Effects of 2-formylpyridine monothiosemicarbazonato copper II on red cell components

Effects of 2-Formylpyridine Monothiosemicarbazonato Copper II on Red Cell Components William E. Antholine and Fumito Taketa Departments

ofRudiolog_v und Biochemistry.

Medial

College of Wisconsin, Milwaukee

ABSTRACT 2-Formylpyridine

monothiosemicarbazonato copper II (CuL+)

is readily taken up by red cells and

is initially bound to glutathione and hemoglobin. Glutathione was depleted within 5 hr of incubation, presumably by oxidation mediated by CuL+ and 0, with concomittant generation of toxic oxygen species. Cupric ion was slowly transferred from CuL+ to hemoglobin within about 7 hr, and hemoglobin was oxidized until the major form prevailing after 10 hr was (~20:. Little increase in hemolysis due to addition of CuL+ dissolved in the radical scavenger dimethyl sulfoxide was observed with prolonged incubation. Strong inhibition of red cell hexokinase by CuL+

was

observed when the enzymes in ted cell lysates and hemoglobin-free ted cell lysates were examined. CuL + was also an effective inhibitor of yeast hexokinase. However, the inhibitory effect of CuL + within the red cells was less pronounced. It is suggested that even though intracellular accumulation of CuL + creates an oxidizing environment and is potentially capable of inhibiting thiol enzymes such as hexokinase, protective effects are exerted in the red ceil by the presence of hemoglobin, of radical scavengers, and of high levels of enzymes that detoxify toxic oxygen species.

INTRODUCTION The tridentate ligand, 2-formylpyridine monothiosemicarbazone (HL), was synthesized for use as a potential antitumor agent by Brockman and French and their co-workers [ 11. It was found to be active as an inhibitor of cell proliferation with an effect localized on ribonucleotide reductase [l-3], presumably because of its interaction with active site metal atoms [4]. Later work showed that its iron and copper complexes, the latter in particular, are even more potent as inhibitors of cell proliferation [5, 61. When tested against Ehrlich ascites tumor cells, the copper complex (CuL+ ) quickly accumulated in the cells due to its rapid uptake and slow efflux [7]. CuL+ also exerted toxic effects on Chinese hamster ovary (CHO) cells and it was found to block the G, /S interphase of the Address reprint requests to Dr. W. E. Antholine,

Department of Radiology, or Dr. F. Taketa, Department of Bio-

chemistry, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226.

Journal of Inorganic Biochemishy 20,69-78 (1984) @ Elsevier Science Publishing Co., Inc. 1984 52 Vanderbilt Ave., New York, NY 10017

69 0162-0134/&4/$3.00

W. E. Antholine and F. T&eta

70

cell cycle [8], which is consistent with the notion that ribonucleotide reductase is affected. However, investigation of the sequelae of events occurring with intracellular accumulation of CuL+ reveals that the compound can interact rapidly with intracellular thiols and promote generation of toxic oxygen species, including superoxide anions and hydroxyl radicals, which concomittant oxidation of the thiols [7]. Thus, this mode of action of the compound may also be significant in causing cellular damage. Studies on the effect of CuL+ on nonproliferating cells such as the red blood cells should be useful to obtain insight concerning potential cellular toxicity arising from a generation of toxic oxygen species. The use of red cells as a model to study cellular effects of CuL+ is suggested from observations of hemolysis in blood of animal and cancer patients treated with the metal-free 5-hydroxy derivative of the ligand 19, lo]. Because red cells contain relatively high concentrations of glutathione as well as oxygen, hemolysis could be a direct consequence of membrane damage due to CuL+ catalyzed generation of oxygen radicals. Furthermore, the presence of hemoglobin and the known sensitivity of its hemes to oxidation by cupric ions provides another basis to examine the intraerythrocyte effects of the compound. In previous work with isolated human hemoglobin, we found that CuL+ is bound to a nitrogen donor atom, presumably on one or more imidazole nitrogens of the protein [ 1 I]. The bound complex demonstrated increased oxygen affinity with little effect on the redox state of the heme irons, indicating that CuL’ can affect protein function allosterically as well. The present study focuses on the effects of CuL+ on intact human red cells, particularly with respect to effects on glutathione oxidation, methemoglobin generation, and hemolysis of cells. Its effect on hexokinase and phosphofructokinase was also examined to assess the possibility that other proteins of the cell, particularly .enzymes that contain essential thiol groups, might be affected.

