Physicochemical Studies of Demetalation of Light-harvestingBacteriochlorophyll Isomers Purified from Green SulfurPhotosynthetic Bacteria

Yuki Hirai1, Hitoshi Tamiaki2, Shigenori Kashimura1 and Yoshitaka Saga*1

1Department of Chemistry, Faculty of Science and Engineering, Kinki University, Higashi-Osaka,Osaka, Japan

2Department of Bioscience and Biotechnology, Faculty of Science and Engineering, Ritsumeikan University,Kusatsu, Shiga, Japan

Received 7 March 2009, accepted 8 April 2009, DOI: 10.1111/j.1751-1097.2009.00580.x

ABSTRACT

Demetalation kinetics of bacteriochlorophylls (BChls) c, d and e

from green sulfur photosynthetic bacteria were studied under

weakly acidic conditions. Demetalation rate constants of BChl e

possessing a formyl group at the 7-position were significantly

smaller than those of BChls c and d, which had a methyl group at

this position. The activation energy of demetalation of 31R-8,12-

diethyl([E,E])-BChl e was 1.5-times larger than that of 31R-

[E,E]-BChl c. 15N-labeled 31R-[E,E]-BChls c and e were purified

from cells of green sulfur bacteria grown in a medium containing15NH4Cl, and their 15N NMR spectra were measured. The

chemical shifts of N21, N22 and N23 atoms of 31R-[E,E]-BChl e

were lower-field shifted than those of 31R-[E,E]-BChl c, respec-

tively, and especially the difference in chemical shifts of N22 was

significantly large. These results suggest that the electron-

withdrawing formyl group at the 7-position of BChl e affected an

electronic state of the chlorin macrocycle and caused BChl e to

be more tolerant for removal of the central magnesium compared

with BChls c and d.

INTRODUCTION

The primary step of photosynthesis is the capture of sunlight

by light-harvesting complexes that contain chlorophylls (Chls),bacteriochlorophylls (BChls), phycobilins, or carotenoids. Theharvested sunlight energy is transferred to reaction centers

where charge separation occurs, and the generated membranepotential is used in the subsequent biosynthesis of high-energychemical compounds.

Among the variety of photosynthetic light-harvesting com-plexes, green photosynthetic bacteria have unique antennacomplexes called chlorosomes (1,2). BChls c, d and e are themajor light-harvesting pigments of chlorosomes in green sulfur

photosynthetic bacteria. These BChls form self-aggregates byonly interaction among pigment molecules in chlorosomes,and little protein would be needed to form the self-assembly

(1–7). This is in sharp contrast to supramolecular structures of

other photosynthetic light-harvesting complexes, where pro-

teins play important roles in their supramolecular structures.Molecular structures of BChls c, d and e are shown in

Fig. 1. They have different substituents directly linked to

chlorin macrocycles at the 7- and 20-position. BChls c and dpossess a methyl group at the 7-position, whereas BChl epossesses a formyl group at this position. BChls c and e havea methyl group at the 20-position, whereas BChl d is

unsubstituted at this position. In addition, these BChls area mixture of different molecular forms that vary the degree ofmethylation on the 8- and 12-alkyl groups (homologs) and the

stereochemistry at the 31-position (epimers). Such variation ofchlorosomal BChl molecular structures affects the physico-chemical properties of both their monomeric and aggregated

forms (1,2,5).Natural Chls and BChls have a magnesium atom in

the tetrapyrrole macrocycle, except Zn-BChl a in an aerobic

photosynthetic bacterium, Acidiphilium rubrum (8). Demetalationof the central magnesium from (B)Chl molecules is called

Figure 1. Molecular structures of bacteriochlorophylls (BChls) c, dand e in green sulfur photosynthetic bacteria. BChl c: R7 = CH3,R20 = CH3. BChl d: R7 = CH3, R20 = H. BChl e: R7 = CHO,R20 = CH3. R8 = C2H5 (E), n-C3H7 (P), iso-C4H9 (I). R12 = CH3

(M), C2H5 (E).*Corresponding author email: saga@chem.kindai.ac.jp (Yoshitaka Saga)� 2009TheAuthors. JournalCompilation.TheAmericanSociety of Photobiology 0031-8655/09

Photochemistry and Photobiology, 2009, 85: 1140–1146

1140

pheophytinization, which is one of the important reactions inthe early processes of (B)Chl degradation (9–13). In thedegradation pathway of Chl a, pheophytinization occurs afterremoval of a long esterifying alcohol at the 17-propionate by

chlorophyllase (9–12,14,15). Hydrolysis of the 17-propionateester by chlorophyllase has been extensively studied and is awell-known step in the degradation pathway of Chl a. In

contrast, the in vivo demetalation mechanism of natural(B)Chls has not been thoroughly unraveled. Shioi et al.reported that a small molecular weight substance had an

activity of magnesium removal from Chls and this substancedid not lose the activity even boiled (16,17). This implies that invivo pheophytinization would not be catalyzed by enzymes that

are high molecular weight proteins, although other worksuggests the existence of an enzyme called Mg-dechelatasecapable of Chl demetalation (18). Therefore, the physicochem-ical properties of (B)Chls with regard to demetalation would

be useful to understand (B)Chl degradation in nature.Some reports are available on demetalation properties of

Chls in higher plants and Zn-BChl in a purple photosynthetic

bacterium (19–21). In the case of light-harvesting BChls ingreen sulfur photosynthetic bacteria, we preliminarily reportedthe kinetic analysis of demetalation of BChls c and e under

weakly acidic conditions, where demetalation of BChl e wasslower than that of BChl c (22). However, no other systematicstudies on the effects of molecular structures on demetalationproperties of chlorosomal BChls are available. In this study,

we perform quantitative kinetic analysis of the demetalation ofpurified BChl c, d and e isomers under weakly acidicconditions and reveal the effects of peripheral substituents at

the 7-position as well as 31-, 82-, and 20-positions of chloros-omal BChls. Additionally, the effects of 7-formyl group ofBChl e on demetalation reaction are discussed based on the

measurements of temperature dependence of demetalation rateconstants and 15N NMR of purified BChl c and e isomers.

