12 AUGUST 2016 • VOL 353 ISSUE 6300 647 SCIENCE sciencemag.org
domains can move independently and
KCNH1 does not contain a conventional S4-
S5 linker, how do conformational changes
in S4 trigger opening and closing of the
S6 gate? Although a definitive answer will
require more experiments, Whicher and
MacKinnon propose that at negative membrane voltages, inward movement of each
S4 helix would allow it to directly interact
with the C-linker helix of the adjacent subunit and thereby stabilize a closed conformation resembling the CaM-inhibited state.
These new structural relations can explain the surprising discovery that voltage-dependent gating in KCNH1 and KCNH2
channels does not require a covalent linkage between S4 and S5 (13), and provide a
framework for exploring how inward motions of the S4 helix open rather than close
the pore in HCN channels. Indeed, the findings from an earlier investigation of state-dependent interactions between the inner
end of S5 and the C-linker in HCN channels are compatible with the new structure
of KCNH1, but not that of conventional Kv
It is truly remarkable how this single new
structure of a voltage-activated ion channel
sparks so many fresh ideas about how voltage, nucleotides, and regulatory proteins
can control the activity of this singular and
diverse family of ion channel proteins. The
structure of KCNH1 reported by Whicher
and MacKinnon will be a windfall for scientists studying members of the KCNH,
CNG, and HCN families of ion channels,
and may help us to understand why the
related hERG (KCNH2) channel interacts
with drugs so often that the pharmaceutical
industry must screen for this off-target interaction early in drug development (15). j
1. T. W. Kaplan, Genetics 61, 399 (1969).
2. J. Warmke, R. Drysdale, B. Ganetzky, Science
252, 1560 (1991).
3. J.H.Morais-Cabral, G.A.Robertson, J. Mol. Biol.
427, 67 (2015).
4. J. R. Whicher, R. MacKinnon, Science 353, 664 (2016).
5. S. B. Long, E. B. Campbell, R. Mackinnon,
Science 309, 897 (2005).
6. K. B. Craven, W. N. Zagotta, Annu. Rev. Physiol.
68, 375 (2006).
7. Y. Haitin, A. E. Carlson, W. N. Zagotta,
Nature 501, 444 (2013).
8. E. C. Gianulis, Q. Liu, M. C. Trudeau, J.Gen.Physiol.
142, 351 (2013).
9. R. Schonherr, K. Lober, S. H. Heinemann,
EMBO J. 19, 3263 (2000).
10. J.Payandeh, T.Scheuer,N.Zheng, W.A.Catterall,
Nature 475, 353 (2011).
11. J. Wu et al. , Science 350, aad2395 (2015).
12. S. B. Long, E. B. Campbell, R. Mackinnon,
Science 309, 903 (2005).
13. E. Lorinczi et al. , Nat. Commun. 6, 6672 (2015).
14. D. C. Kwan, D. L. Prole, G. Yellen, J. Gen. Physiol.
140, 279 (2012).
15. M. C. Sanguinetti, M. Tristani-Firouzi,
Nature 440, 463 (2006).
By Sarah Mizielinska and Adrian M. Isaacs
Repeat expansion mutations cause a range of developmental, neuro- degenerative, and neuromuscular disorders. The repeat sequences generally comprise a 3– to 6–base pair repeat unit that expands above
a critical threshold, leading to disease. Expanded repeats cause disease via a range
of mechanisms, including loss of function
of the repeat-containing protein and production of toxic repeat RNAs and proteins,
making the disorders difficult to treat. In
2011, a hexanucleotide repeat expansion in
the C9orf72 gene was identified as the most
common cause of frontotemporal dementia
and amyotrophic lateral sclerosis (termed
c9FTD/ALS) (1, 2). On page 708 of this
issue, Kramer et al. (3) report that targeting
a single factor, Spt4, reduced production of
C9orf72 repeat expansion–associated RNA
and protein, and ameliorated neurodegeneration in model systems.
Kramer et al.’s use of a single factor to reduce multiple repeat-associated pathologies
is notable in the light of two unexpected
features of repeat expansions. One is that
repeat expansions are transcribed in both
the antisense and sense direction (4). The
other is that repeat-associated non-ATG
(RAN) translation occurs, in which repeat
expansions mediate their own translation
into proteins (5). As no ATG start codon is
required, RAN translation can occur in all
six sense and antisense frames. A major
therapeutic challenge is to target the wide
range of potentially toxic RNA and protein
species that are produced.
The yeast Spt4 (human homolog
SUPT4H1) is a small, evolutionarily con-
served zinc finger protein that forms a
complex with Spt5. The Spt4-Spt5 complex
binds to RNA polymerase II and regulates
transcriptional elongation. Deletion of Spt4
in yeast was shown to reduce the transcrip-
tion of expanded CAG, CTG, and CAA re-
peats, but had little effect on short repeats
(6). Similar effects were observed with CAG
Kramer et al. now extend this work to
C9orf72 GGGGCC repeat expansions. In
c9FTD/ALS, both sense and antisense re-
peat RNA transcripts form aggregates,
termed RNA foci, in patient brains. RNA
foci exert toxicity in other repeat expan-
sion diseases by sequestering essential
RNA-binding proteins (8). Using yeast
models expressing either expanded sense
or antisense C9orf72 repeats, Kramer et al.
found that Spt4 depletion decreased the
number of both sense and antisense repeat
transcripts and RNA foci.
C9orf72 RAN translation leads to the
production of five dipeptide repeat proteins that can cause neurodegeneration in
model systems (9). Kramer et al. showed
that production of one of them, poly(glycine-proline), was substantially reduced by
Spt4 depletion in yeast and worm C9orf72
models, as would be expected if less repeat
RNA was available for translation. Reducing the expression of Spt4 also improved
survival in C9orf72 worm and fruit fly models, indicating a reduction of toxic repeat
To study endogenous expanded repeats,
Kramer et al. used human c9ALS patient
fibroblasts. Reducing the expression of ei-
ther human SUPT4H1 or its binding part-
One target for amyotrophic
lateral sclerosis therapy?
Targeting a single protein reduces both toxic
repeat RNAs and proteins
“A major therapeutic
challenge is to target the
wide range of potentially
toxic RNA and protein
species that are produced.”
Department of Neurodegenerative Disease, UCL Institute of
Neurology, London WC1N 3BG, UK. Email: email@example.com