EXPERIMENTAL Electron paramagnetic resonance (epr) analysis was conducted with a Varian E 109 Century Series X-band spectrometer or a S-band spectrometer using a mircowave bridge operating at 3 GHz located at the National Biomedical ESR Center. Spectra were simulated with a program from John Pilbrow, Department of Physics, Monash University, Australia. CuL+ was obtained from D. H. Petering. A 1.3 mM stock solution of CuL+ was prepared by dissolving in reagent grade dimethyl sulfoxide, (DMSO). Fresh heparinized human blood was obtained from local sources. Red cell suspensions were prepared by removing plasma by aspiration after centrifugation, washing the cells with isotonic saline, and resuspending them in isotonic saline or 0.15 M phosphate buffer, pH 7.4. Cells were incubated at 37°C in a shaker bath for the appropriate number of hours. Hemoglobin solutions were prepared either by hypotonic lysis of washed red cells taken from fresh heparinized blood or by a freeze-thaw process using a dry ice-ethanol bath. Stroma were removed by centrifugation at 14000 rpm for 20 min in a Sorvall RC-5 refrigerated centrifuge. Hemoglobin concentration from red cell suspensions was measured by the cyanomethemoglobin procedure [ 121. Hemoglobin concentrations in the supematant were estimated from their absorbance at 540 nm in a Gilford 240 spectrophotometer. Red cell enzyme activities and glutathione concentrations were measured in the presence and absence of CuL+ using spectrophotometric procedures

Effects of CuL + on Red Cd Components

described MO).

by Beutler [13]. Yeast hexokinase

71

was purchased from Sigma (St. Louis,

Uptake of CuL+ by Red Cells Fresh red cell suspensions prepared by adding two volumes of isotonic saline to packed red cells were incubated in a shaker bath with 50 PM CuL+ or CuSO,, for 40 min at room temperature. The cell suspensions were then centrifuged, and samples of the supematant and packed cells were frozen in liquid nitrogen to measure uptake of CuL+ by epr analysis. The epr parameters (gf = 2.13, Ali = 165 gauss) for the predominant signal found in red cells incubated with CuL+ (Fig. 1) are consistent with the formation of a CuL+ S-glutathione adduct [7]. The major epr signal in the supematant fraction suggests the presence of a CuL+-nitrogen adduct, probably due to CuL+ bound to globin nitrogen of hemoglobin, since a small extent (3%) of hemolysis of red cells had occurred during the incubation. An aliquot of the frozen red cell sample previously exposed to CuL+ was allowed to thaw to lyse the cells and it was exposed to air at room temperature to facilitate oxidation of cuprous ions. The solution was then refrozen and analyzed by epr spectroscopy. There was little increase in the signal height for cupric ion indicated by the difference in maximum and minimum signal heights in the perpendicular region and by the relative shift of intensities for the lines in the gll region. This suggests that cupric ion was the predominant species in the red cell sample frozen initially, i.e., either its reduction is slow or its reoxidation following reduction is very fast. Under the conditions used for the incubation of the red cells, the maximum theoretical intracellular concentration of CuL+ would be 0.43 mM if all of the added compound is taken up by the cells.

FIGURE 1. X-band epr spectra at - 196°C for

CuL + in packed red cells. (A) CuL+ in the supematant after 40 min incubation with red cell suspension and centrifugation;,gain increased to 2 x 10’? 10 set time constant; (B- 1) CuL+ in packed red cells after a 40 min incubation, gain 5 x 105, I SCtimeconstant; (B-2) expansion of B- 1;gain 2 x 106.3 set time constant; (C) packed red cells with CuL+ thawed and refrozen after 2 hr, gain 5 X 105, I set time constant; (D) 0.8 mM Cu-Hb, reference signal indicates expected position of gll lines and absence of these lines in above spectra, gain 5 x 105 I SAC time constant.

