Photochrrnrstry nnd Photobiology, Vol. 58, No. 3, pp. 351-355, 1993 Printed in the United States. All rights reserved

003 1-8655193 $05.00+0.00 0 1993 American Society for Photobiology

PHTHALOCYANINE-INDUCED PHOTOHEMOLYSIS: STRUCTURE-ACTIVITY RELATIONSHIP

AND THE EFFECT OF FLUORIDE

E. BEN-HuR*', Z. MALIK~, T. M. A. R. DlJBBELMAN3, P. MARGAKON4, H. AL14 and J. E. VAN h E R 4

'Nuclear Research Center-Negev, P.O. Box 9001. Beer-Sheva 84190, Israel, 'Bar Ilan University, Department of Life Sciences, Ramat-Can, Israel,

'Sylvius Laboratory, Department of Medical Biochemistry, P.O. Box 9503, 2300 RA Leiden, The Netherlands and 'Sherbrooke University, Faculty of Medicine, MRC Group in the Radiation Sciences, Sherbrooke, Quebec J IH 5N4, Canada

(Received 7 December 1992; accepted 22 February 1993)

Abstract-Phthalocyanine (Pc) containing AI, Ga or Zn as central metal ligand and substituted with a varying number of sulfonic acid residues as well as additional benzene rings were synthesized and their photodynamic activity was assayed using photohemolysis of human erythrocytes as an endpoint. The Pc derivatives vaned >300-fold in their photodynamic activity. Activity corrclated with binding of the dye to the cell, with the exception of some of the arnphiphilic dyes where cell uptake was an order of magnitude higher than expected from the observed activity. Fluoride was shown to inhibit AIPcS,-induccd photohemolysis. This effect occurred also with other AlPc and CaPc derivatives, but the concentration of F required to slow photohemolysis by a factor of two (K,) varied between 4 pM and 10 mM. Ruoresccnce spectral studies indicated complex formation between F-- and the dye, which was stronger for AlPc than GaPc derivatives. Ultrastructural studies using scanning electron microscopy showed that the photosensitized cells were converted to spherocytes and that F- prevented this to a large extent.

INTRODUCTION

Phthalocyanines ( P c ) ~ are second-generation photosensitiz- ers for photodynamic therapy (PDT) of cancer./,2 Recently, Pc have been studied for use in PDT of blood for viral in- activation prior to transfusion to enhance the safety of the blood upp ply.^.^ One of the problems encountered with this approach is that a t the doses required for blood sterilization red blood cells (RBC) are damaged to some cxtcnt. In order to reduce Pc-PDT-induced damage to RBC, yuenchers of type I photodynamic reactions have been used, some ofwhich protect KBC selectively without affecting viral inactivation? WeGY and othersI0,'l have been studying Pc-PDT-induced photohemolysis in order to understand its mechanism of action. However, no systematic study of the relationship be- tween Pc chemical structure and their photohemolytic activ- ity has been reported.

In the present work a series of Pc derivatives was prepared and their photohemolytic activity was assayed. In addition, the ability of fluoride to inhibit photohemolysis was deter- mined for some AlPc and GaPc derivatives. Variations of - 103-fold were found in both respects. In parallel studies (to be reported elsewhere) the activity of these derivatives in viral inactivation was determined, in order to identify a Pc with a high therapeutic ratio (viral inactivation activity di- vided by photohemolytic potency).

*To whom correspondence should be addressed, at The New York Blood Center, 310 East 67th Street, New York, NY 10021, USA.

tAbbreviarzons: Pc, phthalocyanines; MPc, metallophthalocyanines; PcS,,, sulfonated phthalocyanines; PDT, photodynamic therapy; RBC, red blood cell; DPBS, Dulbecco's phosphate-buffered saline; SEM, scanning electron microscopy; HPLC, high performance liquid chromatography.

MATERIALS AND METHODS

Photosensitizers. A mixture of sulfonated aluminum Pc (AIPcS,) was obtained from Ciba-Geigy. All other dyes were synthesized as de~cribed '~- '~ and purified to homogcneity by high performance liq- uid chromatography (HPLC). The disulfonated metallo (M)Pc were enriched in isomers with adjacent substituents (MPcS,,). Figure 1 shows the chemical structures of the dyes used. The Pc were stored at 4°C as 0.5 mM stock solutions prior to use. Quercetin was obtained from Sigma and stored as a 5 mM solution in ethanol. All other chemicals were analytical grade.

