A magnetic switch for the control of cell death signalling in in vitro and in vivo systems

A magnetic switch for the control of cell death signalling in in vitro and in vivo systems

Mi Hyeon Cho 1?,Eun Jung Lee 1,2?,Mina Son 2?,Jae-Hyun Lee 1,Dongwon Yoo 1,Ji-wook Kim 1,Seung Woo Park 3,Jeon-Soo Shin 2,4*and Jinwoo Cheon 1,2*

The regulation of cellular activities in a controlled manner is one of the most challenging issues in ?elds ranging from cell biology to biomedicine 1,2.Nanoparticles have the potential of becoming useful tools for controlling cell signalling pathways in a space and time selective fashion 3,4.Here,we have developed magnetic nanoparticles that turn on apoptosis cell signalling by using a magnetic ?eld in a remote and non-invasive manner.The magnetic switch consists of zinc-doped iron oxide magnetic nanoparticles 5(Zn 0.4Fe 2.6O 4),conjugated with a targeting antibody for death receptor 4(DR4)of DLD-1colon cancer cells.The magnetic switch,in its On mode when a magnetic ?eld is applied to aggregate magnetic nanoparticle-bound DR4s,promotes apoptosis signalling pathways.We have also demonstrated that the magnetic switch is operable at the micrometre scale and that it can be applied in an in vivo system where apoptotic morphological changes of zebra?sh are successfully induced.

Cell signalling is an important process in biological systems for exchanging information through networks of various signal molecules to control cellular activities,such as differentiation,growth,metabolism and death 6.Owing to their newly developed high precision and accuracy,physical stimuli using optical,electrical and magnetic methods have been devised to regulate cell signalling 1–4.Among these,magnetic techniques are uniquely advantageous because magnetic fields can penetrate deeply with negligible attenuation into biological tissues 7,8.Consequently,it has distinctive benefits for in vivo applications.Moreover,when coupled with magnetic nanoparticles,magnetic fields can be transformed into other forms of energy,such as heat and mechanical force 9–16.The magnetic heat induction has been used for gating of the thermosensitive ion channel 9as well as for hyperthermia therapy 10.Although relatively large mechanical force (in the piconewton range)has been used in in vitro and in vivo systems for direct stretching of ion channels and cytoskeletal stimulation 11–13,two recent in vitro studies have revealed that the induction of calcium influx 15and tubulogenesis 16using receptor clustering is also possible by using nanoparticles.Magnetic nanoparticles can exert a gentle force (in the femtonewton range)on membrane receptors to induce their clustering without disturbing the rheological and cytoskeletal properties.Furthermore,the nanoscale dimensions of nanoparticles conjugated with targeting molecules make them beneficial for probing cellular sensory structures and functions at the molecular level and for inducing specific cellular activation processes 7,8.Nonetheless,the nanoscale magnetic switching technique for receptor clustering is still at too

1Department

of Chemistry,Yonsei University,Seoul 120-749,Korea,2Graduate Program for Nanomedical Science,Yonsei University,Seoul 120-749,Korea,

3Department of Internal Medicine,Institute of Gastroenterology,College of Medicine,Yonsei University,Seoul 120-752,Korea,4Department of

Microbiology,Severance Biomedical Science Institute,Institute for Immunology and Immunological Diseases,College of Medicine,Yonsei University,Seoul 120-752,Korea.?These authors contributed equally to this work.*e-mail:jsshin6203@yuhs.ac;jcheon@yonsei.ac.kr.

early a stage of development to guarantee that it will be generally applicable to the control of cell signalling in other biologically important systems.In addition,it is not known whether it will be effective in in vivo systems.

Apoptosis,programmed cell death,is known to be a major factor in maintaining homeostasis and removing undesired cells 17–19.Recently,an extrinsic apoptosis signalling pathway that is initiated by death receptors has emerged as one of the main targets for cancer therapy 20–22.Extrinsic apoptosis signalling is usually activated by clustering of death receptors through docking of biochemical ligands,such as the TNF-related apoptosis inducing ligand (TRAIL)that is a potent inducer of apoptosis 23,24.However,direct use of such ligands in clinical applications is limited by the short plasma half-life and the ease of degradation 25,26.

