Spin-lattice relaxation, nuclear Overhauser enhancements, and long range coupling in chlorophylls and metalloporphyrins

摘要

1150 J.C.S. Perkin I Spin-lattice Relaxation, Nuclear Overhauser Enhancements, and Long Range Coupling in Chlorophylls and Metalloporphyrins By Jeremy K. M. Sanders,* John C. Waterton, and lain S. Denniss, University Chemical Laboratory, Lens- field Road, Cambridge CB2 1EW Methyl groups in the title series have proton spin-lattice relaxation times (T,) which depend largely on the distance of the group from the macrocycle and their steric crowding whereas T, values for meso-protons are dependent on substitution pattern. Nuclear Overhauser enhancements and long-range coupling constants can be used to construct connectivities around the macrocycle. Together, these techniques provide unambiguous assignments of resonances and structures. Systematic linewidth variations are due to predictable contributions from un- resolved couplings, short T, and electron-transfer effects.A non-seasonal preparation of methyl chlorophyllides is described. WE have been using n.m.r. line-broadening to determine spin-density distributions in the radical-cations of chlorophylls and meta1loporphyrins.l. We needed, therefore, reliable assignments of the n.m.r. spectra. It was particularly important to distinguish within (i) sets of methyl groups and (ii) sets of meso-protons which are in similar molecular environments and have similar chemical shifts but experience different unpaired spin densities. Conventional methods of chemical-shift argument using series of closely related compounds are insufficiently rigorous for shift differences of less than 0.1 p.p.m., whilst selective deuteriation is possible only rarely.Recent observations on aggregation shifts may have general utility in assignment^.^ We report here that spin-lattice relaxation times (TI) and nuclear Overhauser enhancements (n.0.e.) provide the necessary rigour as a consequence of the rigidity of the porphyrin nucleus. We also describe some small 1 J. K. M. Sanders and J. C. Waterton, J.C.S. Chem. Comm., 1976, 247; J. C. Waterton and J. K. M. Sanders, J. Amer. Chem. Soc., in the press. J. K. M. Sanders, Chem. SOC.Rev.,1977, 6, 467. long-range coupling constants which we discovered during the n.0.e. work and which can be crucial for assign- ments even when not resolved. These couplings, together with the TIeffects lead to predictable and con- sistent linewidth variations between resonances for protons in the same molecule.THEORY N.m.r. absorption is a transition from a lower to an upper energy level and requires electromagnetic radiation of the correct frequency. The radiationless return to equilibrium is spin-lattice relaxation : this is character-ised by a relaxation time T, and requires fluctuating magnetic fields of the same correct freq~ency.~ These fields can arise from many sources but for protons under normal solution conditions they arise mainly from the intramolecular di pole-dipole interaction, i.e. other protons in the same molecule. The stronger the inter- 3 R. J. Abraham, F. Eivazi, H. Pearson, and K.M. Smith, J.C.S. Chem. Comm., 1976, 698; Tetrahedron, 1977, 53,2277. 4 Preliminarv communication : I. S. Denniss. T. K. M. Sanders. and J. C. Wateiton, J.C.S. Chem. Comm., 1976; i059; 1977, 192: L. D. Hall, Chem. SOC.Rev.,1975, 4, 401. 1978 action, which depends on distance and correlation time, then the greater the relaxation rate R, (= l/Tl). Thus, the relaxation rate of a proton s caused by a proton i is given by equation (1) where ri, is the is distance and T~~is the correlation time of the is vector (effectively the tumbling time). R,(i --+S) cc TtaYis2 (1) The total relaxation rate of proton s relaxed by n neighbouring protons is therefore given by (2). n R,(s) = 2 (Ti8 -Yi?) (2)i=l There are two useful limiting cases.