1 D 1'-H6A
`1 8 ?.
`J-JO . ",J
`3962000
`-------- ~~0 :
`PR L·~ri~G~ n~ Tl ~
`i "~'I -'• L Ar'AnE~~ OF
`'
`
`'
`
`DECEMBER 1988
`VOLUME 85
`NUMBER 23
`
`Proceedings
`OF THE
`National Acaden1y
`of Sciences
`
`OF THE UNITED STATES C)F AMERICA
`
`Th is m at er ia l w as copi e d
`at t h.e, NLM and m ay be
`Subj e ct US Copyright Law .s
`
`TFS1073
`
`1
`
`

`

`Pro c. Na t/ . Acad. Sci. USA
`Vol. 85, pp . 8790- 8794 , D ece mber 1988
`B iochemi stry
`
`Detection of nucleic acid hybridization by nonradiative fluorescence
`resonance energy transfer
`( oligodeoxynucleotide / base pairing/helix structure/ intermolecula r distance/ intercalators)
`RI CHARD A. CARDULLO , SUDH IR AGRAWAL, CA RLOS FLORES , P AUL c. ZAMECNIK, AND D AV ID E. WOLF
`Worcester Fou ndat ion for Experimental Biology, 222 Maple A venue. Shrewsbury. MA 01545
`
`Contrilmted by Pa ul C. Za mecnik , July 28 , 1988
`
`Three approaches were used to study hybrid(cid:173)
`ABSTRACT
`ization of complementary oligodeoxynucleotides by nonradia(cid:173)
`tive lluorescence resonance energy transfer. (i) Fluorescein
`(donor) and rhodamine (acceptor) were covalently attached to
`the S' ends of complementary oligodeoxynucleotides of various
`lengths. Upon hybridization of the complementary oligodeoxy(cid:173)
`nucleotides, energy transfer was detected by both a decrease in
`lluorescein emission intensity and an enhancement in rho(cid:173)
`damine emission intensity. In all cases, lluorescein emission
`intensity was quenched by about 26% in the presence of
`unlabeled complement. Transfer efficiency at soc decreased
`from O.SO to 0.22 to 0.04 as the distance between donor and
`acceptor lluorophores in the hybrid increased from 8 to 12 to
`16 nucleotides. Modeling of these hybrids as double helices
`showed that transfer efficiency decreased as the reciprocal of
`the sixth power of the donor-acceptor separation R , as pre(cid:173)
`dicted by theory with a corresponding R0 of 49 A. (ii)
`Fluorescence resonance energy transfer was used to study
`hybridization of two fluorophore-labeled oligonucleotides to a
`longer , unlabeled oligodeoxynucleotide. Two 12-mers were
`prepared that were complementary to two adjacent sequences
`separated by four bases on a 29-mer. The adjacent S' and 3'
`ends of the two 12-mers labeled with fluorescein and rhodamine
`exhibited a transfer efficiency of = 0.60 at S°C when they both
`hybridized to the unlabeled 29-mer. (iii) An intercalating dye,
`acridine orange, was used as the donor lluorophore to a single
`rhodamine covalently attached to the 5' end of one oligodeoxy(cid:173)
`nucleotide in a 12-base-pair hybrid . Under these conditions, the
`tra nsfer efficiency was = 0.47 at 5°C. These results establish
`that lluorescence modulation a nd nonradiative fluorescence
`~esonan_ce e~~rgy transfer can detect nucleic acid hybridization
`m solution. I hese techniques, with further development, may
`also ~~ove_ us~ful_ f?r detecting and quantifying nucleic acid
`hybmllza hon m llvmg cells.
`
`In thi s pa pe r we descr ibe how fluore sce ntl y labe led o ligo(cid:173)
`deoxynucl eot td es (OD NTs) a nd th e process of no nrad ia ti ve
`flu o resce nce r~sona nce e ne rgy transfer (FRET) can be used
`to stud y nuclet c actd hybridization. When two fluorophores
`whose excJtat to n a nd em iss io n spectra ove rl ap a re in suffi (cid:173)
`c tent ly c lose proxtmtty , th e exctt ed-state e ne rgy of th e donor
`molecule ts transferred by a re so na nce dipole-induced dipo le
`mteractto n to the neighboring acceptor flu o rophore. T he
`res ult s are a decrease in donor lifetime , a quenching of donor
`fluore sce nce , a n enha nceme nt of acceptor flu o resce nce in(cid:173)
`te ~s tl y , and a depolari zation of fluorescence inte nsit y. T he
`eflt c te ncy of e ne rgy tran sfer , £ 1, fall s off ra pidly with th e
`d tstance bet wee n donor a nd acce pto r mo lec ul e , R , and is
`expressed as
`
`.
`. d f
`Th e publicat ion costs of' th ' · · ·t· 1
`. . .
`.
`·
`1s d l 1c e we1 e e rayed 1n part by page charge
`pa yment. I IllS art 1cl e must th erefore be hereby marked " adverli.l'l'll'lent' '
`1n accordan ce Wi th l~ U .S.C. *1734 solely to indicate thi s fact.