W. E. Anthohe and

72

F. Taketa

Calculation of the intracellular concentration by rough approximation of peak to peak heights in the epr spectrum compared with peak heights of spectra of standard solutions of known CuL+ concentration indicates that it was about 0.5 mM. Thus, most of the CuL+ appears to have been taken up by the cells and remained in the cupric state of oxidation. In contrast, no Cu*+ epr signal is observed for red cells incubated with CuSO, (data not shown), indicating that either Cu2+ is not taken up or that free Cu*+ ion is rapidly reduced to Cul+ after its uptake in the cell. Since the nitrogen hyperfine structure from the X-band spectrum was poorly resolved for a frozen solution of CuL+ plus glutathione but well resolved after being taken up by Ehrlich ascites cells [7], improved resolution was sought by adding an excess of glutathione to65 CuL+ in 60 mM bis Tris buffer under nitrogen and then quickly freezing the mixture in a 4 mm outside diameter (o.d.) quartz tube. Under these conditions, the X-band and S-band spectra (Fig. 2) were well resolved in the perpendicular region. The m, = - l/2 line for copper in the g,, region and a portion of the gl region is expanded in Figure 3. The well resolved m, = -l/2 in the gll region follows from a decrease in linewidth with a decrease in frequency due to minimizing the expression for the linewidth that contains an m,-dependent term and a frequency-dependent term with opposite signs, as described by Hyde and Froncisz [14, 151. Even though a second FIGURE 2. eprexperimental and simulated spectra fo#CuL +( 1.3mhQ in 0.05 M bis Ttis buffer plus5 mM glutathione and a final pH of 6.9. X-band spectrum; X-band simulated spectrum; microwave frequency 9. I29 GHz. Q= 2.14, gl= 2.035, A,, = 180 x IO-4cnr~,AL= 30 x lO--4 cm-‘, AN = 12 x 10-4cm-I, parallel linewidth 8.0 gauss, perpendicular linewidth 6.0 gauss; S-band spectrum; S-band simulated spectrum: microwave frequency 3.32 GHz.

73

Effects of CuL + on Red Cell Components

FIGURE line in

3. Expansion of m, = - l/2

g ,, region and gjregion

for

X-band experimental (top) and simulated spectra (bottom). Conditions as in Figure 2.

species that is less than 25% of the CuL-S-glutathione adduct is consistently formed and makes the assignment difficult, the simplest interpretation for the spectrum is a 1:2:3:2: 1 pattern representing the splitting from two equivalent nitrogens. Thus two nitrogen and two sulfur atoms probably comprise the in-plane coordination. The computer simulated spectrum (Fig. 2) fits the experimental data reasonably well in the perpendicular and parallel regions. Differences in Figures 2 and 3 between simulated and experimental spectra may be attributed to the exclusion of inequivalent g values in the planar directions, inequivalent nitrogen donor atoms, and/or another signal due to a second CuL+ configuration. Effect of CuL+ on Hemolysis and on Red Cell Glutathione Concentrations

and Methemoglobin

The time course of changes in glutathione (GSH) and methemoglobin (MetHb) concentrations when red cells are incubated with CuL+ under conditions similar to those described in the previous section are shown in Figures 4 and 5. In the absence of CuL+ , little or no change in GSH concentration (5-8 pmol/mg Hb) or methemoglobin occurred for at least 6 hr of incubation, but in the presence of 30 PM CuL+, GSH declined progressively and it was completely depleted in less than 5 hr. Methemoglobin concentration was increased at 3 hr to a level of about 5% in the presence of the reagent. Samples analyzed by epr at 3 hr of incubation indicate that the predominant species was an adduct of CuL+ with nitrogen donor atoms of protein, presumably hemoglobin. At 5 hr of incubation, however, less than 70% of the CuL+ was in the form of the adduct; more than 30% being cupric ion bound to nitrogen donor atoms on the protein. At this point, the methemoglobin concentration had risen to about 10%. At seven hours of incubation, there was little or no CuL+ -hemoglobin adduct; the predominant epr signal being due to cupric ion bound to hemoglobin and the methemoglobin level was increased to about l&25%. The kinetics of methemoglobin formation are shown in Figure 5. The concentration of methemoglobin slowly increased and then plateaued at about 50% in about 10 hr and remained essentially at this value even after 24 hr. Incubation of intact ted cells with CuSO, under the same conditions resulted in less than 10% methemoglobin at 10 hr. No increase in hemolysis due to CuL+ in DMSO or aqueous CuSO, was observed at 10 hr, or even after 24 hr of incubation. Analysis by isoelectric focusing of the hemoglobin products after 10 hr incubation of red cells with CuL+ revealed that the

