Int. J. Peptidehotein Res. 15, 1980,464-410

N E W R E ACT1 0 N 0 F 4 -CH LO R 0-7-N IT R O B E NZ O F U R A Z A N WITH A M I N E S A T H I G H LOCAL C O N C E N T R A T I O N S

Primitive Enzyme Models

JAKE BELL0 and HELEN PATRZYC

Department of Biophysics, Roswell Park Memorial Institute, Buffalo, New York, U.S.A.

Received 20 August, accepted for publication 10 December 1979

The fluorogenic amine reagent 4-chloro-7-nitrobenzofurazan (CN3Fz) reacts with amines to give one of two products, or a mixture, depending on the con- centration of amine. A t low concentration (e.g. 5 X M) o f unprotonated amine the normal replacement of the chlorine occurs, giving A,, - 470 nm. A t high concentration of unprotonated amine (e.g. 0.5 M), a new product, Amu 380 nm, is produced. At intermediate concentrations, both products are seen. These results were obtained with aqueous ammonia, methylamine and butyla- mine, and with butylomine in dry hernne. The 380nm product also is obtained with high local concentrations o f amines, such as with [Lys], at pH 8.5-10 and micellar dodecylamine a t pH 9.5. These results are suggested to be analogous to those o f very primitive proto-enzyme systems, by which clusters of small mole- cules can produce qualitative changes in “metabolism”. Reaction o f CNBFz with gelatin gave quditative evidence for cross-linking.

Key words: amines; 4-cNoro-7-nitrofurazan; cooperative; evolution; poly(1ysme); proto- enzymes.

4-Chloro-7-nitrobenzofurazan, CNBFz (7- chlor04-nitro-2-oxa-l,3diazole), is a widely used reagent for introducing a fluorescent group into proteins. The standard reaction of CNBFz with an amine is replacement of the chlorine to give the 4-aminO derivative (Ghosh & Whitehouse, 1968; Kenner & Aboderin, 1971 ; Fager et al., 1973). Compounds of this class have Amax near 470nm, and are fluor- escent. Reversible formation of Meisenheimer complexes at C-6 have been reported (Baines etal., 1977; Shipton etal., 1976; D~NUMO et d., 1975). Baines et d. (1977) suggested that when C-4 is covalently bonded to protein, Meisenheimer-adduct formation might occur at C-6 or N-4 with another nucleophilic center in the protein. Allen & Lowe (1973) observed

464

that addition of excess mercaptan to 4-(2‘- hydroxyethylthio)7-nitrobenzofurazan resulted in SWO decrease in hz. Allen & Lowe did not explicitly indicate whether or not a second stable covalent reaction or a reversible Meisenheimer reaction occurred. Nitta et al. (1979) reported that CNBFz reacts with a moderate excess of thiol to give (in addition to the usual 4-alkylthio derivative) a product with A,, at 307nm, which was formed rapidly, and they suggested that this might be a Meisenheimer complex. With a 220-fold excess of thiol a third product was formed, with absorption out to 700nm, in addition to the 307nm product. They suggested that in proteins with more than one thiol group suit- ably juxtaposed sequential intramolecular re-

0367-8377/80/050464-07 %02.00/0 @ 1980 Munksgaard, Copenhagen

4-CHLORO-7-NITROBENZOFlJRAZAN

actions may occur. Here we report that high concentrations of unprotonated amines give a new product with CNBFz.

Two characteristics of enzymically catalyzed reactions are enhanced rate and selectivity of reaction, when more than one reaction is possible with a given substrate. There is much interest in the synthesis of small molecules which can mimic the actions of enzymes. A recent sample of this is the report by Breslow et ul. (1 978) of regioselectivity in the hydroly- sis of the cyclic phosphate of 4-tert.-butyl- catechol by a bis-imidazole derivative of flcyclodextrin, arising from the spatial relation between the imidazole groups as fixed by the cyclodextrin.