`
`[I]
`
`where R 0 is a value that depend s upon the overlap integral of
`the donor emi ssio n spectrum and the acceptor excitatio n
`spectrum , the ind ex of refraction, the qu a ntum yie ld of the
`donor, a nd the o rie ntatio n of the donor e mi ssion a nd the
`accepto r absorbance moments (1, 2). Because of it s 1/ R 6
`dependence , F R ET is extreme ly dependent on mo lecul a r
`distances a nd has been dubbed " the spectroscop ic rul e r' ' (3) .
`To follow ODN T hybridi zat io n the donor a nd accepto r
`fluo ro pho res must be suffi c ien tl y c lose to a llow resonance
`energy transfe r to occur. E nergy tran sfer ca n typically occu r
`up to di stances of about 70 A, o r abo ut twice the he lix repeat
`di sta nce in base-paired nuc le ic ac id s . We have performed
`three t y pes of experim e nt s w ith flu o rescent ly
`labe led
`O DNTs. In the first (Fig. 1a) , two s hort co mpl eme nta ry
`seque nces were labeled wit h a ppropriate donor a nd acceptor
`mol ecules at the ir 5' e nd s. In thi s case resonance e ne rgy
`tra nsfer occurs o nl y between the e nd s of hybridi zed ODNTs.
`T he second type of experime nt (Fig. 1b) in vo lved a n unl a(cid:173)
`beled stra nd of DNA , which was a ll owed to hybridi ze to two
`sho rter, labe led o ligo nu c leot id e seque nce s . Here the " a d(cid:173)
`jace nt " 3' a nd 5' e nd s of the s ho rter seq ue nces we re labeled
`so that FRET could occur over very s ho rt distances upo n
`hybridi zati o n. Fina ll y, expe rimen ts were performed with the
`flu o resce nt dye ac ridine ora nge , wh ic h intercalates into
`do uble- he li ca l nu c le ic ac id s a nd ca n act a s e ith e r a donor o r
`acceptor molecu le to a n a ppropri ate tluorophore cova le ntl y
`attached to a hybridi zed ODNT (Fig. l c).
`We report he re th a t FRET provides a useful mea ns for
`detection o f nucl e ic acid hybridi za ti o n in so luti o n. Us ing thi s
`tec hniqu e, we we re ab le to measure the reso na nce e nergy
`transfer efficie ncy of OD NT seq ue nces of variou s le ngt hs as
`a function o f conce ntrat io n a nd temperature.
`
`MATERIALS AND METHODS
`
`Che micals were purc hased from Aldrich unl ess o th e rwise
`sta ted. OD NT sy nthes is was by the phosphoramidite metho d
`using a Biosearch 8600 ON A synthes ize r a nd {3-cya noe th yl
`ph os phora midit e (G le n Resea rc h , H e rnd o n, VA).
`High-performance liquid c hro matograph y (HPLC) was
`pe rfo rme d with a Waters 600 E grad ie nt programmer a nd
`multi so lve nt delivery syste m , 481 variable wave le ngt h UV
`detector, a nd 745 data mod ul e a nd a Rheod yne 7125 inj ecto r.
`Co lumn s were of th e Rad ia l Pak ca rtridge t y pe of Nova-Pak
`C 18 (reversed-phase) or Wh at ma n 10-SAX (io n-exc ha nge) for
`use with a Z-modu le system. E luti o n buffers a nd pro(cid:173)
`grammed grad ie nts used were id e nti ca l to th ose reported
`ea rlier (4) .
`Preparation of Fluorescently Labeled ODNTs. 5' -(Amino(cid:173)
`hexyi)-ODNTs. A n a mino hexy l linke r wa s intro duced o nt o
`th e 5' e nd of the ODN T by the use of a n ex tra cyc le of
`phosphoramidite s ynthesis with (9-flu ore nyl)met hoxyca r(cid:173)
`bonylaminohexyl {3-cya noeth yl N,N-d ii sopropy lam ino phos-
`
`Abb reviat io ns: ODNT. o ligodeoxy nu cleot ide ; FR ET , nuoresce nce
`reso nance energy tran sfer.
`
`f h is. m ateria I ~7\16oJ) i ed
`atttte NLM an:O m ay be
`Su bj en U.S CoJ)y r ight Law s.
`
`2
`
`

`

`Bioc he mi stry: Ca rdu llo et a/.
`
`Proc. Na t/. A cad. Sci. USA 85 ( 1988)
`
`8791
`
`thi s cuvette , 15 11-l of = 5 11-M donor- labe led or unl abeled
`ODNT in PBS was added in S-11-l ste ps and the flu o resce nce
`intensity was determined. ODNT co ntaining acceptor flu o(cid:173)
`ro pho re was the n add ed in S-11-l steps. Energy tra nsfer was
`observed by bot h donor qu enc hing and accepto r enha nce(cid:173)
`me nt. T ra nsfe r effic ie nc ies we re de term ined fro m the
`qu enching data . T hi s in vo lved correcting th e data fo r dilutio n
`and fo r quenching by unl abe led compl e ment. Inne r filt er
`effects we re negligible. T hu s, if qu.u and qd.a are the quenc h(cid:173)
`ing observed for unlabeled and labeled compl ements, the
`tra nsfer effi cie ncy is give n by
`
`[2]
`
`Acceptor-labeled ODNT was added until £ , was constant.