\

74

w

W. E. Antholine

and F. Taketa

Lysate

RBC suspension

4

Q

Lysate + CUL’

A-’ RBC ;uspensjon \ + cut

\ 0

1

2

-% 3

4

5

6

7

8

9

10

time (HRS)

FIGURE 4. Depletion of GSH in red cell lysates and m intact red blood cell suspensions exposed to CuL’ Initial concentrations in lysate: [Hbj = 0.2 mM, (GSH] = 0.3 mM. [CuL+] = 0.03 mM. Initialconcentrations estimated in the red cells: [Hbj = 5 mM. [GSHI = 2 mM, [f&L+ I = 0.6 mM. GSH assay per Beutler [ 131.

FIGURE 5. Methemoglobin formation in intact red cells after exposure to CuL+ or CU*+. Conditions Were the same as those given in Figure 4 except that the initial concentration of CU*+ was 0.03 mM and Its concentration in red cells was calculated to be 0.60 mM after all the ion is taken up. (4 and (A Methemoglobin formation in the absence ofCuL+

.z -0 $ E z g g

50

-

40

-

30

-

20

-

orCuz+, respectively.

20

25

time (HRS)

75

Effects of CuL + on Red Cell Components

major oxidation product was the half oxidized hemoglobin, (Y&T (not shown), reminiscent of the product that is formed when human hemoglobin is reacted with Cuz+ [ 161. The relatively faster rate of methemoglobin formation with CuL+ compared with CuSO, is apparently due to the higher effective intracellular Cu2+ concentration resulting from the incubation with CuL+ . This is indicated by our observation that when red cells are incubated for 10 hr. with 150 PM CuL+ or CuSO,, GSH is no longer detectable in the former and is reduced but present at about 20 gg/mg of Hb in the latter mixture. At equivalent concentrations, incubation of CuL+ and Cu2+ with red cell lysates resulted in oxidation and depletion of GSH within a few minutes. When dilute hemoglobin solutions (21 PM) “stripped” free of small molecular weight components found in the lysate were reacted with CuL+ or CuSO, (41 PM), little difference in the rates of oxidation of hemoglobin was now observed. Hemoglobin was oxidized to a plateau of about 50% methemoglobin with second-order kinetics (2 x 10m5M-l set-I). Thus, it is concluded that the CuL+ induced oxidation of hemoglobin results from the rapid uptake of the compound by red cells, intracellular binding to glutathione and hemoglobin, followed by redox reactions with thiols leading to a generation of Cu+l . Then, Cu+’ is reoxidized to Cu2+ in the presence of oxygen and Cu *+ is competitively bound to either HL or to the tight or weak binding sites on hemoglobin. Binding to the weak site results in oxidation of the heme as described by Rifkind [ 17, 181.

Effect of CuL+ on Red Cell Enzymes Since CuL+ rapidly accumulates in the cell and interacts with hemoglobin, the effect of this intracellular accumulation on hexokinase (HK) and phosphofructokinase (PFK), thiol containing regulatory enzymes of the glycolytic pathway, was examined. The activity of HK and PFK slowly diminished over a 10 hr period to about 50% of the control value (Fig. 6). When CuL+ was added directly to red cell lysates in which the concentration of CuL+ was increased about twentyfold with respect to hemoglobin and enzymes, a much more rapid rate of inhibition of these enzymes was observed. Similar results were observed when yeast hexokinase was examined. The inhibition is not associated with depletion of substrates in the assay from interaction with CuL+ , since

FIGURE

6. Inhibition of red cell enzyme activity by CuL+. Assays for HK (x) and PFK (a) according to

Beutler [ 131.lnitial conditionsas given in Figure 4. 100 90

80 >

70

c

2 601 5

50

4

40 i

8

30 20 IO OJ 0

2

6

4 TIME

(Hrs

)

8

IO

76

W. E. Antholiie and F. Taketa

>

60

!