MATERIALS AND METHODS

Apparatus. Visible absorption spectra were measured using a Shima-dzu UV-2450 spectrophotometer, where the reaction temperatureswere regulated with a Shimadzu thermo-electric temperature-controlled cell holder TCC-240A. Reverse-phase HPLC was carriedout using a Shimadzu LC-20AT pump and an SPD-M20A photodiodearray detector. 1H and 15N NMR spectra were measured using a JOELJNM-ECA500 NMR spectrometer; chemical shifts were expressed (inppm) relative to chloroform (7.26 ppm) as an internal reference andHCO15NH2 (112.50 ppm) as an external reference, respectively. LC-MS was performed using a Shimadzu LCMS-2010EV system equippedwith an atmospheric pressure chemical ionization (APCI) probe (23).

Materials. Chlorobium (Chl.) tepidum ATCC49652, Chl. vibrioformeNCIB 8327 and Chl. phaeobacteroides 1549 were grown according toprevious reports (24–27), to afford natural BChl c, d and e isomers,respectively. To obtain 15N isotope-enriched samples, Chl. tepidumATCC49652 and Chl. phaeobacteroides 1549 were grown in a modifiedmedium of Wahlund et al., where CH3COONH4 and NH4Cl werereplaced by CH3COONa and 15NH4Cl (15N >98%, CambridgeIsotope Laboratories, Inc.), respectively (28).

Pigments were extracted from the harvested cells with an ace-tone ⁄methanol (1 ⁄ 1, vol ⁄ vol) mixture and the pigment solutions werefiltered. The solutions were diluted with diethyl ether and washed withNaCl-saturated water (neutral pH) to remove water-soluble compo-nents. The solutions were evaporated under reduced pressure and wererecrystallized from hexane. The BChl precipitates were obtained byfiltration and washed with hexane. The extracted light-harvestingBChls were dried in vacuo. The major BChl c, d and e isomerswere assigned by LC-APCI MS on a reverse-phase HPLC column

5C18-AR-II (6 mm/ · 250 mm, Nacalai Tesque) with metha-nol ⁄water (95 ⁄ 5, vol ⁄ vol) at a flow rate of 0.7 mL min)1. Theseisomers were purified on a reverse-phase HPLC column 5C18-AR-II(10 mm/ · 250 mm, Nacalai Tesque) with methanol ⁄water (95 ⁄ 5,vol ⁄ vol) at a flow rate of 1.0 mL min)1. The eluted solutionscontaining BChls were evaporated immediately to prevent BChldegradation and were stored in the dark at )20�C. Just before thepreparation of demetalation reactions, all samples were repurified byreverse-phase HPLC under the same conditions described above.

MS data of purified major epimers of BChls c, d and e are describedas follows: 31R-[E,M]-BChl c: MS (APCI) found: m ⁄ z 793.4, calcd. forC49H61N4O4Mg; MH+, 793.5. 31R-[E,E]-BChl c: MS (APCI) found:m ⁄ z 807.4, calcd. for C50H63N4O4Mg; MH+, 807.5. 31R-[P,E]-BChl c:MS (APCI) found: m ⁄ z 821.4, calcd. for C51H65N4O4Mg; MH+,821.5. 31S-[P,E]-BChl c: MS (APCI) found: m ⁄ z 821.6, calcd. forC51H65N4O4Mg; MH+, 821.5. 31S-[I,E]-BChl c: MS (APCI) found:m ⁄ z 835.4, calcd. for C52H67N4O4Mg; MH+, 835.5. 31R-[E,M]-BChl d:MS (APCI) found: m ⁄ z 779.6, calcd. for C48H59N4O4Mg; MH+, 779.4.31R-[E,E]-BChl d: MS (APCI) found: m ⁄ z 793.6, calcd. forC49H61N4O4Mg; MH+, 793.5. 31R-[P,M]-BChl d: MS (APCI) found:m ⁄ z 793.6, calcd. for C49H61N4O4Mg; MH+, 793.5. 31R-[P,E]-BChl d:MS (APCI) found: m ⁄ z 807.6, calcd. for C50H63N4O4Mg; MH+,807.5. 31S-[P,E] ⁄ 31S-[I,M]-BChl d: MS (APCI) found: m ⁄ z 807.6, calcd.for C50H63N4O4Mg; MH+, 807.5. 31S-[I,E]-BChl d: MS (APCI) found:m ⁄ z 821.6, calcd. for C51H65N4O4Mg; MH+, 821.5. 31R-[E,E]-BChl e:MS (APCI) found: m ⁄ z 821.6, calcd. for C50H61N4O5Mg; MH+,821.4. 31R ⁄S-[P,E]-BChl e: MS (APCI) found: m ⁄ z 835.6, calcd. forC51H63N4O5Mg; MH+, 835.5. 31S-[I,E]-BChl e: MS (APCI) found:m ⁄ z 849.6, calcd. for C52H65N4O5Mg; MH+, 849.5. 15N-labeled 31R-[E,E]-BChl c: MS (APCI) found: m ⁄ z 811.4, calcd. forC50H63

15N4O4Mg; MH+, 811.5. 15N-labeled 31R-[E,E]-BChl e: MS(APCI) found: m ⁄ z 825.1, calcd. for C50H61

15N4O5Mg; MH+, 825.4.Measurements of demetalation kinetics. Soret peak absorbances of

purified BChl c, d and e isomers were adjusted to 1.0 in 3 mL ofacetone solution and the solution was mixed with 1 mL of distilledwater according to previous works (19–21). Final BChl concentrationswere 7.5 · 10)6

MM. In kinetic analysis of demetalation processes, 10 lLof 0.1 MM aqueous hydrochloric acid was added to the BChl solutionsand the absorbance at Soret peak positions was measured by control ofreaction temperatures between 5 and 35�C.