`When th e inte rcalating dye acridine o range was used as a
`do nor flu oroph ore, an unl a be led 12-mer was first add ed to the
`cuvette. Acridine orange was th en add ed at a fina l conce n(cid:173)
`tration of 35 11-g/ ml. T he compl e menta ry 12- me r, e ither
`unlabeled or labeled at its 5' end , was added last. Data were
`treated by using E q. 2.
`Most ex periment s were perfo rmed at soc. Th ermogra ms to
`determine the melting te mperature, tm, were made by mi xing
`th e O DN Ts at 60°C and slow ly cooling to ooc at = 3°C/ min.
`
`F tG. l. Strategies for determining nucleic acid hybridi za tion by
`F R ET. (a ) Fluoresce nt probes are cova lentl y attached to the 5' end s
`of complement ary nucleic acids, all owing transfer to occur between
`a donor (D) and accept or (A) flu orophore over the length of the
`hyb ridi zed co mplex. (b ) Fluoresce nt molecules are covalentl y at(cid:173)
`tac hed to two nuc leic ac ids, one at the 3' end and the other at the 5'
`end. T he flu orop hore-labeled nucleic acids are complementary to
`d isti nct but close l y spaced sequences of a longer, unl abeled nucleic
`ac id. (c) A n int erca lating dye is used as a donor fo r an accept or
`llu o rophore cova lentl y att ached at the 5' end of one of the nucleic
`ac id s.
`
`RESULTS
`Effect of Acceptor Concentration on Transfer Efficiency. To
`determine th e max imum effi cie nc y of transfe r between do nor
`and acce ptor flu o roph ores attached to O DNTs, the emi ssion
`spectrum of acceptor was foll owed as a fun ctio n of increasing
`acceptor co nce ntrati on at a fi xed number of dono r mo lecul es.
`T he first ex periment s were perfo rmed using two compl emen(cid:173)
`tary O DN Ts wi th donor a nd acce pto r flu o ro ph o res attac hed
`at eithe r end of the hybrid ized compl ex: o ne O DN T had
`th e coupling reac tio n (4, 5). Aft er removal of
`phi te in
`flu o rescein attac hed to its 5' e nd (donor), whe reas the other,
`p ro tec ting gro up s with conce ntrated ammo nia solution, the
`complementary O DNT had rhoda mine attached to its 5' end
`a min o hexy l-ODN T was puri fied by reverse- ph ase HPLC.
`(acce pto r). Que nching and transfer efficie ncy we re deter(cid:173)
`3'-(Amin ohexy/am ino)-OD NTs. T he 3'-end derivati zati on
`mined for O DNTs containing 8 nucleotides, 12 nu cleotides,
`o f O DN Ts with an a mino group is based on established
`and 16 nucleotid es.
`c he mi stry fo r 3'-end la be ling of RN A (6, 7). To adapt this
`F ig. 2a shows e mi ssion spectra as a func ti on of increas ing
`c he mi stry fo r la be ling DN A, sy nth es is of a des ired O DN T
`rhodamine- linked 8- mer conce ntrati on to a fixed numbe r of
`se qu e nce was ca rri e d out o n 5'-dim eth oxy trit yl-3'(2')(cid:173)
`flu orescein-link ed 8- mer molec ules. As the amo unt of rho(cid:173)
`ace tylri bo nu c leos id e 2' (3')- linked to long-c hain alkylamino
`da mine-linked 8- mer was inc reased , th e re was a decrease in
`co ntro ll ed-po re glass suppo rt (20 11-mo l/g; G len Research,
`flu orescein emi ss io n inte nsity (5 17 nm) a nd an increase in
`He rnd o n , VA). Aft e r th e sy nth es is, protecting groups we re
`rhodamine emi ssion int ensit y (577 nm). Saturation o f both the
`re moved in co nce ntrated a mmo nia. Crud e ODN Ts were the n
`flu orescein qu enching and th e rhoda mine e nha nce me nt was
`ox id ized with pe ri odate, reacted with 1,6-di aminohexa ne,
`co mpl ete w he n the rati o o f acce ptor to donor exceeded 2: 1.
`a nd redu ced with sodium cya noborohydrid e (4 , 7). T he a mino
`T he max imum quenching of flu o resce in upon saturation was
`ODN Ts we re purifi ed by reverse-phase HPLC since th ey a re
`0.63 in the presence of bo th do no r and acce ptor. Whe n the
`reta rd ed to a sig nifi ca ntl y greate r exte nt th an und eri vati zed
`ex perime nt was repeated with flu o rescein-linked O DN T and
`ODNTs .