Preparation of RBC. Human erythrocytes were prepared as de- The cells were used immediately after preparation, sus-

pended at 0.7% hematocrit in Dulbecco's phosphate-buffered saline (DPBS) containing 10 mM glucose.

Lighl exposure. Prior to light exposure 3 mL RBC suspension was incubated for 30 min with 5 pM dye at 23°C in the dark. The cells were then centrifuged at I500 g for 5 min and resuspended in DPBS. The continuously stirred cell suspension in a 15 mL test tube was exposed to red light emitted from a conventional slide projector equipped with a cutoff filter (A > 605 nm). The spectrum of the emitted light overlapped the absorption spectrum of the different dyes. The light fluence rate was 200 W r r 2 at the surface of the cell suspension. Temperature did not increase by more than 1°C during irradiation. NaF or quercetin, when present during light exposure, was added immediately prior to illumination.

L.vszs of RBC. After light exposure the cells were centrifuged, rcsuspended in DPBS and incubated in the dark at 23°C. At intervals during the dark incubation, 0.3 mL cell suspension was added to 3 mL DPBS and mixed. The cells were centrifuged at 0°C for 5 min, and the hemoglobin content of the supernatant was measured by recording absorbance at 4 15 nm in a spectrophotometer. Percentage hemolysis was then calculated (100% was taken as the absorbance obtained from a sample lysed in distilled water). The time required to obtain 50% hemolysis is designated tSo, and its reciprocal was used to estimate the rate of photohemolysis. The standard errors in triplicate measurements of the same suspension were 1-2%. Vari- ation between experiments on different days was up to 20%. Each experiment was repeated two to three times, and the results shown are of typical experiments.

Binding ofPc to RBC. The RBC suspension (2.5 x 10' cells/mL) was incubated for 30 min at 23°C with 10 gM Pc. The cells were

35 1

352 E. BEN-HUR et al.

D

Figure 1 . The molecular structure of Pc derivatives. M is Zn, AlOH or CiaOlI. A: M-PcSZ,; B: M-PcS,; C: M-NSB,P D: M-NZSB2P; E:

M-N,SBP.

then rinsed 3 x with DPRS to remove unbound dye, and the emission spectrum of the bound dye was recorded using an Aminco SPF 500 spectrofluorimeter. Fluorescence was converted into dye concentra- tion using a standard curve, containing the same cell concentration as above. When fluoride was present, it was added after removal of unbound dye.

.S(anning electron microscopy (SEM) of RBC. The RBC suspen- sion was centrifuged and resuspended in a fixative containing 2.5% glutaraldehyde and 290 paraformaldehyde in 0. I Mphosphate buffer, pH '7, for I h, rinsed in the same buffer and then preparcd for scanning eleciron microscopy by the GTGO method of triple fixation.'! Brief- ly, the fixed cclls were attached to poly-L-lysine-treated glass cov- crslips, the cell monolayer was then postfixed by 2% OsO,, and the linal fixation step was performed in tannic acid/guanidine.HC12n/o. The cclls were then dehydrated in graded ethanol solutions, ethanol was replaced by graded solutions of Freon 1 13, and the preparations werc air dried and gold coated. The cells were examined by a Jeol 840 scanning electron microscope, and 200 cells in each sample were counted to assess cell morphology distribution.

RESULTS

Red blood cells loaded with various Pc derivatives were cxposed to light for 1-30 min (12-360 kJ m-,) and the ki- netics ot'heniolysis dctermincd as described in the Materials and Methods. Fluence-response curves were constructed us- ing the single parametric approach,I6 according t o which the hcmolytic curve can be characterized by tSo, i.e. the time required for 50% hemolysis. The results for the six most potent dycs are shown in Fig. 2. The relative activity of all derivatives, normalized to that of AIPcS,, is shown in Table 1. Evidently, the activity varies more than 300-fold. In most C ~ S C S activity correlated with dye binding to RBC (Table 1). The cxccptions were AIPcS, and AIN,SB2P, which were - 20- fold less active than expected from their cellular binding, and AIPcS2, which was about 10-fold less active.

Figurcs 3 and 4 show the protection by F- and quercetin, respectively, from Pc-induced photohemolysis. From these data K, values were derived (Table 1). The fluoride effect varied by a factor of l o 3 while protection by quercetin varicd only 2.8-fold, among the Pc derivatives.

