Nucleic Acids Research 99NcecAisRsac

Volume 7 Number 7 1979

Voue7Nme

A spectroscopic probe of stacking interactions between nucleic acid bases and tryptophan residues of proteins

Claude Helene, Jean-Jacques Toulne and Trung Le Doan Centre de Biophysique Moleculaire, CNRS, 45045 Orleans Cedex and Laboratoire de Biophysique, Muse'um National d'Histoire Naturelle, 61, rue Buffon, 75005 Paris, France Received 25 September 1979 ABSTRACT The external heavy atom effect of mercury on the spectroscopic properties of the indole ring has been used to investigate stacking interactions of tryptophan with mercurinucleotides in mixed aggregates formed in frozen aqueous solutions as well as in oligopeptide-polynucleotide complexes. This effect is characterized at 77 K by a quenching of the tryptophan fluorescence, an enhancement of the phosphorescence emission and a drastic shortening of the phosphorescence lifetime. These phenomena result from an enhanced spin-orbit coupling due to a close contact between the mercury atom and the indole ring. Dissociation of the complexes leads to a recovery of the spectroscopic properties of the free tryptophan ring. The possible use of this spin-orbit probe to provide evidence for stacking interactions in protein-nucleic acid complexes is discussed.

INTRODUCTION Stacking interactions involving aromatic amino acids and nucleic acid bases might play an important role in the formation of proteinnucleic acid complexes (1, 2). It has been suggested that such interactions might be involved in the selective recognition of single stranded structures of nucleic acids by proteins (3, 4). However it is always difficult to provide direct evidence for the involvement of stacking interactions in a biologically important complex. Nuclear magnetic resonance is probably the most appropriate method since stacking induces upfield shifts of the proton resonances of the stacked aromatic rings (5). However this method can be easily used only with simple systems because usually the presence of several aromatic amino acids in a protein leads to a superimposition of the aromatic proton resonances which preventsany unambiguous analysis of changes in chemical shifts. The replacement of tyrosines by fluorotyrosines or by seC Information Retrieval Limited 1 Falconberg Court London Wl V 5FG England

1945

Nucleic Acids Research lectively deuterated tyrosines has made it possible to study stacking interactions by

9F and

H magnetic resonance in the complexes formed by oligo-

nucleotides with gene 5 protein from phage fd (6). However these methods

require the preparation of labeled proteins and the number of cases where they can be applied is therefore limited. Fluorescence spectroscopy has been widely used to study

protein-nucleic acid complexes (7). Stacking interactions are characterized by a quenching of tryptophyl, tyrosyl or phenylalanyl fluorescence (4, 8). However there might be other mechanisms by which the fluorescence of aro-

matic residues could be quenched. A change of the protein conformation might bring the aromatic residue in the vicinity of a quenching group inside

the protein structure. Also energy transfer from tyrosine and phenylalanine to nucleic acid bases might contribute to fluorescence quenching in the absence of stacking interactions (4).

Recently a photochemical method has been developed which makes use of the photosensitizing properties of tryptophyl residues. It has been shown that tryptophan is able to photosensitize the splitting of thymine dimers in oligopeptide-DNA or protein-DNA complexes (9). This reaction requires a close proximity of tryptophan and dimers. However the prerequisite is that thymine dimers are first made by UV irradiation of DNA. This

might alter the recognition of DNA by the investigated protein.

Photochemical cross-linking of tryptophan to nucleic acid bases in a stacked complex could possibly help identify a stacking interaction.

No such reaction has been described yet with the possible exception of the

photochemical reaction of tryptophan with 5-bromouracil (10). This possibility is presently under investigation. We have tried to use a direct spectroscopic method to provide evidence for stacking interactions. This method is based upon the external heavy atom effect on spin-orbit coupling in aromatic molecules (11). It is known that spin-orbit coupling can be strongly enhanced by Van der Waals contact of the aromatic molecule with a perturbing heavy atom. Such an effect has already been used to investigate intercalation of acridine dyes in polynucleotides containing 5-bromouracil (12) and more recently to probe the presence of a tryptophan residue in the active site of a lectin by using a 1946