`it s unl abeled compl ement , flu orescein emi ssio n int ensit y was
`A twchment r4)1uorescein and rhodamin e to derivatized
`qu enched 0.26 from it s max imum va lue with no detectable
`ODN7\'. Att ac hme nt of flu oresce in , using flu oresce in iso(cid:173)
`increase in inte nsit y at 577 nm (Fig. 2h). T hu s, flu o rescence
`thi ocya na te (S ig ma), or tetra meth ylrhoda mine, using tetra(cid:173)
`was modulated in three ways upo n hybridi zatio n: a decrease
`me th ylrh odamine isothi ocya nate (S igma), to th e de ri vatized
`in flu orescein emi ss io n upo n binding to an unl abe led com(cid:173)
`O DN T a nd subsequ e nt purificati o n were carri ed out acco rd(cid:173)
`pl e mentary O DNT , a la rge r decrease in flu o resce in e mi ssio n
`in g to a pu bli she d proced ure (4, 5).
`intensit y upo n binding to a rhodamine- linked comple me ntary
`Fluorescence Measurements. Fluo resce nce measurement s
`O DN T, and th e detectio n of rhoda mine emi ssio n intensit y
`were made at th e " magic angle" in a Perkin-Elmer spectro(cid:173)
`upon binding to a rhoda mine-linked compl e ment ary O DNT.
`llu o rime te r (mo de l MP F-3) equipped with a tempe rature(cid:173)
`T he first ph eno meno n represent s a qu enc hing of th e flu o ro(cid:173)
`co ntro ll ed c ham be r a nd G la n- T ho mpson po lari ze rs . T he
`phore upon binding to its unl abeled co mpl e me nt , w hil e the
`exc it ati o n wa ve le ngth s used for flu orescein and ac ridine
`latte r two ph enomena represent modul ati o n of flu o resce nce
`o ra nge we re 472 nm a nd 503 nm , res pecti vely. T he emi ssion
`intensit y due to ene rgy tra nsfe r. T he degree of flu o rescein
`wave le ngth s used fo r flu o resce in , ac ridine o range, a nd rho(cid:173)
`qu enching due to ene rgy t ra nsfe r a lo ne was ca lcul ated fro m
`da mine we re 517 nm , 522 nm , and 577 nm , res pec ti ve ly.
`Eq . 2. In the case of th e 8-me r, the tra nsfe r effi ciency
`In a typi ca l ex pe rim e nt th e bac kground flu orescence in (cid:173)
`between flu orescein and rhoda mine was th erefore about 0. 5.
`te nsit y o f 85 11-l of ph os ph ate- bu ffered sa line (PBS: 0.138 M
`Comparabl e ex perime nt s using 12-mers and 16- mers were
`NaCI/0.01 M ph os ph ate, pH 7 .2) in a 200-11-l quartz cuvette
`also perfo rmed (Table 1). In ge ne ral, th e a mo unt of que nch(cid:173)
`(op ti ca l soluti o n pathl e ngth = 0.3 e m) was dete rmined . To
`ing in the abse nce of acce pto r was inde pe nd ent o f cha in
`This m ate ri al w as copi e{j
`at the NLM and m ay be
`Subject US Copy right Law s
`
`3
`
`

`

`8792
`
`Biochem istry: Card ull o et a/.
`
`Proc. Na t/. A cad. Sci. USA 85 ( 1988)
`
`.?;-
`
`()
`
`()
`
`0::
`
`'iii c .,
`s .,
`c .,
`(/) .,
`5
`::J
`G: .,
`. ~ 0.2
`0 a;
`
`1.0
`
`a)
`
`0.8
`
`0.6
`
`0.4
`
`?----=--=-:~
`-~-=..;:---
`
`. . ..... ~ ..
`
`.... ---
`
`0.0
`500
`
`520
`
`540
`560
`Wavelength (nm)
`
`580
`
`600
`
`o---o
`
`6-mer
`12 - mer
`16-mer
`
`1.0
`
`0.8
`
`0 .6
`
`0.4
`
`0 .2
`
`c
`'W -o
`u.,
`"'.c
`"' u
`... c a.,
`.2:::>
`-a
`0~
`c'iii oc
`u~ e·s
`c...
`
`·..;:; Q)
`
`3
`2
`[Rhoda mine labeled OO NT]
`[Fluorescein labeled OONT ]
`
`4
`
`5
`
`F IG. 3. Transfer effic iency of flu orescei n and rhodamine at(cid:173)
`tached to the 5' ends of complementary ODNTs of various lengths .
`Transfer efficiency was ca lculat ed from the difference in flu orescein
`quenching in the presence or absence of rhodamine attac hed to
`complement ary ODNT. The transfer effi ciency. £,, at saturation
`decreased progressively with chain length from the 8-mer (0.37) to
`the 12-mer (0. 16) to the 16- mer (0. 03). refl ecting a decrease Jn
`effi ciency with increas ing di stance . In each case. saturation occ urred
`at an acceptor/ donor ratio of < 2: 1. The data point s are mea ns ± SD
`for four different experime nt s .
`
`detectable rhod a mine signal (data not show n). As the tem(cid:173)
`perature was lowe red the Ouo rescein int e nsit y dec reased and
`the rhodamine intensit y inc reased in a sigmo idal ma nner (Fig.
`4) . This agreed well with the abso rba nce data , which showed
`a characteri sti c sigmoid al decrease in A2M1 with dec reasing
`temperature (Fig. 4).
