Ubiquity and Quantitative Significance of Detoxification Catabolism of Chlorophyll Associated with Protistan Herbivory

Ubiquity and quantitative signiﬁcance of detoxiﬁcation catabolism of chlorophyll associated with protistan herbivory

Yuichiro Kashiyamaa,1,2, Akiko Yokoyamab,1, Yusuke Kinoshitaa, Sunao Shojia, Hideaki Miyashiyac, Takashi Shiratorid, Hisami Sugae, Kanako Ishikawaf, Akira Ishikawag, Isao Inouyeb, Ken-ichiro Ishidab, Daiki Fujinumah, Keisuke Aokih, Masami Kobayashih, Shinya Nomotoi, Tadashi Mizoguchia, and Hitoshi Tamiakia,2

aGraduate School of Life Sciences, Ritsumeikan University, Kusatsu, Shiga 525-8577, Japan; bFaculty of Life and Environmental Sciences and dGraduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan; cGraduate School of Human and Environmental Studies, Kyoto University, Kyoto 606-8501, Japan; eInstitute of Biogeosciences, Japan Agency for Marine-Earth Science and Technology, Yokosuka, Kanagawa 237-0061, Japan; fSystem Analysis Division, Lake Biwa Environmental Research Institute, Otsu, Shiga 520-0022, Japan; gGraduate School of Bioresources, Mie University, Tsu, Mie 514-8507, Japan; hInstitute of Materials Science, University of Tsukuba, Tsukuba, Ibaraki 305-8573, Japan; and iDepartment of Chemistry, University of Tsukuba, Tsukuba, Ibaraki 305-8571, Japan

This Feature Article is part of a series identified by the Editorial Board as reporting findings of exceptional significance.

Edited by David M. Karl, University of Hawaii, Honolulu, HI, and approved August 2, 2012 (received for review May 2, 2012) Chlorophylls are essential components of the photosynthetic appa- as a potentially phototoxic chromophore or fluorophore. It is con- rati that sustain all of the life forms that ultimately depend on solar verted stepwise into an unconjugated and hence, colorless and energy. However, a drawback of the extraordinary photosensitizing nonfluorescent linear tetrapyrrole (SI Text,section1.1). However, efficiency of certain chlorophyll species is their ability to generate many algae and cyanobacteria apparently lack the programmed fi harmful singlet oxygen. Recent studies have clari ed the catabolic detoxifying catabolism of chlorophyll observed in embryophytes (SI fi processes involved in the detoxi cation of chlorophylls in land Text,section1.2). The inability of unicellular phototrophs to detoxify plants, but little is understood about these strategies in aquatic chlorophyll can be rationalized by their lack of the need to remo- ecosystem. Here, we report that a variety of heterotrophic protists accumulate the chlorophyll a catabolite 132,173-cyclopheophorbide bilize nutrients before death, unlike multicellular land plants. Re- a enol (cPPB-aE) after their ingestion of algae. This chlorophyll de- garding heterotrophs, the digestive systems of most terrestrial rivative is nonfluorescent in solution, and its inability to generate herbivores are dark, whereas the digestive systems of most aquatic singlet oxygen in vitro qualifies it as a detoxified catabolite of chlo- herbivores, such as multicellular and unicellular zooplankton, are rophyll a. Using a modified analytical method, we show that cPPB- small and translucent. This finding renders the microscopic aquatic aE is ubiquitous in aquatic environments, and it is often the major herbivores susceptible to damage by the singlet oxygen generated chlorophyll a derivative. Our findings suggest that cPPB-aE metab- when ingested chlorophylls are exposed to light (SI Text,section1.3). olism is one of the most important, widely distributed processes in Thus, strategies to protect against the accumulation of phototoxic aquatic ecosystems. Therefore, the herbivorous protists that con- chlorophyll derivatives should be critical for the survival of the mi- vert chlorophyll a to cPPB-aE are suggested to play more significant fl croscopic aquatic herbivores feeding under illumination. roles in the modern oceanic carbon ux than was previously recog- We conducted feeding experiments on several microorganisms nized, critically linking microscopic primary producers to the mac- to screen for potentially detoxified chlorophyll catabolites. Here, roscopic food web and carbon sequestration in the ocean. we report that 132,173-cyclopheophorbide a enol (cPPB-aE) (Fig. a phototoxicity of chlorophylls | microbial herbivory | phagocytosis | 1) is the dominant product derived from Chl- that accumulates biodiversity of eukaryotes | microbial loop in cells of various aquatic heterotrophic protists (i.e., unicellular eukaryotes) that feed on algae and that cPPB-aE is a virtually fl hlorophylls are crucial to sustaining most life forms on Earth, non uorescent and nonphotosensitizing chlorophyll derivative Cthe majority of which ultimately depend on solar energy. incapable of generating singlet oxygen. These results strongly Photoexcitation of chlorophylls initiates various photosynthetic suggest that herbivorous protists generate cPPB-aE as a detoxi- reactions that convert the energy of photons into chemical poten- fied catabolite of Chl-a. We also show that cPPB-aE is ubiqui- tials, which in turn, drive the full range of metabolic reactions tous in all of the aquatic environments that we tested, frequently throughout the global ecosystem. Chlorophylls play a central role as the most abundant Chl-a derivative in the surface sediments, in the photosynthetic apparatus by absorbing light and trans- and that it is actively generated in the water near illuminated ferring the excitation energy to the reaction centers of photo- surfaces. systems before photosynthetic electron transport. However, without measures to contain the excited energy, chlorophylls can harm organisms because of their high photosensitizing potential. Author contributions: Y. Kashiyama, A.Y., H.M., H.S., K.I., A.I., I.I., K.-i.I., M.K., S.N., T.M., and Photoexcited chlorophylls generate singlet oxygen, a highly re- H.T. designed research; Y. Kashiyama, A.Y., Y. Kinoshita, S.S., H.M., T.S., D.F., K.A., and S.N. active oxygen species that can cause severe cellular damage (1). performed research; Y. Kashiyama and A.Y. analyzed data; and Y. Kashiyama wrote the paper. Therefore, the phototoxicity of chlorophylls has been a continu- The authors declare no conflict of interest. ous concern for the Earth’s ecosystem since the global rise in This article is a PNAS Direct Submission. atmospheric oxygen about 2.3 billion y ago (2). Freely available online through the PNAS open access option. Given its potential risks, chlorophyll metabolism is thought to See Commentary on page 17311. be carefully controlled in the cells of phototrophic organisms. 1Y. Kashiyama and A.Y. contributed equally to this work. Recent works have revealed that the biodegradation of chloro- 2To whom correspondence may be addressed. E-mail: [email protected] or tamiaki@ phyll a (Chl-a) (Fig. 1) in land plants (embryophytes) is regulated fc.ritsumei.ac.jp. as carefully as its biosynthesis (3). The chlorin moiety found in This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. Chl-a is a highly π-conjugated tetrapyrrolic macrocycle that acts 1073/pnas.1207347109/-/DCSupplemental.

