REFERENCES AND NOTES
1. G. O. Wasteneys, Curr. Opin. Plant Biol. 7,651–660 (2004).
2. C. P. Joshi, S. D. Mansfield, Curr . Opin. Plant Biol. 10,220–226 (2007).
3. I. M. Saxena, R. M. Brown Jr., Ann. Bot. 96,9–21 (2005).
4. S. C. Mueller, R. M. Brown Jr., J. Cell Biol. 84,315–326 (1980).
5. T. Desprez et al., Proc. Natl. Acad. Sci. U.S.A. 104,
6. S. Persson et al., Proc. Natl. Acad. Sci. U.S.A. 104,
7. N. G. Taylor, R. M. Howells, A. K. Huttly, K. Vickers, S. R. Turner,
Proc. Natl. Acad. Sci. U.S.A. 100,1450–1455 (2003).
8. A. R. Paradez, C. R. Somerville, D. W. Ehrhardt, Science 312,
9. E. F. Crowell et al., Plant Cell 21,1141–1154 (2009).
10. R. Gutierrez, J. J. Lindeboom, A. R. Paredez, A. M. C. Emons,
D. W. Ehrhardt, Nat. Cell Biol. 11,797–806 (2009).
11. J. C. Gardiner, N. G. Taylor, S. R. Turner, Plant Cell 15,
12. R. Wightman, S. R. Turner, Plant J. 54,794–805 (2008).
13. M. Yamaguchi et al., Plant Physiol. 153,906–914 (2010).
14. A. Sampathkumar et al., Plant Physiol. 162,675–688
15. C. Ambrose, J. F. Allard, E. N. Cytrynbaum, G. O. Wasteneys,
Nat. Commun. 2,430(2011).
16. B. Schneider, W. Herth, Protoplasma 131,142–152 (1986).
17. H. E. McFarlane, A. Döring, S. Persson, Annu. Rev. Plant Biol.
18. S. DeBolt et al., Proc. Natl. Acad. Sci. U.S.A. 104,5854–5859
ACK NOW LE DGM EN TS
Arabidopsis plants containing RFP::TUB6 were the kind gift of
C. Ambrose and G. O. Wasteneys. Supported by CREATE and
Discovery grants (L.S. and S.D.M.); Canadian Natural Sciences and
Engineering Research Council doctoral postgraduate scholarships
(Y.W., M.J.M., and L.M.M.); NSF grant MCB-1158372 (D.W.E. and
H.N.C.); and Japan Society for the Promotion of Science KAKENHI
grants 24114002 and 25291062 (T.D.). We thank the technical
support staff of the UBC Bioimaging Facility and R. White for
statistical advice. The pTA7001 plasmid (VP16-GR vector) is
available from N.-H. Chua of The Rockefeller University under a
material transfer agreement; the VND7-GR line is available from
T.D.; and the GFP::CESA3 and YFP::CESA7 VND7-GR lines are
freely available from L.S. or S.D.M. for noncommercial purposes
under a materials transfer agreement on the GVG system with
N.-H. Chua. Further data are reported in the supplementary
Materials and Methods
Figs. S1 to S5
Tables S1 and S2
Movies S1 to S6
Structure-function analysis identifies
highly sensitive strigolactone
receptors in Striga
Peter J. Stogios,
Strigolactones are naturally occurring signaling molecules that affect plant development,
fungi-plant interactions, and parasitic plant infestations. We characterized the function
of 11 strigolactone receptors from the parasitic plant Striga hermonthica using
chemical and structural biology. We found a clade of polyspecific receptors, including
one that is sensitive to picomolar concentrations of strigolactone. A crystal structure
of a highly sensitive strigolactone receptor from Striga revealed a larger binding pocket
than that of the Arabidopsis receptor , which could explain the increased range of
strigolactone sensitivity. Thus, the sensitivity of Striga to strigolactones from host plants is
driven by receptor sensitivity. By expressing strigolactone receptors in Arabidopsis,we
developed a bioassay that can be used to identify chemicals and crops with altered
trigolactones are small molecules that act
as endogenous plant hormones to influence
plant growth and development (1, 2). Unlike
other plant hormones, multiple forms of
bioactive strigolactones exist and vary in
moieties attached to the A and B rings as well
as the orientation of the D ring (fig. S1) (2). Aside
from their hormonal role, strigolactones are also
exuded through roots into the soil, where they act
as exogenous signals to attract mycorrhizal fungi
for a symbiotic interaction that brings nutrients
to the plant (3). Unfortunately, parasitic plants of
the Striga, Phelipanche,andOrobanche genera also
use root-emitted strigolactones to locate a nearby
host (2). Upon perception of strigolactones pro-
duced by host plants, seeds of parasitic plants
germinate. Resulting seedlings attach themselves
to the host to deplete them of nutrients with dev-
astating consequences. In sub-Saharan Africa
alone, Striga has infested up to two thirds of the
arable land and represents the largest challenge
to food securit y on that continent, affec ting more
than 100 million people in 25 countries (4). A
mechanistic understanding of strigolactone syn-
thesis and signaling particularly in Striga could
open avenues for pest control.