To gain further insight on the protection by F - , the effect on the fluorescence emission spectra was measurcd. The re- sults show (Table 2) that addition of F- to Pc derivatives in solution results in a blue shift whose magnitude was greater for AlPc derivatives than for GaPc. No significant blue shift

0.30

0.20

0.10

0.00 0 50 100 150

LIGHT FLUENCE (kJ/m?

Figure 2. Fluence-response curves of Pc derivatives. ffinetics of hemolysis were determined following various light fluences as de- scribed in the Materials and Methods, from which I/tSo was derived, for ZnPcS,, D; AIPcS2, 0; AIPcS,, A; AINSB,P, 0; GaN2SB2P, 0;

AIN2SB2P, A.

occurred when F- was added to solutions containing dye bound on RBC. In the case of AIPcS, and AINSB3P, which had the lowest K, with NaF there was a pronounced reduction in the intensity of the fluorescence.

Ultrastructural studies of RBC photosensitized with AIPcS, were carried out by SEM. Marked morphological deforma- tions were observed. The normal biconcave discoid cells (Fig. 5a) were gradually modified t o swollen spherocytes still pre- serving the biconcave shape after 30 kJ m - 2 (Fig. 5b). Heavi-

Table 1. Relative activity and binding of Pc derivatives to RBC and the protective cffccts of NaF and quercetin

K, $ Relative Relative NaF Quercetin

Pc activity* binding? (mM) (PLM)

AlPcS2 CraPcS2 ZnPcS? AIPcS, GaPcS, ZnPcS, AIPcS, AINSBIP GaNSB3P

GaN2SB2P GaN,SBP

AlNlSBAP

0.35 0.02 2.8 0.0 I 0.01

<0.01 1 .0 1.7 0.02 0.08 0.20 0.03

3.3 0.004 2.2

6.0 0.24

5.5

I .0 0.23 4.8 1.8 0.025 1.9 0.03 10 I .3 3.8 0.19 1.2 5.2

*Reciprocal of the light fluence required to obtain the same rate of phoiohemolysis (0.005 in Fig. 2) for all dyes, normalized to that of AIPcS,.

?The number of dye molecules bound per cell normalized to that of AIPcS,, (7.4 x lo6).

*K, is the concentration required to increase tSo by a factor of 2.

Photohemolysis by phthalocyanines 353

- c .-

E v

0 0 +-

350

300

250

200

150

100

50

0 ‘ 0 250 500 750 1000

NaF ( p M ) Figure 3 . The effect of fluoride on photohemolysis sensitized with AIPcS, + 180 kJ m-*, B; AINSB,P + 24 kJ m >, 0; AIPcS. + 36 kJ ni - ? , 0; GaN2SB2P + 180 kJ m :, A; AIN,SB2P + 360 kJ

m-z. A.

ly damaged cells exposed to 60 W m-* were converted to spherocytes (Fig. 5c).

The quantitation of the morphological deformations of AlPcS,-sensitized RBC is shown in Fig. 6. There is clearly a light flucnce dependency of RBC deformation of the disco- cytes into spherocytes, The proportion of distorted cells re- mained low in all samples. Fluoride added to the cells prior to light exposure reduced the morphological damage in cor- respondcnce with the reduced lytic effect (Fig. 3). It should be nored that samples were processed for SEM immediately after light exposure, about 30 min prior to commencement of hemolysis following the highest light fluence.

DISCUSSION

The interest in Pc derivatives as photodynamic agents for PDT in general and for blood sterilimtion in particular has

Table 2. The elTect of fluoride on fluorescence of Pc derivatives.

A,, *

Relative fluores- cence? P C

AlPcS,, AIPcS, AlPcS AlNzSB2P

GaN,SB,P GaNSB,P ZnPcS,,

--__

AlNSRIP

A,,

607 601 601 651 668 66 1 680 61 1

5mM 5 m M NaCl NaF

aqueous aqueous solution solution

688 681 693 688 682 616 742 726 135 120 744 142 731 734 680 680

5 mM NaCl RBC

688 69 I 682 134 124 144 131 680

5 mM NaF RBC

688 689 68 1 735 720 142 135 680

0.81 0.80 0.30 0.70 0.29 1.01 1.02 0.95

*Values are the maxima of fluorescence emission spectra of the dyes

tnuorescence of RBC suspension in the presence of 1 mM NaF either in solution or bound to RBC.

divided by fluorescence in the absence of fluoride.