17328–17335 | PNAS | October 23, 2012 | vol. 109 | no. 43 www.pnas.org/cgi/doi/10.1073/pnas.1207347109 Downloaded by guest on September 29, 2021 a Chlorophyll a (Chl-a): hCPLs- may be a product of biotic processing (9, 10) and/or abiotic FEATURE ARTICLE M = Mg; R1 = COOMe; R2 = Phytyl oxidation of cPPB-aE (11). All samples of heterotrophic protists a Pheophytin a (Phe-a): also contain variable amounts of intact Chl- , which may have been M = 2H; R1 = COOMe; R2 = Phytyl derived from undigested algal materials in phagosomes. N N a a Our experiments, thus, suggest qualitatively that the herbivo- Pyropheophytin (pPhe- ): a a M M = 2H; R1 = H; R2 = Phytyl rous protists studied actively modify Chl- into cPPB- E during Pheophorbide a (PPB-a): their course of phagotrophic digestion (Fig. 3). Given that het- N N M = 2H; R1 = COOMe; R2 = H erotrophic and phototrophic cells grew together in the cultures Pyropheophorbide a (pPPB-a): used in our experiments, quantitative estimations of the rate of M = 2H; R1 = H; R2 = H cPPB-aE production from Chl-a as well as the stability of cPPB- SEE COMMENTARY 2 a 13 Methyl pPPB-a: E in vivo were beyond the scope of the present work. None- M = 2H; R1 = H; R2 = Me theless, it seems reasonable to conclude that cPPB-aE is a major R1 O 3 a 17 Cholesteryl pPPB-a: Chl- catabolite produced by the herbivorous protists that we 2 1 2 a O OR M = 2H; R = H; R = Cholesteryl studied. Therefore, cPPB- E could potentially emerge as a biochemical marker of protistan herbivory in aquatic environments.

Photochemical Significance. We determined various properties of cPPB-aE by analyzing an authentic sample that was semisynthesized from Chl-a (Materials and Methods). The most striking property of cPPB-aE is that it is essentially nonfluorescent (4), despite its highly π-conjugated structure (Fig. 4 A and B, SI Text, section 1.5, N NH N NH Fig. S1,andTable S2). This finding is consistent with our

N HN N HN 0% 20% 40% 60% 80% 100%

Heterotrophic stramenopile

fed on Nitzschia sp. MICROBIOLOGY * Nitzschia sp. OH OH O Heterotrophic cercozoan O O fed on Skeletonema sp.

Cyclopheophorbide a enol (132R/S)-Hydroxychlorophyllones a Skeletonema sp. (cPPB-aE) ((R/S)-hCPLs-a)

Fig. 1. Chemical structures of Chl-a and its derivatives discussed in the Heterotrophic heliozoan fed on Pyramimonas sp. present work. Pyramimonas sp.

Results and Discussion Heterotrophic discicristoidean fed on a pennate diatom Feeding Experiments. We showed that three genetically distant heterotrophic protists—a stramenopile, a cercozoan (Filosa), (Pannate diatom) — a and a heliozoan (Centrohelida) accumulated cPPB- E as the Daphnia pulex only major chlorophyll derivative associated with herbivory. We fed on Cryptomonas a tetrapyrenoidosa estimated that levels of cPPB- E in extracts from these protists Cryptomonas tetrapyrenoidosa after they were fed on fresh unialgal cells are as high as 80 mol% of total amounts of chlorins derived from Chl-a (Fig. 2). Howe- Entrobacter aerogenes ver, a heterotrophic discicristoidean did not accumulate any cPPB- fed on green juice powder a a Green juice powder E. No trace of cPPB- E was detected in control experiments, (made from young barley leaves) where the unialgal cultures fed to the protists were incubated under conditions identical to those conditions used for the Chl-a pPhe-a feeding experiments (Table S1). Organic synthesis of cPPB-aE Mg-chelated Chl-a derivatives Steryl chlorin esters – 2 requires a strong base to form the C C bond between 13 and Free base Chl-a derivatives cPPB-aE 173 carbons (4, 5), suggesting that it is difﬁcult to generate cPPB- aE in vitro under normal conditions. In fact, the generation of Fig. 2. Protistan herbivory and associated chlorophyll catabolism/degrada- cPPB-aE has not been reported in so-called dark incubation tion. Relative abundances of chlorophyll derivatives in extracts of cultured experiments involving pure unialgal cultures sealed and pre- heterotrophic protists. Four experiments involved feeding a stramenopile, a cercozoan, a heliozoan, and a discicristoidean with the diatoms Nitzschia sp. served in darkness for months to years (6–8). Therefore, cPPB- a and Skeletonema sp., a prasinophyte Pyramimonas sp., and a pannate diatom, E detected in the present experiments must be conclusively respectively. We observed that, whereas cPPB-aEwasthedominantcomponent attributed to metabolic processes in the heterotrophic protists. in the cultures of the stramenopile, the cercozoan, and the heliozoan, it was Minor products of Chl-a degradation are pyropheophytin a absent from the culture of the discicristoidean. Crustacean zooplankton (pPhe-a) (Fig. 1) as well as various polar chlorins, including (132R)- (Daphnia pulex) was fed a cryptophyte, Cryptomonas tetrapyrenoidosa,from and (132S)-hydroxychlorophyllones a [(R/S)-hCPLs-a](Fig.1).Dark which we did not identify any cPPB-aE production. When the γ-proteobacte- incubation experiments of unialgal cultures revealed substantial rium Entrobacter aerogenes was grown on a media containing green juice powder made from young barley leaves, we did not identify any cPPB-aE pro- accumulation of pPhe-a only after the incubations had proceeded – a duction. None of the unialgal control cultures contained any cPPB-aEeither, for several months (6 8). Given that pPhe- is absent from our suggesting that the Chl-a metabolite was only produced after phagotrophic control experiments involving unialgal cultures, its accumulation can feeding by the three heterotrophic protists. Additional details are provided in also be attributed to the heterotrophic process. However, (R/S)- Table S1, and additional discussions are in SI Text, section 1.4.