Studies from model plants suggest that a fam-
ily of related a/b hydrolases is central to strigo-
lactone perception. For example, D14 (DWARF14)
is a strigolactone receptor important for shoot
branching in rice (5–7). A related a/b hydrolase
gene, HYPOSENSITIVE TO LIGHT/KARRIKIN
INSENSITIVE2 (HTL/KAI2), herein designated
germination (8). Because Striga requires strigolac-
tones to germinate, by inference HTL homologs
may have similar roles in parasitic plant species.
Arabidopsis HTL, however , binds a smoke-derived
germination stimulant, karrikin (9). In contrast,
Striga does not germinate in response to karrikin
but is sensitive to picomolar concentrations of
strigolactones (10). Moreover , Striga appears to dif-
ferentiate among hosts by sensing different com -
binations of strigolactones (10). Because Striga
hemiparasite, the functional roles of Striga HTL/KAI2
homologs (ShHTLs) in strigolactone perception
are difficult to elucidate. He r e, we de s crib e the
roles of ShHTLsinstrigolactoneperception.
Phylogenetic analysis of public databases led
us to identify a nested clade of 11 HTL homologs
in Striga (Fig. 1A). To analyze the function of
these ShHTL genes, we expressed each of the
11 ShHTLs individually in an Arabidopsis loss-
of-function HTL/KAI 2 mutant (htl-3)background
and assayed for germination phenotypes. Seeds of
the Arabidopsis ecotype Columbia germinate
poorly after exposure to high temperature, but
this thermoinhibition is alleviated by strigolac-
tone addition (11). The strigolactone-dependent
rescue of thermoinhibited seeds requires a func-
tional AtHTL gene, which provided the basis
for introducing ShHTL genes into an htl-3 ge-
netic background. We monitored the functional
activity of these genes by assaying for suppres-
sion of thermoinhibited seeds. Thermoinhibited
htl-3 seed constitutively overexpressing AtHTL
driven by the 35S promoter germinated upon
addition of a synthetic strigolactone, GR24 (fig.
S2). Similarly, thermoinhibited htl-3 seed con-
taining different ShHTL transgenes germinated
on GR24 to varying degrees, depending on the
transgene (fig. S2). To compare lines, we de-
termined the effective concentration of GR24
required for 50% germinat io n (EC
ic seed (fig. S3).TheEC
of a transgenic line
overexpressing AtHTL is lower than that of the
wild type, indicating increased strigolactone re-
sponsiveness (Fig. 1B). In contrast, ShHTL1,
ShHTL2,andShHTL3 expression lines were less
responsive to GR24 compared with either wild
type or misexpres sed AtHTL seed (Fig. 1B).
ShHTL8 and ShHTL9 transgenics were less
SCIENCE sciencemag.org 9OCTOBER2015• VOL 350 ISSUE 6257 203
Cell and Systems Biology, University of Toronto, 25
Willcocks Street, Toronto M5S 3B2, Canada.
Chemical Engineering and Applied Chemistry, Banting and
Best Department of Medical Research, University of Toronto,
200 College Street, Toronto M5S 3E5, Canada.
Structural Genomics of Infectious Diseases, contracted by
National Institute of Allergy and Infectious Diseases, National
Institutes of Health, Bethesda, MD, USA.
