.
1 9 JA N UA RY 2 0 1 2 | VO L 4 8 1 | N AT U R E | 2 7 1
© 2012 Macmillan Publishers Limited. All rights reserved
RESEARCH NEWS & VIEWS
a b Satellite galaxy’s e ect
50 Years Ago
At the International Geological
Congress in Copenhagen in 1960,
and at the Russian Trade Exhibition
in London last year, one of the
outstanding scientific exhibits was
a hand-coloured tectonic map
of Europe compiled by Russian
draughtsmen from copy submitted
Figure 1 | Seeking satellites. Vegetti et al.4 have discovered a satellite galaxy around a larger galaxy,
by the various national geological
located at a cosmological distance, by examining how both galaxies act as gravitational lenses and distort
surveys. The map, on a scale of light from a background source as the light travels to Earth. a, Were there no lenses between an observer
1:2,500,000, has been sponsored by on Earth and a source — here illustrated as a kangaroo10 — the observer would, if instrumentation
the Sub-Commission on the Tectonic allowed, see one image of the source. b, If there is a massive lens between the observer and the source,
Map of the World of the International the source is magnified and imaged more than once. In this case, as in the authors’ study, the source
Geological Congress … Western is deformed into a ring that consists mainly of two distorted images of the source. The two images are
geologists who wish to make sure of caused by the larger galaxy, whereas the dent and blob features (labelled) are caused by the satellite galaxy,
receiving a copy should place an order, the mass of which has been exaggerated to demonstrate the effect more clearly.
in advance of publication, with their
national agents for Russian books. In the present study, the gravitational lens those of two other satellite galaxies that have
From Nature 20 January 1962 system, called JVAS B1938 + 666, consists of a been detected7,8 recently using gravitational
very distant galaxy that bends light from an lenses (with the possible exception of another
even more distant background galaxy. Light low-mass satellite8, although the detection of
100 Years Ago from the background galaxy is deformed into
a ring around the lens galaxy — a prime exam-
ple of what is known as an Einstein ring. In this
this object is not yet confirmed).
Vegetti and colleagues4 went on to deter-
mine the mass function of satellite galaxies
The possibility of the discovery of a case, the ring is formed mainly by two lensed — that is, the expected number of satellites
remedy for cancer has been advanced images of the background galaxy5. for a given mass — at distances beyond the
a stage by the preparation by Prof. Vegetti and her team4 obtained a near- local Universe by combining data from the
Wassermann, of Berlin, of a substance infrared image of JVAS B1938 + 666 using new object and another satellite galaxy7. They
which possesses a curative action the 10-metre Keck telescope in Hawaii. They found that the resulting mass function is con-
experimentally on cancer of mice. used an optics system that corrects for the blur- sistent with current galaxy-formation theory.
Prof. Wassermann reasoned that since ring effect of Earth’s atmosphere to improve But owing to the small number of known
the cancer-cells are growing rapidly, the image quality. With such data, they could objects, uncertainties in the deduced function
their oxygen requirements would neatly determine the mass distribution of the are quite large.
be different from, and greater than, lens galaxy, as well as the shape and brightness A natural question to ask is whether the
those of the cells of the body generally. of the background galaxy. satellite galaxy can be observed directly rather
He sought for some substance which And here comes the connection to satel- than by its gravitational effect on the shape of a
might interfere with the oxygen supply lite galaxies. The sophisticated numerical background object. With current instrumenta-
to the cancer-cells, and finally adopted technique6 used by the authors allowed them tion, the answer is no. The object is simply too
selenium as a means to do this. The not only to derive a model of the lens galaxy’s distant to be imaged directly. But the message
next problem was to convey selenium mass, but also to map any excess lens mass that here4 is that it is possible to spot these elusive
to the cancer-cells by means of the could not be accounted for by the galaxy. They objects around distant lens galaxies without
blood stream, and … a compound of found an excess mass near the Einstein ring knowing where to look for them.