Kashiyama et al. PNAS | October 23, 2012 | vol. 109 | no. 43 | 17329 Downloaded by guest on September 29, 2021 rapid generation of singlet oxygen was approximately three times A more rapid for Chl-a than Phe-a. The observed difference be- a a fl Skeletonema sp. tween Chl- and Phe- should partially re ect differences in the never ingested absorption intensities of Qy bands (Table S3). In conclusion, (unfocused) cPPB-aE is likely to be a colored but nonphotosensitizing Chl-a catabolite, which is comparable with the colorless nonfluorescent chlorophyll catabolites (NCCs) of higher plants (SI Text, EP section 1.1) insofar as its generation during the detoxification of chlorophylls protects against the accumulation of singlet oxygen in vivo. LP Previous studies have frequently regarded cPPB-aEasanan- tioxidant that accumulates in marine macrofauna (11, 13–16). However, other workers have failed to highlight the possibility NP that the primary evolutionary advantage of cPPB-aE production by single-celled organisms is the detoxification of the phototoxic chlorophyll derivatives that are inevitably ingested on feeding chlorophyll-containing diets to protists in an illuminated and B aerated environment. The reported antioxidant function of cPPB-aE accumulated in the intestines of metazoans would, thus, Skeletonema sp. be a secondary adaptation if the derivative, indeed, functions as never ingested an antioxidant. We note that its antioxidant activity, which should (unfocused) protect against the oxidative damage exerted by reactive oxygen EP species, has not been shown experimentally. Furthermore, no direct evidence has yet been presented that metazoans produce cPPB-aE. Geochemical evidence that is consistent with the ability of microscopic plankton to produce cPPB-aE is discussed below.

LP Ubiquity and Geochemical Significance of cPPB-aE in Aquatic Environ- ments. To determine the distribution of cPPB-aE metabolism NP among various aquatic settings, we used a geochemical approach that involved detecting cPPB-aE directly from particulate organic matter (POM) in water columns as well as surface sedimentary samples just beneath the columns. We analyzed extracts of glass microfiber filters (GF/F grade; Whatman) that presumably collected Fig. 3. A typical differential interference image (A) and a corresponding any particulate matters larger than 0.7 μmindiameterfromthe fluorescent image (B; excitation light = 400–440 nm) of cercozoans in dif- water samples examined, and therefore, they should have included ferent stages of phagocytosis. The cercozoan cells are indicated by the dashed fl any eukaryotic cells and most prokaryotic cells present as well as lines on the uorescent image. EP, a cell in an earlier stage of phagocytosis fecal materials and any organic aggregates. containing two chains of Skeletonema sp. in different stages of digestion; The results showed that cPPB-aE is ubiquitous and abundant in LP, a cell in a later stage of phagocytosis containing a chain of Sketetonema fi sp. in a later stage of digestion displaying no autofluorescence; NP, a cell all of the aquatic environments tested, con rming the ecological exhibiting no phagosome formation. (Scale bars: 50 μm.) In EP cell, the longer and biogeochemical significance of cPPB-aE metabolism (Fig. 5). chain of diatoms on the left is partially digested, and therefore, the plastids of In summary, we detected cPPB-aE in all POM samples from diverse several diatom cells still display the autofluorescence of chlorophyll (arrows), environments, although its abundance relative to the other Chl-a whereas the shorter chain on the right is in a later stage of digestion and derivatives ranged from 5 mol% (samples collected from Ise Bay) to contains remnant plastids (dark brown grains) displaying no autofluores- 42 mol% (samples collected from a dammed creek). These values cence. Additional information is given in Fig. S2. correspond to ranges in the cPPB-aE/Chl-a ratios from 0.07 to 1.53. However, the relative concentrations of cPPB-aE in the surface – microscopic observation that the autofluorescence of chlor- sediments were generally very high (51 81%), representing virtually the most abundant Chl-a derivatives. In fact, the presence ophylls in the plastids of ingested algae rapidly disappears during of cPPB-aE in sediments has been occasionally reported in pre- their phagocytosis by herbivorous protists (Fig. 3 and Fig. S2). vious works (11, 17–19), although it has been probably completely This observation suggests very rapid and nonradiative quenching a missed in other reports because of analytical artifacts (20). of the photoexcited singlet state of cPPB- E, which is ascribed to The abundance of cPPB-aE relative to Chl-a in water columns internal conversion(s) to the ground singlet state (12) and/or in- implies that a considerable proportion of photosynthetic primary tersystem crossing to the excited triplet state. Considering the fi a production should initially be processed by herbivorous protists biochemical signi cance of producing cPPB- Ebysomeprotists, that produce cPPB-aE. Given that Chl-a is generally regarded as the latter possibility seems unreasonable owing to the highly toxic standing biomass of phototrophic plankton, we similarly expect singlet oxygen that is readily produced from the triplet state that the cPPB-aE content in POM should represent the mass of (Fig. S3). herbivorous protists as well as their excreta in water. Depth C Fig. 4 shows that production of singlet oxygen by photoex- profile patterns of cPPB-aE and the other chlorophyll derivatives cited cPPB-aE is negligible relative to the levels generated from (Fig. 6) are perhaps best explained by (i) in situ feeding activities Chl-a and Phe-a. Here, the generation of singlet oxygen was de- of herbivorous protists in the upper parts of water columns and/ tected using Singlet Oxygen Sensor Green (SOSG), a commercially or (ii) selective preservation into deeper water and sediments, available fluorescent probe. No change in SOSG fluorescence which would explain the elevated relative concentration of cPPB- was observed over the course of the 180-min experiment with aE in the surface sediments. cPPB-aE, which recapitulated the result of the control experi- Evaluations of those data in terms of either the rate of protistan ment. This finding contrasts strikingly with the results for Chl- herbivory or the rate of cPPB-aE production require additional a and Phe-a. The increase in the SOSG fluorescence caused by investigation, including investigation of the rate of cPPB-aEpro-

17330 | www.pnas.org/cgi/doi/10.1073/pnas.1207347109 Kashiyama et al. Downloaded by guest on September 29, 2021 120 2 3 A Chl-a Phe-a cPPB-aE B 13 ,17 -Cyclopheophorbide a enol cPPB-aE FEATURE ARTICLE (cPPB-aE) Chl-a 100 Phe-a