`In general, there was no significa nt difference between 1m
`values obtained by nu oresce in qu enc hing a nd by dec reased
`A260 signal with dec reas ing te mpe rature. T he 1111 values
`obtained by Ouoresce in qu enching were 23 .8 ± 4.2°C, 38.3 :t
`4.SOC, and 47 .2 ± 5.2°C for th e 8- mer, 12-mer. a nd 16- mer,
`res pecti ve ly (mea n ± SO for four determinations). By co m(cid:173)
`pari son, the 1111 values obta ined by a decrease in A 2~oo we re
`24.SOC , 37.5°C, and 46 .0°C (for th e 8- mer, 12-mer. and
`16-mer, res pectively). Hence, in all cases, th e 1111 determined
`by Ou orescence was within 3% of th e 1m dete rmined by A 21o0·
`Hybridization of Two Labeled ODNTs to a Complementary
`Strand. Experim ent s we re also pe rfo rmed with two fluOI·es(cid:173)
`ce ntl y labeled O ONTs h ybridi zed to a longe r co mpl e me nt ary
`strand (Fig. lh) . Whe n these three stra nd s hybridi ze, onl y
`four bases se parate the nu o resce in dono r fro m the rhodamin e
`acce pt or. As in the prev ious expe rime nt , que nchin g of donor
`fluorescence by energy tra nsfe r increased to sa turat io n with
`acce ptor co nce ntratio n (data not show n). Table 2, line A ..
`shows the result s of th ese expe riment s. In th e presence of
`tluoresce in -labeled OONT and unl abe led OO NT hyb ridi zed
`to th e 29-mer, the qu e nching of flu o resce in e mi ss ion was
`about 0.27. In the presence of rh oda mine accepto r. the
`quenching was e nh anced to 0. 71 a nd the re was a large
`fluorescence signal at the rhodamin e peak (577 nm ). He nce .
`the transfer effi ciency, give n by Eq . 2, was about 0.6 .
`Energy Transfer with Acridine Orange. The int erca lating
`dye acridine o range was used as a donor to detec t hyb rid(cid:173)
`ization between a n unl abeled O ONT and its rh odamine(cid:173)
`labeled co mpl eme nt. Flu orescence of ac ridin e ora nge under-
`
`1.0
`
`b)
`
`.?;-
`'iii c
`~ 0.8
`c .,
`c .,
`(/) .,
`5 0.4
`::J
`G: .,
`:g
`a;
`0::
`
`()
`
`()
`
`0.6
`
`:>
`
`0.2
`
`0.0
`500
`
`520
`
`5 40
`560
`Wavelength (nm)
`
`580
`
`600
`
`F I G. 2. Modulat ion of flu orescence intensit y upon 8-mer hybrid(cid:173)
`ization at fi xed numbers of donor molecules and increasing concen(cid:173)
`trati on of the complementary ODNT. (a) Complementary ODNT
`was labeled with rhodamine at the 5' end. Energy transfer occurred
`bel wee n the flu oresce1n donor and the rhodamine acceptor as shown
`by a decrease in flu orescein emission intensit y at 517 nm and an
`increase in rhodam1ne emi ssion intensit y at 577 nm with increasing
`acceptor concentntt1on. (b) Complementary ODNT was unlabeled.
`Curves represent different ratios of rhodamine-linked ODNT to
`co mplement ary flu orescein-linked ODNT: -
`, O: 1; · · · ·, 0.40:1; -··-,
`1.9:1: - --, 3.8: 1. Spectra are not corrected for dilution .
`
`le ngth and had a value of 0.26 ± 0.02 for all OONTs (mean
`:!: S O for four determin ation s of eac h n-mer, where n = 8, 12,
`o r 16 nu cleotides). In th e prese nce of rhodamin e- linked
`co mpl ement ary OONT, the degree of Ouoresce in quenching
`due to e nergy tran sfer alone decreased with increasing chain
`le ngth . As show n in Fig. 3, hybridi za tion was complete for all
`three c ha111 length s at an acceptor/ donor rati o of 2: 1. At
`highe r acce ptor/ donor ratios , no modulation in the corrected
`nu o resce in or rhodamine signal was observed. Subsequent
`expe riment s using these OONTs were done at an acceptor/
`donor ratio of 4:1 to ensure th at hybridi zation was complete.
`Ell'ect of Temperature on Transfer Efficiency. The effect of
`te mperature on hyb ridi zati on was also foll owed for different
`c ha in l ~ngth s (8- , 12- , and 16- mers) at saturating co ncentra(cid:173)
`tion s of acceptor-linked OONT . The resulting melting tem(cid:173)
`peratures (1m) , defin ed as th e midpoint va lues of fluore scein
`qu e nc htng or rhodam111e enhance ment over a temperature
`ra nge of0- 60oC. were co mpared with absorbance data at 260
`nm . Above sooc, there was no nu oresce in qu enching or
`
`Table 1. Modulation of flu orescein intensity at saturating levels of ODNT with and without
`rhodamine attached for ODNTs of chain length 11
`II
`
`R/ R0
`L·,
`C/ r.r
`(/ f.u
`0.99 ± 0.()2
`0.632 ± 0.046
`0.501 ± 0.035
`0.265 ± tl.021
`1.24 - ().05
`0.423 ± 0.030
`0.215 - 0.052
`0.265 ± 0.01 3
`0.295 ± 0.017
`1.66 ± 0.10
`0.045 ± 0.018
`0.262 ± 0.013
`Data represent mean ± SO for four different ex periment s. See Eqs. I and 2. Subscript s f, r, and u
`indi cate flu oresce in-labeled, rhodamine-labeled . and unlabeled ODNT.
