INFORMATION FOR AUTHORS Full details of how to submit a manuscript for publication in Natural Product Communications are given in Information for Authors on our Web site http://www.naturalproduct.us. Authors may reproduce/republish portions of their published contribution without seeking permission from NPC, provided that any such republication is accompanied by an acknowledgment (original citation)-Reproduced by permission of Natural Product Communications. Any unauthorized reproduction, transmission or storage may result in either civil or criminal liability. The publication of each of the articles contained herein is protected by copyright. Except as allowed under national “fair use” laws, copying is not permitted by any means or for any purpose, such as for distribution to any third party (whether by sale, loan, gift, or otherwise); as agent (express or implied) of any third party; for purposes of advertising or promotion; or to create collective or derivative works. Such permission requests, or other inquiries, should be addressed to the Natural Product Inc. (NPI). A photocopy license is available from the NPI for institutional subscribers that need to make multiple copies of single articles for internal study or research purposes. To Subscribe: Natural Product Communications is a journal published monthly. 2015 subscription price: US$2,595 (Print, ISSN# 1934-578X); US$2,595 (Web edition, ISSN# 1555-9475); US$2,995 (Print + single site online); US$595 (Personal online). Orders should be addressed to Subscription Department, Natural Product Communications, Natural Product Inc., 7963 Anderson Park Lane, Westerville, Ohio 43081, USA. Subscriptions are renewed on an annual basis. Claims for nonreceipt of issues will be honored if made within three months of publication of the issue. All issues are dispatched by airmail throughout the world, excluding the USA and Canada.

NPC Natural Product Communications

EDITOR-IN-CHIEF

DR. PAWAN K AGRAWAL Natural Product Inc. 7963, Anderson Park Lane, Westerville, Ohio 43081, USA agrawal@naturalproduct.us

EDITORS

PROFESSOR ALEJANDRO F. BARRERO Department of Organic Chemistry, University of Granada, Campus de Fuente Nueva, s/n, 18071, Granada, Spain afbarre@ugr.es

PROFESSOR ALESSANDRA BRACA Dipartimento di Chimica Bioorganicae Biofarmacia, Universita di Pisa, via Bonanno 33, 56126 Pisa, Italy braca@farm.unipi.it

PROFESSOR DE-AN GUO State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences, Peking University, Beijing 100083, China gda5958@163.com

PROFESSOR YOSHIHIRO MIMAKI School of Pharmacy, Tokyo University of Pharmacy and Life Sciences, Horinouchi 1432-1, Hachioji, Tokyo 192-0392, Japan mimakiy@ps.toyaku.ac.jp

PROFESSOR STEPHEN G. PYNE Department of Chemistry University of Wollongong Wollongong, New South Wales, 2522, Australia spyne@uow.edu.au

PROFESSOR MANFRED G. REINECKE Department of Chemistry, Texas Christian University, Forts Worth, TX 76129, USA m.reinecke@tcu.edu

PROFESSOR WILLIAM N. SETZER Department of Chemistry The University of Alabama in Huntsville Huntsville, AL 35809, USA wsetzer@chemistry.uah.edu

PROFESSOR YASUHIRO TEZUKA Faculty of Pharmaceutical Sciences Hokuriku University Ho-3 Kanagawa-machi, Kanazawa 920-1181, Japan y-tezuka@hokuriku-u.ac.jp

PROFESSOR DAVID E. THURSTON Department of Pharmacy and Forensic Science, King’s College London, Britannia House, 7 Trinity Street, London SE1 1DB, UK. david.thurston@kcl.ac.uk

ADVISORY BOARD

Prof. Viqar Uddin Ahmad Karachi, Pakistan

Prof. Giovanni Appendino Novara, Italy

Prof. Yoshinori Asakawa Tokushima, Japan

Prof. Roberto G. S. Berlinck São Carlos, Brazil

Prof. Anna R. Bilia Florence, Italy

Prof. Maurizio Bruno Palermo, Italy

Prof. César A. N. Catalán Tucumán, Argentina

Prof. Josep Coll Barcelona, Spain

Prof. Geoffrey Cordell Chicago, IL, USA

Prof. Fatih Demirci Eskişehir, Turkey

Prof. Ana Cristina Figueiredo Lisbon, Portugal

Prof. Cristina Gracia-Viguera Murcia, Spain

Dr. Christopher Gray Saint John, NB, Canada

Prof. Dominique Guillaume Reims, France

Prof. Duvvuru Gunasekar Tirupati, India

Prof. Hisahiro Hagiwara Niigata, Japan

Prof. Tsukasa Iwashina Tsukuba, Japan

Prof. Leopold Jirovetz Vienna, Austria

Prof. Vladimir I Kalinin Vladivostok, Russia

Prof. Phan Van Kiem Hanoi, Vietnam

Prof. Niel A. Koorbanally Durban, South Africa

Prof. Chiaki Kuroda Tokyo, Japan

Prof. Hartmut Laatsch Gottingen, Germany

Prof. Marie Lacaille-Dubois Dijon, France

Prof. Shoei-Sheng Lee Taipei, Taiwan

Prof. Imre Mathe Szeged, Hungary

Prof. M. Soledade C. Pedras Saskatoon, Canada

Prof. Luc Pieters Antwerp, Belgium

Prof. Peter Proksch Düsseldorf, Germany

Prof. Phila Raharivelomanana Tahiti, French Polynesia

Prof. Luca Rastrelli Fisciano, Italy

Prof. Stefano Serra Milano, Italy

Prof. Monique Simmonds Richmond, UK

Dr. Bikram Singh Palampur, India

Prof. John L. Sorensen Manitoba, Canada

Prof. Johannes van Staden Scottsville, South Africa

Prof. Valentin Stonik Vladivostok, Russia

Prof. Winston F. Tinto Barbados, West Indies

Prof. Sylvia Urban Melbourne, Australia

Prof. Karen Valant-Vetschera Vienna, Austria

HONORARY EDITOR

PROFESSOR GERALD BLUNDEN The School of Pharmacy & Biomedical Sciences,

University of Portsmouth, Portsmouth, PO1 2DT U.K.