selenium with an anilin dye eosin was that they attributed to the presence of a satellite The satellite should be considered in the
found to fulfil this condition … After galaxy (Fig. 1). This method has been dubbed context of about 50 satellite galaxies that are
two or three injections of the substance gravitational imaging6. Vegetti et al. also used known to exist in the Local Group and that
into a mouse the subject of cancerous an analytical model to test the detected excess have been found by imaging their starlight3,9.
tumours, the tumours are found to mass and found that a satellite galaxy is indeed The number of these nearby satellites has
have softened, and after six to eight required to explain the data. greatly increased over the past decade, and
doses they become cystic, diminish This satellite galaxy is exciting because it often only a hundred (or fewer) stars are
in size, and finally disappear, and no was detected in the excess-mass map despite detected in them. From measurements of the
recurrence takes place. The eosin– its low mass. Assuming that the object is in the velocities of their member stars, the satel-
selenium compound is, however, neighbourhood of the main lens galaxy, it has lites have been shown to be dominated by an
poisonous, and a certain number of a mass of some 113 million solar masses within unseen mass component called dark matter.
mice succumb under the treatment. a radius of about 600 parsecs (2,000 light years) Dark matter is the reason that satellites such as
From Nature 18 January 1912 — values that put it firmly in the realm of sat- the newly discovered one are massive enough
ellite galaxies. The mass is tenfold lower than to be found using gravitational lenses.
2 7 2 | N AT U R E | VO L 4 8 1 | 1 9 JA N UA RY 2 0 1 2
© 2012 Macmillan Publishers Limited. All rights reserved
NEWS & VIEWS RESEARCH
Despite this increase in the observed num- Vegetti and colleagues show convincingly 1. Bullock, J. S. Preprint at http://arxiv.org/
ber of satellites in the Local Group, simulations that gravitational lenses allow low-mass satel- abs/1009.4505 (2010).
2. Kravtsov, A. Adv. Astron. 2010, 281913 (2010).
indicate that there are many more satellites lite galaxies to be discovered regardless of how 3. Willman, B. Adv. Astron. 2010, 285454 (2010).
dominated by dark matter than have been luminous the satellites are. As such, they offer 4. Vegetti, S. et al. Nature 481, 341–343 (2012).
seen. But a fair fraction of the local satellites the means to refine measurements of the mass 5. King, L. J. et al. Mon. Not. R. Astron. Soc. 295,
will probably be found by sensitive searches3, function of faraway satellites. ■ L41–L44 (1998).
6. Koopmans, L. V. E. Mon. Not. R. Astron. Soc. 363,
and theorists have plenty of ideas about why 1136–1144 (2005).
satellites may be lacking in stars1,2. If more Robert W. Schmidt is at the Zentrum für 7. Vegetti, S. et al. Mon. Not. R. Astron. Soc. 408,
satellites are discovered at cosmological dis- Astronomie der Universität Heidelberg, 1969–1981 (2010).
8. Fadely, R. & Keeton, C. R. Mon. Not. R. Astron. Soc.
tances using gravitational lenses, they will have Astronomisches Rechen-Institut, 419, 936–951 (2012).
the potential to independently improve our 69120 Heidelberg, Germany. 9. Richardson, J. C. et al. Astrophys. J. 732, 76 (2011).
knowledge of the satellite mass function4,7,8. e-mail: rschmidt@ari.uni-heidelberg.de 10. Wambsganss, J. Nature 386, 27–28 (1997).
C E LL CYCLE during cytokinesis, the final stage of mitosis.
Potential resolutions of that conundrum
A division duet were set out in early models that explain
the order of events during the cell cycle. In the
domino model, order is a consequence of
the dependency of later events on comple-
The orchestration of cell division requires a programme of events choreographed tion of those occurring earlier; in the clock
by protein modification. A study shows that the relative activity of a phosphatase model, order is imposed by an intrinsic timer5.
enzyme towards its substrates imposes order during the final act of division. Bouchoux and Uhlmann’s study distinguishes
between these two models, at least in the con-
text of cell-cycle events that depend on Cdc14.