NMR measurements. 1H- and 15N NMR spectra of 31R-[E,E]-BChlc and e were measured in a mixed solvent of methanol-d4 ⁄ chloroform-d(1 ⁄ 9, vol ⁄ vol). 31R-[E,E]-BChl c and e concentrations were 10 and7 mMM, respectively, in NMR measurements. 15N signals of 31R-[E,E]-BChls c and e were assigned by heteronuclear multiple-bond correla-tion (HMBC) spectra.

RESULTS

Pigment preparation

Light-harvesting BChl c, d and e isomers were purified frompigment extracts of three green sulfur photosynthetic bacteria,

Chl. tepidum, Chl. vibrioforme, and Chl. phaeobacteroides, andwere assigned by LC-MS analysis and comparison with HPLCelution profiles in previous reports (24–27,29,30). Figure 2

shows elution patterns of BChls c, d and e extracted from thethree green sulfur photosynthetic bacteria. The BChl cfractions c1-c5 in Fig. 2a gave molecular ion peaks at

m ⁄ z = 793.4, 807.4, 821.4, 821.6 and 835.4, respectively, inLC-APCI MS spectra. Compared with the reported data(24,29), these fractions c1-c5 were assigned as 31R-[E,M]-, 31

R-[E,E]-, 31R-[P,E]-, 31S-[P,E]- and 31S-[I,E]-BChl c, respec-

tively. The BChl d fractions d1-d6 in Fig. 2b showed molecularion peaks at 779.6, 793.6, 793.6, 807.6, 807.6 and 821.6,respectively. The LC-MS analysis and the reported characterization

(25,26,30) indicated that d1-d6 were assigned as 31R-[E,M]-,31R-[E,E]-, 31R-[P,M]-, 31R-[P,E]-, 31S-[P,E] ⁄ 31S-[I,M]- and31S-[I,E]-BChl d, respectively. Fractions d1-d4 and d6

Photochemistry and Photobiology, 2009, 85 1141

consisted of single homologs, whereas fraction d5 was amixture of 31S-[P,E]- and 31S-[I,M]-BChl d. In Fig. 2c, the

BChl e fractions e1–e3 showed molecular ion peaks at 821.6,835.6 and 849.6, respectively. Fractions e1 and e3 consisted ofsingle epimers, whereas fraction e2 was a mixture of epimers

(data not shown). The LC-MS analysis and the reportedcharacterization (27) indicated that e1–e3 were assigned as31R-[E,E]-, 31R ⁄S-[P,E]- and 31S-[I,E]-BChl e, respectively.

15N-labeled BChls c and e were prepared from Chl. tepidumand Chl. phaeobacteroides grown in a medium reported byWang et al. (28). The purified 31R-[E,E]-BChl c and e isomersexhibited molecular ion peaks at 811.4 and 825.1, respectively,

by LC-MS analysis. Few molecular ion peaks derived frompartially 15N-labeled or nonlabeled BChls were observed.These revealed that 15N atoms were completely introduced to

N21–N24 positions of BChls c and e. Additionally, in 1H NMRspectra, clear triplet signals were observed on 5- and 10-mesoprotons of the prepared 31R-[E,E]-BChl c and e samples. Such15N-1H couplings were already reported in 15N-labeled BChl c,pheophytin a, and octaethylporphyrins (28,31,32). Hence, 1HNMR analysis also indicated successful incorporation of15N-atoms into both BChls.

Demetalation kinetics

Figure 3 shows spectral changes of homologically and epimer-

ically pure 31R-[E,E]-BChls c, d and e during demetalationprocesses in acetone ⁄water (3 ⁄ 1, vol ⁄ vol) at the proton concen-trations of 1.25 · 10)4, 1.25 · 10)4 and 2.5 · 10)4

MM, respec-tively, at 25�C. 31R-[E,E]-BChl c exhibited intense Soret and Qy

bands at 435 and 666 nm, respectively. These absorption bandsdecreased by incubation under the acidic conditions, and a newSoret absorption band appeared at 412 nm, which was charac-

teristic of bacteriopheophytin (BPhe) c. The isosbestic pointscould be observed at 423, 461 and 575 nm in the spectral change.

By incubating 31R-[E,E]-BChl d and e under the acidic

conditions, essentially the same spectral changes were observed.Soret and Qy absorption bands of 31R-[E,E]-BChl d at 428 and654 nm, respectively, decreased with appearance of a new Soret

band of (31R)-[E,E]-BPhe d at 408 nm. The isosbestic pointswere present at 416, 450 and 556 nm. Soret and Qy bands at 469and 653 nmdecreased in the case of 31R-[E,E]-BChl e, and a newSoret band of 31R-[E,E]-BPhe e appeared at 443 nm with the

isosbestic points at 455, 525 and 558 nm. Demetalation kineticsof BChls c, d and e can be quantitatively analyzed by absorbancechanges at the BChl Soret peak positions.