axuf64@dsl.pipex.com

Revisiting Caffeine Biosynthesis – Speculations about the Proximate Source of its Purine Ring Thomas W. Baumann villacoffea, Dorfstr. 17, CH–8247 Flurlingen, Switzerland thomas.baumann@uzh.ch

Received: February 27th, 2015; Accepted: March 1st, 2015

The prevailing hypothesis of caffeine biosynthesis starting from xanthosine was combined with Kremers’ speculation on NAD as a biochemical precursor of caffeine and trigonelline in coffee. This bold sketch together with a few free-spirited ideas may channel future caffeine biosynthesis studies into novel directions. Keywords: Caffeine, Trigonelline, Biosynthesis, NAD, Metabolic channeling.

Introduction

When looking at textbook chapters and scientific articles reviewing caffeine biosynthesis, we get the impression that this issue has been definitely resolved by the many research groups which chopped up coffee, cocoa , tea, mate, guaraná, or cola in order to track the traces left by radioactive precursors, or to isolate and test enzyme activities. The results have been summarized in a large number of reviews. I refrain here from listing the citations. They can be found elsewhere in this special issue. The today’s ‘caffeine canon’ reads as follows: Xanthosine → 7–methylxanthosine → 7–methylxanthine → 3,7–dimethylxanthine (theobromine) → 1,3,7–trimethylxanthine (caffeine) What is wrong about this sequence? Nothing, unless one postulates that (a) in vitro findings deserve in vivo verification, and (b) a pathway of a plant defense compound must somehow be integrated into the plant physiological processes. For instance, caffeine biosynthesis is strongly (≧ 20x) stimulated by light, and in many experiments we recognize that purine alkaloids (PuA) and chlorogenic acids are influencing each other’s biosynthesis. Frankly, the four-step reaction mentioned above does not provide a firm grip for a plant physiologist. It looks as if caffeine research has not yet plumbed the depths of plant metabolism, even though a large number of researchers have made excellent studies on this matter. In this review, I shall first go back to the now already historical times when studies on caffeine biosynthesis started. Then I will summarize some old and recent hypotheses on it. Finally, I will point to a few findings and publications which were neglected during the search for the needle in the purine haystack, and take those to bring forward new ideas regarding the biogenesis of the most famous stimulant on our globe. Even though all is hypothetical, I hope it will be a modest contribution to ending the impasse.

A Bit of History

Back in my day, the professors were very keen to put, when lecturing, their research into a historical context. That is why Linné,

Lamarck, Darwin, Mendel, Kossel, Emil Fischer, Dobzhansky, Chargaff, and Khorana were for us students like old fellas. Even special lectures on history of science were offered by the faculty. The advantage of plunging into history is threefold: It broadens the horizon, prevents from reinventing the wheel, and last but not least, befits the image of a scientist or politician. At the same time when Sam Cooke (1931–1964) sang his famous “Don’t Know Much About History” (Wonderful World Lyrics), Louise E. Anderson set out to tackle the origin of caffeine’s purine ring. This was at Cornell in the lab of Martin Gibbs (1922–2006), and Louise was his first graduate student. They published the results of these primary studies in 1962 [1]:

“It was concluded that the xanthine base of caffeine is synthesized from the same precursors as are purines in all of the other systems which have been studied. The xanthine is converted to 7 (or 3)-methylxanthine and, in turn, to theobromine and caffeine. The apparent low incorporation into the N-7 methyl carbon atom would suggest an incorporation of a methyl group into this position relatively early in the formation of caffeine. The primary effect of light is an enhancement of purine ring formation.”

More than one decade later, studies on this issue were resumed, coincidentally at two places far away from each other: (a) During his doctoral thesis in the lab of Eiichi Takahashi (1927–) at Kyoto University in Japan, Takeo Suzuki prepared and tested enzyme extracts from leaves of young, rapidly growing tea seedlings. Their catalytic properties were published in 1975 [2]:

“Extracts prepared from tea leaves with Polyclar AT (…) contained two methyltransferase activities catalysing the transfer of methyl groups from S-adenosylmethionine to 7-methylxanthine, producing theobromine, and to theobromine, producing caffeine. (…). Paraxanthine was shown to be most active among methylxanthines, as the methyl acceptor. (…). Xanthine,

NPC Natural Product Communications 2015 Vol. 10 No. 5

793 - 797

794 Natural Product Communications Vol. 10 (5) 2015 Baumann

xanthosine, XMP and hypoxanthine were all inactive as methyl acceptors, (….). Little is known about the pathways leading to the formation of 7-methylxanthine. A good correlation between caffeine synthesis and shoot formation or growth of tea seedlings was shown, suggesting that the methylating systems in caffeine synthesis are closely associated with purine nucleotide and nucleic acid metabolism in tea plants.”