We report a type of cooperative action of even simpler “protoenzymes”, in which the functional groups are not held in a fixed spatial arrangement, but cooperate by virtue of their high local concentrat‘ion.

MATERIALS AND METHODS

CNBFz was obtained from Pierce Chemical Co., [Lys - HBr],, 30000 daltons, from Sigma Chemical Co., and dodecylamine - HCl, butyla- mine and methylamine - HCl from Distillation Products Industries. The butylamine was re- distilled before use, the others were used as received. n-Hexane was dried over Type 4A molecule sieve (Fisher Scientific Co.).

Reactions were carried out at room tempera- ture, usually for 24 h, in solvents and at con- centrations and pH values as described in Results. Reactions with [Lys], were done with solutions maintained at constant pH (pH-stat). Aqueous solutions of butylamine, methylamine and ammonia were adjusted to the desired pH with acetic acid, and were self-buffered. Spectra were measured with a Cary 15 spectrophotometer. The ratio of amine to CNBFz was at least 1O:l in the experiments at low amine concentration, and up to 1OOO:l in the experiments at high amine concentrations. Tryptic digestion of [Lys], was done at pH 7.4 (pH-stat) with 2 pg trypsin per mg of [Lys] ,, for 24 h (Waley & Watson, 1953). The digest was chromato- graphed on Whatman No. 1 paper with n- butanol-acetic acid-water-pyridine (1 5 : 3 : 12 :

lo), and the peptides were detected with ninhydrin.

RESULTS AND DISCUSSION

Reaction of P-acetyllysine in ethanol con- taining sodium acetate gave A,, = 468nm in agreement with &, for the products of other amines reacting in this solvent (Fager etal., 1973). In aqueous solution at pH 8.5 and 10.0 P-acetyllysine, lo-’ M , gave a 468 nm product. But a control reaction of CNBFz gave a hydrolysis product with the same Am=. It is necessary to distinguish hydrolysis from aminolysis. We routinely carried out contrc! reactions without amine to see whether hydroly- sis was an important contributor to the spec- trum. Also, as will be described, some reactions were carried out in dry h e m e in which no hydrolysis occurs.

With [Lys], the results depended on. pH. First, [Lys], reacted much faster than did small amines. Fnhanced reactivity of polymeric reactants is well known. When the reaction was carried out at pH 3.5 or 10, in 0.01 M KCI, at a lysyl residue concentration of lo-’ M and a lysyl residue:CNBFz ratio of 10 or 20, the product had A,, of 382nm (Fig. 1) and was not fluorescent. Reactions of [Lys], with CNBFz at pH 8.5,O.Ol M KCl, gave two peaks, 382 and 463nm in the ratio of 3: l in absor- bance. At pH 7.8 we obtained largely the 463 nm peak. Alkaline hydrolysis also produces a product with A,, in this region, but since [Lys], gives less of the 463 nm peak at higher than at lower pH, the product a t 463 nm is not the result of hydrolysis.

The question arose as to whether the 382 nm product was covalently modified [Lys],, or a low molecular weight product from CNBFz resulting from a reaction catalyzed by [Lys], but not giving a covalent link to [Lys],. It was found that the 3 8 2 m chromophore (from a 2.5-h reaction) was not removed by extensive dialysis against water, 1 M KCl or 6 M guani- dinium chloride. (Dialysis of base-hydrolyzed CNBFz resulted in complete removal of solute.) After extensive dialysis against the sequence KCl, quanidinium chloride and water, the product was trypsinized. The enzymolysis mixture (of pH6.9) had Am= =38Onm.

465

J. EELLO and H. PATRZYC

-! _ - ___ -

0 5~

FIGURE 1 Time dependence of the reaction of [Lys], with CNBFz at pH 9.5. Spectra from bottom to top are for 2, 11, 21, 40, 6 2 and 270 min. Concentrations are: [Lys),, M in residues; CNBFz, 5 X M; KCL 0.01 M.