(i) For protons which have the same 7 because they are attached to the same rigid framework. Equation (2) then simplifies to (3) and differences in relaxation rate between these protons depend only on interproton distances. (ii) For R,(s) K r,is4 (3) methyl protons, relaxation is dominated by the other protons on the carbon to which they are attached, so all riavalues are the same. Equation (2) then simplifies to (4) so that methyl T, values depend on mobility only. Rl(4 CC (4) Where relaxation of s is dependent only on i, then saturation of the i resonance will change the intensity of the s resonance. This is the nuclear Overhauser effect (n.o.e.), and for the proton-proton case in rapidly tumbling molecules the maximum increase in s is 50 .6* * If s is relaxed by more than one proton, then their relative n.0.e.contributions are roughly proportional to v6and the total enhancement is limited to ca. 50. It is possible, in particular geometries of 3 (or more) spin systems, for nearby protons to show little or no n.o.e.6 Thus the absence of an n.0.e. does not necessarily imply that the nuclei concerned do not relax each other. The presence of an n.0.e. proves that they do. The use of T, values and n.0.e. values to assign spectra is most appropriate where T and r can be predicted in advance, i.e. in molecules where a rigid conformation leads to well-defined inter-proton distances and where the motional freedom of a group can be estimated from inspection of the structure.Porphyrins and chloro- phylls are ideally suited to this approach since their substituents are anchored to a rigid macrocycle and the protons of the substituents experience increasing mobil- ity as their separation from the macrocycle increases. Methyl protons in molecules of this size actually relax slightly non-exponentially in a way which enables details of internal rotation to be elucidated.8 This effect is insignificant if initial relaxation rate after the perturbing pulse is measured. * In large molecules n.0.e.s can become negative.' 6 J., H. Noggle and R. E. Schirmer, ' The Nuclear Overhauser Effect , Academic Press, New York, 1971. 7 D. H. Williams and J. R. Kalman, J. Amev. Chew. Soc., 1977, 99, 2768. The spin-spin (transverse) relaxation time, T,, contributes to the resonance linewidth, W, according to R3= Me, R7 = Phytyl R3=CH0, R7= Phytyl R3= R7= Me R3 = CHO, R7= Me R3=Me, R7=H R3 = CHO, R7=H I OR7 C02Me 0 Phytyl ' (9) (12) Phytyl OH (13) Phytyl O.CONHm equation (5) and is equal to or less than Tl.6 Reson-ances with short TI will, therefore, have linewidths greater than those with long T,. W , l/xT, (5) RESULTS AND DISCUSSION Porphyrins (1)-(4) , chlorophylls-a (5) and -b (6), methyl chlorophyllides-a (7)and -b (euro;9, and bacterio- chlorophyll-a (9) were used in this work.For the isol- ation and purification of large quantities of (5),(6), and (9)the dioxan precipitation method was invaluable as a E.Haslinger and R. M. Lynden-Bell, J. Magnetic Resonance, in the press.K. Iriyama, N. Ogura, and A. Takamiya, J.Biochem (Tokyo),1974, 76, 901. preliminary to chromatography. We have developed a new ' non-seasonal ' preparation of (7) and (8). In aqueous acetone, the chlorophyllase of spinach beet (perpetual spinach) hydrolyses the chlorophyll-phytyl ester group to give the free acids (10) and (11) which can J.C.S. Perkin I equilibrium amount, ca. 15, of (5'). This is the C-10 epimer with the carboxymethyl group in the sterically more hindered position cis to the 7-propionate. Spirt-lattice Relaxation Times.-Table 1 summarises TIvalues for compounds (1)-(4) at 310 K. As expected 3a la Phytyl 2b 10-7 L 1 1 I 10 8 6 4 2 0 6 FIGURE 100 MHz 'H N.m.r.spectrum of chlorophyll-a, (5); 20m~in 2H,acetone 1 be separated chromatographically.