In the study described below,we have developed a magnetic switch for apoptosis signalling,and demonstrate its in vivo feasi-bility through the receptor clustering process (Fig.1).The mag-netic switch for cell signalling consists of DR4monoclonal anti-body conjugated to magnetic nanoparticles (Ab–MNPs).DR4s are highly expressed on tumour cells 20–27and magnetic nanoparticles are designed to bind DR4s through a specific monoclonal anti-body interaction.Zinc-doped iron oxide magnetic nanoparticles (Zn 0.4Fe 2.6O 4MNPs,15nm)are chosen (Fig.2a)owing to their high saturation magnetization value (161e .m .u .g ?1),which is essential for effective utilization of magnetic force 5(Supplementary Section S1).For preparation of the Ab–MNPs,protein A is conjugated to thiolated MNPs through a sulpho-SMCC (sulphosuccinimidyl-4-[N -maleimidomethyl]cyclohexane-1-carboxylate)crosslinker.The DR4antibody is then conjugated to protein A on MNPs in a DR4antibody/MNP stoichiometric ratio of 1:1(Fig.2b and Supple-mentary Section S1).

For in vitro magnetic switching On of apoptosis,Ab–MNPs (1pM)are applied to DLD-1colon cancer cells (1.5×104cells per well),expressing DR4s (Supplementary Section S2),and the cells are placed in between two NdFeB mag-nets (Fig.2c).In the absence of magnetic field,Ab–MNPs are observed as an evenly dispersed weak green fluorescence signal (Fig.2d(i))and remain dispersed over time (Supplementary Section S3).However,large aggregated spots exhibiting a strong green fluorescence signal are observed after the application of a magnetic field for 2h (Fig.2d(ii)red circles and Supplementary Section S4).Scanning electron microscope (SEM)images consistently show that the initial evenly distributed Ab–MNPs change to densely populated aggregates (Fig.2e and Supplementary Section S4).Here,the simulated magnetic field is 0.20T at the centre 28and the average

1038

NATURE MATERIALS |VOL 11|DECEMBER 2012|/naturematerials

© 2012 Macmillan Publishers Limited. All rights reserved

MNPs

S

S

N

N

Magnet Magnet

0.0 1 .0

¬1.0¬0.50.5(cm)

0.01 T

t

i ii

5 µm 5 µm

100 nm100 nm Magnet

20 nm Thiolated

MNP

–SH +

O

N

O

O

HN

S

O

O

O

N HN

S

O

O

O

N HN

S

S

N N

Ab¬MNP

Protein A¬MNP

signalling.a,Transmission electron micrograph of15nm zinc-doped

antibody(DR4Ab).A thiolated MNP is linked to protein

The magnetic switch set-up for apoptosis signalling.

magnetic?eld are indicated as a solid line and colour map,

a

M a g n e t i c a l l y i

ii i

ii i ii i

ii

b

body

g

6080

100

a t h (%)

***

Figure 3|In vitro apoptosis induction in the DLD-1colon cancer cell line.a ,Cascade of extrinsic apoptosis signalling pathways.Assembly of DR4s leads to recruitment of DISC,composed of FADD and procaspase-8.Procaspase-8is cleaved to form active caspase-8and leads to subsequent caspase-3

activation.b ,c ,Confocal microscope images of ?uorescently stained active caspase-8(b )and active caspase-3(c )in DLD-1cells for non-activated (i)and magnetically activated (ii)groups.Active caspase-8and caspase-3are immunostained with Alexa-594-labelled secondary antibodies (red)and nuclei are stained with DAPI (blue).d ,e ,Confocal micrographs of membrane inversion (d )and blebbing at the late apoptosis stage (e ),stained with FITC–annexin V (green)and propidium iodide (red),respectively.f ,Differential interference contrast micrographs of DLD-1cell morphology before (i),12h (ii)and 24h (iii)after magnetic ?eld application.Scale bars,10µm.g ,Cell death measured by CCK-8assay.Cells treated with Ab–MNPs (1pM)and magnetic ?eld are compared with the cells alone,Ab–MNPs without magnetic ?eld and TRAIL treatment.Error bars represent standard deviation.(??P <0.01,???P <0.001.)

orders of magnitude weaker and negligible in this set-up where the mid-point of the sample area is 1cm away from the magnet 15(Supplementary Section S5).The Ab–MNPs induce clustering of the DR4s in a similar manner to the biochemical ligand,TRAIL,and they have the unique advantage of being magnetically switched On to activate cell signalling remotely and non-invasively in a spatially and temporally controlled fashion (Fig.2f).