Therefore, the 380nm peak is not the result of high pH per se. Paper chromatography of the tryptic digest showed no large peptides (slow moving), only fast moving small peptides and no fluorescent spots. Thus, the 382 nm peak of the intact polymer is not the result of a 470 nm product having a polymer environment. The 380nm product is unstable to hydrolysis by 6 N HCI at 107", unlike the 470nm product, which is at least partially stable (Fager et al., 1973). Only lysine was seen (ninhydrin) on a thin-layer chromatogram of hydrolyzed [ Lys] , containing the 380nm product, i.e. no lysine with its &-amino free and its €-amino modified was seen.

We considered an explanation of these results as follows. At relatively low pH there are relatively few unprotonated amino groups, and the reaction proceeds largely as with small amines (463nm product). At high pH the proportion of unprotonated amino groups is large and they are relatively close together, as compared with a similar concentration of unconnected amino groups. This provides an opportunity for a cooperative reaction or for sequential reactions to give a different product.

In support of this was a salt effect observed for the reaction with [Lys],. At pH 8.5, in 0.01 M KCI, A380/A463 was 3, but in 0.1 M KCI the major product was the 463 nm chromo- phore. A control with CNBFz alone gave a 470nm peak from hydrolysis with absorbance equal to 14% of the 463nm band from the [Lys] , reaction. (The concentrations of react- ants were 5 x lo-' M CNBFz and M [Lys],.) Also, the rate of formation of the 380nm peak was 2.5 times faster in 0.01 M KCl than in 0.1 M KCI at pH 9 . 5 . Since the pK of

466

WPVE LENGTH, nm

amino groups of [ Lys] , increases with increasing salt (Applequist & Doty, 1962), there would be a higher concentration of unprotonated amino groups in 0.01 M than in 0.1 M salt, leading to the observed results.

This hypothesis was tested in other ways. We treated CNBFz with low and high concen- trations of small amines. With 1 M n-butylamine at pH 10-10.5 (at which the concentration of unprotonated amine is about O S M ) the pro- duct had (after dilution with ethanol) A,, = 372nm. Fig. 2 shows the spectra of products of the reaction of lC4 M CNBFz with IC3, lo-' and 1 M butylamine at pH 10 after a 2 4 h reaction. The shift from the 380 to 470nm products is seen. At pH 8.0 1 M n-butylamine with M CNBFz gave a brown solid, soluble in ethanol. Both the supernate and the dissolved solid had Amax = 468 nm. At this pH the con- centration of unprotonated amine is calculated to be about 3 x M, and, as expected under

WPYE LENGTH. m

FIGURE 2 Dependence of reaction products on mncentration of n-butykmine. A, 1 M; B, lo-' M; C, M. CNBFz mncentration, lO-.M; pH 10.0. Spectra taken on solutions diluted with an equal volume of ethanol.

4-CHLORO-7-NITROBENZOFURAZAN

our hypothesis, the product was the same as that obtained with M n-butylamine at pH 10, the “normal” 468nm product.

Analogous results were obtained with aqueous ammonia and with methylamine at high and low concentrations.

Reactions were also done with a micelle- forming amine, dodecylamine (DDA) at to 3 x M, at pH 9.5, in the absence of added salt, but in the presence of the NaCl produced by adjusting the pH of DDA - HCl with NaOH. At neutral pH DDA * HCl has a critical micelle concentration (c.m.c.) of about 1.2 x M at the ionic strength of our most concentrated solution (Kushner et al., 1957). At the pH of our experiments, 9.5, the c.m.c. is expected to be lower because of lesser charge repulsion, since not all amino groups are protonated. Therefore, our experiments were done above the c.m.c. (except perhaps at the lowest concentration), but some mono- mer and pre-micellar oligomeric aggregates are present as well. As the concentration of DDA is increased the ratio A4,0/A381 decreases from 1.4 to 0.3, as shown in Fig. 3. The spectra of