Nucleic Acids Research ligand containing a mercury atom (13). The heavy atom effect is characterized by a quenching of the fluorescence of the aromatic molecule, an enhancement of the phosphorescence emission and a shortening of the phosphorescence lifetime. It has been shown that mercurinucleotides can be incorporated in polynucleotide chains (14). We have therefore investigated the

possibility of using the heavy atom effect of mercury to provide evidence for stacking interactions of mercurinucleotides with tryptophan. The results presented below show that such a heavy atom effect can be observed in two systems: i) in mixed aggregates formed when aqueous mixtures of tryptophan (or other indole derivatives) and 5-mercuri dUMP are frozen down to

77 K ; ii) in the complexes formed by the oligopeptide Lys-Trp-Lys with mercurated poly(U).

MATERIALS AND METHODS Tryptophan, N-acetyltryptophan amide, indole, were obtained from Sigma. N-methyl indole distilled under vacuum was a gift from Dr. J. P. Privat. N(l)-methyltryptophan was synthesized by Dr. M. Bazin.

5-mercuri deoxyuridine, 5'-monophosphate was obtained from P. L. Biochemicals as the triethyl ammonium salt of the mercuri carbonate derivative of the nucleotide. Two samples of mercurated poly(U) were used: one was a gift from Drs. M. N. Thang and J. L. Drocourt, the

other was obtained from P. L. Biochemicals. Both gave identical results.

The percentage of uracil bases substituted by mercury at carbon 5 was hi-

gher than 70 %. Luminescence measurements at 77 K were carried out with a Jobin-Yvon Bearn spectrofluorimeter modified to correct for lamp fluc-

tuations (8). The solutions were contained in a quartz tube of 2 mm i. d. immersed in a quartz dewar containing liquid nitrogen. Phosphorescence was separated from fluorescence by a rotating can phosphoroscope which was used as a shutter for fast decay time measurements. RESULTS

1. Frozen aqueous solutions of 5-HgdUMP and indole derivatives Freezing an aqueous solution induces the formation of aggregates of the solute molecules (15). When an aqueous mixture of two aroma1947

Nucleic Acids Research tic molecules is frozen down to 77 K mixed aggregates

usually formed.

are

This property has been previously used to study stacking interactions between

aromatic amino acids and nucleic acid bases (16). When

an aqueous

equimolecular mixture of N-Acetyl

phan amide (NATA) and 5 HgdUMP

quenching of indole fluorescence of

a

was

was

frozen down to 77 K,

a

trypto-

complete

observed together with the

appearance

strong phosphorescence emission of short lifetime (figure 1, table 1).

The phosphorescence spectrum had that of NATA dispersed in

(lv/lv) (WPG) although it

a

a

vibronic structure which resembled

frozen glass of water and propyleneglycol

was

red-shifted by about 16

nm.

It should be

ted that NATA aggregates (in the absence of 5 HgdUMP) emit rescence

but only

a very

a

no-

strong fluo-

weak phosphorescence.

In order to determine the origin of these spectroscopic effects

several derivatives of indole and uridine

were

investigated. Uridine itself

quenched the fluorescence of NATA and tryptophan but

no

phosphorescence

IL INATA. SHgdJP 20

-

NATA(x)

5~~~ ~~~HgdUMP

300

500

400

A (nm)

Figure 1 Total luminescence spectra in frozen aqueous solutions at 77 K ), 5 HgdUMP (---) and their equiof N-acetyltryptophanamide (NATA) ( molar mixture (_ ). The concentrations were 3. lxl0-4 M for each compound. Note that the spectra of the individual molecules are multiplied by 3. The spectrum -o-o- represents the phosphorescence emission of NATA in a ...

water-propylene glycol (lv/lv) glass. 1948

Nucleic Acids Research Table 1: Phosphorescence lifetimes at 77 K in frozen aqueous solutions (W) and in frozen equivolume mixture of water and propylene glycol (WPG). The phosphorescence decays were usually not strictly exponential except for 5HgdUMP in water and NATA in WPG. The lifetimes given here correspond to the initial parts of the decay curves. 5 HgdUMP

Tryptophan + 5 HgdUMP

(W)

(W)