Transformative Bio-Molecules, Nagoya University, Japan,
Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602, Japan.
*Corresponding author. E-mail: email@example.com
RESEARCH | REPORTS
effective, whereas ShHTL4, ShHTL5,andShHTL6
were more responsive compared with AtHTL
(Fig. 1B). Complementation results from a study
in which four ShHTL genes were transformed
into an Arabidopsis htl mutant were consistent
with our fin dings (12). ShHTL7 transgenic seed
were four t o five order s o f m agnitude more re-
sponsive than other lines (Fig. 1C). This pico-
molar level of sen si ti vi ty is in the rea lm of Striga
strigolactone sensitivity, which was unexpected
because the remainder of the strigolactone sig-
naling machinery in our transgenic plants is
204 9OCTOBER2015• VOL 350 ISSUE 6257 sciencemag.org SC IENCE
Fig. 1. Striga ShHTL genes.
ShHTL genes from Striga
hermonthica and a variety of
nonparasitic plants. The
circle represents Arabidopsis
HTL/KAI2. Bootstrap values
of 100 replicates are indi-
cated by each node. (B)EC
50% of htl-3 seeds
containing an AtHTL or
ShHTL transgene. Numbers
represent biological repli-
cates from three inde-
pendent transformants. EC
was calculated from fig. S3.
thermoinhibited wild-type or
htl-3 35S::ShHTL7 seed
exposed to different concen-
trations of GR24
17.5 x 10
Fig. 2. Chemical responsiveness
of ShHTL transgenics. (A)Ger-
mination rate of htl-3 , AtHTL,and
ShHTL seed on strigolactones (1 mM)
and strigolactone agonists (10 mM).
(Inset) An expansion of strigolactone
concentrations tested on ShHTL7
seed. Genes were clustered on
the basis of their similarity of
response to all chemicals. Strigol
(STR), 5-deoxy strigol (5DS),
2′-epi-5-deoxy strigol (Epi-DS), and
Sorgolactone (SOR) experiments
were performed by using the
(KAR) and agonists 1 to 7 experiments
were performed by using the
primary dormancy assay. Details
are available in figs. S2 and S4.
0 Germination (%) 100
The far right row represents agonist efficacy to germinate Striga seed. Details are
available in fig. S5. (B) Phylogenetic tree of ShHTLs in which the color of the gene
corresponds to the heat map value of germination on 5DS. Groups are based on
clustering by using chemical responsiveness. Bootstrap values of 100 replicates are
indicated by each node.
RESEARCH | REPORTS
encoded by the Arabidopsis genome. Thus, it
appears that different ShHTL hydrolases are
sufficient to confer the extreme sensitivity of
parasitic plants to strigolactones.
Although the introduction of different ShHTLs
into htl-3 resulted in variable GR24 seed re-
sponsiveness, this strigolactone is synthetic
and exists in two isomeric forms (GR24
test the contribution of each ShHTL more rigo-
rously, we analyzed the response of our transgen-
ics using a collection of naturally occurring
strigolactones, which are all ca p a b l e of ge rm i -
to ShHTL9 were the most sensitive to natural
strigolacto nes (Fig. 2A). This group of g en e s,
group 2, defines a subclade within the ShHTL
phylogeny that arose from a common ancestor
(Fig. 2B). ShHTL7 seed was the most responsive
because it ger mi n a t ed at pi co m o la r con c e n tr a -
tions for 5DS, 2′-Epi-5DS, and sorgolactones but
responded to strigol in the nanomolar range (Fig.
2A). This suggests that the presence of a hydrox-
yl group on carbon 5 of the A-ring of strigol
decreases binding to this receptor and may ex-
plain why crop cultiva rs prod uc ing high levels
of 5DS that lack this hydroxyl group are so sus-
ceptible to infection (10). Last, two ShHTL lines
(ShHTL2 and ShHTL3)respondedtokarrikin,
although this butenolide does not germinate
Striga seed (Fig. 2A). This suggests that activation
of these recepto rs is not suffici ent for Striga ger-
mination. Overall, our functional results correlate
well with enzymatic activities of ShHTLson5DS
with the exception of ShHTL10 and ShHTL11
(13). Although ShHTL10 and ShHTL11 were bio-
chemically active, they could not suppress the
htl mutant phenotype, suggesting a different role
for these ShHTLsotherthangermination.