* Esterification with methyl (and methylene) proton Tl values increase diazomethane before or after separation gives (7) and gradually with distance from the macrocycle, so that, for (8) in good yield. Previous preparations of these com- example, in (2) the propionate methyls (Tl ca. 1.1 s) are pounds required chlorophyllases from the leaves of readily distinguished from ring methyls (Tl ca. 0.5 s).ls2 deciduous (and, in Britain, rare) plants,1deg; whereas spinach This has been confirmed by deuteriation of the ester beet is available throughout the year. methyl group. 10b 5a P-CH 3a UP 8 2a 'P P2 8 P1 4a : I I I I I I I I I I I 11 10 9 8 7 6 5 4 3 2 1 6 FIGURE 100 MHz lH N.m.r. spectrum of chlorophyll-b, (6); 30m~2 in 2H,acetone Proton n.m.r.spectra of (5), (6), and (9) are shown in The meso-protons of (1) and y-H of (3) and (4) have Figures 1-3, together with the assignments determined Tl = 0.8 s whilst a-H, a-H, and 6-H of (3) and (4) have in this work and the conventional numbering scheme. Tl = 1.0 s.7 Whether the shorter values are due to The spectrum of (5) demonstrates the presence of an enforced proximity of the adjacent protons or other * These acids are useful synthetic intermediates2 t Partial meso-signal overlap in (2) prevents accurate measure- F. C. Pennington, H. H. Strain, W. A. Svec, and J. J. Katz, ment. J. Amer. Chem. Soc., 1964, 86, 1418.reasons is not clear, but they are reproducible. They are also manifested by a consistently greater linewidth TABLE1 T,Values for metalloporphyrins TllSSubstituent H CH, Me H 0.8-1.0 Me 0.5-0.6 CH2Me 0.25-0.35 0.7-0.8 CH2CH,C02Me a0.25-0.35 1.1 aP pO.4-0.6 for the y-H signal as predicted by equation (5)(see section on coupling and linewidths). , , I9 8 7 6 reflect interproton distances : this effect is seen clearly in (9) for a-H which is flanked on one side by a carbonyl group. The longer distance to proton neighbours gives a-H a much longer T,value than P-H or 8-H. Similarly, 10-H relaxes slowly in all the chlorophylls; in (ti'), 10-H is in a sterically even less crowded environment and T, is longer still.* The methine protons on the reduced pyrrole rings have very short T,values which reflect their crowded environment.Table 3 gives T, values for methyl groups which should obey equation (4) and simply reflect mobility. The methyls on the aromatic pyrrole rings all have T, ca 0.6 s, but 8a-H and in (9), 3a-HI have T, ca.0.4 s, t! ,Mea 2b IH20 .5 4 3 2 1 6 FIGURE 100 MHz 'H N.m.r. spectrum of bacteriochlorophyll-a (9); 25m~in 2H,acetone under N, atmosphere. 3 sion of 6 8-9 region using 0.12 Hz/data point Results for chlorophylls (5)-(9) are given in Tables 2-4, the assignments anticipating the n.0.e. results. TABLE2 T,Values in seconds for fixed protons in chlorophylls Proton a (5) 1.0 (6) 0.8 (7) 1.3 (8) 1.5 (9) 1.0 8 a 0.9 1.0 0.6 0.8 1.1 1.3 1.1 1.3 0.7 0.6 10 1.4" 1.2 1.7 1.6 0.8 7 0.7 0.5 8 0.6 0.6 0.5 * a For (5') T, = 2.0 s.a For (7'), T, = 2.0 s. Estimate by null point due to signal overlap. TABLE3 T, Values in seconds for methyl groups in chlorophylls Proton (5) (6) (7) (8) (9)la 0.7 0.6 0.9 0.9 0.5 2b 1.3 3a 0.6 0.9 0.3 4b 0.7 0.6 0.9 0.8 0.7 5a 0.7 0.7 1.0 1.2 0.5 8a 0.4 0.5 0.5 0.5 0.4 10b 1.0 0.8 1.1 1.3 0.8 7-C02Me 1.8 " 1.8 " " 3.47 6 Table 2 presents TIvalues for protons directly attached to the macrocycle. These should follow equation (3)and Inset: Expan-presumably due to greater steric hindrance to internal rotation. This effect is particularly clear in (7)and (8). The 4b methyl protons in (5)-(8) relax at the same rate as the ' aromatic ' ring methyls, probably indicating limited mobility of the ethyl group.In contrast the TABLE4 T, Values in seconds for phytyl protons Signal (5) (6) (9) (12) (13)l-P 0.7 0.7 0.5 2.7 1.4 2-P 1.7 1.3 1.2 3.0 2.8 3a-P 1.3 1.4 1.1 3.0 2.2 CH2-P 1.05 1.4 1.1 1.5 1.6 7a-P, lla-P 1.2" 1.4" MI} 1.8 1.8 15a-P, 16-P 2.0a 1.9" Estimated by null point. ethyl group of (9) is out of the plane of the macrocycle and might be expected to be more mobile: this is re- flected in the observed longer T, (relative to the ring methyls). In all the chlorophylls the 10b-methyl ester group relaxes slowly, as does the 2-acetyl group in (9). The 7-Me groups of (7) and (8) have very character- istically long T, values allowing their assignments to be made easily.