(b) A completely different approach was chosen at the University of Zurich in Switzerland in the lab of Hans Wanner (1917–2004), where my first PhD student Elsbeth Looser infiltrated coffee leaf disks with a possible ‘cold’ precursor at high concentration and simultaneously with L-methionine-[methyl-14C]. The outcome was published in 1974 [3]:

“The results permit the identification of theobromine, 7-methylxanthine and 7-methylxanthosine as precursors of caffeine. 7-Methylguanosine seems not to be an intermediate in caffeine formation. (…) The synthesis of radioactive caffeine is increasingly stimulated by 7-methylxanthosine, 7-methylxanthine and 3,7- dimethylxanthine, in this order.”

In essence, both studies [(a) and (b)] excluded xanthosine as a precursor of caffeine: It was neither an acceptor (to be methylated) in the enzymatic studies with Camellia sinensis extracts nor in the in situ stimulation experiments with Coffea arabica leaf disks. They also did not confirm the first report on the ‘natural existence’ of 7–methylxanthosine: Dolapo B. A. Ogutuga, the later inventor of the cola wine, undertook, in the lab of Donald Henry Northcote (1922–2004) at Cambridge University, UK, some pioneering studies on tea callus cultures [4]. The results remain difficult to interpret, but Ogutuga refers to his dissertation work (1969), in which he had mentioned “the presence of 7-methylxanthosine in the supernatant of homogenized tea leaves centrifuged at 105,000g”. I have not yet had the chance to validate this claim. Yet, a decade after the experiments in Kyoto and Zurich, Osamu Negishi published, together with Tetsuo Ozawa and Hiroshi Imagawa, in two ‘Rapid Papers’ and in a ‘Note’ the findings he had collected during his thesis at Tsukuba University in Japan. Based upon (i) feeding excised tea shoots with labeled methionine [Me-14C], xanthosine [2-14C], and 7-methylxanthosine [2-14C] and (ii) separating the products by paper [sic!] chromatography, the following was suggested in the first ‘Rapid Paper’[5]:

‘The methyl group of methionine was incorporated into 7-methylxanthosine (ca. 10%) in the earlier period of incubation after the uptake. About 50% of the radioactivity of xanthosine was rapidly incorporated into caffeine via 7-methylxanthosine, 7-methylxanthine, and theobromine within 24 h. 7-Methylxanthosine was also converted into caffeine at a high rate. The results suggest that the pathway for caffeine biosynthesis is as follows: xanthosine → 7-methylxanthosine → 7-methylxanthine → theobromine → caffeine’.

In the second ‘Rapid Paper’, the authors presented their studies with enzyme extracts from young tea shoots challenged with ‘cold’ xanthosine and [14CH3]SAM [6].

‘To identify the reaction product of methylation by N-methyltransferase, paper chromatographic analysis of

the product and its hydrolysate was done and it was demonstrated that xanthosine was methylated at the N-7 position. The cell-free extracts reproduced [in one and the same assay] all the reactions in the pathway from xanthosine to caffeine, as shown in Figure 1, indicating that they contain the complete enzyme system involved in the biosynthesis of caffeine, that is, three N-methyltransferase activities and a nucleosidase or nucleoside phosphorylase activity.”

Finally in the ‘Note’ the authors reported about similar experiments on young coffee shoots (Coffea arabica) and enzyme extracts from their leaves [7]:

‘Radioactivity was detected in the spot corresponding to 7-methylxanthosine, when xanthosine was used as a methyl acceptor. A small amount of 7-methylxanthine was also found. To confirm the site of methylation in the xanthine ring of the reaction product, the reaction mixture was treated with 1 N HCl at 100°C for 1 h. The hydrolysis yielded only 7-methylxanthine.”

Thus, the researchers also showed for coffee, that xanthosine can act as the first methyl group acceptor of caffeine biosynthesis. However, coffee enzyme extracts did not catalyze the entire reaction chain up to caffeine as was noticed in the related experiment on tea.

The Dilemma – How to Escape it

Incongruity As outlined in the above section, prior to the publications of Negishi et al. in 1985 [5-7] neither the group at Kyoto nor the one at Zurich succeeded in showing that xanthosine played a role in caffeine biosynthesis. Even worse, the in situ existence of 7-methylxanthosine could not be proven in the follow-up experiments by Brigitte Schulthess, my PhD student at the University of Zurich: Caffeine- and non-caffeine-producing coffee cell suspension cultures obtained a ‘biosynthetic stimulus’ by ethephon or/and adenine and were labelled with either [14C]adenine or [methyl-14C]methionine. The cells were analyzed by HPLC-DAD and flow-through radio-detection allowing inter alias the calculation of specific activity. In essence, besides the absence of 7-methylxanthosine in all experimental sets, the specific activity of xanthosine did not fit into the precursor role assigned to it. Photoperiod stimulated the biosynthesis of caffeine by a factor of more than twenty as compared with the control in the dark. Of course, chlorogenic acids also drastically increased during the photoperiod (from 0.1 mM to about 17 mM) [8,9].