50

I

h

401

< c ::

30

8

20

FIGURE 7. Effects of various concentra\

tions of CuL+ \

on the activity of hemo-

globin-free red cell hexokinase (about 0.3

‘\~\

units). Actiwty

‘\

is expressed in terms of

the initial rate of reaction after IO min of incubation of CuL’

IO

with the enzyme.

u..__

j

\ 0

i.~_. 0

--20

40

60

CONCENTRATION

80

100 CUL’

120

140

rJJM )

addition of these components after the inhibition is expressed had no effect on reversing the inhibition. Thus, the agent appears to have a direct effect on the enzyme. The enzyme is protected to some degree by the presence of substrates, since preincubation with substrates diminished the inhibitory effect of CuLf However, even in the presence of a large excess of substrates, complete protection was not observed. When GSH (1 mM) was added before CuL” , hexokinase was completely protected, but addition of GSH after the inhibition by CuL+ had occurred did not restore activity. When red cell hexokinase isolated free of hemoglobin by DE-52 chromatography was assayed, it was found to be more sensitive to inhibition by CuL+ than the same amount of enzyme in the lysate (Fig. 7). The apparent difference in sensitivity of partially purified hexokinase and the enzyme within the lysate was probably due to protection afforded by binding of CuL+ by hemoglobin.

DISCUSSION The CuL-glutathione adduct formed initially in red cells reacts with available gluthationes, GSH, and depletes GSH within about 5 hr under the conditions described herein. A series of reactions, as previously described for Ehrlich ascites cells 171, probably occurs in red cells whereby CuLSG mediates uptake of oxygen and GSH is consumed as follows:

The addition of the sulfur atom from GSH changes the in-plane donor atoms of CuL+ from N-N-S-O in DMSO and H,O or N-N-S-N in human red cells to N-N-S-S. The hyperfine spliting of the m, = - l/2 line in the gll region consists of an odd number of lines for CuGSG. In contrast, an even number of lines attributed to approximately equivalent nitrogen donor atoms plus a proton splitting is clearly resolved for CuL+ adducts for which nitrogen and oxygen atoms are donated to complete the square planar configuration [ 1 I]. The absence of a resolvable proton splitting may indicate that the electron density is polarized toward the thiol ligands. it was previously shown that a steady state concentration of CuL’ remains intact for at

Effects of CuL + on Red Cell Components

77

least 6 hr in Ehrlich ascites cells [7], suggesting high stability for the CuL+ complex. The present work indicates, however, that CuL+ is less stable in red cells. The epr data coupled with the cupric ion catalyzed oxidation of hemoglobin indicate that cupric ion is apparently released from the tridentate ligand more rapidly in these cells. Although CuL+ and/or released Cu2+ ion are effective inhibitors for red cell enzymes in the absence of hemoglobin, the large number of binding sites for CuL+ and Cu*+ on hemoglobin probably protects red cell enzymes. The following reactions between CuL+ and hemoglobin in red cells are proposed on the basis of the electron spin resonance (esr) data and the observation on the slow selective oxidation of the hemes on hemoglobin in red cells: SlOW

CuL-Hb and/or CuL-SC + Hb

2

Cu (site I)-Hb t HL

Cu (site 1)-Hb 2 Cu (site 2)-Hb Cu2+ (site 2)-Hb + Hb (~a&+)-Cu+r The third reaction is considered irreversible after the cell has been transformed from a reducing environment to an oxidizing environment because of depleted GSH. Thus, while immediate effects are due to interaction with CuL+ , long term effects are due to interaction of Cu*+ bound to hemoglobin. Even though Cu2+ is tightly bound to 2-formylpyridine monothioseicarbazone (pH independent log KCuL+ = 5 17) and to hemoglobin at the amino terminal region of its constituent P-chains [ 19, 201, enough Cu*+ becomes slowly available to bind at a weaker site on hemoglobin [ 161 to generate the half oxidized hemoglobin after about 10 hr. Our previous report [ 1 I] that short term. interaction between CuL+ and hemoglobin does not result in heme oxidation must be qualified when considering longer periods of reactions. Although oxygen radicals, including 0; and -OH, arc potentially generated each time Cu+’ is reoxidized [7], i.e., about 1 mM of radical from 2 mM GSH and 10 mM radical from 5 n&I Hb in the red cell, the rate of generation appears to be slow enough such that the cellular enzymes (supcroxide dismutase (SOD) and GSH peroxidase) and dimethylsulfoxide, the solvent for CuL+ , prevent lysis for at least 2 days. This work indicates that the presence of large amounts of hemoglobin, radical scavengers, and of enzymes that detoxify oxygen radicals largely protect the red cell from the potentially severe toxic effects of CuL+ . In addition, free cupric ion and copper amino acid complexes are known to scavenge 0; [24, 251 and cupric ion may also scavenge 0; in red cells. In other types of cells, such as tumor cells that are known to have lower levels of SOD than normal cells [26], the effect of creating an oxidizing environment with CuL+ , where oxygen radicals are generated and thiols depleted, may be quite different. The authors thank Kevin R. Siebenlist. Sandra M. Obrien. and Gregory M. Johnson for technical assistance. This work was supported by National Instihrte of Health Grants AM-15770 and RR-01008.