,--. c . -

E 51

v

4-J

400 1 300

200

100

/,.. 1 / /

t

0 5 10 15 20

QUERCETIN (,u,M) Figure 4. The effect of quercetin on photohemolysis sensitiied with AIPcS, + 180 kJ m-*, A; ZnPcS, + 24 kJ m I , A; AIPcS, + 36 kJ

rn ,, 0; AINSBIP + 24 kJ m ,, 0.

prompted us to explore the structural features that determine the efficacy of Pc in sensitizing photohemolysis. The results (Table 1) show that chemical structure is critical for the abil- i ty of Pc to induce photodamage in this system. No system- atic trend was observed with the structural changes explored, i.e. the metal ligand, degree of sulfonation and the number of benzene rings added to the Pc macrocycle. It is therefore concluded that a unique combination of these parameters allows the dye to bind effectively to a critical target (protein) in the RBC membrane, leading to efficient photohemolysis. It is well established that the conformation of tetrapyrroles depends on the kind and number of substituents, central metal, axial ligands, charge and aggr~gation.’~ As the con- formation of the chromophore affects its interaction with proteins it is not surprising that a simple correlation with a single determinant was not observed in this study. The total amount of bound dye appears generally to be a good predictor for activity, with the exception of AIPcS, and AIN2SB,P, which bind -20-fold more than predicted from their pho- toactivity, and AlPcS2, which binds 10-fold more than pre- dicted. Presumably, most of the binding in these cases is to nonsensitive targets. Alternatively, most of the bound dye is in an aggregated, photochemically inactive state. Both AIN,SB,P and AIPcS, are amphiphilic dyes that readily pen- etrate cell membranes to redistribute among intracellular components resulting, in the case of mammalian tumor cells, in augmented photodynamic activity.’2.’8 Other possible ex- planations for the differential activity of Pc derivatives in sensitizing photohemolysis are differences in absorption and triplet quantum yield. These differences, however, are about two-fold’J and although they may contribute to a minor cxtent they cannot account for the more than 300-fold vari- ation in activity.

The protective effect of F- against Pc-induced photohemo- lysis was explored in detail because it could be used to en- hance the selectivity ofviral inactivation in blood.4 The basis for this effect is complex formation between F- and the metal

354 E. BEN-HUR et al.

100

80 s

t;

1 w 0 60

I- z 0

W

40 a a

20

0

Figure 5. Scanning electron micrographs of RBC. (a): Control, un- treated cells. (b): Cells incubated with 5 pMAIPcS, and then exposed to 30 kJ m ?. (c): Cclls incubated with 5 pM AIPcS, and exposed

to 60 kJ m ?. Cclls were processed immediately after light exposure.

-

ligand, resulting in modificd binding of Pc to proteins and inhibition of type I photodynamic reactions.’Y The IO’-fold variation in K, for protection against photohcmolysis by F ~

(Tablc I ) indicates that not only the mctal ligand is important for lhis effect (although i t can takc place only with Al and

L

A B C D

TREATMENT E

Figure 6. Quantitation of the morphological changes due to AIPcS, sensitiration. Empty bars are biconcave cells. Full bars, spherocytes. Half full bars are distorted cells. (A): Control, untreated cells. (B): Cells incubated with 5 p M AlPcS, in the dark. (C): Cells incubated with dye plus 30 kJ m 2 . (D): Exposure to dye plus 60 kJ m light. (E): Exposure to dye plus 60 kl m * light in the presence of 5 mM

NaF.

Ga). What other factors are involved remains to be deter- mined. Clearly, complex formation is more pronounced for Al than for Ga, as expected from their positions in the pe- riodic table (see the blue shifts in Table 2). In spite of this AIN,SB,P and GaN,SB,P have K, of 3.8 and 1.2 mM, re- spectively. Thus, substitution of the macrocycle appears to affect complex formation with F-, as reflected in the pho- todynamic activity. In general, Pc derivatives with high ac- tivity in our assay display a large F effect (low K,). An exception is AIPcS,, which displays intermediate activity but has the lowest K, (Table 1). The reason for this is not known but may reflect the amphiphilic nature of the dye, resulting in a different cellular targct. The dramatic reduction in flu- orescence intensity in the presence of fluoride in the case of AIPcS, and AINSB,P, which had the highest fluoride effect, is probably due to dye aggregation. Aggregated dye molecules have reduced fluorescence yield and reduced photochemical activity.