The Adoption of Xanthosine Shortly after Negishi’s work, Ashihara’s group published two papers, in which the ‘xanthosine pathway’ was taken for granted, despite scanty evidence after feeding [14C]adenine to tea seedlings [10, 11]. However, in the next paper, an exciting study on purine alkaloids in Camellia flowers, the reaction scheme was omitted by the authors and luckily, we find the first note related to the dilemma [12]: ‘In our tracer experiments, little or no radioactivity was found in 7-methylxanthosine and 7-methylxanthine (data not shown)’. We should not rejoice too quickly because in a following paper in collaboration with Takeo Suzuki [13], the above-mentioned pathway remains undisputed, even though the key compounds were

Revisiting caffeine biosynthesis Natural Product Communications Vol. 10 (5) 2015 795

more than difficult to determine by paper chromatography. Instead of citing further examples we conclude that all the many in-situ-related studies failed to properly ‘round up and enqueue xanthosine and 7-methylxanthosine into the caffeine sequence’ of the live cell.

Metabolic Channeling Directs to the Last Resource As seen for many biosynthetic reactions, the enzymes of the de novo purine biosynthesis have also been shown to aggregate reversibly into a transient multi-enzyme complex, the purinosome [14]. Currently, there is an interesting debate about the nature and advantages of such compartmentalization [15, 16]. To escape from the deadlocked situation we had proposed in 1996 a caffeine biosynthesis route starting with the metabolically-channelled formation of 7-methyl-XMP [17] based on the well-known failure to label 7-methylxanthosine in situ and on the proof of a N7-specific XMP N-methyltransferase activity. However, experiments with a recombinant [18] protein named CmXRS1 showed methylation of xanthosine, but not of XMP and at first glance this may rule out the channeled pathway we had proposed. At second glance one must agree with these authors’ statement [18]: ‘The present results support the presence of the former pathway [starting with xanthosine], although we cannot exclude the possibility that unidentified CtCSs are related to the latter pathway, i.e., methylation of XMP’. This is the fatal crux. Irrespective of how pure either the enzyme protein or the substrate is, the in vitro test mirrors only one constellation out of many. In the target ‘caffeine plant’ we may find about one hundred potential related methyltransferases and an unattainably high number of possible substrates. This situation is like an equation with two unknown variables and leads us to the next section.

The superb substrates All who were wrestling with PuA biosynthesis came sooner or later across the phenomenon of paraxanthine (1,7-dimethylxanthine): It is the first compound which was detected as a superb precursor of caffeine in the many labs dealing with this pathway, be it in vivo or in vitro. Since it is not or merely formed from either of the two possible mono-methylxanthines, its presumed role was easily and officially debunked [19], but persists in patents aiming at efficiently producing caffeine. However, recently a caffeine synthase (PcCS) was found in young seeds of Paullinia cupana with higher affinity for theobromine than paraxanthine [20], indicating that seeking the appropriate (highly specific) methyltransferase in the two last steps of caffeine biosynthesis remains a task! It could well be that xanthosine is another superb substrate. Surprisingly and uniquely, it had been reported to be transformed in one and the same enzymatic assay through four steps into caffeine [6], which is really something! Conversely, we are faced with the in-situ setting, where one side records xanthosine’s role as meagre [5-7], and the other side that specific radioactivity rules out [9] its position attested by ex-situ experiments. As mentioned in the last section, the only way out is metabolic channeling, but it does not solve the problem of xanthosine: Is it a superb substrate like paraxanthine, to which one member of the scantily tested large methyltransferase family has been assigned [21]?

Off the Beaten Track

Instead of clinging to the ‘classical caffeine plants’ the research should be expanded to include lesser known PuA-producing

species. Historically, methyluric acids were discovered [22, 23], because the analytical focus was set on the ‘liberio-excelsoids’, a taxonomic group unimportant to consumers' needs. A challenging example may be Theobroma bicolor (patashte, pataxte, pataste). The young leaves of it accumulate in parallel to theobromine considerable concentrations of 1-methyluracil [24]. Is the methylated uracil a product of theobromine synthase activity and/or does it get us to hit the road to pseudouridylation in RNA [25-27] and methylation of Ψ [28]?

Old and New Hypotheses

Perhaps, Camargo [29] was the first who put a distinct-compound-related hypothesis into the ‘caffeine world’: In 1924 he reported about ‘relatively large quantities of vernine’ (=guanosine) in green coffee fruits. Leaves killed by frost contain more caffeine nitrogen than green leaves. Similarly, the so-called alloxuric bases’ nitrogen (mostly free guanine) is increased by frost. His keen hypothesis derived from few results reads as follows: ‘… formation of caffeine in the leaves and berries of the coffee tree is brought about by the transformation of guanosine to guanine, of the latter to xanthine, and lastly, of xanthine to caffeine, due to the action of the enzymes’. Quite recently, guanosine obtained new significance in caffeine biosynthesis by the discovery of an alternate catabolic route mediated by a plant specific deaminase which generates xanthosine [30]. Moreover, we should mention here the precursor role of guanosine in young tea leaves discovered by Ashihara and Suzuki in 1991 (unpublished): About 55% of the label of [8-14C]guanosine appears in theobromine and caffeine after 24 h; see Figure 3 in [31]. Roland E. Kremers (1894-1990), until 1951 Research Associate at General Foods Corporation in Hoboken, and thereafter staff member at the Institute of Paper Chemistry at Lawrence College in Appleton Wisconsin, designed, in 1954, for coffee a most intriguing hypothesis, in which similar reactions are thought to transform adenine and nicotinamide of NAD into caffeine and trigonelline, respectively [32]. In the following years, several researchers have focused on trigonelline metabolism. The first who examined Kremers’ hypothesis – which implies a stoichiometric (molar) caffeine/trigonelline ratio of 1 to 1 in the coffee bean – was Hans Kende (1937-2006), Professor at MSU-DOE Plant Research Laboratory, who studied for his thesis at our institute [33]. He showed, based on his own results and values from literature, that the two alkaloids neither in seeds nor in leaves of the coffee plant follow the postulated rule, but instead exhibit a strict reciprocal behavior: High trigonelline goes along with low caffeine and vice versa. Already in 1931, Nottbohm and Mayer noticed that high caffeine is always accompanied by low trigonelline [34]. Unfortunately, the values for the latter were systemically too low. In retrospect, unambiguous data supporting the ‘caffeine-trigonelline-rule’ were published by Hadorn and Suter in 1956 [35]. Finally we should keep in mind that Kremers’ values fitted badly in his own hypothesis. Luckily, this incongruence did not stop him from publishing his idea. In experiments from 1979 till 1983, my graduate student Peter Morath thoroughly followed the label from [8-14C]adenine and [L-methyl-14C]methionine to purine alkaloids in leaf disks and suspension cultures of Coffea arabica using HPLC and affinity chromatography for cis-diols. He synthesized 7-methylxanthosine monophosphate (5’), 7-methylinosine monophosphate (5’), MTA,