`
`!l
`12
`16
`
`This. m aterial w as.copoie.d
`atth e NLM and m ay be
`Subj e-ct USCopoy r ight Laws
`
`4
`
`

`

`Bioc he mi stry: Ca rdull o et a /.
`
`Proc. Nat/. A cad. Sci. USA 85 ( 1988)
`
`8793
`
`t .O -·""'
`.
`.
`"'
`"'
`.
`~'.
`\
`
`~·-·
`/J •
`o--o A260
`~·
`Fl s 17
`Flsn
`
`j
`
`""' o.
`
`._
`·~.
`
`40
`30
`20
`Temperature (°C)
`
`5 0
`
`60
`
`0
`c:
`a>
`Ui
`"0
`"'
`N
`0
`E
`0
`z
`
`0 .8
`
`0.6
`
`0.4
`
`0.2
`
`0.0
`0
`
`I
`
`•/o
`6~/
`t O
`
`1.0
`
`:;:;. 0.8
`'iii
`c: .,
`:E 0 .6
`.,
`u
`c .,
`Ill .,
`0
`:;J 0 .2
`G:
`
`0 .4
`
`u
`
`0.0
`0
`
`3
`2
`[Rhodamine labeled OD NT)
`(Unlabeled OD NT)
`
`4
`
`5
`
`F IG. 5. C ha nges in flu o rescence in te ns it y a t 517 nm (o ) a nd at 577
`nm (e ) as a fun ct io n of rhodamine-lin ked ODNT concentrat io n at
`fi xed concentrati o ns o f unl a beled compl ementary ODNT a nd ac ri(cid:173)
`dine o ra nge.
`
`hybridi zatio ns (9, 10) . In thi s re port we have show n th at
`nonradiativ e F RET provid es a se nsitive method for detecti on
`of bi nding between compl e me ntary ODNTs .
`FRET has been used to qu antify the distance between
`donor and acceptor fluoropho res and accurately pred icts the
`distance between donor and acceptor in pol yme rs of vari ous
`lengt hs (3) . Our studies using complement ary ODNTs cova(cid:173)
`lentl y tagged wi th flu orescent probes indicate th at thi s
`technique can be used to in vest igate ce rt ain physica l param(cid:173)
`eters of nucleic acid hybridi zati on . In parti c ula r , the di sta nce
`between th e donor and the acceptor, R. can be calcul ated if
`it is assumed that the hybridi zed ODNTs fo rm a he li x and if
`th e o ri entation of th e donor and acceptor tluo rop hores
`relati ve to the heli cal ax is is know n . Fro m the transfer
`effici encies in Table 1, we have ca lc ul ated the dista nce
`between the flu o resce in donor and the rhodamine acce ptor
`mol ecules as a fu nctio n of base position by using Eq. 1.
`As regard s the geometry of the fluoropho res re lati ve to the
`helica l ax is, we have considered three simpl e mode ls (Fig. 6).
`In ge ne ral, the di stance be twee n donor and accepto r ca n be
`ex pressed as
`
`13]
`
`R = {2(r + d cos cp)"{ 1 + cos [ (~N + 1)7T/ 5J}
`112.
`+ (3 .4~N + 2d sin cp) 2
`}
`wh ere r is the rad iu s of the double heli x (10 A), ~N is th e base
`separati on (0 , 1, 2, ... ), d is the le ngth of th e probe
`re presenting the six-ca rbon spacer and a phosphate group
`(= 12.8 A) , and cp is the angle of the spacer from th e helical
`axis (- 7T/ 2 ~ cp ~ 7T/ 2). Model 1 (Fig. 6) assumes that th e
`probes are perpendicular to the heli cal ax is (cp = 0). In model
`2 the probes are parallel to th e heli cal ax is and pointing away
`from the bases at the 5' end (cp = 7T/2). and in model 3 the
`probes are parallel to the he li cal ax is bu t lie alo ng the heli x
`so that the probes are at closest a pproac h (cp = - 7T/ 2). A plot
`of these ca lc ul ated R values (from Eq. 3) as a fu nct ion of
`1]116 should give a straight line wi th the slope
`[(1 / £ 1)
`-
`corresponding to R 0 for th e donor a nd acceptor pa ir as
`predict ed by Eq . 1.