Time-dependency of Soret peak absorbances of 31R-[E,E]-BChls c, d and e incubated in acetone ⁄water (3 ⁄ 1, vol ⁄ vol) atthe same proton concentration of 2.5 · 10)4

MM at 25�C are

depicted in Fig. 4. The logarithm of Soret absorbancesthrough the demetalation processes of all the BChls showeda linear time-dependence. Therefore, BChl demetalation reac-tions can be regarded as pseudo first-order reactions under the

(a)

(b)

(c)

Figure 2. HPLC elution patterns of BChl c (a), BChl d (b) and BChl e(c) isomers. The pigments were eluted on a COSMOSIL 5C18-AR-II(6 mm/ · 250 mm) with methanol ⁄water (92 ⁄ 8, vol ⁄ vol) at the flowrate of 1.0 mL min)1. Chromatograms of BChl c, d and e isomers wererecorded at 435, 428 and 475 nm, respectively. The BChl c fractionsc1–c5 were assigned as 31R-[E,M]-, 31R-[E,E]-, 31R-[P,E]-, 31S-[P,E]-and 31S-[I,E]-BChl c, respectively. The BChl d fractions d1-d6 wereassigned as 31R-[E,M]-, 31R-[E,E]-, 31R-[P,M]-, 31R-[P,E]-, 31S-[P,E] ⁄31S-[I,M]- and 31S-[I,E]-BChl d, respectively. The BChl e fractions e1–e3 were assigned as 31R-[E,E]-, 31R ⁄S-[P,E]- and 31S-[I,E]-BChl e,respectively.

Figure 3. Spectral changes of 31R-[E,E]-BChl c (a), 31R-[E,E]-BChl d(b) and 31R-[E,E]-BChl e (c) in acetone ⁄water (3 ⁄ 1, vol ⁄ vol) at theproton concentration of 1.25 · 10)4, 1.25 · 10)4 and 2.5 · 10)4

MM,respectively, at 25�C. Spectra from 0 to 300 min at 10-min intervals.Arrows show the direction of absorbance changes.

1142 Yuki Hirai et al.

present conditions, since the proton concentration(2.5 · 10)4

MM) was much higher than the BChl concentration

(7.5 · 10)6MM) (19–22). Figure 4 clearly indicates significantly

slower kinetics of the demetalation reaction of BChl epossessing a formyl group at the 7-position relative to BChlsc and d possessing a methyl group at this position.

The demetalation rate constants, k, could be obtained byfitting the time courses of the BChl Soret absorbance to thefollowing kinetic Eq. (1):

ln½ðA� A1Þ=ðA0 � A1Þ� ¼ �kt ð1Þ

whereA0,A andA¥ are Soret absorbances of BChls at the onsetof measurements, at time t, and at the complete demetalation,respectively. The demetalation rate constants of 31R-[E,E]-

BChls c, d and e were determined to be 1.4 · 10)1, 8.9 · 10)2

and 1.7 · 10)2 min)1, respectively, under the reaction condi-tions in Fig. 4. These values were the averages of 6–7 indepen-

dent measurements, and the standard deviations were 9.7%,5.2% and 3.3% of the averages, respectively. The rate constantsof 31R-[E,E]-BChl d and 31R-[E,E]-BChl ewere 1.5- and 8-times

smaller than that of 31R-[E,E]-BChl c.Figure 5 shows a comparison of demetalation kinetics of

31R-[E,E]- and 31R-[P,E]-homologs of chlorosomal BChls atthe proton concentration of 2.5 · 10)4

MM at 25�C to examine

the effects of methylation at the 82-position. The rate constantsof 31R-[P,E]-BChls c, d and e were determined to be 1.3 · 10)1,7.5 · 10)2 and 1.5 · 10)2 min)1, respectively. These values

were the averages of 3–4 independent measurements, and thestandard deviations were 3.0%, 6.9% and 2.1% of theaverages, respectively. The relative ratios of the rate constants

of 31R-[E,E]-homologs of three types of BChls to those of31R-[P,E]-homologs were ranged between 1.1 and 1.2.

31R-[P,E]- and 31S-[P,E]-BChls c were major isomers of

BChls c inChl. tepidum, and these epimeric pairs can be isolatedby preparative HPLC more easily than those of BChls d and e.Therefore, effects of 31-configuration of chlorosomal BChls ondemetalation kinetics were investigated using 31R-[P,E]- and

31S-[P,E]-BChl c as shown in Fig. 6. From the kinetics, thedemetalation rate constant of 31S-[P,E]-BChl c was determinedto be 1.6 · 10)1 min)1 (the average of 4 independent measure-

ments, the standard deviation was 4.9% of the average), whichwas 1.2 (±0.1) times larger than that of 31R-[P,E]-BChl c.

Temperature dependence of demetalation rate constants

Temperature dependence of the demetalation rate constants of31R-[E,E]-BChls c and e at the proton concentration of

2.5 · 10)4MM was measured between 5 and 35�C at a 10�C

interval, and is shown in Fig. 7. In this temperature range, thedemetalation reactions of 31R-[E,E]-BChls c and e could be

regarded as pseudo first-order reactions. Demetalation kineticsof both 31R-[E,E]-BChls c and e became slower with decreas-ing reaction temperatures. The demetalation rate constantswere obtained by fitting the time course of BChl Soret

Figure 4. Kinetic plots of demetalation of 31R-[E,E]-BChl c (opencircle), 31R-[E,E]-BChl d (open triangle) and 31R-[E,E]-BChl e (opendiamond) in acetone ⁄water (3 ⁄ 1, vol ⁄ vol) at the proton concentrationof 2.5 · 10)4

MM at 25�C. Absorbance changes were monitored at 435,428 and 469 nm for BChls c, d and e, respectively. A0, A and A¥ areSoret absorbances of BChls at the onset of measurements, at time t andat the complete demetalation, respectively.