* Other T, values which can be measured in (5') are not significantly different from (5). The remaining protons of the macrocycle, are neither fixed nor in methyl groups but they can be interpreted. The diastereotopic methylene protons 7a-H,7b-H and in (9), 4a-H have 0.2 TI bsol; 0.4 s, probably reflecting both large and short 7. 2b-H and 2a-H in (5)-(8) have T, ca. 1 and 1.5 s respectively; the slower relax- ation in the latter case reflects lack of a geminal partner. The partially resolved signals due to the methyl groups of the phytyl chains in chlorophylls relax at markedly non-uniform rates in 2H6acetone (Table 4). The 6 H doublet at 6 0.85 relaxes slo'wly, with TI ca.2 s, and is tentatively assigned to 15a-P and 16-P. The 8 7 4FIGURE Portion of 270 MHz 'H n.1n.r. spectrum of (5) in the presence of sufficient (5)+.to broaden 7-H and 8-H 3 H doublets (7a-P, Ila-P) at 6 0.82 and 0.80 are conformationally more restricted, giving T, ca. 1.3 s. I-P, 2-P, and 3a-P are also conformationally restricted by more than is characteristic for esterification of the phytyl group cf. (13). In (12) and (13), 7a-, lla-, 15a-, and 16-P are equivalent at 100 MHz (6 0.87) and appa- rently relax with the same T1.* We conclude, therefore, that when the chlorophylls are dissolved in 2H6acetone, the phytyl group takes up a preferred conformation. This idea is supported by observation of ca. 0.1 p.p.m.non- equivalence of the 1-P protons (J = 11 Hz) in the chlorophylls but not in the C-10 epimers, (12) or (13) (Figure 4). Metal-free porphyrins and metalloporphyrins in non- polar solvents are aggregated.3 We find (J. K. M. S., unpublished results) that this leads to shorter absolute TI values but within a molecule relative T, values remain unchanged. In summary, relaxation times provide a method for sorting methyl and methylene groups according to distance from the macrocycle, and in favourable cases can help assign meso-protons. Nuclear Overhauser Enhancements.-Irradiation of the * The decay is a good exponential for 2T,. J.C.S. Perkin I methylene resonance of (1)causes an increase of ca. 40 in both the peak height and integral intensity of the meso resonance. This is close to the theoretical maxi- mum effect.The correspondence of integral and peak height increases rules out decoupling effects. Table 5 lists the n.0.e. values observed in (3) when high power irradiation was used to saturate all the methyl resonances simultaneously. The assignments of y-H and 6-H follow directly from the results; those of a-H and P-H are found as follows. Irradiation of the lowfield (6 3.81 in C2H6acetone) 6 H singlet with a field too weak to saturate those at 6 3.66 and 3.68 gives a small n.0.e. (ca. 10) on the 6 10.40 and 10.27 signals and simultane- ously sharpens the vinyl methine signals at 6 8.5 by removing long-range coupling (see next section). The 6 3.81 methyls must therefore be la-H and 3a-H and hence the 6 10.40 proton is a-H.Similar results are obtained in CDC1,-pyridine (Tables 5 and 6) for (3) and TABLE5 Assignments and n.0.e.s for meso-protons of (3) and (4) a-H a-H y-H 6-H 6 (CDC1,-pyridinc) (4)6 (CDC1,-pyridine) (3)6 (zH6acetone) (3) 10.24(i 10.31 (1 10.40 10.15 10.22 10.32 9.96 9.97 10.15 10.07 10.14 