N NH

) 80 -1 N HN cm

-1 60

(mM 40 O

Extinction coefficient 20

0 SEE COMMENTARY 300 350 400 450 500 550 600 650 700 750 800 Wavelength (nm)

C 2000

1500

1000 ence intensity ence intensity

crement 500 c c in

0 Fluores -500 0 30 60 90 120 150 180 Irradiation time (min)

Fig. 4. Photochemical properties of cPPB-aE. (A) Chl-a, Phe-a, and cPPB-aE in anisole (50 μM in a 4-mL quartz cuvettete) under white (Upper) and UV (Lower) light, showing the absence of red ﬂuorescence from cPPB-aE, which is typical of chlorophyll derivatives. (B) Absorption spectra of Chl-a, Phe-a, and cPPB-aEin

anisole. (C) Experiments detecting production of singlet oxygen in vitro using SOSG. When Chl-a is codissolved with SOSG in methanol-anisole (1:1, vol/vol; ), MICROBIOLOGY relatively rapid generation of singlet oxygen was indicated on illumination of red light (>630 nm). Phe-a (■) also acted as a photosensitizer for the generation of singlet oxygen. By contrast, no generation of singlet oxygen was evident from cPPB-aE( ) or the control experiment (only SOSG; ). In each experiment, the concentrations of pigments and SOSG used were 10 and 1 μM, respectively.

duction by various herbivorous protists, its lifetime in their cells and the possibility that protistan catabolism responsible for producing the environment (e.g., in the excreta), and the sinking rate of POM cPPB-aE would occur even in darkness (e.g., in deep water columns in each environment. In particular, a better understanding of the and sediments), despite our hypothesis that the process evolved chemical properties of cPPB-aE in vivo is required to explain its primarily to avoid the phototoxicity of chlorophyll derivatives. apparently high preservation potential in environmental samples, ﬁ despite its instability in organic solutions. In fact, we do not exclude Evolutionary Signi cance of Protistan Herbivory. We postulate that the ingestion of Chl-a by small, transparent, or translucent microbes in illuminated and oxygenated environments places an evolutionary constraint on microbial herbivory by necessitating appropriate 0% 20% 40% 60% 80% 100% biochemical strategies to contain levels of singlet oxygen-gener- Kumano-Nada in the Pacific Ocean Water (50 m) ating molecules after feeding on microalgae, cyanobacteria, or phototrophic bacteria. The evolution of aquatic microbial her-

Water (8 m) bivory should have been accompanied by the establishment of Ise Bay some enzymatic pathways that catabolize chlorophylls into non- Sediment fluorescent substances. Unlike the NCCs of higher plants, cPPB- aE is a colored catabolite, despite its nonphotosensititizing prop- Water (5 m) erties. Given that conversion of Chl-a to cPPB-aE apparently Tokyo Bay requires fewer enzymatic steps than the generation of NCCs, it Sediment has emerged as a simple and efficient strategy for rapid detoxification of chlorophyll among aquatic protists. Water (10 m) a Lake Biwa Our evidences regarding the synthesis of cPPB- E by protistan Sediment herbivores are comparable with the results of similar feeding experiments in the work by Goericke et al. (17) using heterotrophic fi a Dammed creek water (BKC) protists. The work by Goericke et al. (17) also identi ed cPPB- E as a major Chl-a derivative in the fecal material of three protists, the ciliate Strombidinopsis acuminatum and the heterotrophic Paddle water in a garden rock dinoflagellates Amphidinium sp. and Noctiluca scintillans. Fig. 7 illustrates that cPPB-aE producers, thus, distribute widely among Chl-a pPhe-a the protistan lineage, at least in two supergroups. These are the Mg-chelated Chl-a derivatives Steryl chlorin esters Stramenopile-Alveolate-Rhizaria (SAR) clade and the Cryptophyte- Free base Chl-a derivatives cPPB-aE Centrohelid-Telonemid-Haptophyte (CCTH) clade (21, 22). In Fig. 5. Relative abundances of chlorophyll derivatives in extracts of POM contrast, one species belonging to the opistkonta clade lacks the from various aquatic samples and surface sediments. The samples analyzed ability to synthesize cPPB-aE. Significantly, over 70% of hetero- are summarized in Table S4. trophic protists in the marine environment are known to belong

Kashiyama et al. PNAS | October 23, 2012 | vol. 109 | no. 43 | 17331 Downloaded by guest on September 29, 2021 (pmol/L) (pmol/L) (pmol/L) A 0.1 1 10 100 1000 B 0 500 1000 1500 2000 2500 C 0 200 400 600 800 1000 1200 0 0 0 Chl-a 5 10 100 (pmol/L) Mg-chelated 0200400 20 10 Chl-a derivatives 200 0 30 10 Free base 20 15 40 300 30 Chl-a derivative 40 20 50 50 pPhe-a 400 60 60 25 Water depth (m) Water depth (m) 70 Water depth (m) 70 Steryl chlorin esters 500 80 30 90 80 100 cPPB-aE 600 35 90 D 0% 20% 40% 60% 80% 100% E 0% 20% 40% 60% 80% 100% F 0% 20% 40% 60% 80% 100% 0 m 0 m 0 m 10 m Chl-a 20 m 5 m 30 m Mg-chelated 8 m 40 m Chl-a derivatives 50 m 10 m 60 m Free base 70 m 20 m 15 m Chl-a derivatives 80 m pPhe-a 90 m 20 m 100 m 32 m Steryl chlorin esters 150 m 80 m 200 m 300 m Sediment Sediment cPPB-aE 500 m

Fig. 6. Depth proﬁles of absolute and relative abundances of chlorophyll derivatives in water columns. (A–C) Absolute abundances. (D–F) Relative abundances. POM in the pelagic Paciﬁc Ocean (Kumano-Nada; A and D), POM and surface sediments from Ise Bay (B and E), and POM and surface sediments from freshwater Lake Biwa (C and F) are shown.

to either the SAR or CCTH clade (23), which is consistent with presumably heterotrophic. Nonetheless, it is still not certain our geochemical data. whether these protists lost plastids during the course of evolution The SAR and CCTH clades include various phototrophic and (i.e., secondary evolution of heterotrophy) or had never possessed mixotrotrophic protists other than the obligate heterotrophs plastids (i.e., multiple origins of secondary phototrophy). Our evi- analyzed here. Many of these phototrophic and mixotrophic dence implies that the generation of cPPB-aE, a key metabolic protists possess plastids that had been derived from ancestral red step in the heterotrophic lifestyle, is likely to have evolved at algae through secondary symbiosis (Fig. 7). They include major the root of the two clades, if not at an earlier evolutionary marine algae, such as diatoms and dinoﬂagellates, as well as di- stage. The ability to synthesize cPPB-aE must, thus, be retained verse picophytoplankton species (24). The rest of the SAR and in heterotrophic protists of distant lineages, regardless of the CCTH protists are colorless and lack plastids, and thus, they are reason for the absence of plastids.