2.7 ms

2.9 ms

Indole + 5 HgdUMP

NATA + 5 HgdUMP

N-methyl indole + 5 HgdUMP

(W)

(W)

2.4 ms

2.7 ms

NATA

NATA + 5 HgdUMP

(W)

(WPG)

(WPG)

3.7 ms

6.s2

6.s1

N(1) methyl tryptophan + 5 HgdUMP

(W) 2.9 ms

enhancement was observed as already reported (16). A similar result was obtained with 5-bromodeoxyuridine. Fluorescence quenching was previously ascribed to the formation of electron donor-acceptor complexes in aggregates (16). The following indole derivatives were also investigated: indole,

N-methylindole, tryptophan,N(l)-methyltryptophan. Equimolar mixtures of these compounds with 5-HgdUMP all gave rise to the spectroscopic effects described above for NATA. They were characterized by a quenching of the

indole fluorescence, the appearance of an intense phosphorescence with a vibronic structure characteristic of indole derivatives and a very short phosphorescence lifetime of the order of 3 ms (table 1). In the absence of 5-HgdUMP the phosphorescence emitted by indole derivatives in frozen aqueous solutions was too weak to permit a precise measurement of its de-

cay. However this phosphorescence was easily measured in a frozen glass, e. g., a mixture of water and propylene glycol (lv/lv). Its decay was expo-

nential with a lifetime of 6-7 s for all indole derivatives (17). The interaction with 5 HgdUMP in aggregates therefore reduced the lifetime by a factor of about 2, 000. It is known that addition of organic solvents to aqueous solutions before freezing prevents the formation of aggregates (15). Addition of an equivolume of propylene glycol to an equimolar mixture of NATA and 5 HgdUMP suppressed the

spectroscopic effects described above. The fluo-

1949

Nucleic Acids Research rescence

of NATA

and lifetime

were

was

not quenched, the normal phosphorescence spectrum

restored. This experiment demonstrates that aggregate

formation is required to observe the heavy

atom effect of 5 HgdUMP

on

NATA. 2. Complexes of Lys-Trp-Lys with mercurated poly(U) We have previously shown that the tryptophyl ring of the tri-

peptide Lys-Trp-Lys forms stacked complexes with nucleic acid bases when it binds to polynucleotides (5, 7). We have therefore investigated the complexes

formed by Lys-Trp-Lys with mercurated poly(U). As shown in figure 2,

the formation of this complex is characterized by rescence, an

with

a

a

quenching of Trp fluo-

enhancement of the phosphorescence quantum yield together

drastic shortening of the phosphorescence lifetime (table 2). It should

be noted that the phosphorescence spectrum is red-shifted by 8

nm.

This

red-shift is smaller than that observed in aggregates (see above). Complex formation between Lys-Trp-Lys and polynucleotides involves electrostatic interactions between amino

20

groups

(N-terminal and

ys-Trp.Lys + PolyHgU

L

Lys-Trp-Lys

10

Poly Hg

300

400

U

500

X

(nm)

Figure 2 Total luminescence spectra at 77 K in a pH 6 buffer (B) containing 1 mM sodium cacodylate and 1 mM sodium chloride of 8x10-5 M Lys) 3. 3x10-4 M poly HgU ( ---) and the mixture of 8x10-5 M Trp-Lys ( Lys-Trp- Lys with 3 .3x10 4 M poly HgU ( 1950

Nucleic Acids Research Table 2: Phosphorescence lifetimes of poly(5 HgU) (1. 7x1O M), LysTrp-Lys (3.4x10-5 M) and their mixture in a buffer (B) containing 1 mM sodium cacodylate, 1 mM sodium chloride, pH 6 or in an equivolume mixture of buffer and propylene glycol (WPG)

poly(5 HgU)

Lys-Trp-Lys Lys-Trp-Lys Lys-Trp-Lys Lys-Trp-Lys + poly(5 NaCl M HgU) g ) +0.4 P Y((5 HgU) + poly(5 HgU)~+ poly +poly(5HgU)

(B)

(B)

5.9 ms (82%)

1. 8 ms

+ 5.1 s

(WIPG)

(WlPG)

6. 2 s

5. 6 ms (85%) +6.Os (15%)

6.s0

(18%)

lysyl side chains) and phosphate bases (5, 8). These complexes creases.