To examine the extrinsic apoptosis signalling process with concurrent assembly of DR4s promoted by using the magnetic switch,biologically important intermediate species of the signalling

cascades are monitored.It is known that clustering of DR4s forms the death-inducing signalling complex (DISC)containing the Fas-associated death domain (FADD)and procaspase-8(refs 20–22).In the DISC,procaspase-8is cleaved to active caspase-8(initiator caspase),which leads to further activation of caspase-3(Fig.3a).Treatment of Ab–MNPs (1pM)on DLD-1cells for 2h with a 0.20T magnetic field results in a strong red fluorescence signal arising from the active caspase-8in the cytoplasm,but no fluorescence is observed in the control group not exposed to a magnetic field (Fig.3b).Subsequent generation of active caspase-3,

1040

NATURE MATERIALS |VOL 11|DECEMBER 2012|/naturematerials

© 2012 Macmillan Publishers Limited. All rights reserved

Magnet

Magnet Magnet Magnet Magnet

Magnet Magnet Magnet

100 µm

100 µm

200 µm

200 µm

0.01 T

control of apoptosis signalling.Ab–MNPs (1pM)applied to a cell culture slide containing DLD-1cells with DR4s.a ,c ,(c )of magnetic ?eld focused on the cell culture Fluorescence micrographs of the active caspase-3(red)signal nuclei (blue).Each image is created by stitching 50–120?uorescence images pictured by ×200magni?cation confocal microscope.

is strongly amplified in cooperation with other members of the caspase family 31.Entry of the cell into the demolition phase is associated with membrane inversion and blebbing at the late apoptosis stage as major phenomena 17–19.The occurrence of these changes is confirmed by using FITC (fluorescein isothiocyanate)–annexin V and propidium iodide staining,respectively.The observation of a bright fluorescence signal,which is not observed in magnetically non-activated cells,is a clear indicator of this apoptotic outcome (Fig.3d,e).Moreover,temporal morphological changes of the cells are also monitored during the magnetically activated apoptotic process.In comparison with that of normal cells,1.5×104cells treated with Ab–MNPs (1pM;Fig.3f(i))gradually shrink and fragment after 12and 24h application of the magnetic field (Fig.3f(ii,iii)).Quantitative analysis of apoptosis,using a cell counting kit-8(CCK-8)assay,shows that about 52%of the cells treated with Ab–MNPs are dead after magnetic activation,whereas the cells in the non-activated group remain viable (Fig.3g).This level of efficacy is slightly higher than that of TRAIL,the biochemical ligand previously employed as an extrinsic apoptosis signalling agent and potential cancer drug 20–26,which induces cell death in about 45%of the cells treated under the same conditions.We also observe that the cell death rate is dependent on the concentration of Ab–MNPs (Supplementary Section S9).Apoptosis signalling,induced in this manner by the magnetic switch,requires

culture slide containing DLD-1cells by using the magnetic set-up shown in Fig.4a.On application of the magnetic field for 4h,activation of caspase-3,seen in the form of an immunostained red fluorescence signal,is observed to occur exclusively in the area of about 200µm ×100µm between the magnets (Fig.4b).Similarly,active caspase-3is consistently observed in an array comprising magnetic focusing on three local spots (Fig.4c,d),which testifies to the spatial controllability of the magnetic switching technique.

Although this magnetic switching strategy for apoptosis signalling is effective at the in vitro level,its further extension to in vivo activation is both significant and challenging because of the complexity of live biological systems.We choose zebrafish as an in vivo model not only because of the benefits associated with optical imaging and convenient screening 32–34,but also because of the genetic closeness of the zebrafish ovarian TNF receptor (OTR)to the human DR4for apoptosis signalling 35,36.Zebrafish OTRs are targeted and magnetically manipulated by using zebrafish OTR antibody conjugated to MNPs (zAb–MNPs).First,fluorescein-labelled zAb–MNPs (2.5ng per embryo)are micro-injected into the yolk of the zebrafish embryo at the one-cell stage of growth.After 24h,pronase is used to hatch the zebrafish embryos,which are then divided into two groups including those that will and will not be magnetically activated.For the magnetically activated group,a magnetic field with a strength of