We then probed the promiscuity of ShHTL
receptors using a collection of synthetic strigo-
lactone agonists that show different potencies for
germinating Striga (Fig. 2A and fig. S4) (14). In
this experiment, we assayed germination of un-
derripened seed. This dormancy assay, similar
to thermoinhibition, is alleviated by strigolac-
tone and requires a functional HTL gene (11).
Within group 2, we found that ShHTL4 to
ShHTL7—but not ShHTL8 and ShHTL9—showed
the broadest range of sensitivity to various ag-
onists. Of the 11 strigolactone receptors in Striga,
it appears that ShHTL4 to ShHTL7 form a min-
imal set of receptors sufficient for Striga germi-
nation. Clustering analysis based on germination
efficiencies on natural strigolactones and syn-
thetic agonists confirmed the formation of the
ShHTL4 to ShHTL7 subclade and identified
ShHTL7 as the most functionally sensitive receptor
SCIENCE sciencemag.org 9OCTOBER2015• VOL 350 ISSUE 6257 205
Fig. 3. Structural analysis of ShHT L5. (A)StructuresofShHTL5and
AtHTL (PDB 4JYM). Spheres indicate solvent accessible volume comprising
activ e site cavities. (B)Comparisonofactivesiteresiduesofstrigolactone
receptors. Sticks are shown and labeled for active site residues showing notable
residue substitutions between ShHTL5 (yellow) and AtHTL (blue) plus the
catalytic serine and histidin e residues; equivalent ShHTL7 resid ues are labeled in
black. Karrikin1 (KAR, blue) bound to AtHTL (PDB 4JYM) and GR24 (red; A and D
rings of GR24 are labeled) in silico docked into the active sites of ShHTL5.
Single-letter abbr eviations for the amino acid residues are as follows: A, Ala; C,
Cys; D , Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Le u; M, Met; N, Asn; P,
Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V , V al; W, Trp; and Y ,Tyr . (C)Sequencealignment
of solven t-e xposed active site amino acid residues for ShHTL proteins. Amino
acids that differ fr om AtHTL are colored orange. AtHTL is blue, ShHTL5 is yellow ,
and ShHTL7 is black. A fully expanded alignment can be found in fig. S6. The far
right column is a heat map of germination of each transgenic in response to 5DS
from Fig. 2A. Color brackets represent groups based on Fig. 2B.
RESEARCH | REPORTS
(Fig. 2A). Although the phylogenetic tree based
on receptor sequence is mostly reflected by clust-
ering according to functional response, the use of
natural strigolactones and synthetic agonists
increases the functional resolution within a clade
of receptors (Fig. 2B). Thus, chemical probes
can be used to dissect functions of individual
Striga receptors not obvious through genetic
or evolutionary studies. Our result s indic a te that
this set of four receptors should be the focus of
chemical screens in the development of com-
pounds to combat Striga.
To establis h the molecularbasisforstrigolac-
tone sensitivity by ShHTL proteins, we determined
and characterized the crystal structure of ShHTL5
at a resolution of 2.48 Å because ShHTL5 be-
longs to the most sensitive group of receptors
(group 2) (table S1). Structural conservation
between ShHTL5 and AtHTL [Protein Data Bank
(PDB) 4JYM, PMID 23613584] structures (root
mean square deviation of 1.0 Å over 268 matching
Ca atoms) allowed comparative analysis of their
active site s (Fig. 3A). In both ShHTL5 and AtHTL ,
the active site is localized to a cavity formed by
respective a4, a5, and a7helices(Fig.3A),which
was more than twice as large for ShHTL5 as
compared with AtHTL (998 versus 403 Å
3A). There is also substantial variation among
residues localized to the active site cavity, with
from their counterpar ts in AtH T L (F i g. 3B ) . Pa r -
ticularly, the substitutions of Y124 and S196 in
AtHTL to V124 and Y196, respectively, in ShHTL5
deepened the active site cleft in the latter pro-
tein, making it more complementary to the D-
ring of the GR24 molecule (Fig. 3B, blue boxes).