(WPG)

groups

are

a

a

stacking of Trp with

dissociated when the ionic strength in-

Figure 3 shows the results of

vent WPG which forms

together with

an

experiment conducted in the

sol-

transparent glass at 77 K and does not prevent

complex formation between Lys -Trp-Lys and polynucleotides. Increasing the ionic strength from 1 mM to 400 mM leads to described above

a

reversal of the effects

the original fluorescence of the tryptophyl residue is

re-

IL 15

A

Poly HgU + Lys-TrpLyvs Lys-Trp.Lys

Poly HgU+ Lys Trp Lys +O.4M NoCI 10

350

450

A(nm)

550

Figure 3 Total luminescence spectra in an equivolume mixture of propylene glycol and buffer B (see legend of figure 2) at 77 K of 4x10-5 M LysTrp-Lys (...), and the mixture of 4x10-5 M Lys-Trp-Lys with 1. 6xl0-4M poly HgU in the absence (-) and in the presence (--- ) of 0. 4 M NaCl. It should be noted that the phosphorescence of HgU is superimposed on that of Lys-Trp-Lys when the complex with poly(HgU) is dissociated in the presence of 0. 4 M NaCl (-- -). 1951

Nucleic Acids Research covered while the phosphorescence quantum yield is decreased and the phosphorescence lifetime returns to that of the free peptide (table 2). It should be noted that the phosphorescence decay of

Trp-Lys complexes with poly 5 HgU shows

superimposed

on

or

residues

component

(Nv 5-6 s)

to be explained only by the

presence

are

are

not fully mercurated it is likely that

some

tryptophyl

stacked with non-mercurated uracils and that the triplet state

of these residues is efficiently populated via triplet

neighboring bases. However due to the absence of

energy

transfer from

a mercury

atom in their

immediate vicinity the phosphorescence of these tryptophyl residues is

pected to have the an

of

externally bound peptides (8). Since the poly U samples used in

experiments

our

long-lived

the main short-lived decay. The contribution of this long-

lived component (Au 15 %) is too high

unbound

a

Lys-

efficient triplet

same

lifetime

energy

as

ex-

unperturbed tryptophan (table 2). Such

transfer from adenine to tryptophan

was

already

observed in the complexes formed by Lys-Trp-Lys with poly(A) (20).

DISCUSSION AND CONCLUSION The results obtained both in mixed aggregates and in Lys-Trp-

Lys complexes with mercurated poly(U) clearly demonstrate the influence of the

mercury

atom

on

the spectroscopic properties of the indole ring. The

enhancement of spin orbit coupling due to the heavy atom effect is characterized by

a

complete quenching of the indole fluorescence,

of the phosphorescence emission and rescence

a

a

strong increase

drastic shortening of the phospho-

lifetime. The fact that the

same

phenomenon is observed in mixed

gregates when the N(1) atom of the indole ring

is

ag-

methylated or when the

amino and carboxyl groups of the amino acid function of tryptophan

are sub-

stituted, indicates that the spectroscopic changes are due to a direct effect of the mercuri nucleotide on the indole ring, and do not involve binding of mercury

to the amino group

as

reported for complexes of tryptophan

tryptamine with CH3HgOH (18).

These

spectroscopic changes are

or

charac-

teristic of the heavy atom effect of mercury on spin-orbit coupling in the

excited indole derivative. They require the

1952

mercury atom and the indole

a Van der Waals contact between

ring (11). Since it is known that tryptophan

Nucleic Acids Research forms stacked complexes in mixed aqueous

solutions it is

very

aggregates with nucleosides in frozen

likely that the

and the indole ring results from such

a

contact between the

mercury

atom

stacking interaction.

