Photochemical reactions of fentichlor with soluble proteins

Chem.-Biol. Interactions, 52 (1984) 213-222 Elsevier Scientific Publishers Ireland Ltd.

PHOTOCHEMICAL PROTEINS

REACTIONS

213

OF FENTICHLOR

WITH SOLUBLE

DIANE M. RICKWOOD and MARTIN D. BARRATT* Environmental Safety Laboratory, Bedfordshire, MK44 ILQ (U.K.)

Unilever

Research,

Colworth

House,

Sharnbrook,

(Received April 26th, 1984) (Revision received August 17th, 1984) (Accepted August 20th, 1984)

SUMMARY

The photochemical reactions of the photoallergen fentichlor with soluble proteins have been studied. [ ?‘S] Fentichlor was shown to bind covalently to human serum albumin (HSA) when irradiated with UV light (313 nm). HSA had the ability to bind at least eight molecules of fentichlor per molecule of protein. Fractionation of fentichlor-HSA photoadducts after (a) treatment with cyanogen bromide and (b) reduction, carboxymethylation and digestion with trypsin showed that the bound fentichlor was distributed fairly evenly throughout the sequence of the HSA molecule. Fentichlor was also shown to form photoadducts with human r-globulin and with bovine insulin. Its binding to insulin was restricted to the B chain of the molecule. Fundamental differences between the photochemical reactions of the photoallergens fentichlor and tetrachlorosalicylanilide (T&S) with soluble proteins are discussed. The reactions of fentichlor with soluble proteins are not restricted to specific binding sites (unlike T&S). Fentichlor has the potential .to react photochemically with a wide range of proteins in the epidermis and dermis, to form antigens. Key words: Fentichlor Photoadduct

Bithionol -

Photoallergen -

Protein-binding -

INTRODUCTION

In 1961, Wilkinson [I], described photodermatitis due to the interaction of the germicide, T4CS with human skin under the influence of UV light. In some cases the skin was found to remain abnormally sensitive to UV light *To whom all correspondence should be sent. Abbreviations: HSA, human serum albumin; T,CS, tetrachlorosalicylanilide.

214 for months or even years after all known contact with the photosensitiser [2] had ceased. Similar photoallergic reactions have been described for a number of structurally-related compounds, including other substituted salicylanilides, hexachlorophene, fentichlor and bithionol [ 31. The photochemical reaction of T4CS with isolated soluble proteins is well established [4-61 and the involvement of a T4CS-protein conjugate in the photoallergic response to T&S has been demonstrated in guinea pigs using serum albumin [ 71. Photochemically-induced protein conjugate formation by photoallergens as a general phenomenon in contact photodermatitis has been proposed [8], but is yet to be firmly established. In the course of our work in this area, we have previously identified the location of a major strong binding site for the photochemical reaction of T,CS with HSA [9]. The present study is mainly concerned with the photochemical interactions of the fungicide and germicide fentichlor (bis[ 2-hydroxy-5-chlorophenyllsulphide) with soluble proteins, with some reference to the related compound bithionol (bis[ 2-hydroxy3,5-dichlorophenyl] sulphide). A number of cases of photocontact dermatitis have been reported from the use of antifungal preparations containing fentichlor [lo-121 and bithionol [ 131. In this paper the stoichiometry and specificity of their photochemical reactions with human serum albumin, y-globulin and bovine insulin are discussed and compared with those of T&S. EXPERIMENTAL

Materials

HSA (fraction V, fatty acid free) bovine insulin, human y-globulin, trypsin (bovine pancreas type XI, treated with diphenyl carbamyl chloride chymotrypsin inactivator) and cyanogen bromide were purchased from the Sigma Chemical Co. (Poole, U.K.). Fentichlor (bis[2-hydroxy-&chlorophenyl] sulphide) (Cocker Chemical Co. Ltd., Oswaidtwistle, U.K.) was recrystallised from benzene and had m.p. 174-176°C (lit. 173°C [14]). Bithionol (bis[2hydroxy3,5dichlorophenyl] sulphide) (Hopkin and Williams Ltd., U.K.) was recrystallised from xylene and had m.p. 187-189°C (lit. 188°C [ 151). Elemental. sulphur-35 (solution in toluene) was obtained from Amersham International PLC (Amersham, U.K.). All other chemicals were the best grades available. Synthesis

of [““S] fentichlor

[15]