`Lin ear regress ion of each of the three ca lc ulated R va lues
`gave a good fit (coeffi cient of dete rmination r 2 > 0 .9) o nl y fo r
`
`1-· 1G. 4. C ha nges in flu o rescence int e ns it y of do no r- a nd acce pto r(cid:173)
`li nk e d X- m e rs a s a fun c ti o n o f te mpe ra ture . Whe n th e te mpe rature
`w~" d ec reased fro m 60°C to 0°C, the lluo rescein e mi ssio n intensit y
`t e ) m o n o to ni call y dec reased, indicat ing a n inc rease in transfer
`c:tfi c ie ncy w ith th e rh odamine accep to r o n th e comple me nt ary 8-me r .
`In <tdditi o n to th e dec rease in flu orescein e missio n inte nsit y with
`in c reas ing te mpe ra ture . th e re was a conc urre nt inc rease in rho(cid:173)
`d ~1min c em issio n int e ns it y (6 ). Am1 (o ) showed a cha racte risti c
`d ec rease w ith d ec reas ing te mpe ra ture . indicatin g hyb ridi zati o n o f
`co mrl e m e nt a ry O DNTs.
`
`goes a s pec tra l s hift upon binding of the dye to double(cid:173)
`~ tra nd e d DNA. a nd thi s modul ati on can be used to determine
`th e ra ti o of single-stranded to double-stranded species in
`so luti o n (8). In o ur hybridi zatio n studi es, when acridine
`o ra nge int e rca lated int o compl ement ary 12-mers it ex hibited
`a n exc ita ti o n max 1mum at 503 ± 3 nm and an em1 ss1o n
`m ax imum a t 522 ± 2 nm.
`In th ese studie s, two ODNTs were used : one was unla(cid:173)
`b e le d a nd it s co mpl e me nt was labeled with rhoda mine at the
`') ' e nd . Ac ridine o ra nge was add ed to the c uvett e along with
`~he unl abe led strand at a fi xed co ncentration , and the
`rh o d a m in e- la be led ODNT was added in various amou nt s .
`Th e d eg ree of flu o rescence quenching alo ng with the en(cid:173)
`ha nce m e nt o f rhoda min e flu o rescence at 577 nm increased
`a nd th e n leveled off with inc reasing acce ptor concentratio n
`(F ig. 5) . As ex pected , unlike th e case "':' here single donor a nd
`acce pt o r mo lecules we re covalentl y ltnked to ODNTs, the
`deg ree o f e ne rgy transfe r was much highe r when the int er(cid:173)
`ca la tin g d ye was used. Under saturating conditi ons of ac(cid:173)
`ce pto r co nce ntrati on , the qu enc hing of acridine ora nge ap(cid:173)
`proac hed 0 .57 (Tab le 2, line 8) . T he a mo unt of quenching
`ob~e r ve d w ith two unl a beled 12- mers was only about 0.11 , so
`that th e resulting transfer e fficiency wa s 0 .52. T hi s can be
`co mpa red to th e case of th e 12-mers with flu orescein and
`rhoda min e at th e ir respective 5' end s, which gave a transfer
`e ffi c ie ncy of o nl y 0 .22 (T a bl e 1). Hence, the presence of an
`int e rca la tin g d ye alo ng the entire le ngth of the ODNT
`e nh a nced th e tra nsfe r e ffi c ie ncy by a factor of = 2 in th e
`12-mc r .
`
`DISCUSSION
`
`Fo r de tec ting nu c leic ac id hybridi zati o n , an idea l flu oresce nt
`O DNT pro be would (i) be easil y att ac hed to an ODNT , (ii)
`he de tec tabl e a t low conce ntrati ons , (iii) produce a mod u(cid:173)
`la ted signa l w he n the la be led ODNT hybridi zed to a com(cid:173)
`pl e me nt ary OD NT , and (iv) be sta bl e at te mperat ures used in
`
`Table 2. Q ue nc hin g a nd tra nsfe r e ffi c ie ncy o f two 12- me rs att ac hed to a 29- mer (li ne A) a nd o f
`two h yb ridi zed 12-me rs in the presence of ac ridin e o range (line B)
`
`A
`13
`
`(/d .r
`0 .712 :±: 0 .024
`0 .573 :±: 0 .027
`
`0 .276 :±: 0 .0 12
`0.109 :±: 0 .009
`
`0.602 - 0.017
`0 .520 :±: 0.016
`
`R / R u
`
`0.933 :±: O.Oll
`
`Data rep rese nt th e mea n :±: S D for four diffe re nt ex perime nt s . Subscript d represents lluo rescein in
`line I\ a nd ac ridin e o ra nge in line B.
`
`This m at eri al w asco~ i e·d
`at the NLM and m ay b.e
`Subject USCo~yr ight Laws
`
`5
`
`

`

`8794
`
`Biochem istry: Cardull o et a/.
`
`Proc. N at/. A cad. Sc i. USA 85 ( 1988)
`
`...
