forces operating during mitosis. In this context, the pervasively depicted mechanism of
chromatin compaction assuming an ordered
hierarchical folding should perhaps be discarded. Instead, the unwound state of the
30-nm chromatin fiber containing the linker
histone could adopt alternative modes of
nucleosome stacking—either a helix corresponding to the canonical 30-nm fiber or a
network of more continuous stacks containing nucleosomes from one, or even several,
individual fibers (see the figure).
The details of the structure elucidated
by Song et al. raise several important ques-
tions. Does it correspond to the “native”
30-nm fiber with fewer nucleosomes per
turn or to an alternative structure? And does
unwinding of the fiber involve change to
internucleosomal interactions? The alternat-
ing stacking suggests the potential for fine-
tuning epigenetic marks and array organi-
zation. A prime candidate for this would be
the interface of the globular domains of the
linker histone within the tetranucleosomal
unit. Similarly, the details of the epigenetics
and the transition to mitotic chromatin are
poorly understood. These, and related, ques-
tions should ensure that, despite its pur-
ported demise, the 30-nm fiber, in what-
ever form, will continue to be a structure of
intense interest for biologists.
1. J. T. Finch, A. Klug, Proc. Natl. Acad. Sci. U. S.A. 73, 1897
2. P. T. Lowary, J. Widom, J. Mol. Biol. 276, 19 (1998).
3. F. Song et al., Science 344, 376 (2014).
4. T. Schalch, S. Duda, D. F. Sargent, T. J. Richmond, Nature
436, 138 (2005).
5. D. Z. Staynov, Nucleic Acids Res. 28, 3092 (2000).
6. P. J. J. Robinson, L. Fairall, V. A. Huynh, D. Rhodes, Proc.
Natl. Acad. Sci. U.S.A. 103, 6506 (2006).
7. D. Z. Staynov, C. Crane-Robinson, EMBO J. 7, 3685
8. S. H. Syed et al., Proc. Natl. Acad. Sci. U. S.A. 107, 9620
9. L. Fan, V. A. Roberts, Proc. Natl. Acad. Sci. U. S.A. 103,
10. A. C. Lennard, J. O. Thomas, EMBO J. 4, 3455 (1985).
11. F. Livolant et al., Philos. Trans. A Math. Phys. Eng. Sci.
364, 2615 (2006).
12. M. P. Scheffer, M. Eltsov, J. Bednar, A. S. Frangakis, J.
Struct. Biol. 178, 207 (2012).
13. E. Fussner et al., Trends Biochem. Sci. 36, 1 (2011).
14. P. Recouvreux et al., Biophys. J. 100, 2726 (2011).
15. Y. Cui, C. Bustamante, Proc. Natl. Acad. Sci. U.S.A. 97,
For almost a century, neuroscientists have tried to understand how patterns of neuronal activity generate behavior.
Many of the early studies turned to the “
simple” systems of invertebrates in the hope of
discovering the components of circuits and
their connections. A striking finding was the
existence of “command neurons” in arthropods and molluscs that produce complex and
coordinated movements when stimulated (1,
2). The challenge was then to identify the
connections between these neurons and other
neurons important for those behaviors. This
approach was always limited to examining
only a few neurons from the tens or hundreds
of thousands in the animal (3). On page 386
of this issue, Vogelstein et al. (4) usher in a
new era of integrated methods for deciphering how an entire nervous system generates
The nervous systems of even simple
organisms like the fruit fly typically com-
prise thousands of densely interconnected
neurons with a variety of specialized sig-
naling properties. Mapping neural activity
patterns onto behavior therefore requires
unambiguous identification and control of
small, distinct sets of neurons. Moreover,
the behavioral repertoire itself is complex—
how many different “behaviors” does an
animal have? Can they be delineated in a
principled, practical way?
Three new technologies enabled Vogel-
stein et al. to attack these questions in a way
that was unimaginable only a few years ago.
Genetic tools (5) developed in Drosophila
melanogaster enable reliable genetic tag-
ging of the same sets of neurons in different
animals. Optogenetics, in particular light-
activated ion channels (6), allow precisely
timed, noninvasive activation and deactiva-
tion of individual neurons. And advances
in machine learning algorithms and com-
puter hardware facilitate the daunting
task of turning thousands of hours of
video-recorded behavior into a manageable
data set. What makes the study of Vogel-
stein et al. unique and important is that
it successfully combines these technolo-
gies to analyze the function of an entire
Over 1000 different genetic lines of
Drosophila were used to target specific neurons
in the larval nervous system. Vogelstein et
al. then systematically stimulated the identified neurons and recorded behavioral output, resulting in a large data set. A key step
in handling the data was to reduce their
dimension in a way that preserved features
(such as forward-backward locomotion
and body curvature) such that they could
be sorted by an unsupervised classification
algorithm into distinct classes of behaviors
(see the figure).
Mapping Neural Activation onto
Behavior in an Entire Animal
Timothy O’Leary and Eve Marder
A combination of technologies reveals which
neurons constitute circuits for specific behaviors in Drosophila larvae.
Volen Center and Biology Department, Brandeis University,
Waltham, MA 02453, USA. E-mail: firstname.lastname@example.org
Neuronal activity Behavior
Neurons and behavior.
Neural activity induced in
single cells (left) is associated with behavioral output
(right), shown in the color-coded schematic. In general,
this relationship is probabilistic and not necessarily one-to-one: Multiple behaviors
can be associated with activation of the same neurons and
multiple neurons can trigger
the same behavior.