Iodine (0.1 g) and flowers of sulphur (0.32 g) in dry carbon tetrachloride (15 ml) were placed in a round bottomed flask with a side-arm and delivery tube and a stirrer in an ice-salt bath at -5°C. To this mixture was added a solution of sulphur-35’(approx. 2.6 mCi) in toluene (0.1 ml). A stream of chlorine gas (0.7 g) prepared by adding concentrated hydrochloric acid to potassium permanganate crystals, was bubbled slowly into the stirred reac-

215 tion mixture. When the addition of chlorine was complete, the mixture was allowed to warm up to room temperature, at which point a solution of 4-chlorophenol (2.57 g) in carbon tetrachloride (15 ml) was added with stirring. After stirring for a further 3 h in the dark and standing overnight, the resulting white precipitate of [“S] fentichlor was filtered off and washed with cold carbon tetrachloride. After recrystallisation from toluene, the yield of [35S] fentichlor was 0.87 g (30.3% of theory). This had m.p. 176.5179°C and a specific activity of 0.778 pCi/mg. It was found to be pure by thin-layer chromatography (Silica gel H) in chloroform/methanol (19 : 1) and by isotope dilution analysis. Irradiation

of protein

solutions

with fentichlor

Stock solutions of fentichlor (and bithionol) in ethanol were prepared and added to solutions of protein in 0.1 M Tris-HCl buffer (pH 8.1) using a microsyringe. The final concentration of ethanol in a sample never exceeded 1%. Samples were irradiated either in a quartz cuvette (3 ml) or in quartz vials (20 ml) with stirring at a distance of 5 cm from the light source. No attempt was made to exclude oxygen. The light source was a Hanovia medium pressure Hg-arc lamp fitted with Schott UG-5 (2 mm; UV-transmitting) and WG-310 (1 mm) filters giving 58% transmission at 313 nm. The intensity of the light source was measured as 4.1 pW/cm’ at a distance of 25 cm using an Oriel Spectroradiometer. After irradiation, samples were passed through a column of Sephadex G-10 (5 cm2 X 33 cm) equilibrated with 0.05 M ammonium bicarbonate to separate the photoadduct from unbound fentichlor. Methods

Monomer HSA was prepared from the commercial sample by the method of Pedersen [16]. Reduction, carboxymethylation and tryptic digestion of fentichlor-protein photoadducts were carried out by the procedures of Wiman [ 171. Monomer-HSA-fentichlor photoadduct was treated with cyanogen bromide according to the method of Meloun et al. [ 181. The major two fractions obtained by Sephadex G-100 chromatography were designated C and N, respectively, according to Meloun and Kusnir [19]. Separation of reduced and carboxymethylated A and B chains of the fentichlor-insulin photoadduct was effected by precipitation of the B chain in ammonium acetate at pH 6.5 [20]. Protein concentrations were determined by the method of Lowry et al. [21] using HSA as a reference standard. Liquid scintillation ,counting was performed using a Packard Prias P.L.D. liquid scintillation spectrometer. Packard MI-97 liquid scintillant (3 ml) was added to O.l- or 0.2-ml aliquots of the sample to be counted. UV spectra were obtained using a Cary 118C UV-VIS spectrophotometer. Samples for amino acid analysis were hydrolysed in 6 N HCl at 1ld”C for 18 h under Nz. Amino acid analyses were carried out on a Rank-Hilger Chromaspek amino acid analyser.

216 RESULTS

Irradiation

AND DISCUSSION

of fentichlor

and bithionol

with HSA

The UV spectra of equimolar mixtures of HSA (3.52 X lo-’ M) with fentichlor and bithionol before and after irradiation for 30 min by UV light are shown in Figs. 1 and 2, respectively. Spectra of fentichlor and bithionol alone are also shown for comparison. Addition of HSA to fentichlor and bithionol resulted in small shifts of the absorption maxima of both compounds to longer wavelengths. Irradiation with UV light over 30 min resulted

I:.:\ .

0

280

320

34o

400

320

340

400

B l-4

l-2

.

o-a;

O-6.

o-4.

o-2.

01

280

Fig. 1. UV Spectra of (A) 3.52 x 10.’ M fentichlor in 0.;: TrisHCl (pH 8.1) and (B) 3.52 x 10.’ M HSA and 3.52 x 10m5 M fentichlor before (solid line) and after (broken line) irradiation with UV light (30 min).