796 Natural Product Communications Vol. 10 (5) 2015 Baumann

MTR and their S-oxides, as well as [methyl-14C]7-XMP. Based upon the labeling kinetics and the scarce label in the methyl group of 7-methylxanthine (cf. also Anderson & Gibbs), he came to the conclusion that the starting point of PuA synthesis could be SAM: Two SAM molecules react with each other by transferring the amino-carboxypropyl group from one SAM molecule to the N7 position of the other. Subsequently, the acp3-modified adenine moiety is transformed to 7-methylxanthine [36]. This hypothesis was not tested further, but in 2001 Koshiishi et al. showed for tea, that a large fraction of the radioactivity of [adenosyl-14C] SAH is reappearing in purine alkaloids [37]. As plant physiologists we, Paulo Mazzafera and I, felt unhappy about the prevailing hypothesis of a caffeine pathway starting in the nowhere of the purine pool. Therefore, we discussed together alternative pathways. For example, since the cellular SAM concentration is highly regulated, excess SAM would be ‘detoxified’ by a short-circuit reaction deviating the purine ring and methyl groups into purine alkaloids. Of course, all the above-mentioned variants are highly speculative and could, before we are leaving this playground, easily be supplemented e.g. by including (a) nucleic acid methylation simultaneously considering adenosine-to-inosine editing [38, 39] or (b) cytokinin metabolism (evidence suggests that kinetin applied to coffee leaves was readily transformed into caffeine in [40]). Last but not least, one could aim at an ‘All-in-One-Approach’ by considering also the formation of isoguanosine/isoguanine [41] and its methylated derivatives, this uniting all PuA formed in higher plant clades in one and the same metabolic origin.

In Harmony and Disharmony with Prevailing Hypotheses, or Let us Reduce with NADH/NADPH!

Since paper does not blush, I have the courage to complete this article with a hypothesis based on Kremers’ idea, that NAD is processed in parallel steps to both PuA and trigonelline [32]. In Figure 1, the already mentioned interplay of trigonelline and caffeine is substantiated (Maro R. Sondahl, Michael Stutz and Thomas W Baumann, 2002, not published). The data suggest a

Figure 1: Correlation caffeine/trigonelline in green coffee; n = 3.

Figure 2: A highly speculative scheme of alkaloid formation in coffee. competition between the pathways leading to these compounds; the correlation is negative (r = –0.9394): High caffeine goes along with low trigonelline and vice versa, this covering an array of cultivars and species. The interpretation in metabolic terms may be difficult. For my part, I have not yet found the Columbus’ egg. However, the aim of this final section is firstly, to provide a backbone leading us to the xanthosine moiety compatible with the many in vitro observations discussed above. Secondly, it is suggested that all the enzyme activities (and the related proteins) as well as the cofactors mentioned below are clustered and so accelerate the processing of the nicotine amide and adenine residues while preventing intermediates from leaving the metabolic channel. Thirdly, the hypothesis was simplified as much as possible in a manipulative way and by ‘creating dual activities’ for at least one of the enzymes involved. In the following scheme (Figure 2) ‘NAD’ stands for all states and conditions of these cofactors, NADP included. In a first step (1,2), NAD is transformed to nicotinic acid-hypoxanthine-dinucleotide (NaHD) by the simultaneous action of nicotinamidase and a deaminase, the latter being effective specifically on adenosine derivatives as described by Kaplan et al. [42]. In a second step (3), the ‘hypoxanthine part’ is oxidized by a NAD+-dependent IMP dehydrogenase to yield nicotinic acid-xanthine-dinucleotide (NaXD) [43]. Finally (4), by the dual action of one and the same N–methyltransferase trigonelline and 7-methylxanthine are liberated, as inspired by [44]. For reasons of simplicity, possible salvage cycles [45] were omitted, but may eventually explain the asymmetric alkaloid formation (Figure 1). Therefore, I feel the urge to point to a recent work showing that a xanthosine phosphorylase is capable of using nicotinamide as a substrate for nicotinamide riboside formation [46]. Acknowledgments - I am thankful to Fredi Müller, Zurich, Switzerland, for his valuable mental assistance.