Quercetin, a plant flavonoid, has been shown to protect RBC against photohemolysis induced by hematoporphyrin*O and AIPCS,,~’ by quenching singlet oxygen. As shown here quercetin can also protect against photohemolysis sensitized by other Pc derivatives. Unlike F , Ki of quercetin varies only by a factor of 2.8 with the various Pc derivatives, con- sistent with a different mechanism of action.

The ultrastructural studies show that the effect of Pc-in- duced photodamage to RBC on the morphological level is manifestcd mainly by cellular swelling. Interestingly, the bi- concave shape of the cells was detected even after swelling (Fig. 5c). In contrast, hematoporphyrin photosensitization of RBC caused conversion of the discoid to the echinocytic cell type, together with an increase in cell volume.22 Previous studies have confirmed that hemin (Fe-protoporphyrin) can causc dissociation of the membrane-associated cytoskeletal erythrocyte proteins.23 Any weakness in the mechanical flex- ibility of the membrane cytoskeleton as a result of chemical modifications of the proteins will lead to alterations in the biconcave structure and will increase the sensitivity to he- ~ n o I y s i s . ~ ~ This was demonstrated by the appearance of

Photohemolysis by phthalocyanines 355

echinocytes a n d spherocytes pr ior to hemolysis. It is possible that thc morphological alteration photosensitized by hema- toporphyrin is d u e to those membrane-cytoskeleton changes induced by protoporphyrin, bo th in the dark a n d after light e x p o s ~ r e . ~ ~ , ~ ~ T h e present results indicate that t h e mecha- nism o f morphological alteration and swelling induced by Pc photosensitization m a y be d u e mainly to water influx rather than t o alterations in membrane-associated cytoskeleton.

In parallel studies with the same dyes, their photodynamic activity in viral inactivation varied qui te differently than for RBC hemolysis (results t o be published elsewhere). As a re- sult, some MPc derivatives with a high therapeutic ra t io can be used t o enhance the efficacy of blood decontaminat ion with PDT.

REFERENCES

1. Ben-Hur, E. ( 1 992) Basic photobiology and mechanisms of action of phthalocyanines. In Photodynamic Therapy; Basic Principles and Clinical Applications (Edited by B. W. Henderson and T. J. Dougherty), pp. 63-77. Marcel Dekker, New York.

2. van Lier, J. E. (1990) Phthalocyanines as sensitizers for PDT of cancer. In Photodynamic Therapy of Neoplastic Disease, Vol. 1 (Edited by D. Kessel), pp. 279-29 1. CRC Press, Boca Raton, FL.

3. Horowitz, B., B. Williams, S. Rywkin, A. M. Prince, D. Pascuall, N. Ceacintov and J. Valinsky (1991) Inactivation of viruses in blood with aluminum phthalocyanine derivativcs. Transfu- sion 31, 102-108.

4. Ben-Hur, E.. R. C . Hoeben, H. van Ormondt, T. M. A. R. Dubbelman and J. Van Steveninck (1992) Photodynamic in- activation of retroviruses by phthalocyanines: the effect of sul- fonation, metal ligand and fluoride. J . Photochem. Photobiol. B

5 . Rywkin, S., L. Lenny, J. Goldstein, N. E. Geacintov, H. Mar- golis-Nunno and B. Horowitz (I 992) Importance of type I and type I 1 mechanisms in the photodynamic inactivation of viruses in blood with aluminum phthalocyanine derivatives. Photo- chcm. I’hotobiol. 56, 463-469.

6. Ben-Hur, E. and I. Rosenthal (1986) Photohemolysis of hu- man erythrocytes induced by aluminum phthalocyanine tetra- sulfonate. Cancer Lett. 30, 321-327.

7 . Ben-Hur, E., A. Orenstein, A. Livne and I. Rosenthal (1990) Photosensitized oxidation of human red blood cells: cation ef- fects on volume changes and relevance to blood vessel occlusion. Lusers Life Sci. 3, 245-253.

8. Ben-Hur, E., A. Freud, A. Canfi and A. Livne (1991) The role of glycolysis and univalent ions in phthalocyanine-sensitized photohemolysis of human erythrocytes. Int. J . Radiat. Biol. 59,

9. Roscnthal, I. and E. Ben-Hur (1992) Ascorbate-assisted, phthalocyanine-sensitized photohemolysis of human erythro- cytes. Int. J . Radiat. Biol. 62, 481-486.