217

a

0

O-8

\ 0.6

o-4

o-2

0 2

320

360

Fig. 2. UV Spectra of (A) 3.52 X lo-’ M bithionol 3.52 x 10“ M HSA and 3.52 X 10“ M bithionol line) irradiation with UV light (30 min).

400 nm

in 0.1 M TrisHCI (pH 8.1) and (B) before (solid line) and after (broken

in a gradual broadening of the absorption maxima at 315 and 330 nm, respectively and an increase in absorption at 280 nm; at the same time, the solutions became brown in colour. After passing through a column of Sephadex G-10, the irradiated samples retained their brown colour and gave spectra with identical contours to those obtained prior to the column treatment. Passage of unirradiated germicide-HSA samples through the column resulted in quantitative removal of the germicide from the protein. These results suggest that both fentichlor and bithionol can react photochemically with HSA. Since we do not know the extinction coefficients of the photo-

218 products, UV spectroscopy can give us only a qualitative assessment of the photochemical binding of fentichlor or bithionol to HSA. In order to determine the stoichiometry of the interaction, it is necessary to use a radiotracer method. [ 35S] Fentichlor was conveniently synthesized directly from elemental sulphur-35 [15]. Unfortunately, repeated attempts at the synthesis of [ 35S] bithionol by the same method, were unsuccessful. Table I shows the effect of UV-irradiation on various ratios of fentichlorl HSA. These results show that the photochemical reaction of fentichlor with HSA is not restricted to one major binding site on the I-ISA molecule, as is the case for the binding of T,CS to HSA [9]. It is quite possible that with longer irradiation times and higher initial ratios of fentichlor/HSA, even higher binding levels could be achieved. Cyanogen bromide cleavage of fentichlor-HSA photoadduct Elution of the cyanogen bromide digest of [ 35S] fentichlor-ESA (molar ratio, 0.68:1) on Sephadex G-100 (Fig. 3) yielded the two main fractions (C, residues 299-585 and N, residues l-123 and 124-298 [17]) of the HSA molecule as well as some polymerised and incompletely-digested protein. Radioactive counts showed that the [ 35S] fentichlor was distributed between the C and N fractions in a ratio of 1: 0.98, respectively. When a similar experiment was carried out previously on a [ ‘“C]T,CSHSA photoadduct, the distribution of the radioactivity between the C and N peaks was 1: 2.8 [9]. Subsequently the binding of T4CS was shown to take place at a single major site in the N-fraction of the HSA molecule [9,22]; the labelling of the C fraction was shown to arise from incomplete protein digestion by cyanogen bromide. The present results demonstrate that unlike T4CS, fentichlor does not bind preferentially to any one specific site on the HSA molecule. Tryptic digestion of carboxymethylated fentichlor/HSA The elution profile of the tryptic digest of *carboxymethylated fentichlor/HSA (molar ratio, 7.75 : 1) on Sephadex G-50 (superfine) TABLE

I

COVALENT

BINDING

HSA concentration:

1.5

OF [ =S] FENTICHLOR X 1O-4

Irradiation

1.0 1.0 5.0 5.0 10.0 10.0

0 10 10 20 10 20 irradiation,

unbound

TO HSA

M.

Initial molar ratio (Fentichlor/HSA)

aAfter

[35S] is shown

fentichlor

time (min)

Binding ratioa HSA, M/M)

(Fentichlori

0.13 0.76 2.21 3.13 3.86 7.75 was removed

by passing

through

Sephadex

G-l 0.

219

1000

250

Fraction

No.

Fig. 3. Elution profile of the cyanogen bromide digest of [ “Slfentichlor-HSA ratio, 0.68: 1) on Sephadex G-100 in 0.1 M ammonium formate (pH 2.9).

(molar

in Fig. 4. The elution profile consisted of two major peaks - fractions 20-23, containing undigested protein and large peptides and fractions 50-55, consisting of very small fragments -with a broad spread of material in between. There was no evidence of specific labelling of any peptide fragments in the tryptic digest. All of the applied radioactivity was recovered with the protein. The distribution of the [3sS] fentichlor throughout the elution profile supports the view that fentichlor binds generally over the surface of the HSA molecule rather than to any specific site or sites. Irradiation of fentichlor with human y-globulin A solution of human r-globulin (1.16 X 10e4 M) was irradiated with an equimolar ratio of [35S] fentichlor for 30 min. After passing through a column of Sephadex G-10, the molar ratio of bound fentichlor/r-globulin was 0.54. A solution of HSA (1.16 X 10e4 M) with equimolar fentichlor irradiated under identical conditions gave a molar ratio of bound fentichlorl HSA of 0.63. The observation that fentichlor can bind to r-globulin almost as efficiently as to HSA is in direct contrast to the behaviour of T4CS. Kochevar and Harber were unable to demonstrate covalent bonding between T&S and y-globulin after UV-irradiation [5]. In our laboratory the photochemical binding of [ 14C] T4CS to -y-globulin has been shown to be about 100 times less efficient than its binding to HSA (M.D. Barratt, unpublished). Irradiation of fentichlor with bovine insulin Irradiation of a solution of bovine insulin