References

[1] Anderson L, Gibbs M. (1962) The biosynthesis of caffeine in the coffee plant. Journal of Biological Chemistry, 237, 1941–1944. [2] Suzuki T, Takahashi E. (1975) Biosynthesis of caffeine by tea-leaf extracts. Enzymic formation of theobromine from 7-methylxanthine and of

caffeine from theobromine. Biochemical Journal, 146, 87-96. [3] Looser E, Baumann TW, Wanner H. (1974) The biosynthesis of caffeine in the coffee plant. Phytochemistry, 13, 2515–2518. [4] Ogutuga DBA, Northcote DH (1970) Biosynthesis of caffeine in tea callus tissue. Biochemical Journal, 117, 715–720. [5] Negishi O, Ozawa T, Imagawa H. (1985) Conversion of xanthosine into caffeine in tea plants. Agricultural and Biological Chemistry, 49, 251–253. [6] Negishi O, Ozawa T, Imagawa H. (1985) Methylation of xanthosine by tea-leaf extracts and caffeine biosynthesis. Agricultural and Biological

Chemistry, 49, 887–890.

Revisiting caffeine biosynthesis Natural Product Communications Vol. 10 (5) 2015 797

[7] Negishi O, Ozawa T, Imagawa H. (1985) The role of xanthosine in the biosynthesis of caffeine in coffee plants. Agricultural and Biological Chemistry, 49, 2221–2222.

[8] Schulthess BH, Baumann TW. (1995) Stimulation of caffeine biosynthesis in suspension-cultured coffee cells and the in situ existence of 7-methylxanthosine. Phytochemistry, 38, 1381–1386.

[9] Schulthess BH, Baumann TW. (1995) Are xanthosine and 7-methylxanthosine caffeine precursors? Phytochemistry, 39, 1363–1370. [10] Ashihara H, Kubota H. (1986) Patterns of adenine metabolism and caffeine biosynthesis in different parts of tea seedlings. Physiologia Plantarum,

69, 275–281. [11] Ashihara H, Kubota H. (1987) Biosynthesis of purine alkaloids in Camellia plants. Plant & Cell Physiology, 28, 535–539. [12] Fujimori N, Ashihara H. (1990) Adenine metabolism and the synthesis of purine alkaloids in the flowers of Camellia. Phytochemistry, 29, 3513–

3516. [13] Fujimori N, Suzuki T, Ashihara H. (1991) Seasonal variations in biosynthetic capacity for the synthesis of caffeine in tea leaves. Phytochemistry,

30, 2245–2248. [14] Kyoung M, Russell SJ, Kohnhorst CL, Esemoto NN, An S. (2014) Dynamic architecture of the purinosome involved in human de novo purine

biosynthesis. Biochemistry, 54, 870−880. [15] Castellana M, Wilson MZ, Xu Y, Joshi P, Cristea IM, Rabinowitz JD, Gitai Z, Wingreen NS. (2014) Enzyme clustering accelerates processing of

intermediates through metabolic channeling. Nature Biotechnology, 32, 1011–1018. [16] Zhao A, Tsechansky M, Ellington AD, Marcotte EM. (2013) Revisiting and revising the purinosome. Molecular BioSystems, 10, 369–374. [17] Schulthess BH, Morath P, Baumann TW. (1996) Caffeine biosynthesis starts with the metabolically-channelled formation of 7-methyl-XMP – a

new hypothesis. Phytochemistry, 41, 169-175. [18] Mizuno K, Kato M, Irino F, Yoneyama N, Fujimura T, Ashihara H. (2003) The first committed step reaction of caffeine biosynthesis: 7–

methylxanthosine synthase is closely homologous to caffeine synthases in coffee (Coffea arabica L.) FEBS Letters, 547, 56–60. [19] Kato M, Mizuno K. (2004) Caffeine synthase and related methyltransferases in plants. Frontiers in Bioscience, 9, 1833–1842. [20] Schimpl FC, Kiyota E, Mayer JLS, Gonçalves JFC, da Silva JF, Mazzafera P. (2004) Molecular and biochemical characterization of caffeine

synthase and purine alkaloid concentration in guarana fruit. Phytochemistry, 105, 25–36. [21] McCarthy AA, McCarthy JG. (2007) The structure of two N-methyltransferases from the caffeine biosynthetic pathway. Plant Physiology, 144,

879–889. [22] Wanner H, Pesakova M, Baumann TW, Charubala R, Guggisberg A, Hesse M, Schmid H. (1975) O(2),1,9–Trimethyluric acid and 1,3,7,9–

tetramethyluric acid in leaves of different Coffea species. Phytochemistry, 14, 747–750. [23] Petermann JB, Baumann TW. (1983) Metabolic relations between methylxanthines and methyluric acids in Coffea L. Plant Physiology, 73,

961–964. [24] Bernet S. (2000) 1-Methyluracil in Theobroma bicolor and related species; its relationship to caffeine biosynthesis. Master Thesis University of

Zurich, Institute of Plant Biology (in German) [25] Jaffrey S. (2014) An expanding universe of mRNA modifications. Nature Structural & Molecular Biology, 21, 945–946. [26] Carlile TM, Rojas-Duran MF, Zinshteyn B, Shin H, Bartoli KM, Gilbert WV. (2014) Pseudouridine profiling reveals regulated mRNA

pseudouridylation in yeast and human cells. Nature, 515, 143–146. [27] Lovejoy AF, Riordan DP, Brown PO. (2014) Transcriptome-wide mapping of pseudouridines: Pseudouridine synthases modify specific mRNAs in