10. Sonada, M., C. Murali Krishna and P. Riesz (1987) The role of singlet oxygen in the photohemolysis of red blood cells sen- sitked by phthalocyanine sulfonates. Photochem. Photobiol. 46, 625-63 1.

Bid. 13, 145-152.

797-806.

11. Deuticke, B., U. Henseleit, C. W. M. Haest, K. B. Heller and T. M. A. R. Dubbelman (1989) Enhancement of transbilayer mobility of a membrane lipid probe accompanies formation of membrane leaks during photodynamic treatment of erythro- cytes. Biochim. Biophys. Acta 982, 53-61.

12. Margaron, P., R. Langlois, J . E. van Lier and S. Gaspard (1992) Photodynamic therapy of naphtosulfobenzoporphyrazines, novel asymmetric amphiphiIic derivatives. J. Photochcnz. Phufobiol. B Bid. 14, 187-199.

13. Gaspard, S., P. Margaron, C. Tempete and T. H. Tran Thi (1 990) Mixed acennalated metallotetraazoporphyrins: a new class of amphiphilic photosensitizers for photodynamic therapy of cancer. J. Photochem. Photobiol. B Biol. 4, 4 19-423.

14. Ah, H., R. Langlois, J. R. Wagner, N. Brasseur, B. Paquette and J . E. van Lier (1988) Biological activities of phthalocyanines X. Synthesis and analysis of sulfonated phthalocyanines. Pho- tochem. Photobiol. 47, 7 13-7 17.

15. Gamliel H. (1985) Optimum fixation conditions may allow air drying of soft biological specimens with minimal cell shrink- age and maximum preservation of surface features. Scanning Electron Microsc. 4, 1649-1 664.

16. Valenzeno, D. P. and J. P. Pooler (1982) The concentration and Ruence dependence of delayed photohemolysis. Photochem. Photobiol. 35, 427-429.

17. Senge, M. 0. (1992) The conformational flexibility of tetra- pyrroles-current model studies and photobiological relevance. J . Photochem. Photohiol. B Bid. 16, 3-36.

18. Paquette, B. and J. E. van Lier (1992) Phthalocyanines and related compounds: structure-activity relationships. In Photo- dynamic Therapy: Basic Principles and Clinical Applications (Edited by B. W. Henderson and T. J. Dougherty), pp. 145-1 56. Marcel Dekker, New York.

19. Ben-Hur, E., T. M. A. R. Dubbelman and J. Van Steveninck (1991) The effect of fluoride on binding and photodynamic action of phthalocyanines with proteins. Photochem. Photobiol.

20. Sorata, Y . , U. Takahama and M. Kimura (1984) Protective effect of quercetin and rutin on photosensitized lysis of human erythrocytes in the presence of hematoporphyrin. Biochirn. Bio- phys. Acta 799, 3 13-3 17.

2 I . Ben-Hur, E., I. Rosenthal and Y. Granot(Graziani) (1993) In- hibition of phthalocyanine sensitized photohemolysis of human erythrocytes by quercetin. Photochem. Photohiol. 57, 984-988.

22. Malik, Z. and B. Lev (1991) Erythrocyte photohemolysis and its relation to deformability and cell morphology. A commen- tary. Blood Cells 17, 570-574.

23. Solar, I . , U. Mullereberhard, Y . Shviro and N. Shaklai (1991) Long-term intercalation of residual hemin in erythrocyte mem- branes distorts the cell. Biochim. Biophys. Acta 1062, 5 1-58.

24. Miller, F. N., G. J. Tangelder, D. W. Slaaf and R. S. Reneman ( I 99 I ) Use of microphotohemolysis to distinguish dilferences in erythrocyte treatments. Blood Cclls 17, 555-556.

25. Dubbelman, T. M. A. R., A. F. P. M. de Goeij and J. Van Steveninck (1978) Protoporphyrin-sensitized photodynamic modification of proteins in isolated human red cell membranes. Photochem. Photobiol. 28, 197-204,

26. Dubbelman, T. M. A. R., A. F. P. M. de Goeij, K. Christianse and J. Van Steveninck (1 98 1) Protoporphyrin induced pho- todynamic effects on band 3 protein ofhuman erythrocyte mem- branes. Biochim. Biophys. Acta 649, 3 10-3 16.

54, 703-708.