(2 X lob3 M) [35S]fentichlor

220 o-0.

1000

0.6. 1

750

00 6 04 0

O-2. 250

OF.”

*<,aI

20

30

40 Fraction

50

60

No.

Fig. 4. Elution profile of the tryptic digest of reduced and carboxymethylated [35S]fentichlor/HSA (molar ratio, 7.75 :l) on Sephadex G-50 (superfine) in 0.05 M ammonium bicarbonate solution.

(initial molar ratio, 1.4: 1) for 2 h resulted in the photochemical binding of a 0.17 molar ratio of fentichlor. The fentichlor-insulin photoadduct had a distinct brown colour. After reduction and carboxymethylation, the A and B chains of the [ 35S] fentichlor-insulin photoadduct were separated by precipitation of the B chain [20] and purified by chromatography on Sephadex G-10. It was found that 94% of the radioactivity (average of 2 experiments) was associated with the B chain. The B chain had the same brown colour as the initial photoadduct whilst the A chain was colourless. The photochemical binding of fentichlor to bovine insulin appears to be very similar to that of T&S, where specific binding to the B chain was also observed [6]. In the case of T4CS, the covalent binding site on insulin was identified as one of the two histidine residues on the B chain, via the isolation of a [ 14C] T4CS-labelled tryptic peptide. Possibly due to its low aqueous solubility, we were unable to obtain a sufficiently high ratio of fentichlor/insulin in the photoadduct to identify the site of reaction by amino acid analysis. General discussion The photochemical reaction of fentichlor with proteins probably takes place via loss of one or both chlorine atoms. A similar mechanism is well established for the photochemical reactivity of T4CS [23]. From consideration of its chemical structure, bithionol might be expected to react in a similar manner. The symmetrical nature of the fentichlor molecule also suggests the possibility that it could act as a protein crosslinker, a property

221 which has been observed when T&S reacts photochemically with erythrocyte membrane proteins [ 241. The photochemical binding of fentichlor to proteins differs fundamentally from that of T4CS. In the case of T,CS; strong non-covalent binding to the protein appears to be a prerequisite for photochemical (covalent) binding [ 5,9]. This requirement is exemplified by the difference between the reactivity of T.&S with HSA and with 7-globulin. The photochemical reaction of fentichlor with proteins appears to be unconstrained by such requirements, e.g. it reacts non-specifically with HSA and with almost equal efficiency with both HSA and y-globulin. Since a molecule of fentichlor apparently may not require to be non-covalently bound to protein for photochemical binding to occur (unlike T&S), it seems likely that the activated species produced by UV-irradiation of fentichlor may have a much longer lifetime than the corresponding one from T,CS. Evidence of the formation of long-lived (5-10 min) free radicals from the irradiation of fentichlor and bithionol has recently been obtained by electron spin resonance spectroscopy (SK. Jackson et al., unpublished). A further consequence of the differences in specificity and reactivity between fentichlor and T4CS may be for the two photoallergens to form antigens by reacting with different groups of proteins from the epidermis and dermis. The ability of fentichlor and bithionol to conjugate with pro.teins when irradiated with UV light greatly improves the correlation between known photoallergens and their ability to form protein conjugates. In addition to T&S, the formation of protein conjugates has previously been demonstrated for the photoallergen buclosamide(4-chlorosalicyl-n-butylamide) [25]. Work is in progress in this laboratory to test the ability of other known photoallergens to form photoadducts with proteins, with a view to establishing an in vitro screening procedure for potential photoallergens. ACKNOWLEDGEMENTS