10.27 (signal irradiated) -CH,vin yl-CH -CH, 15 SO ( increase) 30 15 SO 30 The a-P distinctions are made by analogy with acetone results for (3). TABLE6 Assignment of methyl groups in (3) la,3a 5a,8a Estero 6 (CDC1,-pyridine) 3.77 3.65, 3.63 3.60 6 (zH6acetone) 3.81 3.68, 3.66 3.65 4 Assigned by long TI. (4). These results agree with the limited assignments made by other methods3 Analogous results are obtained for chlorophyll-a in 2H,acetone.Irradiation of the vinyl-CH signal (6 8.08) gives a 15 increase in the 6 9.34 signal which must, therefore, be a-H. a-H also experiences a 20 n.0.e. from high power irradiation at ca. 6 3.3 which is due to the 3a-H methyl group. Low power irradiation (see next section) pinpoints this resonance at 6 3.26. Similarly 6-H (6 8.51) experience 20y0n.0.e. from la-H (8 3.32), 15 from 8a-H (8 1.75, d, J = 7), whilst (3 (6 9.67) is enhanced 25 by 4a-H (6 3.74, q, J = 7.5). These assignments are indicated in Figure 1. Assignments for (6) (Figure 2) follow closely those for (5). For bacteriochlorophyll-a (9) there are no unambigu- ous starting points since the recognisable signals for the macrocyclic periphery consist entirely of apparent singlets and a pair of doublets due to 3a-H and 8a-H.N.0.e.s at meso protons are observed as follows: 8a-H, la-H --t 6-H; 5a-H P-H; a-H3a-H -(Figure 5). These assignments are the only ones consistent with the structure and are confirmed by long- range coupling constants. Long-range Coufiling and Lirtewidtks.-In the previous section we noted the presence of a small, ca. 0.3 Hz, coupling-,r ,18 I I 3 5a 8a 8 47 1 L -r I n.0.e. 5FIGURE Schematic representation of coupling and n.0.e.s in (9) unresolved five-bond coupling between methyl and vinyl methine protons attached to the same pyrrole ring.* 'ie Increase in 2a tL 80 -60 -40 -20 -0'/a -I 1 I 3.4 3.3 3.2 6 6FIGURE Assignment of 3a-H (6 3.26) and la-H (6 3.32) in (5)by long-range coupling (--O-) and n.0.c.(--n-) This is illustrated for (5) in Figure 6, which plots peak- heights for 2a-H and a-H as a function of irradiation * There is similar six bond coupling to the terminal vinyl portons.t This is not due to long-range coupling between these protons. It is also not due to non-equivalence of the 6-CHz and 7-CHz protons as we observe the same effect in the symmetrical uropor- phyrin-11-octamethyl ester (J. K. M. S., unpublished). frequency. Integration shows that the 2a-H,la-H effect is due solely to coupling, whilst the a-H,3a-H effect is a n.0.e. Irradiation of 2a-H sharpens la-H as expected.In addition, irradiation of the 8-H methine resonance affects 6-H; the linewidth decreases by ca. 0.2 Hz, whilst there is no significant n.0.e. This 4-bond allylic coupling is most apparent in (9): when a large number of data points is used to define the spectrum, a 0.8 Hz coupling to 4-H can be directly observed on P-H (see inset, Figure 3). Smaller couplings of ca. 0.2 Hz are inferred for 3-H -a-H and 8-H 8-H from irradi- ation experiments which simultaneously assign the 3-H,- 3a-H and 8-H,8a-H pairs (Figure 5). An unresolved long-range 7-HI10-H coupling of ca. 0.4 Hz is also inferred in all the chlorophylls but not in the C-10 epimers. These allylic couplings and associated linewidth effects are most useful, and it is perhaps surprising that they have not been reported previously.In all our spectra the linewidth variations between different methyl groups is very striking. This is in part due to long-range coupling (e.g. la-H vs. 3a-H) and in part due to inherent differences imposed by equation (5). Thus in (5), lob-H is always the sharpest singlet methyl; 2b-H in (9) is even sharper as it has a longer Tl value; and the added methyls in (7) and (8) are sharp for the same reason. Similarly, in (3) and (4), y-H and the propionate