SAR Dinoflagellates Chromerids

Perkinsids Flabellinida/Discosea Amebozoa Apicomplexa cercozoans

Ellobiopsids

Entamoebae

Colpodellids

LabyrinthuridsBicosoecids

Endomyxa Pelobionts Filosa Dictyostelids stramenopiles Cochliopodids Myxogastrids Foraminifera Protostelids Sloomycetes Radiozoa Stramenochomes Archamoebae Acanthamoebae Ciliates Tublinea Chytridiomycetes Discicristoidea Thecamoebae Placidids BreviatesApusozoa Glomeromycetes Slopalinids Basidiomycetes Ascomycetes Picobiliphytes Rozellids Microsporids Blastocladiomycetes Cryptomonads Zygomycetes Corallochytrium Nephridiophagids Kathablepharids IchthyosporidsChoanoflagellatesMetazoa Haptophytes Capsaspora owczarzaki

Centrohelids Malawimonads (heliozoans) Telonemids Glaucophytes Oxymonads Jakobids CCTH Trimastix Opistokonta Euglenozoa Parabasalids Cyanidiophytes Rhodophytes Heterolobosea Carpediemonads Retortamonads Streptophytes

Chlorophytes Diplomonads Archaeplastida + Enteromonads Excavata

Fig. 7. Unrooted tree of eukaryotes comprising six distinctive supergroups. Yellow stars beside taxonomic groups denote that they contain protists that produce cPPB-aE, which was reported in this study or a previous report (17). Examined protists belonging to discicristoidea did not produce cPPB-aE (red x). The taxonomic groups with green circles contain phototrophs with true chloroplasts. Note that the stramenopiles include not only the cPPB-aE–producing heterotrophs but also phototrophic diatoms used as diets that did not produce any trace of cPPB-aE (Fig. 2). This ﬁnding suggests that cPPB-aE metabolism evolved primarily for heterotrophy. Modiﬁed from Walker et al. (22).

17332 | www.pnas.org/cgi/doi/10.1073/pnas.1207347109 Kashiyama et al. Downloaded by guest on September 29, 2021 FEATURE ARTICLE

cPPB-aE

Chl-a Phagocytosis by herbivorous protists SEE COMMENTARY White light VU thgil Chlorophyll etihW VUthgil thgil detoxification

N N catabolism N NH Mg

N N N HN

Highly fluorescent and phototoxic Non-fluorescent and non-phototoxic MeOOC O Chlolophyll a (Chl-a) OH 2 3 O O O 13 ,17 -Cyclopheophorbide a enol (cPPB-aE)

(solar energy) Phototrophy Nutrients Phytoplankton Nekton (N, P, S, etc.)

(>3 μm) MICROBIOLOGY

CO2 Metazoan Protistan zooplankton Heterotrophic Pico- herbivory phytoplankton protists Dissolved organic (<3 μm) matter

Protistan Dissolved bacteriovory Export (sinking) inorganic Bacteria matter Archea

Benthos

Fig. 8. Inferences on microbial food web and material/energy ﬂows in aquatic environments. During their phagocytosis, herbivorous protists uniquely catabolize Chl-a into cPPB-aE(Upper). The product cPPB-aE was substantially abundant relative to other degradative products Chl-a in all aquatic environments that we studied, which strongly suggests that herbivorous protists play much more quantitatively signiﬁcant roles (red arrows) in the microbial loop (blue arrows) than ever recognized (Lower) in these environments.

Ecological Significance of Protistan Herbivory. We postulate that never been studied because of an inability to maintain them in herbivory, particularly of the picophytoplankton, by cPPB-aE– culture. Indeed, only recent advances showed the tremendous producing protists is one of the most important processes in the diversity of protists in aquatic environments (32, 35, 36), thanks modern aquatic food web, but it has been poorly understood largely to molecular biological approaches, such as PCR surveys until now (Fig. 8). Importance of protistan herbivory as a part of of the diversity of 18S rRNA sequences in environmental samples. the microbial loop in aquatic environments has been recognized Therefore, the ecological functions of aquatic colorless protists for nearly three decades (25). During this time, the quantitative remain largely unexplored (37, 38). significance of picophytoplankton (ϕ ≤ 3 μm) has also been Our evidence suggests that picophytoplankton-based protistan revealed (23, 24). Picophytoplankton includes coccoidal cyano- herbivory is a quantitatively important process in aquatic envi- bacteria and eukaryotic picophytoplankton. The former belong ronments. In analyses of POM samples from the water column of to the two major genera Prochlorococcus and Synechococcus, Lake Biwa, cPPB-aE was substantially enriched in the fine POM which account for up to 50% of pelagic marine oxygenic pho- fractions, which contained POM with diameters of 5–0.7 μm tosynthesis (26–28). Eukaryotic picophytoplankton are believed (Fig. S4), which indicates that cPPB-aE was either accumulated to account for 20–50% of aquatic photosynthesis (29–32). The in picozooplanktonic protists (ϕ ≤ 3 μm) and/or concentrated in use of the PCR to analyze environmental samples has also very fine particulate excreta most likely derived from variously revealed the amazingly wide diversity and activity of eukaryotic sized zooplanktonic protists. Therefore, picophytoplankton are picophytoplankton species (33, 34). probably consumed by these protistan herbivores (Fig. 8). In fact, However, we are still largely ignorant of abundance as well as the small size of picophytoplankton precludes their consumption ecological and biogeochemical roles of colorless protists in the by metazoan zooplankton such as copepods, suggesting that they oceans, in large part because the majority of these protists has are instead likely to be preyed on by protists (23, 39).