The authors wish to thank Miss Ruth Pendlington, Miss Sally Cole and Mr. Paul Clifton for providing excellent technical assistance during the course of this work. REFERENCES 1 D.S. Wilkinson, Photodermatitis due to tetrachlorosalicylanilide, Br. J. Dermatol., 73 (1961) 213. 2 I. Willis and A.M. Kligman, The mechanism of the persistent light reactor, J. Invest. Dermatol., 51 (1968) 385. 3 S.Z. Smith and J.H. Epstein, Photocontact dermatitis to halogenated salicylanilides and related compounds, Arch. Dermatol., 113 (1977) 1372. 4 F.P. Jenkins, D. Welti and D. Baines, Photochemical reactions of tetrachlorosalicylanilide, Nature, 201 (1964) 827. 5 I.E. Kochevar and L.C. Harber, Photoreactions of 3,3’,4’,5tetrachlorosalicylanilide with proteins, J. Invest. Dermatol., 68 (1977) 151.

222 6 M.D. Barratt, The photochemical reaction of tetrachlorosalicylanilide with bovine insulin, Photobiochem. Photobiophys., 3 (1981) 59. 7 P.S. Herman and W.M. Sams, Requirement for carrier protein in salicylanilide sensitivity, The migration-inhibition test in contact photoallergy, J. Lab. Clin. Med., 77 (1971) 572. 8 L.C. Harber, H. Harris and R.L. Baer, Photoallergic contact dermatitis, Arch. Dermatol., 94 (1966) 255. 9 D.M. Rickwood and M.D. Barratt, Evidence for a major strong binding site for tetrachlorosalicylanilide on human serum albumin, Photochem. Photobiol., 35 (1982) 643. 10 J.N. Burry, Photoallergies to fentichlor and multifungin, Arch. Dermatol., 95 (1967) 287. 11 R. Clayton and M. Feiwel, From fentichlor sensitivity to actinoid reticuloid?, Proc. Roy. Sot. Med., 69 (1976) 379. 12 CA. Ramsay, Skin responses to ultraviolet radiation in contact photodermatitis due to fentichlor, J. Invest. Dermatol., 72 (1979) 99. 13 O.F. Contact photo-dermatitis to bithionol, Arch. Jillson and R.D. Baughman, Dermatol., 88 (1963) 409. B. Dunning and E.W. Drake, Preparation and bacteriological study of 14 F. Dunning, some symmetrical organic sulphides, J. Am. Chem. Sot., 53 (1931) 3466. Dihydroxydiphenyl sulphide derivatives, Japanese patent 15 T. Ueda and K. Tsutsui, (1958) 776 [Chem. Abs., 53 (1959) 1254d]. Pedersen, Exclusion chromatography, Arch. Biochem. Biophys., Suppl. 1 16 K.O. (1962) 157. Primary structure of peptides released during activation of human 17 B. Wiman, plasminogen by urokinase, Eur. J. Biochem., 39 (1973) 1. M.A. Saber and J. Kusnir, Structure of N-terminal cyanogen bromide 18 B. Meloun, fragment of human plasma albumin, Biochim. Biophys. Acta, 393 (1975) 505. and J. Kusnir, Cleavage of human plasma albumin by cyanogen bromide 19 B. Meloun and character&&ion of the fragments, Collect Czech. Chem. Commun., 38 (1973) 148. of oxidised insulin, Biochem. J., 44 (1949) 126. 20 F. Sanger, Fractionation 21 O.H. Lowry, N.J. Rosebrough, A.L. Farr and R.J. Randall, Protein measurement with the Folin phenol reagent, J. Biol. Chem., 193 (1951) 265. 22 D.M. Rickwood and M.D. Barratt, Identification of the major covalent binding site for tetrachlorosalicylanilide on human serum albumin, Photobiochem. Photobiophys., 5 (1983) 365. 23 J.A. Coxon, F.P. Jenkins and D. Welti, The effect of light on halogenated salicylanilide ions, Photochem. Photobiol., 4 (1965) 7 13. J.C. Evans, C.A. Lewis and C.C. Rowlands, Comparison of the photo24 M.D. Barratt, dynamic action of Rose Bengal and tetrachlorosalicylanilide on isolated porcine erythrocyte membranes, Chem.-Biol. Interact., 38 (1982) 215. 25 E.G. Jung, U. Dummler and H. Immich, Photoallergie durch 4-chlor-2-hydroxybenzoesaure-n-butylamid I. Lichtbiologische Untersuchungen zur Antigenhildung, Arch. Klin. Exp. Derm., 232 (1968) 403.