NATURE MATERIALS |VOL 11|DECEMBER 2012|/naturematerials

1041

© 2012 Macmillan Publishers Limited. All rights reserved

Fluorescein- conjugated B r i g h t -f i e l d i m a g e

M N P

b

a

d

e

f

ii

i

iii

ii

i

iii

ii i iii

ii i iii

1 mm 1 mm

1 mm

100 μm 100 μm 100 μm

C a s p a s e -3

Zebrafish OTR

Magnetic field

Figure 5|In vivo magnetic apoptosis signalling for zebra?sh.a ,The apoptosis experimental scheme of zebra?sh.Fluorescein-labelled zAb–MNPs

(2.5ng embryo ?1)are micro-injected into yolk of embryo at one-cell stage to label extrinsic apoptosis receptor (OTR).At 24h post-fertilization (h.p.f.),the magnetic ?eld is applied to zebra?sh.b ,Bright-?eld microscope images of three different groups of zebra?sh,control (i),non-activated (ii)and

magnetically activated (iii).The magnetically activated group of zebra?sh shows morphological alterations in the tail region compared with other groups.c ,Quantitative analysis on tail bending by measuring the angle between the line on the pronephros (PR)and the line of tail tip (TT)for each group.

d –f ,Fluorescenc

e images o

f zebra?sh in which zAb–MNPs are green and active caspase-3is immunostained as red.Green and red ?uorescence are only observed in the tail region of the magnetically activated group (d (iii),e (iii)).Insets in d and e are magni?ed images of tail region.f ,Magni?ed ?uorescence micrographs of active caspase-3in zebra?sh tail region.

g ,A grap

h of caspase-3activity measured by using ?uorescence intensity of zebra?sh.Error bars are standard deviation.(???P <0.001).

0.50T is applied to a zebrafish chamber for 24h (Fig.5a).After application of the magnetic field,characteristic features associated with apoptosis,such as morphological deformation of embryos and caspase-3activation 35,36,are examined at the 48h post-fertilization (h.p.f.)stage.Inspection of bright-field microscope images shows that both the control and non-activated groups exhibit normal ontogenic zebrafish embryo development (Fig.5b(i,ii)).In contrast,morphological alterations in the tail region are clearly observed in images of the magnetically activated group (Fig.5b(iii)).This morphological change is visualized by determining the angle of tail tip bending.For this purpose,a straight line is drawn along the pronephros and the angles between this line and tail tip line are measured in three different groups,as shown in Fig.5c.The magnetically activated group shows an approximately 3.5-fold larger angle (about 22?)than that of the control and non-activated groups (about 6?).This morphological change is proposed to be a consequence of apoptosis signalling 35.Besides,zebrafish embryos

without zAb–MNPs injection are not affected by magnetic field application alone (Supplementary Section S7).

To gain further evidence for the occurrence of apoptosis,the location of zAb–MNPs and the apoptosis signalling products are examined using an optical method.zAb–MNPs are fluorescently labelled and the strong green fluorescence signal is observed in the yolk for both the magnetically activated and non-activated groups owing to residual zAb–MNPs after the injection.Interestingly,clumps of strong green fluorescence signal are seen in zebrafish tail regions that are magnetically activated,whereas this region of non-activated zebrafish shows only a dispersed and weak green fluorescence signal (Fig.5d(ii,iii)).These observations are caused by the combination of higher OTR expression in the tail region of zebrafish in a manner that is consistent with previous observations 35,36,and subsequent magnetic clustering of the OTRs.The activation of caspase-3at the tail region is confirmed by using immunostaining.Both the control and non-activated

1042

NATURE MATERIALS |VOL 11|DECEMBER 2012|/naturematerials

© 2012 Macmillan Publishers Limited. All rights reserved

groups of zebrafish embryos show no apparent signals associated with active caspase-3(Fig.5e(i,ii)).In contrast,strong red spots associated with active caspase-3are clearly observed in the tail region for the magnetically activated group(Fig.5e(iii),f(iii)), which match closely with the region of green fluorescence signal of clustered zAb–MNPs.Note,the faint red signals seen for the control and non-activated groups in this immunostaining process are due to the naturally occurring mild apoptosis process of zebrafish(Fig.5f(i,ii)).The results of quantitative measurements of active caspase-3show that the magnetically activated group has an approximately6-fold higher fluorescence intensity level than those of the control and non-activated groups(Fig.5g).It has been documented that a bent tail is regarded as one of the most representative traits of apoptosis35,36.Consistent with this,the highly localized presence of OTR receptors,clustered MNPs and activated caspase-3in the lower part of the tail stem is likely to cause the tail to bend up through localized apoptotic cell death in the tail. The combined results demonstrate that the magnetic switch can be effectively applied to in vivo live vertebrates.