Figure 5. Kinetic plots of demetalation of 31R-[E,E]- (open circle) and31R-[P,E]- (closed circle) homologs of BChls c (a), d (b) and e (c) inacetone ⁄water (3 ⁄ 1, vol ⁄ vol) at the proton concentration of2.5 · 10)4

MM at 25�C.

Figure 6. Kinetic plots of demetalation of 31R-[P,E]- (closed circle)and 31S-[P,E]- (open triangle) homologs of BChl c in acetone ⁄water(3 ⁄ 1, vol ⁄ vol) at the proton concentration of 2.5 · 10)4

MM at 25�C.

Photochemistry and Photobiology, 2009, 85 1143

absorbances to Eq. (1). The rate constants were measuredmore than 3 times at each temperature, and the logarithms ofthese values ks were plotted against reciprocal of the absolute

temperature T of the demetalation reactions as shown inFig. 8. As a result, the activation energies, Es, of thedemetalation reactions of 31R-[E,E]-BChls c and e wereestimated by fitting the following Arrhenius Eq. (2):

ln k ¼ �ðE=RTÞ þ B; ð2Þ

where R is the gas constant. The Arrhenius plots in Fig. 8 gavethe activation energies of 35 and 53 kJmol)1 for 31R-[E,E]-BChls c and e, respectively. The activation energy of 31R-[E,E]-

BChl e was 1.5-times larger than that of 31R-[E,E]-BChl c.

15N NMR of 3

1R-[E,E]-BChls c and e

15N-labeled 31R-[E,E]-BChls c and e were obtained from cellsgrown in a medium containing 15NH4Cl. Purified 31R-[E,E]-BChls c and e were proved to be fully 15N-labeled by LC-MS

and 1H NMR as described above. Figure 9 shows 15N NMRspectra of 15N-labeled 31R-[E,E]-BChls c and e in a mixedsolvent of methanol-d4 ⁄ chloroform-d (1 ⁄ 9, vol ⁄ vol). Each

nitrogen signal of 15N-labeled 31R-[E,E]-BChls c and e wasassigned by 1H-15N HMBC, and the chemical shifts ds of thefour nitrogen atoms are summarized in Table 1. The ds of N21,

N22, N23 and N24 of 15N-labeled 31R-[E,E]-BChl c wereassigned as 194.95, 207.61, 191.45 and 245.70 ppm, respec-tively, whereas those of 15N-labeled 31R-[E,E]-BChl e were

assigned as 196.55, 215.02, 192.72 and 245.15 ppm, respec-tively. The d of N22 atom of BChl e was considerably low-fieldshifted (7.41 ppm) compared with the corresponding d of31R-[E,E]-BChl c. The ds of N21 and N23 atoms were also low-

field shifted (1.60 and 1.27 ppm, respectively). Only the d ofN24 atom was slightly high-field shifted by 0.55 ppm. The low-field shifts of nitrogen atoms in BChl e molecules suggest a

decrease of electron density of these atoms in BChl e and ⁄ ordifference of ring-current effects between BChls c and e. It isworthy to note that the difference in chemical shifts of 2-CH3,

17-H, 18-H, and 20-CH3, which were related to ring-currenteffects, were 0.06–0.09 ppm between 15N-labeled 31R-[E,E]-BChls c and e.

Figure 7. Kinetic plots of demetalation of 31R-[E,E]-BChl c (a) and31R-[E,E]-BChl e (b) in acetone ⁄water (3 ⁄ 1, vol ⁄ vol) at the protonconcentration of 2.5 · 10)4

MM at 5�C (open circle), 15�C (opentriangle), 25�C (open diamond) and 35�C (open square).

Figure 8. Arrhenius plots over the temperature range of 5–35�C forthe demetalation rate constants of 31R-[E,E]-BChl c (open circle) and31R-[E,E]-BChl e (open triangle).

(a)

(b)

Figure 9. Proton-decoupled 15N NMR spectra for the fully15N-labeled 31R-[E,E]-BChl c (a) and 15N-labeled 31R-[E,E]-BChl e(b) in methanol-d4 ⁄ chloroform-d, (1 ⁄ 9, vol ⁄ vol). The concentrations ofBChls c and e were 10 and 7 mMM, respectively. Chemical shifts ds werereferenced to 2% 15N-formamide (112.5 ppm).

Table 1. Assignments of 15N NMR spectra of 15N-labeled 31R-[E,E]-BChls c and e in methanol-d4 ⁄ chloroform-d (1 ⁄ 9, vol ⁄ vol).

Position d(BChl c) d(BChl e) Dd

N21 194.95 196.55 +1.60N22 207.61 215.02 +7.41N23 191.45 192.72 +1.27N24 245.70 245.15 )0.55

Dd = d(BChl e) - d(BChl c).