methy- lene nearer the ring are always broader t because of their short T, value. Solutions of (9) in 2H,acetone often contain low (ca. ~O-*M) steady-state concentrations of the radical-cation (9)+' even when thoroughly degassed.This leads to electron transfer broadening of the hyperfine-coupled resonances and accounts for the exceptional broadness of la-H and 5a-H in Figure 3. Knowledge of the relevant coupling enables ready identification of this source of br0adening.l. Observation of the above described coupling and line- width phenomena requires relatively dilute solutions and therefore Fourier transform techniques. High-field studies at 2200 MHz are unlikely to reveal these effects as computer word length usually limits digitisation to ca. 0.5 Hz/data point. In contrast, Figures 1-3 were obtained at 100 MHz with 0.25 Hz/point and the inset to Figure 3 with 0.12 Hz/point. The four-bond allylic coupling is clearly dependent on very subtle geometrical factors.In (9) the relationsliip 4-H,p-H is apparently the same as 3-H,a-H and 8-H,- 8-H but the coupling constant, 0.8 Hz, is ca. four times larger. An analogous 0.9 Hz coupling has been used in the structure elucidation of sirhydrochlorin monolac- tone.ll Conclusions.-These results taken together provide an overdetermined set of assignments for (5)-(9) without resort to chemical modification or biosynthetic labelling. Assignments are in agreement with those most recently l1 A. R. Battersby, K. Jones, E. McDonald, J. A. Robinson, and H. R. Morris, Tetrahedron Letters, 1977, 2213. published.12 Detailed relaxation behaviour in these compounds remains to be elucidated but TImeasure-ments provide a useful tool in natural product structure elucidation.ll EXPERIMENTAL Proton n.m.r.spectra were obtained at ca. 310 K at 80 MHz (Varian CFT 20), 100 MHz (Varian XL-loo), or 270 MHz (Bruker WH270) in the Fourier transform mode. Five to one hundred free induction decays following 90" pulses were collected and transformed without exponential weighting. Chemical shifts are in 6 units (p.p.m.) from SiMe, and were measured from internal SiMe, or calculated via the solvent deuterium field-frequency lock. For relaxation and intensity measurements intervals between pulses were ca. four times the longest T, measured. Integral intensities were determined from these fully relaxed spectra by cutting and weighing spectral peaks. Spin-lattice relaxation times were determined at 100 MHz using the inversion-recovery ( 180"-t-90") method,5 only the initial exponential recovery being used in the calculation.Nuclear Overhauser enhancements were measured by comparing integral intensities of the s resonance when the i resonance and a blank spectral region were being separ- ately irradiated. Quoted n.0.e.s are the average of several experiments and are f5. Porphyrin spectra were obtained from dilute (10-25 mM) solutions in deuteriochloroform containing sufficient pyri- dine to disaggregate the p~rphyrin,~ in 2H,acetone. or The chlorophyll spectra were all obtained in 2H,acetone. Solutions of (9) were degassed with dry N, to minimise production of (9)+.. Preliminary experiments on the other compounds revealed no dependence of T, on degassing and it was discontinued.Field desorption mass spectra of all the chlorophylls were kindly obtained for us by Dr. H. R. Morris of Imperial College. Each showed only the expected molecular ion. Compounds (l),(3),and (4)were gifts from Dr. E. McDonald of this Department, (12) was obtained from Sigma, (13) was a gift from Professors E. H. Ahrens and A. R. Battersby, and (2) was synthesised as described elsewhere.13 Sucrose Chromatography :General Procedure.