RJWERENCES 1. K. C. Agrawal and A. C. Stiorelli, in Antineoplastic and Immunosuppressive and D. G. Johns, Eds., Springer-Verlag, New York, 1975.

Agent II, A. C. Sartorelh

78

W. E. Antholine

and F. Tnketa

2. A. C. Sartorelli, K. C. Agrawal, and E. C. Moore, Bwchem. Phurmacol. 20,3 I 19 ( 1971). 3. F. A. French, E. J. Blanz, Jr., J. R. DoAmaraJ, and D. A. French, J. Med. Chm. 13, I I 17 (1970). 4. L. Thelander and P. Reichard, Annu. Rnr Bwchem. 48, 133 (1979). 5. W. E. Antholine. J. Knight, and D. H. Petering, J. Mrd C‘hrm. 19, 339 (1976). 6. L. A. Saryan, E. Ankel, C. Krtshnamurti, D. H. Petering, and H. Elford,J. Med. C‘henz. 22. 1218 (1979). 7. L. A. Saryan, K. Mailer, C. Krishnamurti. W. Antholine, and D. H. Petering, Brochen~. PharmLrcol. 30. 1595(1981). 8. W. E. Antholine, P. Gunn. and L. E. Hopwcod, Inr. J. Radrar. Orrcoi. Htol. PIzu. 7.491 ( 1981). 9. W. A. Creasey, K. C. Agrawal, R. L. Capizzi. K. K. Stinson. and A. C. Sartorelli. (.~rt<,rzrRr3. 32, 565 ( 1972). IO. R. C. DeContt, B. R. Toftnesa. K. C. Agrawal, R. Tomchick, 1. A. K. Mead, I. K. Bettino, A. C. Sartorelh, and W. A. Creascy, t‘uncer Res. 32. 14.55 ( 1972). I I. W. Antholine and F. Taketa. J. lmq. Brorhm. 16. 145 (1982). I?_.A. Hainline, in .S/andurri Merhod.~ of C/rm.u/ C‘hermrn. vol. 2, D. Seligson, Ed, Academic Press, New York, 195X. 13. E. BeutJer, in Red Cell .Wer&~/rsm: A Mmuo/ oj B~o~~hemw/ Merhod.~. 2nd ed.. Grune & Stratton, New York, 1975, pp. 38-40. 14. W. Fronciszand J. S. Hyde. J. Chem. P1ry.t. 73,3123 (1980). 15. J. S. Hyde and W. Froncisz, Alznu. Rnv Bwphw. Biorq. 11, 391 (1982). 16. F. Taketaand W. E. Antholine, J. /nor,q. Bwchrm. 17. I09 (1982). 17. J. M. Rifkind. Brochrmisr~ 18 (19). 3860 (1979). 18. J. M. Ritkind, in Mew/ fom m Biological S~srems, Vol. 12, H. Sigel, Ed, Marcel Dekker, New York, 1981. 19. W. E. Antholine, J. M. Knight, and D. H. Petering, Inorg. Chrm. 6,569 (1977). 20. J. M. Rifkind, L. D. Lauer, S. C. Chiang, and N. C. Li, Biochrmisrq 15 (24), 5337 ( 1976). 21. R. E. Viola, J. F. Morrison, and W. W. Cleland, Bwchernrstr~ 19 (14) (1980). 22. K. E. Neet, T. C. Fuirman, and W. J. Hueston, Arch. Biorhem. Bwphw 213. I4 (1982). 23. B. A. Peters and K. E. Neet, J. Brol. C‘hrm. 253, 6826 (1978). 24. K. E. Joester, G. Jung, 0. Weber, and U. Weser, t’EB.5 Lerr. 25.25 (1972). 25. D. Klug-Roth and J. Rabani, J. P/ILS. Chem. X0, 588 (1976). 26. L. W. Oberly and G. R. Buettner, Ccrnc er Kc>. 39, I I4 I (1979).