S. cerevisiae. PLoS ONE 9, e110799. [28] Wurm JP, Meyer B, Bahr U, Held M, Frolow O, Kötter P, Engels JW, Heckel A, Karas M, Entian K-D, Wöhnert J. (2010) The ribosome assembly

factor Nep1 responsible for Bowen–Conradi syndrome is a pseudouridine-N1-specific methyltransferase. Nucleic Acids Research, 38, 2387-2398. [29] Camargo T de A. (1924) The presence of vernine (guanosine) in the green leaves and berries of the coffee tree (Coffea arabica L.) and its relation

to the origin of caffeine in this plant. Journal of Biological Chemistry, 58, 831-834. [30] Dahncke K, Witte C-P. (2013) Plant purine nucleoside catabolism employs a guanosine deaminase required for the generation of xanthosine in

Arabidopsis. Plant Cell, 25, 4101–4109. [31] Suzuki T, Ashihara H, Waller GR. (1992) Purine and purine alkaloid metabolism in Camellia and Coffea plants. Phytochemistry, 91, 2575–2584. [32] Kremers RE. (1954) Speculation on DPN as a biochemical precursor of caffeine and trigonelline in coffee. Journal of American Pharmaceutical

Association, 43, 423-424. [33] Kende H. (1960) Untersuchungen über die biosynthese und bedeutung von trigonellin in Coffea arabica. Bericht der Schweizerischen Botanischen

Gesellschaft, 70, 232-267. [34] Nottbohm FE, Meyer F. (1931) Trigonellin gehalt des Kaffees. European Food Research and Technology, 61,429–435. [35] Hadorn H, Suter H. (1956) Über die analyse und zusammensetzung von rohkaffee verschiedener provenienzen. Mitteilungen

Lebensmitteluntersuchung und Hygiene, 47, 33–51. [36] Morath P. (1983) Untersuchungen zur Coffein-Biosynthese in Coffea arabica L. unter besonderer Berücksichtigung der Rolle des Adenins.

Inaugural-Dissertation der Philosophischen Fakultät II der Universität Zürich. [37] Koshiishi C, Kato A, Yama S, Crozier A, Ashihara H. (2001) A new caffeine biosynthetic pathway in tea leaves: utilisation of adenosine released

from S-adenosyl-L-methionine cycle. FEBS letters, 499, 50–54. [38] Zhou W, Karcher D, Bock R. (2013) Importance of adenosine-to-inosine editing adjacent to the anticodon in an Arabidopsis alanine tRNA under

environmental stress. Nucleic Acids Research, 41, 3362–3372. [39] Zhou W, Karcher D, Bock R. (2014) Identification of enzymes for adenosine-to-inosine editing and discovery of cytidine-to-uridine editing in

nucleus-encoded transfer RNAs of Arabidopsis. Plant Physiology, 166, 1985–1997. [40] Wanner H, Hagemann P. (1969) Über den Einfluss von Kinetin auf den Coffeingehalt der Blätter von Coffea arabica L. Planta, 86, 42–49. [41] Cheng Q, Gu J, Compaan KR, Schaefer HF, III. (2012) Isoguanine formation from adenine. Chemistry - A European Journal, 18, 4877–4886. [42] Kaplan NO, Colowick SP, Ciotti MM. (1952) Enzymatic deamination of adenosine derivatives. Journal of Biological Chemistry, 194, 579–591. [43] Kulikowska E, Kierdaszuk B, Shugar D. (2004) Xanthine, xanthosine and its nucleotides: solution structures of neutral and ionic forms, and

relevance to substrate properties in various enzyme systems and metabolic pathways. Acta Biochimica Polonica, 51, 493–531. [44] Mizuno K, Matsuzaki M, Kanazawa S, Tokiwano T, Yoshizawa Y, Kato M. (2014) Conversion of nicotinic acid to trigonelline is catalyzed by N-

methyltransferase belonged to motif B′ methyltransferase family in Coffea arabica. Biochemical and Biophysical. Research Communications, 452, 1060–1066.

[45] Ashihara H. (2015) Plant biochemistry: Trigonelline biosynthesis in Coffea arabica and Coffea canephora. In Coffee – In Health and Disease Prevention, Reedy VR. (Ed). Academic Press, 19-28.

[46] Dong W-R, Sun C-C, Zhu G, Hu S-H, Xiang L-X, Shao J-Z. (2014) New function for Escherichia coli xanthosine phosphorylase (xapA): genetic and biochemical evidences on its participation in NAD+

salvage from nicotinamide. BMC Microbiology, 14, 29.