-The chloro-phylls were isolated, separated, and purified by chromato- graphy in the dark on powdered sucrose (British Sugar Corporation CP icing sugar containing l* calcium phosphate). The absorbent was dry-packed into 7 cm (1 kg) or 9 cm (2 kg) diameter columns using a steel tamper. The chlorophylls were applied to the dry column dissolved in a suitable solvent and the chromatography was run with water pump suction at the foot of the column.Once the pigments were well separated, the column was sucked dry and the required bands were dug out with a long-handled spoon. The chlorophylls were eluted from sucrose with an appropriate solvent. Chlorophylls-a and -b (5), (6).-Chlorophylls-a and -b were prepared from leaf beet perpetual spinach (Beta vulgaris var. cicla)by an improved procedure based on the methods of Strain and Svec,14 and Iriyama, Ogura, and Takami~a.~Chlorophyll solutions and precipitates were kept in the dark as much as possible. The spinach was grown in the University Botanic Garden.l2 H. Scheer and J. J. Katz in Porphyrins and Metallo- porphyrins ', ed. K. M. Smith, Elsevier, Amsterdam 1975, p. 599. l3 J. K. M. Sanders, C. G. Newton, and J. C. Waterton, J. Mag-netic Resonance, in the press. J.C.S. Perkin I Solvents were laboratory reagent grade and were used as received, except that dioxan and diethyl ether were re- distilled from, and stored over, sodium. Spinach was freshly picked and the leaf midribs were removed, only the soft green parts of growing leaves being used; 200 g (fresh weight) of this preparation were washed first with water and then with methanol. The leaves were then homogenised for 2 min in a Waring blender using 1 000 ml methanol previously cooled to -25 "C, and filtered through a Biichner funnel.The precipitate was washed until it was pink with about 500 ml cold methanol. The supernatant was combined with 4 its own volume of dioxan and stirred magnetically, while ice-cold water was added dropwise. On addition of a little water the red fluorescence of the solution disappeared and on addition of more water a green precipitate was observed. The addition of water was continued until the supernatant liquid was yellow-green in colour. About 200 ml water was required, although this varied considerably between preparations. The precipitate was collected by centrifugation at 0 "C. It was redissolved in 7 : 1 methanol-dioxan (300 ml) and reprecipitated with water (50 ml), again added dropwise with stirring.The second precipitate was collected by centrifugation, redis- solved in ether (100 ml), and washed with ice-cold water (3 x 300 ml). It is important to use cold water for washing chlorophyll solutions in ether as this reduces the formation of stable emulsions. The ether solution was evaporated and stored at 0 "C in the dark. The chlorophylls were separated from each other, from their pheophytins (formed in small amounts during the preparation) and from other lipid pigments by column chromatography. The extract was dissolved in ether (25 ml), diluted with light petroleum (b.p. 30-40 "C; 75 ml) and applied to a sucrose column with a loading ratio of 10 000 : 1. The yellow pigments were eluted with 9 : 1 light petroleum : ether. Chlorophylls were then separated by elution with light petroleum containing 0.2y0,increasing to 0.6 of n-propanol.The chlorophylls were separately eluted from the isolated sucrose with ether which was then concentrated to 100 ml, washed with 3 x 300 ml cold water, and evaporated. Each chlorophyll was then rechromato- graphed in the same way using a loading ratio of 7 500 : 1. The solids obtained were dried in vucuo to yield ca. 120 mg of chlorophyll-a and 40 mg of chlorophyll-b. The u.v./visible spectra Amax. (log,, E) in ether were as follows. Chlovophyll-a. Found: 410(4.95), 430(5.12), 535(3.68), 579( 3.94), 619(4.21) and 664(4.94) lit.,15 410(4.88), 430( 5.07), 534( 3.58), 578( 3.92), 615(4.16), and 