1144 Yuki Hirai et al.

DISCUSSION

The demetalation kinetics dependent on molecular structures

of chlorosomal BChls in green sulfur photosynthetic bacteriawere unraveled in the present physicochemical measurements.The 7-formyl group in the chlorin macrocycle of BChl e hadthe largest effect in peripheral substituents of natural BChls in

green sulfur bacteria. 20-Methylation in chlorin macrocyclesalso affects the demetalation kinetics. The effect of 20-methylation on demetalation kinetics of BChl c would

originate from electron-donating ability of a 20-methyl groupdirectly linked to chlorin macrocycles. A slight distortion ofthe chlorin macrocycle by the 20-methylation might also affect

the demetalation kinetics. The difference of demetalation rateconstants (BChl c > d > e) can be qualitatively rationalizedby invoking inductive effects of 7- and 20-substituents of

chlorosomal BChls. In addition, 82- and 31-configuration ofchlorosomal BChls slightly influenced their demetalationkinetics. This might result in a difference in microenviron-mental structures including solvent molecules around homo-

logs or epimers, which have a different polarity and areseparated by reverse-phase HPLC. Slight difference ofdemetalation rate constants between Chl epimers was reported

for 132-configuration in Chls a and a’ (19,20).To examine the 7-substitution effects in detail, the activa-

tion energies of demetalation reaction and 15N NMR spectra

were compared between homologically and epimerically pure31R-[E,E]-BChl c and e. The activation energy of 31R-[E,E]-BChl e was larger than that of 31R-[E,E]-BChl c, indicatingthat the tolerance of BChl e to demetalation was derived from

its higher activation energy of demetalation reaction relative toBChl c. In 15N NMR spectra of 15N-labeled 31R-[E,E]-BChls cand e, the chemical shift of N22 atom located on the pyrrole-

ring B of BChl e was significantly low-field shifted and those ofN21 and N23 atoms were also low-field shifted relative to thoseof BChl c. These suggest that the electron-withdrawing

7-formyl group affect the electronic state of the chlorinmacrocycle, especially N22 in the ring B.

Difference of the physicochemical properties of the demeta-

lation between BChl c and e would mainly originate fromsubstitution at the 7-position directly linked to chlorinmacrocycles. Similar 7-formyl effect on demetalation kineticshas been reported for Chl b possessing the formyl group at the

7-position, where Chl b was demetalated more slowly than Chla possessing a methyl group at this position (20). Thedemetalation rate constant of Chl b was about 10-times

smaller than that of Chl a under similar reaction conditions,indicating that the effect of the 7-formyl group of the chlorinmacrocycle on demetalation properties would be comparable

for BChl e and Chl b. Hence, the tolerance of BChl e and Chl bto removal of central magnesium is ascribable to electron-withdrawing ability of the formyl group at the 7-position of

chlorin macrocycles. The electron-withdrawing property of the7-formyl group in chlorin macrocycles has been proposed todecrease the electron density of pyrrole nitrogen atoms (33).This is consistent with the observed low-field shifts of N21, N22

and N23 atoms in the 15N NMR spectrum of BChl e comparedwith those of BChl c. These effects would prevent electrophilicattack of protons to pyrrole nitrogen atoms, providing 7-

formyl-type (B)Chls resistance to removal of the central

magnesium. The increase of Lewis acidity of central magne-sium by the electron-withdrawing effect of the 7-formyl groupmight be another possible reason for such tolerance of BChl eand Chl b. The acidity increase might result in stronger

magnesium-nitrogen coordination bonds and ⁄or stronger axialcoordination of a solvent molecule such as a water molecule(34). Actually, peripheral electron-withdrawing substituents in

synthetic zinc chlorins enhanced the axial coordination tocentral metal (35). Molecular structures including peripheralsubstituents and central metals affect electronic states of

chlorophyllous pigments (36,37), leading to the variation oftheir demetalation properties.

The present study implies that BChl e or bacteriochloro-

phyllide (BChlide) e, which has no farnesol by enzymatichydrolysis of BChl e, would be more difficult to be demeta-lated in vivo than BChls c and d through the degradationprocess. Such possibility is supported by appearance of

BChlide e in ancient sediments (38), suggesting that demeta-lation of BChlide e is actually slow in nature. To make in vivodemetalation of BChlide e smooth, green sulfur photosynthetic

bacteria might convert 7-formyl group of BChlide e into7-methyl or 7-hydroxymethyl group in early steps of thedegradation pathway. Higher plants have such a conversion

system of 7-substituents called the Chl cycle, and use it inbiosynthesis and degradation of Chl b possessing a 7-formylgroup (39–42). The conversion of 7-formyl-type (B)Chls (or(B)Chlides) into 7-methyl-type (B)Chls (or (B)Chlides) might

be important for in vivo demetalation, which is one of the keysteps in the (B)Chl degradation pathway.

Acknowledgements—We thank Dr. Toshie Minematsu, Kinki Univer-sity and Dr. Tadashi Mizoguchi, Ritsumeikan University, for NMRmeasurements and LC-MS measurements, respectively. This work waspartially supported by Grants-in-Aid for Scientific Research forYoung Scientists (B) (No. 20750143) and for Scientific Research (B)(No. 19350088) from the Japan Society for the Promotion of Science,and by the Salt Science Research Foundation (No. 0818). Y. H. wassupported by the Sasagawa Research Grant from the Japan ScienceSociety.

REFERENCES

1. Olson, J. M. (1998) Chlorophyll organization and function ingreen photosynthetic bacteria. Photochem. Photobiol. 67, 61–75.

2. Blankenship, R. E., J. M. Olson and M. Miller (1995) Antennacomplexes from green photosynthetic bacteria. In AnoxygenicPhotosynthetic Bacteria (Edited by R. E. Blankenship,M. T. Madigan and C. E. Bauer), pp. 339–435. Kluwer AcademicPublishers, Dordrecht.

3. Niedermeier, G., H. Scheer and R. G. Feick (1992) The functionalrole of protein in the organization of bacteriochlorophyll c inchlorosomes of Chloroflexus aurantiacus. Eur. J. Biochem. 204,685–692.