Kashiyama et al. PNAS | October 23, 2012 | vol. 109 | no. 43 | 17333 Downloaded by guest on September 29, 2021 Conclusions trifluoroacetic acid, and (iii) use of stabilizing agents. In particular, the The singlet oxygen generated when oxygen is photosensitized by addition of imidazole in the mobile phase dramatically improved the quantitative analysis of cPPB-aE using HPLC. In addition, cPPB-aE was found to be chlorophylls is probably a major concern among the organisms particularly stabilized in anisole solution. Therefore, the use of the standard living on the oxygenated and illuminated surface of the Earth. solution in anisole permitted accurate calibration on the HPLC analysis (SI However, all organisms have evolved to somehow protect them- Text, section 2.2). Consequently, the current methods significantly improved selves against the potential phototoxicity of the chlorophylls in the quantitative analysis of cPPB-aE(Fig. S7, Fig. S8, and SI Text, section 2.3). their respective environments. However, the significance of these Detection limit of cPPB-aE was ∼30 fmol per injection on the current detoxification processes is only readily apparent in such striking HPLC method. manifestations as autumnal tints. In the present work, we have Analytical HPLC (Fig. S9) was performed using a Shimadzu Prominence liquid shown a cryptic process of chlorophyll detoxification that is widely chromatograph system, which comprised a CBM-20A communications bus distributed among the SAR and CCTH protists, where colored but module, a DGU-20A3 degasser, two LC-20AD pumps constituting a binary nonfluorescent cPPB-aE is the catabolite of Chl-a. The ubiquitous pumping system, an SIL-20AAC auto sampler, a CTO-20AC column oven, and an SPD-M20Avp diode array detector. The system was coupled to a personal occurrence of cPPB-aE in all aquatic environments indicates that fi a fi computer con gured to run Shimadzu LC Solution software. All solvents used cPPB- E catabolism is one of the major chlorophyll detoxi cation for the analytical HPLC were of HPLC-grade quality, and they were purchased mechanisms in the modern global ecosystem. from Nacalai Tesque. The standard solutions of various Chl-a derivatives used in If we consider cPPB-aE a biomarker of protistan activity, our the following analytical HPLCs were prepared using anisole (SI Text, section 2.2). results suggest the quantitative importance of herbivory by the Analytical HPLC involved the use of a reverse-phase monomeric column SAR and CCTH protists in aquatic environments. Picophyto- ZORBAX Eclipse Plus C18 (4.6 × 30 mm, 1.8-μm silica particle size; Rapid Reso- planktons are particularly likely to be an important prey. Although lution HT). The solvent gradient program used is summarized in Table S5,in- the microbial loop tends to be pictured as a material flow starting cluding the conditions used to precondition the column. Because only a binary automated programming technique with two pumps is available in our system, from bacteria and archaea that are fed on POM or dissolved or- solvents B and C are introduced to the same second pump using a switching ganic matter (i.e., protistan bacterivory), nano- and microplanktonic valve in the line. Therefore, switching of the solvents and purging of the pump protists should also play primary roles in the picophytoplankton- system by the solvent were performed promptly within 1 min after pre- based microbial loop (37, 40, 41) that, hence, makes substantial conditioning of the column. All three solvents were degassed in vacuo with contributions to the flow of carbon and energy in aquatic envi- ultrasonication and sealed under argon. The solvent reservoir bottles were ronments (42). Additional studies of cPPB-aE, including its suit- specially customized for the convenience of degassing or sealing with argon as ability as a proxy for protistan herbivory, may provide approaches well as preventing the solvent from contacting the air during analysis. The flow −1 for quantifying the contribution of picophytoplankton to flux in rate of the mobile phase was 1.00 mL min . The column oven was set to 25 °C. the aquatic geochemical cycle. The auto sampler tray was set to 15 °C. Given that standard solutions and a samples were prepared in anisole, which is not a component of the eluent, the A better understanding of cPPB- E metabolism within and injection volume was generally 1.00 μL or smaller. outside the SAR and CCTH clades of protists may provide valu- able insights into the evolutionary origin of the modern aquatic Feeding Experiments. Four combinations of heterotrophic protists with algae, ecosystem, founded largely on protistan herbivory. Feeding on a combination of a crustacean zooplankton with an alga, and a combination of photosynthetic organisms in situ in the presence of molecular a bacterium with plant materials were examined. Experiments, thus, comprise six oxygen and light, would have enabled much more efficient bio- feeding experiments as well as six control experiments without heterotrophs/ geochemical recycling and bioenergetic redistribution within envi- crustacean zooplankton/bacteria. All protistan strains as well as the copepod ronments near the water surface. Interestingly, chemical structures Daphnia magna were maintained by feeding unialgal clones (Table S1). After of fossil biomarkers that could be derived from cPPB-aEhavebeen certain incubation periods, these strains were collected on sterile glass microfiber filters (GF/F grade; 47 mm ϕ; Whatman), which were then extracted and reported from various aquatic sedimentary rocks as old as the Early ca. – analyzed according to the analytical procedures used for natural samples de- Jurrasic ( 183 Ma) (Fig. S5)(10,43 50). Interestingly, fossil- scribed below. Information including specific names of organisms examined, based evidence suggests that this timing correlates roughly with their phylogenetic position, strain identities, and experimental conditions are the emergence of secondary algae containing red algae-derived summarized in Table S1. plastids. These plastids include phototrophic protists within the SAR and CCTH clades, including dinoflagellates, coccolitho- Optical Spectrometry and Singlet Oxygen Detection Experiments. Electronic phores, and diatoms (51). An understanding of aquatic microbial absorption spectra were measured using a Hitachi U-3500 spectrophotometer. herbivory gleaned from combining biochemical, molecular Fluorescence spectra were measured using a Hitachi F-4500 spectrophotometer, biological, and geochemical evidences would provide critical and fluorescence quantum yields were obtained using a photoluminescence insights into the evolutionary dynamics of the aquatic protistan method with an absolute photoluminescence quantum yield measurement system model C9920-02 comprising an excitation xenon light source, a lineages as well as the global geochemical cycles. monochromator, an integral sphere, and a multichannel CCD spectrometer (Hamamatsu Photonics). OD was about 1.0/10 mm at the Soret absorption Materials and Methods maximum in both anisole and tetrahydrofuran (THF) for electronic absorption Preparation of Authentic Samples. Schemes for preparation of authentic measurements. THF used for spectroscopy was distilled from a regent pur- samples of Chl-a, Phe-a, pPhe-a, pPPB-a, cholesteryl pPPB-a, cPPB-aE, and (R/S)- chased from Nacalai Tesque. Anisole (RegentPlus grade) and tert-butyl hCPLs-a are summarized in Fig. S6. Additional details are described in SI Text, methyl ether (ACS regent grade) were purchased from Sigma-Aldrich. All of section 2.1. the other solvents used herein were spectrometry-grade regents purchased from Nacalai Tesque, and they were used without additional purification. In Development of HPLC Methods. We developed improved analytical methods optical spectroscopy, absorption and emission properties of chlorophyll that enable quantitative identification of unstable cPPB-aE from microbio- derivatives in THF and anisole are listed in Table S2. In experiments involving logical and environmental samples with high sensitivity. Analytical difficul- cPPB-aE dissolved in THF, we, therefore, prepared the solution under argon ties arose as a consequence of the instability of cPPB-aE during handling in freshly distilled THF and measured all optical properties immediately. In during extraction and analysis involving HPLC (11, 13, 17). Given that cPPB- singlet oxygen detection experiments, generation of singlet oxygen was de- aE is rather unstable in most organic solvents in the presence of molecular tected using SOSG (Invitrogen), a commercially available fluorescent probe oxygen, it was rapidly degraded, especially when present at low concen- (52). A chlorophyll derivative (10 μM) and SOSG (1 μM) were dissolved in trations. Our results, thus, depend largely on the development of improved anisole and methanol (1:1, vol/vol), placed in a quartz cell, irradiated with analytical methods that required the availability of the semisynthesized red light that was provided by a 250 W metal halide lamp (LS-250–7500; authentic standard. In short, the analysis required (i) careful removal of Sumita Optical Glass), and passed through a colored glass filter that failed to molecular oxygen from the solvents for extraction and the mobile phases of transmit light with a wavelength shorter than ∼630 nm (AGC Techno Glass). HPLC, (ii) use of an end-capped reverse-phase HPLC column as well as its Therefore, the irradiated light selectively excited chlorophyll derivatives but deactivation by preconditioning using mobile phase with 0.5% (wt/vol) not SOSG. The aerated solutions were continuously stirred during irradiation