In the study described above,we have demonstrated that apoptosis signalling can be turned On in vitro and in a zebrafish in vivo model by using a magnetic switch.Our magnetic switch may be broadly applicable to any type of surface membrane receptors that exhibit cellular functions on clustering.With apoptosis being one of the main cancer research targets,the development of an extrinsic apoptosis agonist that can avoid p53mutation-induced drug resistance is important and here our magnetic switch can serve as a selective inducer19,27.Likewise,the applications can be extended to other clinically useful membrane receptors,such as the vascular endothelial growth factor receptor for regenerative medicine37and the Toll-like receptor for immune potentiation38.Although at an early stage of development,the spatially and temporally controlled magnetic switch system has the potential to be a useful tool for the activation of various cell signals at the target region.

Received20March2012;accepted23August2012;

published online7October2012

References

1.Hoffman,B.D.,Grashoff,C.&Schwartz,M.A.Dynamic molecular processes

mediate cellular mechanotransduction.Nature475,316–323(2011).

2.Gorostiza,P.&Isacoff,E.Y.Optical switches for remote and noninvasive

control of cell signaling.Science322,395–399(2008).

3.Lee,S.E.,Liu,G.L.,Kim,F.&Lee,L.P.Remote optical switch for localized

and selective control of gene interference.Nano Lett.9,562–570(2009).

4.Chung,I.et al.Spatial control of EGF receptor activation by reversible

dimerization on living cells.Nature464,783–787(2010).

5.Jang,J-t.et al.Critical enhancements of MRI contrast and hyperthermic

effects by dopant-controlled magnetic nanoparticles.Angew.Chem.Int.Ed.48, 1234–1238(2009).

6.Jordan,J.D.,Landau,E.M.&Iyengar,R.Signaling networks:The origins of

cellular multitasking.Cell103,193–200(2000).

7.Pankhurst,Q.A.,Connolly,J.,Jones,S.K.&Dobson,J.Application of

magnetic nanoparticles in biomedicine.J.Phys.D36,R167–R181(2003).

8.Gupta,A.K.&Gupta,M.Synthesis and surface engineering of iron oxide

nanoparticles for biomedical applications.Biomaterials26,3995–4021(2005).

9.Huang,H.,Delikanli,S.,Zeng,H.,Ferkey,D.M.&Pralle,A.Remote control

of ion channels and neurons through magnetic-field heating of nanoparticles.

Nature Nanotech.5,602–606(2010).

10.Creixell,M.et al.EGFR-targeted magnetic nanoparticle heaters kill cancer cells

without a perceptible temperature rise.ACS Nano5,7124–7129(2011).

11.Wang,N.,Butler,J.P.&Ingber,D.E.Mechanotransduction across the cell

surface and through the cytoskeleton.Science260,1124–1127(1993).

12.Kanczler,J.M.et al.Controlled differentiation of human bone marrow

stromal cells using magnetic nanoparticle technology.Tissue Eng.16,

3241–3250(2010).

13.Hughes,S.,Haj,A.J.E.&Dobson,J.Magnetic micro-and nanoparticle

mediated activation of mechanosensitive ion channels.Med.Eng.Phys.27, 754–762(2005).

14.Dobson,J.Remote control of cellular behavior with magnetic nanoparticles.

Nature Nanotech.3,139–143(2008).15.Mannix,R.J.et al.Nanomagnetic actuation of receptor-mediated signal

transduction.Nature Nanotech.3,36–40(2008).

16.Lee,J-H.et al.Artificial control of cell signaling and growth by magnetic

nanoparticles.Angew.Chem.Int.Ed.49,5698–5702(2010).

17.Danial,N.N.&Korsmeyer,S.J.Cell death:Critical control points.Cell116,

205–219(2004).

18.Williams,G.T.Programmed cell death:Apoptosis and oncogenesis.Cell65,

1097–1098(1991).

19.Storey,S.Targeting apoptosis:Selected anticancer strategies.Nature Rev.7,

971–972(2008).

20.Fulda,S.&Debatin,K-M.Extrinsic versus intrinsic apoptosis pathways in

anticancer chemotherapy.Oncogene25,4798–4811(2006).

21.Ashkenazi,A.Targeting death and decoy receptors of the tumour-necrosis

factor superfamily.Nature Rev.Cancer2,420–430(2002).