I O - ~ ~ x I O - ~ ~ X I O - ~ DODECYLAMINE, MOLESILITER

FIGURE 3 Ratio of A4,0/A380 as a function of dodecylamine Concentration. Conantration of CNBFz, 10.’ M at

and 5 x M dodecylamine and M at loT3 and 3 X M dodecylamine.

the DDA reaction mixtures were measured on solutions diluted with an equal volume of ethanol. (Ethanol was required to make a clear solution, since DDA solutions were turbid except at the lowest concentration.) Since alcohols destabilize micelles, the spectra are probably not distorted by a micellar environ- ment for the chromophore, but are comparable to the spectra for the derivatives of the smaller amines. The high A470/A381 ratio at low DDA may arise from a significant fraction of the DDA being non-micellar. As the fraction of micellar DDA increases, the proportion of 381 nm product increases, but it levels off at a ratio of 0.3, significantly higher than for 1 M butylamine. If the pK of DDA is typical of primary aliphatic amines, at pH 9.5 about half is in the protonated form, DDAH*. We assume that mixed micelles of DDA and DDAH’ form. Electrostatic repulsions would disperse the DDAH’ through the micelle. This would inter- rupt the continuity of the DDA, limiting the proportion of aggregates suitable for carrying out the multiple or cooperative reaction. Also, it is likely that hydrogen bonded association between - N H f and -NH2 head groups occurs, resulting in diminished reactivity of the -NH1 groups, effectively reducing the concentration of free amine in the micelle surface.

Between the two lowest concentrations of DDA a distinct decrease in A-470/A381 takes place. Some hydrolysis of CNBFz occurs at pH 9.5, contributing to the magnitude of the 470nm peak, particularly a t the lowest DDA concentration, at which the reaction with DDA is slow. This may be a factor in the higher &,,,/A381. The main point is that an amine at low average concentration, but at high local concentration, produces the 381 nm product.

Detergent reactions are open to another interpretation, namely, that reaction takes place with the CNBFz in the less polar transi- tion zone between the apolar and polar regions, and that the result arises from the effect of the less polar medium. For this reason, and others t o be noted below, we studied the reaction of n-butylamine in the apolar solvent n-hexane. Butylamine, at M and 1 M, was treated with M CNBFz, at room tempera- ture overnight. The solutions were diluted 10-fold with hexane. The spectrum from lo-’

467

J . BELL0 and H. PATRZYC

M amine had A,, = 430 nm, and the product from 1 M amine had A,, = 390. When the solutions were diluted 10-fold with ethanol instead of hexane, the wavelengths shifted to 465 and 380nm (with a shoulder at 465), respectively. The 465nm product was fluor- escent under ultraviolet illumination. The 380nm product contained some of the 465 nm product, as indicated by the shoulder and a weak fluorescence when spotted on paper. After a 7-h reaction (instead of 24-h one) only the product from the lo-* M amine showed fluorescence.

Taking all the results together, it is clear that when several amino groups can act together or successively the 380 nm product is obtained. But when they are widely spaced the 470nm product is obtained. The novel reaction to generate the 380 nm product does not require both unprotonated and protonated amino groups, since it is formed in the reaction with n-butylamine in dry hexane, a condition in which the amine is not protonated.

Ferguson et al. (1 974) reported that phenolic hydroxyl gives a 370nm product with CNBFz. Our results show that a 370nm absorption band for modified protein does not necessarily demonstrate modification of a tyrosyl residue; it may indicate a cluster of amino groups, or, perhaps, one amino group cooperatively assoc- iated with a cluster of other groups which may serve to polarize CNBFz so as to alter the reaction. Also, when labeling protein amino groups it should not be assumed that only the 470nm product is obtained; the 370nm pro- duct should be looked for.