17334 | www.pnas.org/cgi/doi/10.1073/pnas.1207347109 Kashiyama et al. Downloaded by guest on September 29, 2021 experiments using a magnetic stirrer. The increase in ﬂuorescence intensity a glove bag. Acetone and anisole were carefully degassed and purged with

(i.e., initial fluorescence ascribable to the purchased SOSG was subtracted) argon gas before use. The sample dissolved in anisole was then transferred FEATURE ARTICLE was plotted against the irradiation time to show and compare the relative to the vial for the autosampler of the HPLC system. abilities of singlet oxygen generation between different Chl-a derivatives as well as a control SOSG solution without the addition of any photosensitizers ACKNOWLEDGMENTS. We thank the crew and scientific party of the (Fig. 4C). Training/Research Vessel Seisui-maru (cruise number 1127) and Research Vessel Hakken. We also thank Mr. Hokuto Tsutamoto and Dr. Nobutaka Analytical Procedures Used for Natural Samples. Water samples were filtered Imamura for kindly supplying the clone of D. pulex and Dr. Hiroshi Endoh using a glass microfiber filter (GF/F grade; 47 mm ϕ; Whatman). The filters for favorably providing the clone of E. aerogenes; Mr. Ryuzou Narita, Dr. were then stored below −20 °C before analysis. Sediment samples were Satoshi Ohkubo, Ms. Hiroko Usui, Mr. Koichi Fujita, and Dr. Shinnosuke Machida for technical assistance; and Dr. Jun Yokoyama and Dr. Jonathan stored at 4 °C before analysis. A wet filter sample (containing POM) or a wet J. Tyler for critical comments. This study was supported in part by Grants-in- sediment sample was extracted in acetone and ultrasonicated for 5 min at Aid for Scientific Research 23870028 (to Y. Kashiyama), 20570081 (to A.Y.), SEE COMMENTARY 0 °C in the dark, with the extraction and sonication procedures repeated 24657060 (to A.Y.), 21247005 (to H.M.), and 21247010 (to I.I.), a Research a total of three times. The combined extracts were then dried under Fellowship for Young Scientists (to Y. Kashiyama), and a grant from the a stream of argon in the dark. The residue was then dissolved in anisole. All Ritsumeikan Global Innovation Research Organization (to Y. Kashiyama of the above operations were performed in an argon atmosphere using and H.T.).