`
`:''
`
`--
`
`--
`
`c ~
`
`0
`
`100
`· ~ E
`..------ - - - .1'
`"' o 80 ~
`Q) L
`•
`L ~
`Q(ll
`::>0>
`:;::::~
`c ~ 60
`.,.,
`Q) c
`~ · -Q)E 40
`.DO
`.,u
`00 cE 20
`2-o
`"'c
`oO
`
`0
`
`0
`
`5
`
`10
`Base Separation, LIN
`
`15
`
`20
`
`F tG. 6. Dista nce between dono r a nd acce ptor mo lec ul es as a
`function of base sepa ration for different -le ngth ODNTs . Three
`mode ls of differe nt s pace r o ri ent ation were tes ted (lnsl!l ). The three
`cu rves represe nt th e ca lculated di sta nce between th e nu orescein (F)
`d o no r and th e rhodam ine (R) acce ptor as a fun cti o n o f base positi on
`, model I (cp = 0): ---, model 2 (cp
`fo r eac h o f th e three modeb: -
`= 7T/ 2); .. .. mode l 3 (cp = - 7T/ 2) . Linea r regress io n of th e ca lcul a ted
`R va lues fo r the 8- mer. 12- mer. and 16-mer as a fun cti on of 10/ £1)
`6 gave a best fit for modcl 2 (coefficient o f de te rminati o n r 2 =
`- Jj 1
`/
`0. 94) with a corresponding R0 of 49.1 A. By co mpariso n . co rrec ti o n
`of the published Ro va lue for nuo rescein a nd rh odam ine att ached to
`long-chain hydrocarbo ns of 54.2 A (II) by our observed qu e nc hing
`factor of 0 .26 res ult ed in a va lue of 52.2 A. 'l'he data point s (0 )
`represent th e dista nce between th e don or a nd acce pt o r molec ules
`obta ined by using th e co rrec ted R0 va lue o f 52.2 A. Agai n. th ese dat a
`gave a reaso nab le fit
`to mode l 2, sugges ting th at upon ODNT
`h yb rid iza ti on. the probes are pointing away from the 5' e nds of th e
`he li x.
`model 2 (Fig. 6) with a corresponding R0 of 49.1 A. By
`co mparison, th e publi shed va lue for flu orescein donors and
`rhod amine acceptors att ached to long-chain hydrocarbons in
`solution is 54.2 A (11). However , when one account s for the
`quenching that we observed in the absence of acceptor-linked
`ODNT, a co rrec ted R0 va lue of 52.2 A 154.2/ (1.26) 1101 is
`obtained . Use of thi s correc ted va lue for 1?0 along with our
`energy- transfer data for th e 8-mer . 12-mer , and 16-mer gave
`a reaso nable fit to th e va lues ca lculated for model 2 (open
`c ircles in Fig. 6). Hence, the data support a model in which
`th e probes are pointing away from the 5' end s of the heli x.
`In co nstructing fluorophore end-labeled nucleic acid s, it is
`poss ible to choose different space rs that will give a range of
`transfer efficiency va lues upon hybridi za ti on. Multiple de(cid:173)
`terminal ions of distances with different- length spacers might
`allow th e secondary stru cture of nucleic acids to be ascer(cid:173)
`ta ined in solution.
`We have also shown th at th e technique of FRET allows
`hyb ridi za tion to be monitored by cova lently att aching flu o(cid:173)
`rophores to th e 3' and 5' end s of ODNTs that ca n be
`hybridi zed to a longer ODNT. In thi s case , th e donor and
`acceptor mol ecules ca n theoreti ca ll y be separated by onl y a
`few base pairs so th at energy transfer occurs with ve ry high
`effi ciency upon hybridi zati on. Interca lating dyes prov ide
`a not her method for detecting nucleic acid hybridi zation by
`FRET. Th e usc of an interca lat ing dye allows a short
`sepa rati on di stance be twee n donor and acceptor molecules
`and . in th e case of a helica l structure. prese nt s many dye
`molecules at different di stances and ori entations . T he result
`
`is an "ante nna effect .. th at gives a high degree of transfer
`effi ciency. Some interca lating dyes , such as acridine orange
`and ethidium brom ide , have th e added benefit of changing
`th eir spectral properties upon bi nding to double-s tra nded
`DNA . T he widespread use of ot her intercalating dyes in cell
`and molec ular biology allows for a variety of potent ial donor/
`acceptor pairs to be chosen for a particu lar bi ologica l appli(cid:173)
`cation.
`On th e basis of these studies and our ex perience so far. it
`is not apparent th at FRET offers increased sensiti vit y over
`existing met hods for detec ting nucleic acid hybridi zati on (9) .
`However, FRET has a feature not shared by any oth er
`method. In all existing meth ods fo r de tecting nucleic ac id
`hybridi zation , th e hybrid s ca n onl y be detected after removal
`of nonh ybridi zed nucleic acid. FRET is different. and pow(cid:173)
`erfull y so, in thi s respec t. Resonance energy transfer creates
`a distinctiv e signal in the presence of nonh yb ri dized nucleic
`acid strands. T he implicati ons of thi s for analyti cal and
`diagnosti c methods and for detec ting hyb ridi zati on eve nt s in
`l'i vo are obv ious.
`Finall y, the approach we have described here in solution
`studi es shou ld be applicab le to li vi ng ce ll s by microinjec ting
`tluorophore- tagged pairs of nucleic acid s for an endogenous.
`com pl ementary, nucleic acid sequence of