In contrast, the Y124 in AtHTL would occlude
this region of GR24. Other differences in active
site composition included alteration at positions
(W153, F194, and A2 1 9 ) invol v e d in posi t i oning
the karrikin ligand in AtHTL but that are rep-
resented as M153, S194, and L219, respectively, in
ShHTL5 (Fig. 3B, pink boxe s) (9). These struc-
tural differences provide a s u bs t an t i al l y dif fe r e nt
chemical environment in the ShHTL5 active site,
suggesting tailoring of the active site to recognize
the strigolactone scaffold and discriminate against
Taking advantage of the structural character-
ization of ShHTL5, an analysis of the variation
among active-site residues across the ShHTL pro-
tein family rationalized the enhanced response
of ShHTL7- versus ShHTL5-harbouring transgenic
plants to strigolactones (Fig. 3B). Structural analy-
ses revealed that the active-site pocket of ShHTL1
was most similar to that of AtHTL (14 identical
and one chemically similar active-site residues);
the ShHTL7 active site was most distinct (two
identical and five chemically similar active-site
residues); and the ShHTL5 active site was inter-
mediary (four identical and five chemically sim-
ilar residues). The degree of similarity of ShHTL
active sites to those of AtHTL is consistent with
low, intermediate, and high response of ShHTL1-,
ShHTL5-, and ShHTL7-expressing transgenic plants
to GR24 (Fig. 3C). This observation prompted us to
(Fig. 4, A and B) hypothesize that ShHTLshave
evolved differential strigolactone-binding affin-
ities through modulation of their active site
Success in controlling Striga infestations will
require integrated strategies involving both chem-
ical control and plant breeding (15, 16). Striga
itself is difficult to study and, as a noxious weed,
is generally inaccessible to most researchers.
Arabidopsis seed expressing ShHTL7 transgenic
seed could function as a bioassay for strigolac-
tone. Indeed, thermoinhibited ShHTL7 seed re-
sponded to GR24 concentrations as low as 2 pM.
High and low germination rates of ShHTL7 seeds
sprinkled on rice roots reflected cultivars emitting
high (IAC165) and low levels (TN1) of strigolac-
tone, respectively (Fig. 4C) (16).
Striga receptors possess larger and modified
active site architectures with high sensitivities
to strigolactone compared with that of Arabi-
dopsis HTL. We hypothesize that these changes
have contributed to the ability of Striga to ex-
pand its host range through recognition of a
wide range of strigolactones (12). Although broad
specificity and high sensitivity are not gener-
ally associated, biochemical studies involving
soluble multidrug transporters demonstrate that
the evolution of large binding pockets to ac-
commodate multiple ligands does not necessar-
ily result in a loss of sensitivity (17). Multidrug
transporters possess broad specificity because
of large hydrophobic binding pockets but also
form high-affinity interactions through hydro-
phobic and electrostatic interactions. It will be
interesting to determine whether Striga her-
monthica strigolactone receptors have similar
REFERENCES AND NOTES
1. C. Ruyter-Spira, S. Al-Babili, S. van der Krol, H. Bouwmeester,
Trends Plant Sci. 18,72–83 (2013).
2. X. Xie, K. Yoneyama, K. Yoneyama, Annu. Rev. Phytopathol. 48,
3. K. Akiyama, K. Matsuzaki, H. Hayashi, Nature 435,824–827
206 9OCTOBER2015• VOL 350 ISSUE 6257 sciencemag .org SC IENCE
Fig. 4. Strigolactone bioassay. (A)Germinationofthermoinhibitedhtl-3 seed containing different
ShHTL transgenes (ShHTL4 to ShHTL10). The center filter disc was impregnated with GR24
labels are the calculated final concentrations of GR24
in the plate. (B)Germinationoftransgenichtl-3
seed containing ShHTL7 on varying concentrations of GR24
,andblacklabelsareasin(A).(C)Germinationofthermoinhibitedhtl-3 35S::ShHTL7 in the
presence of a high- (IAC165) or low- (TN1) strigolactone–emit ting rice cultivars. (Left) R epr esentativ e
picture of Arabidopsis htl-3 35S::ShHTL7 seeds sprinkled around rice roots. (Right) Quantification of
thre e independent experiments T SD. Scale bar , 1 mm.