Aboderin etal. (1973) reported that 4- alkylamino derivatives of CNBFz have a pK near 10, and that on raising the pH from 7 to 11 the 480nm band decreases and a band appears at about 390 nm. Several workers (Baines et d., 1977; DiNunno eta/ . , 1975) have discussed Meisenheher complex for- mation of CNBFz with alkali (Ama 380), but thi’s does not occur under our pH conditions. In our experiments the shorter wavelength band is generated not only at pH 10 but also at relatively low pH in the case of [Lys],, and persists at neutral pH, so that the alkali effect is not a factor. Furthermore, at about pH 8 both loe3 M and 1 M butylamine give the

468

470nm product, while at pH 10, 1 M, but not M, butylamine gives the 380nm product.

Therefore, it is not pH per se which determines the nature of the product, but it is rather the concentration, and more important, the local concentration, of unprotonated arnine which is dominant.

We did not see any indication of the 470 nm product being converted to the 380nm product. In reactions such as with 1 M butyl- amine at pH 10.5 the 380nm product is formed from the beginning with no indication of an initial 470nm product being formed and then vamshing. Either the 380 nm product is formed from the 470nm product via a very rapid reaction, or is formed via a separate route. This was investigated in another way. [Lys], was treated with CNBFz at pH 8.5 to give the 470nm product. After dialysis the solution was brought to pH 10. Spectra taken immed- iately and after 24 h showed no 380nm peak, nor did one appear on neutralization to pH 7. Thus, the 380nm product arises from a path which does not go by way of the 470nm product. We did see some reversible decrease in absorbance at 480 and increase at 370. These effects arise from the titration of the 4-amino derivative.

The phenomenon reported here may be related to the evolution of primitive enzymes. We normally consider a catalyst in terms of accelerating a reaction which is slow in the absence of the catalyst. But by accelerating one reaction relative to another, or by acceler- ating a second sequential reaction, a catalyst can change the nature of the product. Thus, by incorpomting (covalently or non-covalently) a multiplicity of reactive groups into a micelle, membrane, polymer or colloidal particle, all of which might lack a fixed or regular structure, a change in primitive metabolism could occur which would be of a qualitative nature. The cooperative interaction might involve some groups binding the substrate, binding water or ions, polarizing certain bonds, or forming electron-rich complexes of altered reactivity. Such a mechanism provides a path for converting substrate t o product via an intermediate which is thermodynamically disfavored relative to the substrate, and has other advantages of enzyme clusters (Gaertner, 1978).

4CHLORO-7-NITROBENZOFURAZAN

Considering small molecules as proto- enzymes, a cooperative or a sequential route is analogous to the mechanism of an enzyme, either an enzyme which catalyzes a new re- action, or a cluster of enzymes which traps the product of one enzyme as substrate for the next. The acceleration of sequential reactions is, in effect, a qualitative change, since the demands of a successful protolife system are not compatible with indefinitely prolonged conversions, which would risk the loss of inter- mediates by diffusion or side reactions. Although in our case the substrate residue becomes attached covalently to the "enzyme", the same principle would apply to the formation of transient covalent intermediates.

In exploratory experiments of CNBFz with gelatin we obtained qualitative evidence of cross-linking. Gelatin in 1 : 1 HzOethanol, con- t d g 0.05M sodium phosphate buffer, pH 9.5, was treated with CNBFz at a ratio of 100 amino acid residues per CNBFz molecule. After standing overnight at room temperature, the gelled brown solution, and the gelled colorless control, were broken into lumps, extracted with a 50-fold volume of 50% ethanol, and washed for several days with water. On warming to SO", the control melted to a low viscosity fluid, while the modified gelatin partially melted to a viscous, lumpy fluid. On prolonged heating the lumps became fluid, but the solu- tion remained much more viscous than the control. A second experiment was done with gelatin flakes swollen in a solution of CNBFz in 1 : 1 HzOethanol, pH 9.5. After 20h, the supernate was decanted from the now-brown gelatin, which was extracted with 1 : 1 ethanol- water and washed extensively with water. A control was carried through a mock reaction lacking CNBFz; this did not tum brown. The control dissolved readily at SO", as is normal for swollen gelatin heated to SO". But the CNBFz-treated gelatin did not dissolve after 3 h at 50"; on heating to 75" it gradually dissolved. Relatively few nitrobenzofurazan groups could have been bound covalently, as gelatin contains only about 3 lysyl residues per 100 residues, no sulfhydryl and very small amounts of histidine or tyrosine (Eastoe & Leach, 1977). Therefore, the retarded solu- bility is not likely to have arisen merely from