4. Miyatake, T. and H. Tamiaki (2005) Self-aggregates of bacterio-chlorophylls-c, d and e in a light-harvesting antenna system ofgreen photosynthetic bacteria: Effect of stereochemistry at thechiral 3-(1-hydroxyethl) group on the supramolecular arrange-ment of chlorophyllous pigments. J. Photochem. Photobiol.C: Photochem. Rev. 6, 89–107.

5. Tamiaki, H. (2005) Self-aggregates of natural and modified chlo-rophylls as photosynthetic light-harvesting antenna systems: Sub-stituent effect on the B-ring. Photochem. Photobiol. Sci. 4, 675–680.

6. Tamiaki, H., R. Shibata and T. Mizoguchi (2007) The 17-propi-onate function of (bacterio)chlorophylls: Biological implication oftheir long esterifying chains in photosynthetic systems. Photochem.Photobiol. 83, 152–162.

Photochemistry and Photobiology, 2009, 85 1145

7. Mizoguchi, T., T.-Y. Kim, S. Sawamura and H. Tamiaki (2008)Pressure-induced red shift and broadening of the Qy absorption ofmain light-harvesting antenna chlorosomes from green photo-synthetic bacteria and their dependency upon alkyl substituents ofthe composite bacteriochlorophylls. J. Phys. Chem. B 112, 16759–16765.

8. Wakao, N., N. Yokoi, N. Isoyama, A. Hiraishi, K. Shimada, M.Kobayashi, H. Kise, M. Iwaki, S. Itoh, S. Takaichi and Y. Sakurai(1996) Discovery of natural photosynthesis using Zn-containingbacteriochlorophyll in an aerobic bacterium Acidiphilium rubrum.Plant Cell Physiol. 37, 889–893.

9. Krautler, B. and P. Matile (1999) Solving the riddle of chlorophyllbreakdown. Acc. Chem. Res. 32, 35–43.

10. Krautler, B. (2000) Chlorophyll breakdown and chlorophyllcatabolites. In The Porphyrin Handbook, vol. 13 (Edited by K. M.Kadish, K. M. Smith and R. Guilard), pp. 183–209. AcademicPress, San Diego.

11. Krautler, B. and S. Hortensteiner (2006) Chlorophyll catabolitesand the biochemistry of chlorophyll breakdown. In Chlorophyllsand Bacteriochlorophylls: Biochemistry, Biophysics, Functions andApplications (Edited by B. Grimm, R. J. Porra, W. Rudiger and H.Scheer), pp. 237–260. Springer, Dordrecht.

12. Krautler, B. (2008) Chlorophyll breakdown and chlorophyllcatabolites in leaves and fruit. Photochem. Photobiol. Sci. 7, 1114–1120.

13. Iturraspe, J., N. Engel and A. Gossauer (1994) Chlorophyllcatabolism. Isolation and structure elucidation of chlorophyll bcatabolites in Chlorella prototecoides. Phytochemistry 35, 1387–1390.

14. Jacob-Wilk, D., D. Holloand, E. E. Goldschmidt, J. Riov and Y.Eyal (1999) Chlorophyll breakdown by chlorophyllase: Isolationand functional expression of the Chlase1 gene from ethylene-treated Citrus fruit and its regulation during development. Plant J.20, 653–661.

15. Tsuchiya, T., H. Ohta, K. Okawa, A. Iwamatsu, H. Shimada, T.Masuda and K. Takamiya (1999) Cloning of chlorophyllase, thekey enzyme in chlorophyll degradation: Finding of a lipase motifand the induction by methyl jasmonate. Proc. Natl Acad. Sci. USA96, 15362–15367.

16. Suzuki, T. and Y. Shioi (2002) Re-examination of Mg-dechelationreaction in the degradation of chlorophylls using chlorophyllina as a substrate. Photosynth. Res. 74, 217–223.

17. Kunieda, T., T. Amano and Y. Shioi (2005) Search for chlorophylldegradation enzyme, Mg-dechelatase, from extracts of Chenopo-dium album with native and artificial substrates. Plant Sci. 169,177–183.

18. Langmeier, M., S. Ginsburg and P. Matile (1993) Chlorophyllbreakdown in senescent leaves: Demonstration of Mg-dechelataseactivity. Physiol. Plant. 89, 347–353.

19. Mazaki, H. and T. Watanabe (1988) Pheophytinization of chlo-rophyll a and chlorophyll a’ in aqueous acetone. Bull. Chem. Soc.Jpn 61, 2969–2970.

20. Mazaki, H., T. Watanabe, T. Takahashi, A. Struck and H. Scheer(1992) Pheophytinization of eight chlorophyll derivatives inaqueous acetone. Bull. Chem. Soc. Jpn 65, 3212–3214.

21. Kobayashi, M., M. Yamamura, M. Akiyama, H. Kise, K. Inoue,M. Hara, N. Wakao, K. Yahara and T. Watanabe (1998) Acidresistance of Zn-bacteriochlorophyll a from an acidophilic bacte-rium Acidiphilium rubrum. Anal. Sci. 14, 1149–1152.

22. Saga, Y., Y. Hirai and H. Tamiaki (2007) Kinetic analysis ofdemetalation of bacteriochlorophyll c and e homologs purified fromgreen sulfur photosynthetic bacteria. FEBS Lett. 581, 1847–1850.

23. Mizoguchi, T., J. Harada and H. Tamiaki (2006) Structuraldetermination of dihydro- and tetrahydrogeranylgeranyl groups atthe 17-propionate of bacteriochlorophylls-a. FEBS Lett. 580,6644–6648.