RESEARCH | REPORTS
4. G. Ejeta, in Integrating New Technologies for Striga Control,
G. Ejeta, J. Gressel, Eds. (World Scientific Publishing,
Singapore, 2007), pp. 3–16.
5. T. Arite et al., Plant Cell Phys iol. 50,1416– 1424
6. L. Jiang et al., Nature 504,401–405 (2013).
7. F. Zhou et al., Nature 504,406–410 (2013).
8. M. T. Waters et al., Development 139,1285–1295
9. Y. Guo, Z. Zheng, J. J. La Clair, J. Chory, J. P. Noel, Proc. Natl.
Acad. Sci. U.S.A. 110,8284–8289 (2013).
10. A. A. Awad et al., Plant Growth Regul. 3,221(2006).
11. S. Toh et al., Plant Cell Physiol. 53,107–117
12. C. E. Conn et al., Science 349,540–543 (2015).
13. Y. Tsuchiya et al., Science 349,864–868 (2015).
14. S. Toh, D. Holbrook-Smith, M. E. Stokes, Y. Tsuchiya,
P. McCourt, Chem. Biol. 21,988–998 (2014).
15. N. H. Fischer, J. D. Weidenhamer, J. L. Riopel, L. Quijano,
M. A. Menelaou, Phytochemistry 29,2479–2483 (1990).
16. M. Jamil, J. Rodenburg, T. Charnikhova, H. J. Bouwmeester,
New Phytol. 192,964–975 (2011).
17. A. A. Neyfakh, Mol. Microbiol. 44,1123–1130 (2002).
ACK NOW LE DGM EN TS
We thank A. Babikier for providing the Striga seeds and T. Nakagawa
for the pGWB611 binary vector. Rice lines were provided by the National
Genetic Resources Program at the U.S. Department of Agriculture.
This work was supported by a grant from the National Science
and Engineering Research Council of Canada to P.M. We thank R. Di Leo
for technical assistance. Nucleotide and amino acid sequences
corresponding to ShD14 and ShHTLs have been deposited with
GenBank under accession KR013120 to KR013131. S.T., D.H.-S., P.M.,
and the University of Toronto have filed a U.S. provisional patent
application (US62/151,701) that relates to the specific topic
“Strigolactone biosensors.” We declare no financial conflicts
of interest in relation to the work. The supplementary
materials contain additional data.
Materials and Methods
Figs. S1 to S6
Genomic correlates of response to CTLA-4
blockade in metastatic melanoma
Eliezer M. Van Allen,
* Diana Miao,
* Bastian Schilling,
* Sachet A. Shukla,
Marnix H. Geukes Foppen,
Simone M. Goldinger,
Jessica C. Hassel,
Katharina C. Kaehler,
Catherine J. Wu,
† Levi A. Garraway
Monoclonal antibodies directed against cytotoxic T lymphocyte–associated antigen-4
(CTLA-4), such as ipilimumab, yield considerable clinical benefit for patients with
metastatic melanoma by inhibiting immune checkp oint activity, but clinical predic tors of
response to these therapies remain incompl etely characterized. To inves tigate the roles
of tumor-specific neoantigens and alt erations in the tumor microenvironment in the
response to ipilimumab, we analyzed whole exomes from pretreatment melanoma tumor
biopsies and matchin g germline tissue samples from 110 patients. For 40 of these
patients, we also obtained and analyzed transcri ptome dat a from the pretreatment tumor
samples. Overall mutational load, neoantigen load, and express ion of cytolytic markers in
the immune mi croenvironment were significantly associa ted with clinical benefit.