the introduction of the few hydrophobic groups. Cross-linking is more likely. A low degree of cross-linking of gelatin is charac- terized by insolubility at SO", and gradual dissolution at higher temperature as a few sensitive bonds are hydrolyzed (Bello & Bello, 1958). It is not apparent whether or not the cross-linking involves the 370 nm reaction path. The spectrum of the cross-linked gelatin, after dissolution, had a 470nm band in a broad spectrum, but not a distinct one at 370nm. Since formation of an infinite network requires only two cross-links per chain, a small amount of the 370nm product might not be easily detected. The low frequency of lysine in gelatin is not likely to lead to a cooperative reaction. The cross-links might arise from yet another route.

It appears likely that reaction of CNBFz with biopolymers can produce at least three products: the 4-amino-substituent product (Imu - 470 nm), the cross-linked product, and the new product with A,, - 370nm. The latter appears not to be the same as the second, as judged by the two-step reaction with [LYs],, first at pH 8.5 then at pH 10.

REFERENCES

Abode&, A.A., Semakula, R.E.K., Boedefeld, E. & Kenner, R.A. (1973) FEBSLett. 34,90-94

Allen, G. & Lowe,G. (1973) Biochem. J. 133, 679- 686

Applequist, J . & Doty, P. (1962) in Polyamino Acids, Polypeptides and Proteins (Stahmann, M.A., ed.), pp. 161-182, Univ. Wisconsin Press, Madison

Baines, B.S., Allen, G . & Brocklehurst, K . (1977) Biochem. J. 163,189-192

Bello, J. & Riese-Bello, H. (1958) Sci. Ind. Photo- graph. Ser. 2, 29, 361-364

Breslow, R., Doherty, J.B., Guillot, G . & Lipsey, C. (1978) J. Am. Chem. SOC. 100,3227-3229

DiNunno, L., Flono, S. & Todesco, P.E. (1975) J. Chem. SOC. Perkin Trans. 11,1469-1472

Eastoe, J.E. & Leach, A.A. (1977) in The Science and Technology of Gehtin (Ward, A.G. & Courts, A., eds) , pp. 73-107, Academic Press. New York

Fager, R.S., Kutina,C.B. & Abraharnson, E.W. (1973) Anal. Biochem. 53,290-294

Ferguson, S.J., Lloyd, W.J. & Radda, G.K. (1974) FEBS Lett. 38,234-236

Gaertner, F.H. (1978) Trends Biochem. Sci. 3,63-65 Ghosh, P.B. & Whitehouse, M.W. (1968) Biochem. J.

469

J. BCLLUand H. Y A I K L Y C

108,155-156 Kenner, R.A. & Aboderin, A.A. (1971) Biochemistry

Kushner, L, Hubbard,W.D. & Parker, R.A. (1957) J. Res. Natl. Bur. Standards US59.113-119

Nit ta , K., Bratcher, S.C. & Kronman, M.J. (1979) Biocbem, J. 177,385-392

Shipton, M., Stuchbury, T., Brocklehurst, K., Herbert, J.A.L & Suschitzky, H. (1976) Biocbem J. 161,

10,4433-4440

627-637

Waley, S.G. & Watson,J. (1953) Biochem J. 55, 328-337

Address: Jake Bell0 Biophysics Department Roswell Park Memorial Institute Buffalo, New York 14263 U.S.A.

470