In the complexes formed by Lys-Trp-Lys with mercurated

poly(U) at pH 6 the amino

groups

of lysyl residues

are

positively charged

and interact strongly with the negatively charged phosphates (rather

form complexes with the xes

is

very

mercury

atom) (8). The behavior of these comple-

similar to that observed with other polynucleotides, including

their dissociation at high ionic strength (7). Since ce

proton magnetic

resonan-

experiments have clearly shown that the tryptophyl ring of the tripeptide

is stacked with nucleic acid bases (5), it is tions on

than

are

very

likely that the

same

interac-

involved in the complexes described here. The heavy atom effect

the spectroscopic properties of the tryptophyl residue is thus due

close contact between the

mercury atom

to the

and the indole ring brought about by

the stacking of tryptophan and uracil. The red shift of the phosphorescence spectrum

as

interaction as

well

as

compared to free tryptophan is very likely due to this stacking

as

already observed in Lys-Trp-Lys complexes with poly(A) (20)

to an effect of mercury itself (18).

This heavy atom effect should be easily detected in proteinnucleic acid complexes if stacking interactions presence

are

involved. Of

course the

of mercuri nucleotides in the nucleic acid is a prerequesite. It

will be necessary to prevent reaction of sulfhydryl groups of the protein (if

present) with the mercurinucleotide. This

can

be achieved either by reaction

of mercaptans with the mercuri-substituted nucleic acid before complex formation

or

by protection of the protein SH groups if this does not prevent bin-

ding to the nucleic acid.

Acknowledgements We widh to thank Drs. M. N. Thang and J. L. Drocourt for a gift of mercurated polyU and Dr. M. Bazin for a gift of N(l)-methyltrypto-

phan. This work

was

supported in part by a grant (MRM 77-7-1741) of the

Delegation Gen6rale a la Recherche Scientifique et Technique (DGRST).

1953

Nucleic Acids Research REFERENCES 1 H,61ene, C. (1976) Studia Biophysica 57, 211-222 2 Gabbay, E. J., Adawadkar, P. D. and Wilson, W. D. (1976) Biochemistry 15, 146-151 3 Toulme, J. J. and Helene, C. (1977) J. Biol. Chem. 252, 244-249 4 Mayer, R., Toulme, F., Montenay-Garestier, T. and Helene, C. (1979) J. Biol. Chem. 254, 75-82 5 Dimicoli, J. L. and He"lne, C (1974) Biochemistry 13, 714-723 6 Coleman, J.E. and Armitage, I. M. (1978) Biochemistry 17, 5038-5045 7.Hel'ne, C. (1977) in Excited States in Organic Chemistry and Biochemistry, pp. 65-78, B. Pullman and N. Goldblum Eds, Reidel 8 Brun, F., Toulmg, J. J. and Hel'ene, C. (1975) Biochemistry 14, 558-563 9 He'lbne, C. and Charlier, M. (1977) Photochem. Photobiol. 25, 429-434 10 Saito, L, Ito, S. and Matsuura, T. (1978) J. Am. Chem. Soc. 100, 2901 -2902 11 McGlynn, S. P., Azumi, T. and Kinoshita, M. (1969) Molecular Spectroscopy of the triplet state, Prentice Hall, N. J. 12 Galley, W. C. and Purkey, R. M. (1972) Proc. Nat. Acad. Sci. USA 69, 2198-2202 13 Monsigny, M., Delmotte, F. and Helene, C (1978) Proc. Nat. Acad. Sci. USA 75, 1324-1328 14 Dale, R. M. K., Livingston, D.C. and Ward, D.C. (1973) Proc. Nat. Acad. Sci. USA 70, 2238-2242 15 Montenay-Garestier, T. (1973) J. Chim. Phys. 70, 1379-1384 16 Montenay-Garestier, T. and Hel'ne, C. (1971) Biochemistry 10, 300306 17 Longworth, J. (1971) in Excited states of protein and nucleic acids, pp. 319-484, R.F. Steiner and I. Weinryb Eds., Plenum Press 18 Svejda, P., Maki, A.H. and Anderson, R.R. (1978) J. Am. Chem. Soc. 100, 7138-7145 19 Rabenstein, D. L., Ozubko, R., Libich, S., Evans, C.A., Fairhurst, M. T. and Suvanprakorn, C. (1974) J. Coord. Chem. 3, 263-271 20 H61e'ne, C. (1973) Photochem. Photobiol. 18, 255-262.

1954