CURT WITTENBERG is sequestered in a region of the nucleus called The authors engineered yeast strains in
the nucleolus, and so does not have access to which the activities of Cdc14 and CDK could
P
rotein phosphorylation plays a central its protein targets. However, it has remained be modulated. Using careful in vivo analy-
part in the intricate choreography of unclear how, after its release, Cdc14 triggers sis, they found that Cdc14 dephosphorylates
cell-cycle events 1 . A minimalistic the events that it regulates in a specific order. CDK substrates in a sequence consistent with
model that holds that simple oscillations in Indeed, elongation of the mitotic spindle the order of cell-cycle events. That is, dephos-
the activity of a kinase enzyme called CDK occurs subsequent to spindle attachment phorylation events that promote spindle elon-
— which phosphorylates cell-cycle pro- to chromosomes, but before cell separation gation occur after those that promote spindle
teins — is sufficient to promote the attachment to chromosomes, but before
orderly progression of the cycle has those that promote spindle disassembly.
been tested and confirmed2. However, Cdc14
Cdc14 activity rises upon release from
protein phosphorylation, like other the nucleolus, allowing it to compete
reversible cellular processes, depends with CDK activity6. Bouchoux and Uhl-
not only on enzymes that catalyse the mann3 performed careful analysis using
addition of phosphate groups, but also purified proteins in vitro and show that
on those that catalyse dephosphoryla- the timing of dephosphorylation in vivo
tion — the protein phosphatases. In a PP2 PP3 PP4 is a consequence of Cdc14’s differential
PP1
paper published in Cell, Bouchoux and activity towards specific substrates.
Uhlmann3 show that the intrinsic ability That is, ‘early’ Cdc14 substrates — those
E2 E3 E4
of a protein phosphatase to dephospho- E1 dephosphorylated in vivo during the ini-
rylate specific substrates is crucial for CDK M tial Cdc14-triggered events — are more
establishing the proper order of events efficiently dephosphorylated by this
during the final stages of mitotic cell G2 G1 enzyme in vitro than are ‘late’ substrates
division*. (Fig. 1). Furthermore, by establishing a
The cell cycle proceeds from the G1 competition between CDK and Cdc14
gap phase through the S phase, in which in vitro — a situation that exists dur-
chromosomes are duplicated, and then ing anaphase, the penultimate stage of
S
on to the G2 gap phase and M phase mitosis, when both enzymes are highly
(mitosis). At this point, the duplicated active — they showed that dephospho-
chromosomes are separated, leading to rylation of early substrates is favoured.
the creation of two daughter cells back Figure 1 | Imposing order through protein modification. The These findings establish that differences
at G1 phase. The protein phosphatase cell cycle proceeds from the G1 phase to the S, G2 and, finally, in the intrinsic activity of Cdc14 for spe-
Cdc14 is required for the completion M phases. At the end of M phase, the cell divides into two daughter cific substrates is sufficient to explain
of chromosome segregation and subse- cells, which enter the G1 phase. Several events (hypothetical events the order of dephosphorylation events
quent events that lead to the resetting E1–E4) mediate the passage from M phase into G1 phase, and, for observed in vivo; they also show that
of the ‘cell-cycle clock’ such that the them to occur in an orderly manner, proteins phosphorylated Cdc14 is an important component of a
by CDK (hypothetical phosphoproteins PP1–PP4) must be
G1-phase daughter cells can initiate a dephosphorylated by Cdc14 in the right order (from PP1 to PP4). cell-cycle clock.
new cycle4 (Fig. 1). Those events are pre- Bouchoux and Uhlmann3 show that the differential activity Could mechanisms besides the
vented before M phase because Cdc14 of Cdc14 towards CDK-phosphorylated proteins determines intrinsic activity of Cdc14 towards its
*This article was published online on their orderly dephosphorylation, and so the order of events substrates determine the order of pro-
11 January 2012. that depend on them. tein dephosphorylation? Bouchoux and
1 9 JA N UA RY 2 0 1 2 | VO L 4 8 1 | N AT U R E | 2 7 3
© 2012 Macmillan Publishers Limited. All rights reserved