Natural Product Communications Vol. 10 (5) 2015 Published online (www.naturalproduct.us)

A Stepwise Protocol for Induction and Selection of Prominent Coniferous Cell Cultures for the Production of -Thujaplicin Shinjiro Ogita, Masahito Shichiken, Chizuru Ito, Toshiyuki Yamashita, Taiji Nomura and Yasuo Kato 783

Low Caffeine Content in Novel Grafted Tea with Camellia sinensis as Scions and Camellia oleifera as Stocks Wei-Wei Deng, Min Li, Chen-Chen Gu, Da-Xiang Li, Lin-Long Ma, Yang Jin and Xiao-Chun Wan 789

Accounts/Reivews

Revisiting Caffeine Biosynthesis – Speculations about the Proximate Source of its Purine Ring Thomas W. Baumann 793

Biosynthesis of Caffeine Underlying the Diversity of Motif Bʹ Methyltransferase Fumiyo Nakayama, Kouichi Mizuno and Misako Kato 799

Occurrence, Biosynthesis and Metabolism of Theanine (γ-Glutamyl-L-ethylamide) in Plants: A Comprehensive Review Hiroshi Ashihara 803

Involvement of Allelopathy in the Formation of Monospecific Colonies of Ferns Hisashi Kato-Noguchi 811

Plant Cell, Tissue and Organ Culture: the Most Flexible Foundations for Plant Metabolic Engineering Applications Shinjiro Ogita 815

Natural Product Communications 2015

Volume 10, Number 5

Contents

Gerald Blunden Award (2014) Page

Stereocontrolled Total Synthesis of Tetrodotoxin from myo-Inositol and D-Glucose by Three Routes: Aspects for Constructing Complex Multi-Functionalized Cyclitols with Branched-Chain Structures Ken-ichi Sato, Shoji Akai and Juji Yoshimura 691

Metabolism and Biological Function of Natural Products in Plants (Guest editor: Hiroshi Ashihara)

Original Paper

(a) Biosynthesis

Occurrence and De novo Biosynthesis of Caffeine and Theanine in Seedlings of Tea (Camellia sinensis) Wei-Wei Deng and Hiroshi Ashihara 703

Metabolism of Purine Alkaloids and Xanthine in Leaves of Maté (Ilex paraguariensis) Yuling Yin, Riko Katahira and Hiroshi Ashihara 707

Comparative Analysis of Two DOPA Dioxygenases from Phytolacca americana Kana Takahashi, Kazuko Yoshida and Masaaki Sakuta 713 Biochemical Analysis of Phytolacca DOPA Dioxygenase Kana Takahashi, Kazuko Yoshida, Kei Yura, Hiroshi Ashihara and Masaaki Sakuta 717

Unraveling the Biosynthesis of Pilocarpine in Pilocarpus microphyllus Alexandra Christine Helena Frankland Sawaya, Yanna Dias Costa and Paulo Mazzafera 721

(b) Allelopathy

Transcriptomic Evaluation of Plant Growth Inhibitory Activity of Goniothalamin from the Malaysian Medicinal Plant Goniothalamus andersonii Naoya Wasano, Tomoko Takemura, Raihan Ismil, Baki Bakar and Yoshiharu Fujii 725

Momilactone Sensitive Proteins in Arabidopsis thaliana Hisashi Kato-Noguchi and Shinya Kitajima 729

Effect of Caffeine on the Expression Pattern of Water-Soluble Proteins in Rice (Oryza sativa) Seedlings Wei-Wei Deng, Hamako Sasamoto and Hiroshi Ashihara 733

Short Term Effect of Caffeine on Purine, Pyrimidine and Pyridine Metabolism in Rice (Oryza sativa) Seedlings Wei-Wei Deng, Riko Katahira and Hiroshi Ashihara 737

Cyanamide Phytotoxicity in Soybean (Glycine max) Seedlings involves Aldehyde Dehydrogenase Inhibition and Oxidative Stress John S. Maninang, Shin Okazaki and Yoshiharu Fujii 743

Allelopathy in a Leguminous Mangrove Plant, Derris indica: Protoplast Co-culture Bioassay and Rotenone Effect Aya Inoue, Daisuke Mori, Reiko Minagawa, Yoshiharu Fujii and Hamako Sasamoto 747

Effect of Purine Alkaloids on the Proliferation of Lettuce Cells Derived from Protoplasts Hamako Sasamoto, Yoshiharu Fujii and Hiroshi Ashihara 751

A Protocol for Axenic Liquid Cell Cultures of a Woody Leguminous Mangrove, Caesalpinia crista, and their Amino Acids Profiling Aya Inoue, Shinjiro Ogita, Shinpei Tsuchiya, Reiko Minagawa and Hamako Sasamoto 755

Phytotoxic Substance with Allelopathic Activity in Brachiaria decumbens Ai Kobayashi and Hisashi Kato-Noguchi 761

Isolation and Identification of an Allelopathic Substance from Hibiscus sabdariffa Prapaipit Suwitchayanon, Piyatida Pukclai, Osamu Ohno, Kiyotake Suenaga and Hisashi Kato-Noguchi 765

Angelicin as the Principal Allelochemical in Heracleum sosnowskyi Fruit Maryia Mishyna, Nikolai Laman, Valery Prokhorov and Yoshiharu Fujii 767

Identification of Octanal as Plant Growth Inhibitory Volatile Compound Released from Heracleum sosnowskyi Fruit Maryia Mishyna, Nikolai Laman, Valery Prokhorov, John Solomon Maninang and Yoshiharu Fujii 771

Identification of Safranal as the Main Allelochemical from Saffron (Crocus sativus) Hossein Mardani, Takayuki Sekine, Majid Azizi, Maryia Mishyna and Yoshiharu Fujii 775 (c) Biotechnology

The Biosynthetic Activities of Primary and Secondary Metabolites in Suspension Cultures of Aquilaria microcarpa Shinjiro Ogita, Jung-Bum Lee, Fumiya Kurosaki and Yasuo Kato 779

Continued inside backcover