1. Wagner D, et al. (2004) The genetic basis of singlet oxygen-induced stress responses of 27. Ting CS, Rocap G, King J, Chisholm SW (2002) Cyanobacterial photosynthesis in the Arabidopsis thaliana. Science 306:1183–1185. oceans: The origins and significance of divergent light-harvesting strategies. Trends 2. Farquhar J, Zerkle AL, Bekker A (2011) Geological constraints on the origin of oxy- Microbiol 10:134–142. genic photosynthesis. Photosynth Res 107:11–36. 28. Yooseph S, et al. (2010) Genomic and functional adaptation in surface ocean plank- 3. Hörtensteiner S, Kräutler B (2011) Chlorophyll breakdown in higher plants. Biochim tonic prokaryotes. Nature 468:60–66. Biophys Acta 1807:977–988. 29. Andersen RA, Bidigare RR, Keller MD, Latasa M (1996) A comparison of HPLC pigment 4. Falk H, Hoornaert G, Isenring HP, Eschenmoser A (1975) Enol derivatives in the signatures and electron microscopic observations for oligotrophic waters of the North chlorophyll series. Preparation of 132,173-cyclopheophorbide enols. Helv Chim Acta Atlantic and Pacific Oceans. Deep Sea Res Part 1 Oceanogr Res Pap 43:517–537. 58:2347–2357. 30. Lovejoy C, Massana R, Pedrós-Alió C (2006) Diversity and distribution of marine mi- 5. Ma L, Dolphin D (1996) Stereoselective synthesis of new chlorophyll a related anti- crobial eukaryotes in the Arctic Ocean and adjacent seas. Appl Environ Microbiol 72: oxidants isolated from marine organisms. J Org Chem 61:2501–2510. 3085–3095. 6. Louda JW, Li J, Liu L, Winfree MN, Baker EW (1998) Chlorophyll-a degradation during 31. Not F, et al. (2008) Phytoplankton diversity across the Indian Ocean: A focus on the – cellular senescence and death. Org Geochem 29:1233–1251. picoplanktonic size fraction. Deep Sea Res Part 1 Oceanogr Res Pap 55:1456 1473. 7. Louda JW, Liu L, Baker EW (2002) Senescence- and death-related alteration of 32. Not F, del Campo J, Balagué V, de Vargas C, Massana R (2009) New insights into the chlorophylls and carotenoids in marine phytoplankton. Org Geochem 33:1635–1653. diversity of marine picoeukaryotes. PLoS One 4:e7143. MICROBIOLOGY 8. Louda JW, Mongkhonsri P, Baker EW (2011) Chlorophyll degradation during senes- 33. Liu H, et al. (2009) Extreme diversity in noncalcifying haptophytes explains a major – cence and death-III: 3–10 yr experiments, implications for ETIO series generation. Org pigment paradox in open oceans. Proc Natl Acad Sci USA 106:12803 12808. Geochem 42:688–699. 34. Cuvelier ML, et al. (2010) Targeted metagenomics and ecology of globally important – 9. Aydin N, Daher S, Gülaçar FO (2003) On the sedimentary occurrence of chlorophyllone uncultured eukaryotic phytoplankton. Proc Natl Acad Sci USA 107:14679 14684. a. Chemosphere 52:937–942. 35. Moon-van der Staay SY, De Wachter R, Vaulot D (2001) Oceanic 18S rDNA sequences – 10. Mawson DH, Keely BJ (2008) Novel functionalised chlorins in sediments of the Mes- from picoplankton reveal unsuspected eukaryotic diversity. Nature 409:607 610. 36. Baldauf SL (2003) The deep roots of eukaryotes. Science 300:1703–1706. sinian Vena del Gesso evaporitic sequence: Evidence for a facile route to reduction for 37. Jürgens K, Massana R (2008) Protistan grazing on marine bacterioplankton. Microbial biomarkers. Org Geochem 39:203–209. Ecology of the Oceans, ed Kirchman DL (Wiley, New York), 2nd Ed, pp 383–441. 11. Louda JW, Loitz JW, Rudnick DT, Baker EW (2000) Early diagenetic alteration of 38. Worden AZ, Not F (2008) Ecology and diversity of picoeukaryotes. Microbial Ecology chlorophyll-a and bacteriochlorophyll-a in a contemporaneous marl ecosystem; Flor- of the Oceans, ed Kirchman DL (Wiley, New York), 2nd Ed, pp 159–205. ida Bay. Org Geochem 31:1561–1580. 39. Sherr E, Sherr B (2008) Understanding roles of microbes in marine pelagic food webs: 12. Scheer H, Katz JJ (1978) Peripheral metal-complex - Chlorophyll isomers with mag- A brief history. Microbial Ecology of the Oceans, ed Kirchman DL (Wiley, New York), nesium bound to ring-E beta-keto ester system. J Am Chem Soc 100:561–571. 2nd Ed, pp 27–44. 13. Khalesi M, Louda JW (2011) Hemisynthesis of 132,173-cyclomesopheophorbide-a-enol. 40. Sherr EB, Sherr BF (1988) Role of microbes in pelagic food webs: A revised concept. Tetrahedron Lett 52:1078–1081. Limnol Oceanogr 33:1225–1227. 14. Karuso P, et al. (1987) 132,173-Cyclopheophorbide enol, the first porphyrin isolated 41. Sherr EB, Sherr BF (1994) Bacterivory and herbivory: Key roles of phagotrophic protists from a sponge. Tetrahedron Lett 27:2177–2178. in pelagic food webs. Microb Ecol 28:223–235. 15. Ma L, Dolphin D (1999) The metabolites of dietary chlorophylls. Phytochemistry 50: 42. Richardson TL, Jackson GA (2007) Small phytoplankton and carbon export from the 195–202. surface ocean. Science 315:838–840. 16. Louda JW, Neto RR, Magalhaes ARM, Schneider VF (2008) Pigment alterations in the 43. Junium CK, Mawson DH, Arthur MA, Freeman KH, Keely BJ (2008) Unexpected oc- brown mussel Perna perna. Comp Biochem Physiol B Biochem Mol Biol 150:385–394. currence and significance of zinc alkyl porphyrins in Cenomanian–Turonian black 17. Goericke R, Strom SL, Bell MA (2000) Distribution and sources of cyclic pheophorbides shales of the Demerara Rise. Org Geochem 39:1081–1087. – in the marine environment. Limnol Oceanogr 45:200 211. 44. Junium CK, Keely BJ, Freeman KH, Arthur MA (2011) Chlorins in mid-Cretaceous black 18. Ocampo R, Sachs JP, Repeta DJ (1999) Isolation and structure determination of the shales of the Demerara Rise: The oldest known occurrence. Org Geochem 42:856–859. 2 3 unstable 13 ,17 -Cyclopheophorbide a enol from recent sediments. Geochim Cos- 45. Keely BJ, et al. (1994) Porphyrin and chlorin distributions in a Late Pliocene lacustrine – mochim Acta 63:3743 3749. sediment. Geochim Cosmochim Acta 58:3691–3701. 19. Kusch S, et al. (2010) Implications for chloro- and pheopigment synthesis and pres- 46. Sundararaman P, Boreham CJ (1993) Comparison of nickel and vanadyl porphyrin fi δ13 δ15 Δ14 ervation from combined compound-speci c C, N, and Canalysis.Biogeosci distributions of sediments. Geochim Cosmochim Acta 57:1367–1377. – 7:4105 4118. 47. Kashiyama Y, Ogawa NO, Nomoto S, Kitazato H, Ohkouchi N (2010) Nitrogen and 20. Kashiyama Y, et al. (2008) Evidence of global chlorophyll d. Science 321:658. carbon isotopic compositions of copper, nickel, and vanadyl porphyrins in Cretaceous 21. Cavalier-Smith T (2010) Kingdoms Protozoa and Chromista and the eozoan root of black shales. Earth, Life, and Isotopes, eds Ohkouchi N, Tayasu I, Koba K (Kyoto the eukaryotic tree. Biol Lett 6:342–345. University Press, Kyoto), pp 313–335. 22. Walker G, Dorrell RG, Schlacht A, Dacks JB (2011) Eukaryotic systematics: A user’s 48. Prowse WG, Chicarelli MI, Keely BJ, Kaur S, Maxwell JR (1987) Characterisation of guide for cell biologists and parasitologists. Parasitology 138:1638–1663. fossil porphyrins of the “di-DPEP” type. Geochim Cosmochim Acta 51:2875–2877.

23. Massana R (2011) Eukaryotic picoplankton in surface oceans. Annu Rev Microbiol 65: 49. Sundararaman P, et al. (1994) Temporal changes in the distribution of C33cyclo- – 91 110. heptenoDPEP and 17-nor C30DPEP in rocks. Org Geochem 21:1051–1058. 24. Vaulot D, Eikrem W, Viprey M, Moreau H (2008) The diversity of small eukaryotic 50. Kashiyama Y, et al. (2008) Reconstruction of the biogeochemistry and ecology of phytoplankton (< or =3 microm) in marine ecosystems. FEMS Microbiol Rev 32: photoautotrophs based on the nitrogen and carbon isotopic compositions of vanadyl 795–820. porphyrins from Miocene siliceous sediments. Biogeosci 5:797–816. 25. Azam F, et al. (1983) The ecological role of water-column microbes in the sea. Mar 51. Falkowski PG, et al. (2004) The evolution of modern eukaryotic phytoplankton. Sci- Ecol Prog Ser 10:257–263. ence 305:354–360. 26. Field CB, Behrenfeld MJ, Randerson JT, Falkowski P (1998) Primary production of the 52. Gollmer A, et al. (2011) Singlet Oxygen Sensor GreenÒ: Photochemical behavior in biosphere: Integrating terrestrial and oceanic components. Science 281:237–240. solution and in a mammalian cell. Photochem Photobiol 87:671–679.

Kashiyama et al. PNAS | October 23, 2012 | vol. 109 | no. 43 | 17335 Downloaded by guest on September 29, 2021