22.Sayers,T.J.Targeting the extrinsic apoptosis signaling pathway for cancer

therapy.Cancer Immunol.Immunother.60,1173–1180(2011).

23.Mahalingam,D.,Szegezdi,E.,Keane,M.,Jong,S.de&Samali,A.TRAIL

receptor signaling and modulation:Are we on the right TRAIL?Cancer Treat.

Rev.35,280–288(2009).

24.VondálováBlanárová,O.et al.Cisplatin and a potent platinum(IV)

complex-mediated enhancement of TRAIL-induced cancer cells killing is

associated with modulation of upstream events in the extrinsic apoptotic

pathway.Carcinogenesis32,42–51(2011).

25.Schneider-Brachert,W.et partmentalization of TNF receptor1

signaling:Internalized TNF receptosomes as death signaling vesicles.Immunity 21,415–428(2004).

26.Kelley,S.K.&Ashkenazi,A.Targeting death receptors in cancer with

Apo2L/TRAIL.Curr.Opin.Pharmacol.4,333–339(2004).

27.Ashkenazi,A.&Herbst,R.S.To kill a tumor cell:The potential of proapoptotic

receptor agonists.J.Clin.Invest.118,1979–1990(2008).

28.Meeker,D.C.Finite element method magnetics.

(accessed Feb9,2011).

29.Türkcan,S.et al.Observing the confinement potential of bacterial pore-forming

toxin receptors inside rafts with nonblinking Eu3+-doped oxide nanoparticles.

Biophys.J.102,2299–2308(2012).

30.Pralle,A.,Keller,P.,Florin,E-L.,Simons,K.&Hörber,J.K.H.

Sphingolipid-cholesterol rafts diffuse as small entities in the plasma

membrane of mammalian cells.J.Cell Biol.148,997–1007(2000).

31.Porter,A.G.&Jänicke,R.U.Emerging roles of caspase-3in apoptosis.

Cell Death Differ.6,99–104(1999).

32.Pyati,U.J.,Look,A.T.&Hammershmidt,M.Zebrafish as a powerful

vertebrate model system for in vivo studies of cell death.Semin.Cancer Biol.17, 154–165(2007).

33.Eimon,P.M.&Ashkenazi,A.The zebrafish as a model organism for the study

of apoptosis.Apoptosis15,331–349(2010).

34.Yamashita,M.Apoptosis in zebrafish p.Biochem.Phys.B

136,731–742(2003).

35.Eimon,P.M.et al.Delineation of the cell-extrinsic apoptosis pathway in the

zebrafish.Cell Death Differ.13,1619–1630(2006).

36.Bobe,J.&Goetz,F.W.Molecular cloning and expression of a TNF receptor

and two TNF ligands in the fish p.Biochem.Phys.B129,

475–481(2001).

37.Matsuda,H.The current trends and future prospects of regenerative medicine

in cardiovascular n Cardiovasc.Thorac.Ann.13,101–102(2005).

38.Oblak,A.&Jerala,R.Toll-like receptor4activation in cancer progression and

therapy.Clin.Dev.Immunol.475,1–12(2011). Acknowledgements

This work was financially supported by grants from the Creative Research Initiative (2010-0018286),WCU Program(R32-10217),National Research Foundation of Korea (2011-0017611)and the second stage BK21for Chemistry and Medical Sciences of Yonsei University.S.W.P.was supported by the National Research Foundation,

Mid-career Researcher Program(72011-0043).M.H.C.was supported by a Hi Seoul Science/Humanities Fellowship from the Seoul Scholarship Foundation.

Author contributions

J.C.and J-S.S.conceived and designed the experiments.M.H.C.,E.J.L.,M.S.and J-w.K. performed the experiments.S.W.P.provided advice on the in vivo zebrafish experiments. M.H.C.,E.J.L.,J-H.L.,D.Y.,J-S.S.and J.C.wrote the manuscript.

Additional information

Supplementary information is available in the online version of the paper.Reprints and permissions information is available online at /reprints. Correspondence and requests for materials should be addressed to J-S.S.or J.C. Competing?nancial interests

The authors declare no competing financial interests.

NATURE MATERIALS|VOL11|DECEMBER2012|/naturematerials1043

© 2012 Macmillan Publishers Limited. All rights reserved

本文来源：https://www.bwwdw.com/article/wuyi.html