However, no recurren t n eoantigen peptide sequences predicted responder patient
populations. Thus, detailed integrated mole cular characterization of large patient cohorts
may be needed to identify robust determinants of response and resistance to immune
lockade of cytotoxic T lymphocyte anti-
gen-4 (CTLA-4), an inhibitor of T cell
activation, with the monoclonal antibody
ipilimumab yields improvements in over-
all survival in patients with metastatic
melanoma as a monotherapy (1, 2)orincombi-
nation with other T cell immune checkpoint in-
hibitors (3, 4). Although overall single-agent
response rates are low, a long-term clinical ben-
efit is consistently observed for ~20% of treated
patients (5, 6). Preclinical and clinical studies
have suggeste d that tumor -specific missense muta-
tions may generate individual neoantigens that
mediate response to ipilimumab and other im-
mune checkpoint inhibitors (7–10). Clinical stud-
ies of exceptional responders (11)andofsmall
cohorts of melanoma patients have highlighted
NRAS mutation status, total neoantigen load,
and a neoantigen-derived tetrapeptide signature
as possible correlate s of res pons e to ipilimumab in
metastatic melanoma (12, 13). RNA-based studies
have also identified gene expression signatures
linked to immune infiltration within the tumor
microenvironment that correlate with overall sur-
vival, neoantigen load (14, 15), and resistance to
immunotherapy (16). To date, however, compre-
hensive genomic studies of tumor- and immune-
related factors in larger (i.e., >100 patients) clinical
cohorts have not been reported.
We hypothesized that both tumor-specific
neoantigens and the tumor immune micro-
environment might influence clinical benefit from
ipilimumab. To test this, we performed whole-
exome sequencing (WES) on a cohort of 110 pa-
tients with metastatic melanoma from whom
pretrea tm ent tumor biopsies were availab le for
study (Fig. 1A). Tumor whole-transcriptome se-
quencing was performed in 42 of these patients,
of whom 40 had matched WES. This cohort in-
cluded 92 cutaneous, 4 mucosal, and 14 occult
melanomas. After WES of matched tumor and
germline samples (17), quality-control metrics were
applied to ensure sensitive mutation detection
(18). Average exome-wide target coverage was
183.7-fold for tumor samples and 157.2-fold for
germline samples. We performed somatic muta-
tion identification (table S1) and germline human
lymphocyte antigen (HLA) typing (table S2) using
established methods (14, 19). The median non-
synonymous mutational load was 197 per sample
(range: 7 to 5854), which is consistent with the
known high mutational loads in cutaneous mel-
anoma (13, 20).
To stra tify our cohort, “clinical benefit” was de-
fined using a composite end point of complete
response or partial response to ipilimumab by
RECIST criteria (21)orstablediseasebyRECIST
criteria with overall survival greater than 1 year
(n = 27). “No clinical benefit” was defined as
progressive disease by RECIST criteria or stable
disease with overall survival less than 1 year (n =
73). The basis for these designations stems from
clinical trials demonstrating that ipilimumab sig-
nificantly improves median overall survival, with
(~20%), but does not affect progression-free survival
SCIENCE sciencemag.org 9OCTOBER2015• VOL 350 ISSUE 6257 207
Department of Medical Oncology, Dana-Farber Cancer
Institute, Boston, MA 02215, USA.
Broad Institute of MIT
and Harvard, Cambridge, MA 02142, USA.
Center for Cancer
Precision Medicine, Dana-Farber Cancer Institute, Boston,
MA 02215, USA.
Department of Dermatology, University
Hospital, University Duisburg—Essen, 45147 Essen, Germany.
German Cancer Consortium (DKTK), 69121 Heidelberg,
Department of Medical Oncology, Netherlands
Cancer Institute, 1066 CX Amsterdam, Netherlands.
Department of Dermatology, University Hospital Zurich,
8091 Zurich, Switzerland.
Skin Cancer Unit, German Cancer
Research Center (DKFZ), 69121 Heidelberg, Germany.
Department of Dermatology, Venerology, and Allergology,
University Medical Center, Ruprecht-Karls University of
Heidelberg, 68167 Mannheim, Germany.
Dermatology, University Hospital, Ruprecht-Karls University
of Heidelberg, 69120 Heidelberg, Germany.
Dermatology, University Hospital Tübinge n, 72076 Tübingen,
Department of Dermatology, University Hospital
Kiel, 24105 Kiel, Germany.
Department of Dermatology,
University Medical Center Mainz, 55131 Mainz, Germany.
Department of Dermatology, Elbe-Kliniken, 21614
Department of Dermatology and
Allergy, Skin Cancer Center Hannover, Hannover Medical
School, 30625 Hannover, Germany.
*These authors contributed equally to this work. †Corresponding
author. E-mail: firstname.lastname@example.org (L.A.G.);
RESEARCH | REPORTS