662(4.95). Chlorophyll-b. Found: 429(4.73), 453(5.18), 548(3.76), 595(4.01), 64q4.67) lit.,15 430(4.76), 455(5.20), 549(3.81), 595(4.06), and 644(4.75).Methyl chlorophyllides (7), (8). Healthy leaves of per- petual spinach were stripped of large stalks and blended for 1 min with 3 volumes/g leaves of acetone-water (4 : 1). The mixture was degassed on a water pump and set aside in the dark for 3-4 h with occasional swirling. Chlorophyll hydrolysis can be followed by extracting aliquots of clear supernatant liquid with equal volumes of light petroleum : the chlorophyllide acids are soluble in the acetone layer whilst the chlorophylls and yellow pigments dissolve in the upper petroleum layer. When the reaction was completed, 14 H. H. Strain and W. A. Svec in ' The Chlorophylls' ed.L. P. Vernon and G. R. Seely, Academic Press, London, 1966, ch. 2. l5 Ref. 12, p. 880. the mixture was filtered under vacuum through a large Buchner funnel and extracted with petroleum (0.7 vol) to remove yellow pigments. The aqueous acetone was then extracted with ether (0.7 vol).* The ethereal extract was concentrated and treated with an excess of ethereal diazo- methane until methylation was complete. On further concentration of the solution, wet (7) and (8) were precipi- tated; they were collected by centrifugation and dried in vacuo. Chromatography was performed as described for the chlorophylls but using similar conditions to those given by Chow et. aZ.16 Mixed product (100 mg) was dissolved in pyridine (20 ml), diluted to 700 ml with light petroleum, applied to the column (1 kg), and then washed with 0.5 pyridine in petroleum (200 ml).The column was eluted with 0.5 pyridine and 0.5 n-propanol in petroleum and then washed with petroleum when the pigments had separated. The bands were dug out and (7) and (8) eluted separately from the sucrose with dry distilled acetone. These solutions were evaporated and each pigment re-chromatographed separately as above. Finally the solid residues were triturated with dry, distilled light petroleum, filtered, dried in vacuo, and stored at 0 "C in the dark. U.v./visible spectra were in agreement with structures (7) and (8). BacteriochZorup~zyZl-a. Bacteriochlorophyll was obtained * Evaporation of this wet solution yields a precipitate of almost pure mixed chlorophyllides (0.05yield based on net weight of leaves).1157 as a by-product from a corrin extraction 179 from Clzro-matium D. From 500 g of bacteria a semisolid precipitate from aqueous ethanol at pH 4 was obtained. This was re-suspended in water buffered at pH 7, filtered through a Biichner funnel and washed with water. The lipid pre-cipitate was dissolved in methanol-dioxan (7 : 1; 3 1) and centrifuged at 0 "C to remove any insoluble debris. It was then precipitated with water (ca. 800 ml) and collected by centrifugation at 0 "C. The precipitate was resuspended in water and recentrifuged. The yield was ca. 15-20 g. This extract was three times chromatographed on sucrose as described for chlorophylls-a and -b. The first column had a loading ratio of ca. 2 000 : 1 and served simply to remove the most polar and non-polar material. Uv./ visible spectrum (log,, E) in ether 359(4.81), 394(4.58), 540sh(3.52), 572(4.26), 705sh(3.96), 768(4.89) (lit.,I5 359-(4.87), 392(4.62), 530sh(3.44), 577(4.32), 697sh(3.96), and 773(4.96). We thank Dr. I. A. Rubinson for the CJiroitzntiuwz D. extract, Dr. R. M. Lynden-Bell for valuable discussion, the Biophysics Department of Portsmouth Polytechnic for access to the 270 MHz instrument, and the S.1I.C. for financial support. 7/1836 Received, 19th October, 19771 l6 H. C. Chow, K. Serlin, and C. E. Strouse, J. Amer. Chem. SOC.,1975, 97, 7230. V. B. Koppenhagen, F. Wagner, and J. J. Pfiffner, J. Biol. Chem., 1973, 248, 7999.