24. Saga, Y., T. Wazawa, T. Mizoguchi, Y. Ishii, T. Yanagida and H.Tamiaki (2002) Spectral heterogeneity in single light-harvestingchlorosomes from green sulfur photosynthetic bacterium Chloro-bium tepidum. Photochem. Photobiol. 75, 433–436.

25. Saga, Y., H. Oh-oka, T. Hayashi and H. Tamiaki (2003) Presenceof exclusively bacteriochlorophyll-c containing substrain in the

culture of green sulfur photosynthetic bacterium Chlorobiumvibrioforme strain NCIB 8327 producing bacteriochlorophyll-d.Anal. Sci. 19, 1575–1579.

26. Saga, Y. and H. Tamiaki (2004) Comparison between chloro-somes containing bacteriochlorophyll-c and chlorosomes con-taining bacteriochlorophyll-d isolated from two substrains ofgreen sulfur photosynthetic bacterium Chlorobium vibrioformeNCIB 8327. J. Photochem. Photobiol. B: Biol. 75, 89–97.

27. Saga, Y., K. Matsuura and H. Tamiaki (2001) Spectroscopicstudies on self-aggregation of bacteriochlorophyll-e in nonpolarorganic solvents: Effects of stereoisomeric configuration at the 31-position and alkyl substituents at the 81-position. Photochem.Photobiol. 74, 72–80.

28. Wang, Z.-Y., M. Umetsu, M. Kobayashi and T. Nozawa (1999)13C- and 15N-NMR studies on the intact bacteriochlorophyllc dimers in solutions. J. Am. Chem. Soc. 121, 9363–9369.

29. Miyatake, T., T. Oba and H. Tamiaki (2001) Pure and scrambledself-aggregates prepared with zinc analogues of bacteriochloro-phylls c and d. ChemBioChem 2, 335–342.

30. Mizoguchi, T., Y. Saga and H. Tamiaki (2002) Isolation andstructure determination of a complete set of bacteriochlorophyll-dhomologs and epimers from a green sulfur bacterium Chlorobiumvibrioforme and their aggregation properties in hydrophobicsolvents. Photochem. Photobiol. Sci. 1, 780–787.

31. Boxer, S. G., G. L. Closs and J. J. Katz (1974) The effect ofmagnesium coordination on the 13C and 15N magnetic resonancespectra of chlorophyll a. The relative energies of nitrogen np*states as deduced from a complete assignment of chemical shifts.J. Am. Chem. Soc. 96, 7058–7066.

32. Kawano, K., Y. Ozaki, Y. Kyogoku, H. Ogoshi, H. Sugimoto andZ. Yoshida (1978) Nuclear magnetic resonance spectra of nitro-gen-15 enriched octaethylporphyrin and its derivatives. J. Chem.Soc. Perkin II, 1319–1325.

33. Eggink, L. L., R. LoBrutto and J. K. Hoober (2004) Chlorophyllsb and c: Why plants make them. In Agricultural Applications inGreen Chemistry (Edited by W. M. Nelson), pp. 23–37. OxfordUniversity Press, New York.

34. Chen, M., L. L. Eggink, J. K. Hoober and A. W. D. Larkum(2005) Influence of structure on binding of chlorophylls to peptideligands. J. Am. Chem. Soc. 127, 2052–2053.

35. Tamiaki, H., S. Yagai and T. Miyatake (1998) Synthetic zinctetrapyrroles complexing with pyridine as a single axial ligand.Bioorg. Med. Chem. 6, 2171–2178.

36. Yerushalmi, R., I. Ashur and A. Scherz (2006) Metal-substi-tuted bacteriochlorophylls: Novel molecular tools. In Chlorophyllsand Bacteriochlorophylls: Biochemistry, Biophysics, Functions andApplications (Edited by B. Grimm, R. J. Porra, W. Rudiger andH. Scheer), pp. 495–506. Springer, Dordrecht.

37. Noy, D., L. Fiedor, G. Hartwich, H. Scheer and A. Scherz (1998)Methal-substituted bacteriochlorophylls. 2. Changes in redoxpotentials and electronic transition energies are dominated byintramolecular electronic interactions. J. Am. Chem. Soc. 120,3684–3693.

38. Mallorquı, N., J. B. Arellano, C. M. Borrego and L. J. Garcia-Gil(2005) Signature pigments of green sulfur bacteria in lower Pleis-tocene deposits from the Banyoles lacustrine area (Spain).J. Paleolimnol. 34, 271–280.

39. Ito, H. and A. Tanaka (1996) Determination of the activity ofchlorophyll b to chlorophyll a conversion during greening of eti-olated cucumber cotyledons by using pyrochlorophyllide b. PlantPhysiol. Biochem. 34, 35–40.

40. Tanaka, A., H. Ito, R. Tanaka, N. K. Tanaka, K. Yoshida and K.Okada (1998) Chlorophyll a oxygenase (CAO) is involved inchlorophyll b formation from chlorophyll a. Proc. Natl Acad. Sci.USA 95, 12719–12723.

41. Rudiger, W. (2000) The last steps of chlorophyll synthesis. In ThePorphyrin Handbook, vol. 13 (Edited by K. M. Kadish, K. M.Smith and R. Guilard), pp. 71–108. Academic Press, San Diego.

42. Rudiger, W. (2006) Biosynthesis of chlorophylls a and b: The laststeps. In Chlorophylls and Bacteriochlorophylls: Biochemistry, Bio-physics, Functions andApplications (Edited byB.Grimm,R. J. Porra,W. Rudiger and H. Scheer), pp. 189–200. Springer, Dordrecht.

1146 Yuki Hirai et al.